Mckean Sylvia Ross John Dressler Daniel D Scheurer Danielle

Principles and Practice
of Hospital Medicine
NOTICE
Medicine is an ever-changing science. As new research and clinical experience broaden our
knowledge, changes in treatment and drug therapy are required. The authors and the publisher
o this work have checked with sources believed to be reliable in their e orts to provide in orma-
tion that is complete and generally in accord with the standards accepted at the time o publica-
tion. However, in view o the possibility o human error or changes in medical sciences, neither
the authors nor the publisher nor any other party who has been involved in the preparation or
publication o this work warrants that the in ormation contained herein is in every respect accu-
rate or complete, and they disclaim all responsibility or any errors or omissions or or the results
obtained rom use o the in ormation contained in this work. Readers are encouraged to con irm
the in ormation contained herein with other sources. For example, and in particular, readers are
advised to check the product in ormation sheet included in the package o each drug they plan to
administer to be certain that the in ormation contained in this work is accurate and that changes
have not been made in the recommended dose or in the contraindications or administration. This
recommendation is o particular importance in connection with new or in requently used drugs.
Principles and Practice
of Hospital Medicine
Second Edition
Editors
Sylvia C. McKean, MD, SFHM, FACP
Deputy Editor or Editorial Projects, UpToDate
Formerly:
Leave o absence:
Associate Pro essor o Medicine,
Harvard Medical School
Hospitalist
Brigham and Women’s Hospital
Boston, Massachusetts
John J. Ross, MD, CM, FIDSA

Assistant Pro essor o Medicine
Hospitalist Service
Daniel D. Dressler, MD, MSc, SFHM, FACP

Pro essor o Medicine
Director, Internal Medicine Teaching Services
Emory University Hospital
Associate Program Director
J. Willis Hurst Internal Medicine Residency Program
Co-Director, Semmelweis Society
Emory University School o Medicine
Atlanta, Georgia
Danielle B. Scheurer, MD, MSCR, SFHM

Chie Quality O icer and Hospitalist
Associate Pro essor o Medicine
Medical University o South Carolina
Charleston, South Carolina
New York Chicago San Francisco Athens London Madrid Mexico City
Milan New Delhi Singapore Sydney Toronto
Principles and Practice o Hospital Medicine, Second Edition
Copyright © 2017 by McGraw-Hill Education. All rights reserved. Printed in the United States
o America. Except as permitted under the United States Copyright Act o 1976, no part o this
publication may be reproduced or distributed in any orm or by any means, or stored in a data
base or retrieval system, without the prior written permission o the publisher. Previous edition
copyright © 2012 by The McGraw-Hill Companies, Inc.
1 2 3 4 5 6 7 8 9 DOW 21 20 19 18 17 16
ISBN 978-0-07-184313-3
MHID 0-07-184313-2
This book was set in Myriad pro by Cenveo Publisher Services.

The editors were Amanda Fielding and Kim J. Davis.
The production supervisor was Richard Ruzycka.
Project management was provided by Vastavikta Sharma, Cenveo Publisher Services.
The designer was Alan Barnett; the cover designer was Dreamit, Inc.
RR Donnelley was printer and binder.
Library o Congress Cataloging-in-Publication Data
Names: McKean, Sylvia C., editor. | Ross, John J. (John James), 1966-editor.
| Dressler, Daniel D., editor. | Scheurer, Danielle, editor.
Title: Principles and practice o hospital medicine / editors, Sylvia C.
McKean, John J. Ross, Daniel D. Dressler, Danielle B. Scheurer.
Description: Second edition. | New York : McGraw-Hill Education Medical, [2017]
| Includes bibliographical re erences and index.
Identi iers: LCCN 2016022668 (print) | LCCN 2016023825 (ebook)
| ISBN 9780071843133 (hardcover : alk. paper) | ISBN 0071843132 (hardcover :
alk. paper) | ISBN 9780071843140 (ebook)
Subjects: | MESH: Hospital Medicine—methods | Hospitalization | Inpatients |
Hospitalists Classi ication: LCC RA972 (print) | LCC RA972 (ebook) | NLM WX 21 |
DDC 362.11—dc23
LC record available at https://lccn.loc.gov/2016022668
McGraw-Hill Education books are available at special quantity discounts to use as premiums
and sales promotions, or or use in corporate training programs. To contact a representative,
please visit the Contact Us pages at www.mhpro essional.com.
CONTENTS
Editors ...............................................................................................................xi 18 Standardization and Reliability . . . . . . . . . . . . . . . . . . . . . . . 113
Contributors ................................................................................................. xiii 19 Tools to Identify Problems and Reduce Risks . . . . . . . . . . 118
Section Reviewers ..................................................................................xxxix 20 Preventing and Managing Adverse Patient
C
Foreword ......................................................................................................... xli Events: Patient Safety and the Hospitalist . . . . . . . . . . . . . 124
O
N
Preface ...........................................................................................................xliii 21 Principles and Models of Quality Improvement:
T
Plan-Do-Study-Act . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
E
Acknowledgments ..................................................................................... xlv
N
22 The Role of Information Technology in
T
S
Hospital Quality and Safety . . . . . . . . . . . . . . . . . . . . . . . . . . 134
PART I: THE SPECIALTY OF HOSPITAL

MEDICINE AND SYSTEMS OF CARE SECTION 5 Practice Management
23 Building, Growing and Managing a

The Value and Values of Hospitalist Practice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
SECTION 1 Hospital Medicine
24 Best Practices in Physician Recruitment
1 The Face of Health Care: Emerging Issues and Retention . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150
for Hospitalists . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 25 Teamwork in Leadership and Practice-Based
2 Value-Based Health Care for Hospitalists . . . . . . . . . . . . . . . 10 Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
3 Racial/Ethnic Disparities in Hospital Care . . . . . . . . . . . . . . . 18 26 Negotiation and Conflict Resolution . . . . . . . . . . . . . . . . . . 164
4 Comanagement of Orthopedic Patients ............... 23
5 Professionalism in Hospital Medicine . . . . . . . . . . . . . . . . . . 29 Billing, Coding, and Clinical
SECTION 6 Documentation
6 Principles of Leadership . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
27 Professional Coding and Billing Guidelines
Critical Decision Making at the for Clinical Documentation . . . . . . . . . . . . . . . . . . . . . . . . . . 173
SECTION 2 Point of Care 28 Consultation, Comanagement, Time-Based,
and Palliative Care Billing . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182
7 Principles of Evidence-Based Medicine and
Quality of Evidence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 29 Billing for Procedures and Use of Modifiers
in Inpatient Practice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192
8 Diagnostic Reasoning and Decision Making . . . . . . . . . . . . 45
30 Billing in the Teaching Setting and Billing
9 Principles of Evidence-Based Prescribing . . . . . . . . . . . . . . . 56 with Advanced Practice Providers . . . . . . . . . . . . . . . . . . . . 198
10 Summary Literature: Practice Guidelines 31 Hospital-Driven Documentation ..................... 204
and Systematic Reviews . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
32 Taming the ICD-10 Monster ......................... 210
11 Practical Considerations of Incorporating
Evidence into Clinical Practice . . . . . . . . . . . . . . . . . . . . . . . . 70
Principles of Medical Ethics and
SECTION 7 Medical-Legal Concepts
SECTION 3 Transitions of Care
33 Common Indications for Ethics Consultation .......... 217
12 Care Transitions into the Hospital: Health Care
Centers, Emergency Department, Outside 34 Medical-Legal Concepts: Advance Directives
Hospital Transfers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 and Surrogate Decision Making . . . . . . . . . . . . . . . . . . . . . . 224
13 Care Transitions within the Hospital: The Hand-Off ..... 84 35 Medical Malpractice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
14 Care Transitions at Hospital Discharge . . . . . . . . . . . . . . . . . 90
SECTION 8 Professional Development
Patient Safety and Quality
SECTION 4 36 Principles of Adult Learning and Continuing
Improvement
Medical Education . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243
15 Principles of Patient Safety: Intentional 37 Cultural Competence ............................... 246
Design and Culture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
38 Career Design and Development in Academic
16 Patient-Centered Care . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104 and Community Settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253
17 Harnessing Data to Make Quality Improvement 39 Mentorship of Peers and Trainees . . . . . . . . . . . . . . . . . . . . 257
Decisions: Measurement and Measures . . . . . . . . . . . . . . . 110
40 Research in the Hospital . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 264
v
41 For Individuals and Practices: Career 65 Management of Common Perioperative
Sustainability and Avoiding Burnout ................. 273 Complications in Orthopedic Surgery. . . . . . . . . . . . . . . . . 437
66 Transplant Surgery. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 447
67 Urology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 453
PART II: MEDICAL CONSULTATION
SECTION 1 Surgery PART III: REHABILITATION AND SKILLED

42 Physiologic Response to Surgery . . . . . . . . . . . . . . . . . . . . . 283 NURSING CARE
68 Postacute Care Rehabilitation Options . . . . . . . . . . . . . . . . 463
43 Perioperative Hemostasis . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286
C
69 Physical Therapy and Rehabilitation . . . . . . . . . . . . . . . . . . 469
O
44 Postoperative Complications ........................ 292
N
70 The Role of Speech/Language Pathologists in
45 Surgical Tubes and Drains . . . . . . . . . . . . . . . . . . . . . . . . . . . 297
T
Dysphagia Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . 476
E
N
46 Surgical Critical Care ................................ 300
71 Incontinence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 482
T
S
72 Pressure Ulcers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 491
SECTION 2 Anesthesia
73 Patient Safety and Quality Improvement
47 Anesthesia: Choices and Complications . . . . . . . . . . . . . . . 309 in Postacute Care . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 497
48 Perioperative Pain Management ..................... 313 74 Hospice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 505
Perioperative Risk Assessment and

SECTION 3 Management PART IV: APPROACH TO THE PATIENT AT
THE BEDSIDE
49 Role of the Medical Consultant . . . . . . . . . . . . . . . . . . . . . . . 325
75 Acute Abdominal Pain .............................. 513
50 Preoperative Cardiac Risk Assessment and
Perioperative Management . . . . . . . . . . . . . . . . . . . . . . . . . . 329 76 Acute Back Pain .................................... 522
51 Perioperative Pulmonary Risk Assessment and 77 Evaluation of Anemia ............................... 533

Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 336 78 Bleeding and Coagulopathy . . . . . . . . . . . . . . . . . . . . . . . . . 539
52 Perioperative Risk Assessment and Management 79 Chest Pain ......................................... 547
of the Diabetic Patient . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 342
80 Constipation ....................................... 556
53 Preoperative Evaluation of Liver Disease . . . . . . . . . . . . . . 348 81 Delirium ........................................... 563
54 Preoperative Assessment of Patients with 82 Diarrhea ........................................... 572
Hematologic Disorders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 352
83 Disorders of the Eye . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 579
Prevention, Assessment, and 84 Dizziness and Vertigo ............................... 587

SECTION 4 Management of Common 85 Dyspnea ........................................... 595
Complications in Noncardiac Surgery
86 Edema . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 604
55 Antimicrobial Prophylaxis in Surgery ................. 361 87 Falls . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 613
56 Venous Thromboembolism (VTE) Prophylaxis for 88 Fever and Rash ..................................... 620
Nonorthopedic Surgery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 368
89 Headache . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 625
57 Postoperative Blood Transfusion . . . . . . . . . . . . . . . . . . . . . 373
90 Hemoptysis ........................................ 636
58 Nutrition and Metabolic Support . . . . . . . . . . . . . . . . . . . . . 377
91 Hypertensive Urgencies and Emergencies ............ 641
59 Cardiac Complications after Noncardiac Surgery. . . . . . . 385
92 Hyperthermia and Fever. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 647
60 Management of Postoperative Pulmonary
93 Hypotension ....................................... 657
Complications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 391
94 Hypothermia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 665
61 Assessment and Management of Patients with
Renal Disease . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 397 95 Hypoxia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 669
62 Postoperative Neurologic and Psychiatric 96 Sleep Disturbance in the Hospitalized Patient . . . . . . . . . 676
Complications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 412 97 Nausea and Vomiting ............................... 689
98 Numbness: A Localization-Based Approach ........... 694
Specialty Consultation—What the
SECTION 5 99 Pain ............................................... 701
Consulting Hospitalist Needs to Know
100 Suspected Intoxication and Overdose ................ 709
63 Surgical Management of Obesity . . . . . . . . . . . . . . . . . . . . . 421 101 Syncope ........................................... 714
64 Common Postoperative Complications 102 Tachycardia ........................................ 729
in Neurosurgery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 430
103 The Geriatric History and Physical Examination . . . . . . . . 740
vi
104 The Neurologic Examination . . . . . . . . . . . . . . . . . . . . . . . . . 747 136 Pacemakers, Defibrillators, and Cardiac
105 Using Prognosis to Guide Treatment ................. 754 Resynchronization Devices in Hospital Medicine . . . . . 1025
106 Weakness: How to Localize the Problem . . . . . . . . . . . . . . 763

SECTION 2 Critical Care
137 Inpatient Cardiac Arrest and Cardiopulmonary

PART V: DIAGNOSTIC TESTING AND Resuscitation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1037
PROCEDURES 138 Acute Respiratory Failure . . . . . . . . . . . . . . . . . . . . . . . . . . 1047
139 Pain, Agitation and Delirium in the Critical
SECTION 1 Interpretation of Common Tests Care Setting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1055
C
140 Mechanical Ventilation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1065
O
107 Basic Diagnostic Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 771
N
108 The Resting Electrocardiogram . . . . . . . . . . . . . . . . . . . . . . . 776 141 Sepsis and Shock . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1074
T
E
142 Acute Respiratory Distress Syndrome . . . . . . . . . . . . . . . 1085
N
109 Elevated Liver Biochemical and Function Tests . . . . . . . . 796
T
143 Prevention in the Intensive Care Unit Setting . . . . . . . . 1094
S
110 Pulmonary Function Testing . . . . . . . . . . . . . . . . . . . . . . . . . 805
111 Urinalysis and Urine Electrolytes ..................... 814
SECTION 3 Dermatology
SECTION 2 Radiology 144 Flushing and Urticaria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1105

145 Adverse Cutaneous Drug Reactions ................ 1114
112 Introduction to Radiology . . . . . . . . . . . . . . . . . . . . . . . . . . . 825
146 Psoriasis and Other Papulosquamous Disorders ..... 1125
113 Patient Safety Issues in Radiology . . . . . . . . . . . . . . . . . . . . 831
147 Diabetic Foot Infections . . . . . . . . . . . . . . . . . . . . . . . . . . . 1131
114 Basic Chest Radiography (CXR) . . . . . . . . . . . . . . . . . . . . . . . 838
148 Venous Ulcers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1138
115 Advanced Cardiothoracic Imaging ................... 853 149 Dermatologic Findings in Systemic Disease ......... 1145
116 Basic Abdominal Imaging ........................... 864
117 Advanced Abdominal Imaging . . . . . . . . . . . . . . . . . . . . . . . 870 SECTION 4 Endocrinology
118 Neurologic Imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 876
150 Glycemic Emergencies ............................ 1171
119 Interventional Radiology . . . . . . . . . . . . . . . . . . . . . . . . . . . . 884 151 Inpatient Management of Diabetes and
Hyperglycemia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1178
SECTION 3 Procedures 152 Thyroid Emergencies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1184
120 Vascular Access . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 891 153 Adrenal Insufficiency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1191
121 Intubation and Airway Support ...................... 895 154 Pituitary Disease . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1198
122 Arterial Blood Gas and Placement of A-line . . . . . . . . . . . . 901

SECTION 5 Gastroenterology
123 Feeding Tube Placement . . . . . . . . . . . . . . . . . . . . . . . . . . . . 905
124 Thoracentesis ...................................... 909 155 GERD and Esophagitis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1209
125 Lumbar Puncture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 914 156 Upper Gastrointestinal Bleeding ................... 1217
157 Acute Pancreatitis ................................ 1227
126 Paracentesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 918
158 Jaundice, Obstruction, and Acute Cholangitis ....... 1232
127 Arthrocentesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 921
159 Acute Liver Disease ............................... 1239
160 Cirrhosis and Its Complications .................... 1253
PART VI: CLINICAL CONDITIONS IN THE 161 Acute Lower Gastrointestinal Bleeding ............. 1269
INPATIENT SETTING 162 Small Bowel Disorders. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1279
163 Large Bowel Disorders ............................ 1290
SECTION 1 Cardiovascular Medicine 164 Inflammatory Bowel Disease . . . . . . . . . . . . . . . . . . . . . . . 1308
128 Acute Coronary Syndromes . . . . . . . . . . . . . . . . . . . . . . . . . . 929
129 Heart Failure ....................................... 941 SECTION 6 Geriatrics
130 Myocarditis, Pericardial Disease, and 165 Principles of Geriatric Care . . . . . . . . . . . . . . . . . . . . . . . . . 1323
Cardiac Tamponade . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 954
166 Agitation in Older Adults . . . . . . . . . . . . . . . . . . . . . . . . . . 1330
131 Valvular Heart Disease .............................. 965
167 Elder Mistreatment ............................... 1336
132 Supraventricular Tachyarrhythmias . . . . . . . . . . . . . . . . . . . 980
168 Malnutrition and Weight Loss in Hospitalized
133 Bradycardia ........................................ 996 Older Adults. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1346
134 Ventricular Arrhythmias . . . . . . . . . . . . . . . . . . . . . . . . . . . 1003
135 Cardioversion .................................... 1015
vii
SECTION 7 Hematology SECTION 10 Neurology
169 Abnormalities in Red Blood Cells . . . . . . . . . . . . . . . . . . . 1353 207 Coma and Disorders of Consciousness . . . . . . . . . . . . . . 1667
170 Disorders of the White Cell ........................ 1373 208 Intracranial Hemorrhage and Related Conditions ... 1674
171 Quantitative Abnormalities of Platelets: 209 Transient Ischemic Attack and Stroke . . . . . . . . . . . . . . . 1681
Thrombocytopenia and Thrombocytosis . . . . . . . . . . . . 1381 210 Parkinson’s Disease and Related Disorders . . . . . . . . . . 1690
172 Approach to Patients with Bleeding Disorders . . . . . . . 1392 211 Seizures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1700
173 Hypercoagulable States . . . . . . . . . . . . . . . . . . . . . . . . . . . 1399 212 Multiple Sclerosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1712
174 Hematologic Malignancies ........................ 1405 213 Peripheral Neuropathy . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1719
C
O
SECTION 8 Oncology
N
SECTION 11 Palliative Care
T
E
175 Overview of Cancer and Treatment . . . . . . . . . . . . . . . . . 1431
N
214 Principles of Palliative Care ........................ 1727
T
176 Oncologic Emergencies ........................... 1436
S
215 Communication Skills for End-of-Life Care .......... 1733
177 Approach to the Patient with 216 Domains of Care: Physical Aspects of Care .......... 1740
Suspected Malignancy . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1443
217 Care of the Dying Patient ......................... 1756
178 Breast, Ovary, and Cervical Cancer. . . . . . . . . . . . . . . . . . 1454
179 Men’s Cancers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1458 SECTION 12 Pregnancy
180 Cancers of the Kidney, Renal Pelvis, and Ureter . . . . . . 1463
181 Oncologic Issues of the Aerodigestive Tract . . . . . . . . . 1468 218 Overview of Physiologic Changes of Pregnancy ..... 1767
182 Gastrointestinal Cancers . . . . . . . . . . . . . . . . . . . . . . . . . . . 1474 219 Medication Management ......................... 1771
183 Immune-Related Adverse Events (irAEs) 220 Critical Care of the Pregnant Patient . . . . . . . . . . . . . . . . 1781
in Cancer Patients . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1477 221 Common Medical Problems in Pregnancy. . . . . . . . . . . 1786
222 Postpartum Consultation for Common Complaints .. 1807
SECTION 9 Infectious Disease
SECTION 13 Psychiatry
184 Fundamentals of Antibiotics . . . . . . . . . . . . . . . . . . . . . . . 1489
185 Antibiotic Resistance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1498 223 Mood and Anxiety Disorders . . . . . . . . . . . . . . . . . . . . . . . 1813
186 Community-Acquired Pneumonia . . . . . . . . . . . . . . . . . . 1503 224 Combat Stress and Related Disorders . . . . . . . . . . . . . . . 1827
187 Health Care and Hospital-Acquired 225 Assessment and Management of Psychosis ......... 1833
Pneumonia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1514 226 Eating Disorders ................................. 1841
188 Intravascular Catheter-Related Infections: 227 The Suicidal Patient . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1848
Management and Prevention . . . . . . . . . . . . . . . . . . . . . . 1519 228 The Difficult Patient .............................. 1853
189 Infective Endocarditis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1528 229 Approach to the Patient with Multiple Unexplained
190 Clostridium difficile–Associated Somatic Symptoms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1861
Disease (CDAD) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1537
191 Peritonitis and Intra-Abdominal Abscess . . . . . . . . . . . . 1542 SECTION 14 Pulmonary and Allergy Immunology
192 Meningitis and Encephalitis. . . . . . . . . . . . . . . . . . . . . . . . 1549
230 Allergy and Anaphylaxis . . . . . . . . . . . . . . . . . . . . . . . . . . . 1871
193 Osteomyelitis and Septic Arthritis . . . . . . . . . . . . . . . . . . 1557
231 Asthma . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1876
194 Prosthetic Joint Infections . . . . . . . . . . . . . . . . . . . . . . . . . 1564
232 Chronic Obstructive Pulmonary Disease ............ 1887
195 Sexually Transmitted Infections . . . . . . . . . . . . . . . . . . . . 1574
233 Interstitial Lung Diseases/Diffuse Parenchymal
196 Skin and Soft Tissue Infections . . . . . . . . . . . . . . . . . . . . . 1582 Lung Diseases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1898
197 Urinary Tract Infections. . . . . . . . . . . . . . . . . . . . . . . . . . . . 1589 234 Cystic Fibrosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1905
198 Viral Infections ................................... 1596 235 Sleep Apnea and Obesity Hypoventilation
199 Tickborne Infections .............................. 1604 Syndrome . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1915
200 Tuberculosis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1612 236 Pleural Diseases .................................. 1923
201 Candida and Aspergillus ........................... 1618 237 Pulmonary Hypertension ......................... 1932
202 Histoplasmosis, Blastomycosis, Coccidioidomycosis,
and Other Dimorphic Fungi. . . . . . . . . . . . . . . . . . . . . . . . 1625 SECTION 15 Renal
203 The Hospitalized Patient with HIV . . . . . . . . . . . . . . . . . . 1634
238 Acid-Base Disorders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1943
204 Infections of the Immunocompromised Host ....... 1646
239 Acute Kidney Injury . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1952
205 Fever in the Returning Traveler .................... 1653
240 Calcium Disorders ................................ 1961
206 Undiagnosed Fever in Hospitalized Patients ........ 1659
viii
241 Potassium and Magnesium Disorders . . . . . . . . . . . . . . . 1972 SECTION18 Vascular Medicine
242 Disorders of Sodium and Water Balance ............ 1982
252 Venous Thromboembolism Prophylaxis for
243 Kidney Stones . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1993
Hospitalized Medical Patients . . . . . . . . . . . . . . . . . . . . . . 2077
244 Secondary Hypertension .......................... 2000
253 Diagnosis and Treatment of Venous
245 Chronic Kidney Disease and Dialysis . . . . . . . . . . . . . . . . 2008
Thromboembolism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2081
254 Anticoagulant Therapy ........................... 2093
SECTION 16 Rheumatology
255 Diseases of the Aorta . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2107
246 Rheumatologic Emergencies. . . . . . . . . . . . . . . . . . . . . . . 2017 256 Acute and Chronic Lower Limb Ischemia . . . . . . . . . . . . 2115
247 Gout, Pseudogout, and Osteoarthritis 2023
C
.............. 257 Vasculitis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2121
O
248 Systemic Lupus Erythematosus .................... 2033
N
T
249 Rheumatoid Arthritis and Other Inflammatory Online Chapters
E
N
Arthritides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2046
T
e1 Global Health and Hospital Medicine
S
SECTION 17 Toxicology and Addiction e2 The Economics of Hospital Care
e3 Principles of Medical Ethics
250 Drug Overdose and Withdrawal . . . . . . . . . . . . . . . . . . . . 2057 e4 The Core Competencies in Hospital Medicine
251 Addiction of Prescription and e5 Bioterrorism
Nonprescription Drugs . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2070
Index ................................................. 2129
ix
EDITORS
Sylvia C. McKean, MD, SFHM, FACP
xi
CONTRIBUTORS
Numbers in brackets refer to the chapters written or co-written by the contributor.
Samer Abdel-Aziz, MD Daniel A. Anaya, MD

Pain Medicine Fellowship Associate Pro essor o Surgery
Medical College o Wisconsin Division o Surgical Oncology
C
Milwaukee, Wisconsin [99] Michael E. DeBakey Department o Surgery
O
N
Research Scientist
Meredith C. B. Adams, MD, MS
T
The Houston Center or Quality o Care &Utilization Studies
R
Assistant Pro essor o Anesthesiology and Pain Medicine
I
Director
B
Director
U
Liver Tumor Program, Michael E. DeBakey VAMC
T
Pain Medicine Fellowship Baylor College o Medicine
O
Medical College o Wisconsin Houston, Texas [55]
R
S
Milwaukee, Wisconsin [99]
Douglas S. Ander, MD
Aaron W. Aday, MD Pro essor o Emergency Medicine
Division o Cardiovascular Medicine Emory University School o Medicine
Department o Medicine Atlanta, Georgia [121]
Boston, Massachusetts [255] Eddy Ang, MD
Instructor in Medicine
Bhavin Adhyaru, MD Harvard Medical School
Emory University School o Medicine Division o Gerontology
Atlanta, Georgia [10] Beth Israel Deaconess Medical Center
Kush Agrawal, MD Boston, Massachusetts
Advanced Endovascular and Structural Interventional Fellow Department o Medicine, Hebrew SeniorLi e
VIVA Physicians Roslindale, Massachusetts [68]
El Camino Hospital Kelly Armstrong, PhD
Mountain View, Cali ornia [101] Senior Clinical Ethicist
Mikhail Akbashev, MD Memorial Health System
Assistant Pro essor o Medicine Adjunct Assistant Pro essor
Emory University School o Medicine Department o Medical Humanities
Atlanta, Georgia [11] SIU School o Medicine
Spring eld, Illinois [34]
A saneh Alavi, MD, MSc, FRCPC
Department o Medicine (Dermatology) Vineet M. Arora, MD, MAPP
University o Toronto Associate Pro essor and Assistant Dean or Scholarship and
Toronto, Ontario, Canada [148] Discovery
Director
G. Caleb Alexander, MD, MS GME Clinical Learning Environment Innovation
Associate Pro essor o Epidemiology and Medicine Pritzker School o Medicine
Bloomberg School o Public Health University o Chicago
Johns Hopkins University Chicago, Illinois [13]
Baltimore, Maryland [e3]
Cameron Ashbaugh, MD
Anne E. Allan, MD Assistant Pro essor
Miraca Li e Sciences Harvard Medical School
Irving, Texas [146] Division o In ectious Diseases
Ashwin Ananthakrishnan, MD Boston, Massachusetts [196]
Attending Physician
Massachusetts General Hospital Saima Aslam, MD, MS
Instructor in Medicine, Harvard Medical School Assistant Pro essor
Boston, Massachusetts [183] Director
Solid Organ Transplant In ectious Diseases service
Division o In ectious Diseases
University o Cali ornia, San Diego
San Diego, Cali ornia [188]
xiii
Mark J. Ault, MD Robert B. Baron, MD, MS
Pro essor o Medicine Pro essor o Medicine
University o Cali ornia Associate Dean or Graduate and Continuing Medical Education
Los Angeles School o Medicine Vice Chie , Division o General Internal Medicine
Department o Medicine University o Cali ornia
Cedars-Sinai Medical Center San Francisco School o Medicine
Los Angeles, Cali ornia [125] San Francisco, Cali ornia [36]
Patrick Avila, MD, MPhil, MPH Tom Baudendistel, MD, FACP

Internal Medicine Resident Physician Internal Medicine Residency Program Director
Brigham and Women’s Hospital Kaiser Permanente
C
Boston, Massachusetts [160] Oakland, Cali ornia [39]
O
N
Vasilis C. Babaliaros, MD Mihaela H. Bazalakova, MD, PhD
T
Pro essor o Medicine and Surgery Assistant Pro essor
R
I
B
Co-Director Department o Neurology
U
Emory Structural Heart and Valve Center Center or Sleep Medicine and Sleep Research
T
O
Emory University Hospital University o Wisconsin-Madison
R
Atlanta, Georgia [131] Madison, Wisconsin [96]
S
Lindsey R. Baden, MD Joshua A. Beckman, MD, MSc
Associate Pro essor Section o Vascular Medicine
Harvard Medical School Cardiovascular Division
In ectious Diseases Division Vanderbilt University Medical Center
Brigham and Women’s Hospital and Dana-Farber Cancer Institute Nashville, Tennessee [255]
Boston, Massachusetts [204]
Nicole M. Bedi, RD, CNSC
Meridale V. Baggett, MD Birmingham, Michigan [58]
Harvard Medical School Laurence Beer, MD, SFHM
Inpatient Clinician Educator Service Emory University School o Medicine
Department o Medicine Atlanta, Georgia [8]
Massachusetts General Hospital Michael Belkin, MD
Boston, Massachusetts [78] Division o Vascular Surgery
James L. Bailey, MD Brigham and Women’s Hospital
Pro essor Boston, Massachusetts [256]
Emory University School o Medicine Elie F. Berbari, MD
Atlanta, Georgia [241] Pro essor o Medicine
Stephen J. Balevic, MD Mayo Clinic College o Medicine
Adult and Pediatric Rheumatology Fellow Rochester, Minnesota [194]
Duke University Medical Center Colm Bergin, MD, FRCPI, FRCP, FIDSA
Durham, North Carolina [248] Clinical Pro essor o Medicine
Peter A. Banks, MD Trinity College Dublin
Pro essor o Medicine Consultant Physician in In ectious Diseases
Harvard Medical School Associate Director
Director o the Center or Pancreatic Disease Wellcome-Health Research Board Clinical Research Facility
Division o Gastroenterology, Hepatology and Endoscopy St. James’s Hospital
Department o Medicine Dublin, Ireland [195]
Brigham and Women’s Hospital Aaron L. Berkowitz, MD, PhD
Boston, Massachusetts [157] Department o Neurology
Aditya Bardia, MD, MPH Brigham and Women’s Hospital
Assistant Pro essor Harvard Medical School
Harvard Medical School Boston, Massachusetts [98]
Attending Physician Rachelle E. Bernacki, MD, MS
Massachusetts General Hospital Cancer Center Assistant Pro essor o Medicine
Boston, Massachusetts [177, 183] Harvard Medical School
Maria F. Barile, MD Director o Quality Initiatives
Clinical Instructor in Radiology Palliative Care
Harvard Medical School Dana-Farber Cancer Institute
Thoracic Radiologist, Brigham and Women’s Hospital
Brigham and Women’s Hospital Ariadne Labs
Boston, Massachusetts [114, 115] Boston, Massachusetts [105, 214]
xiv
Robert A. Bessler, MD Ghada Bourjeily, MD
CEO Associate Pro essor o Medicine
Sound Physicians The Warren Alpert Medical School o Brown University
Tacoma, Washington [23, 25] The Miriam Hospital
Pulmonary, Critical Care, Obstetric Medicine
John P. Bilezikian, MD Department o Medicine
Dorothy L. and Daniel H. Silberberg Pro essor o Medicine Providence, Rhode Island [220]
Pro essor o Pharmacology
Columbia University Medical College John M. Braver, MD
Chie , Division o Endocrinology Assistant Pro essor
Director Harvard Medical School
C
Metabolic Bone Diseases Program Director
O
Columbia University Medical Center Gastrointestinal Radiology
N
New York, New York [240] Brigham and Women’s Hospital
T
R
Boston, Massachusetts [116, 117]
I
Courtney Bilodeau, MD, FACP
B
U
Assistant Pro essor Ursula C. Brewster, MD
T
O
The Warren Alpert Medical School o Brown University Associate Pro essor o Medicine
R
Department o Obstetric Medicine, Women’s Medicine Section o Nephrology
S
Collaborative Yale University School o Medicine
Miriam Hospital New Haven, Connecticut [245]
Providence, Rhode Island [222]
Joseph Brito, MD
Kenneth D. Bishop, MD, PhD Division o Urology
Assistant Pro essor o Medicine The Warren Alpert Medical School o Brown University
Division o Hematology/Oncology Providence, Rhode Island [180]
Rhode Island Hospital
The Warren Alpert Medical School o Brown University Jared R. Brosch, MD, MS
Providence, Rhode Island [176] Assistant Pro essor o Neurology
Indiana University School o Medicine
Ioannis A. Bliziotis, MD, PhD, MSc Indianapolis, Indiana [192]
Internal Medicine and In ectious Diseases Specialist
Senior Researcher Katherine L. Brown, MD, MPH
Al a Institute o Biomedical Sciences Suncoast Dermatology
Athens, Greece [184] Orlando, Florida [148]
Arline D. Bohannon, MD Tod A. Brown, MD

Associate Pro essor o Internal Medicine Assistant Pro essor
Virginia Commonwealth University Anesthesia and Perioperative Medicine
Richmond, Virginia [103] Medical University o South Carolina
Charleston, South Carolina [62]
Peter A. Boling, MD
Pro essor o Internal Medicine Avery L. Buchholz, MD, MPH
Virginia Commonwealth University Health System Department o Neurosurgery
Richmond, Virginia [103] Medical University o South Carolina
Marcy B. Bolster, MD
Associate Pro essor Tina Budnitz, MPH, MHM
Harvard Medical School Senior Advisor
Director Society Hospital Medicine
Rheumatology Fellowship Training Program Philadelphia, Pennsylvania [e4]
Massachusetts General Hospital Robert Burakof , MD, MPH
Boston, Massachusetts [248, 249] Associate Pro essor o Medicine
Diego F. Bonilla Arcos, MD Division o Gastroenterology, Hepatology, and Endoscopy
Pulmonary Critical Care Brigham and Women’s Hospital
Boston University Harvard Medical School
Pulmonary Center Boston, Massachusetts [155]
Boston Medical Center T. Karl Byrne, MD, FACS
Boston, Massachusetts [237] Pro essor o Surgery
Joanna M. Bonsall, MD, PhD Director
Assistant Pro essor o Medicine Bariatric Surgery Program
Division o Hospital Medicine Medical University o South Carolina
Emory University School o Medicine Charleston, South Carolina [63]
Atlanta, Georgia [12]
xv
Amanda Caissie, MD, PhD, FRCPC Olga S. Chajewski, MD
Department o Radiation Oncology Department o Pathology and Laboratory Medicine
Dalhousie University Medical University o South Carolina
Saint John Regional Hospital Charleston, South Carolina [57]
Saint John, New Brunswick, Canada [215]
Walter W. Chan, MD, MPH
Evelyn Cantillo, MD, MPH Assistant Pro essor o Medicine
Clinical Instructor o Obstetrics and Gynecology Division o Gastroenterology, Hepatology, and Endoscopy
Program in Women’s Oncology Harvard Medical School
Women &In ants’Hospital o Rhode Island Brigham and Women’s Hospital
The Warren Alpert Medical School o Brown University Boston, Massachusetts [155]
C
O
Arjun S. Chanmugam, MD, MBA
N
Stephanie M. Cantu, MD Associate Pro essor o Emergency Medicine
T
Department o Medicine Johns Hopkins University School o Medicine
R
I
B
Brigham and Women’s Hospital Baltimore, Maryland [79]
U
T
Helen Chen, MD
O
Mitchell S. Cappell, MD, PhD Assistant Pro essor o Medicine
R
S
Pro essor o Medicine Harvard Medical School
Oakland University William Beaumont School o Medicine Division o Gerontology
Chie , Division o Gastroenterology and Hepatology Beth Israel Deaconess Medical Center
Department o Medicine Chie Medical O cer
William Beaumont Hospital Hebrew SeniorLi e
Royal Oak, Michigan [163] Boston, Massachusetts [73]
Alexander R. Carbo, MD, FACP, SFHM Kenneth K. Chen, MD, FRACP

Assistant Pro essor o Medicine Assistant Pro essor o Medicine and OB/GYN
Harvard Medical School Division o Obstetric and Consultative Medicine
Hospitalist The Warren Alpert Medical School o Brown University
Beth Israel Deaconess Medical Center Providence, Rhode Island [221]
Steven T. Chen, MD, MPH
Teresa L. Carman, MD Instructor in Dermatology
Assistant Pro essor o Medicine Harvard Medical School
Case Western Reserve University School o Medicine Massachusetts General Hospital
Director Boston, Massachusetts [145]
Vascular Medicine
University Hospitals Case Medical Center Xi Chen, MD, PhD
Cleveland, Ohio [86] Neurology Department
Atrius Health
Patrick J. Cawley, MD, MHM Boston, Massachusetts [96]
CEO
MUSC Health Nishay Chitkara, MD
Vice President or Health A airs Assistant Pro essor o Medicine
Medical University o South Carolina NYU Langone Medical Center/Bellevue Hospital
Charleston, South Carolina [1] Department o Medicine
Division o Pulmonary
Laura K. Certain, MD, PhD Critical Care and Sleep Medicine
Instructor in Medicine New York, New York [143]
Division o In ectious Diseases Louisa W. Chiu, MD
Massachusetts General Hospital Assistant Pro essor o Surgery
Boston, Massachusetts [193] Michael E. DeBakey Department o Surgery
Baylor College o Medicine
Matthew E. Certain, MD Houston, Texas [55]
Interventional and Peripheral Cardiologist
Southeast Georgia Health Systems Elbert B. Chun, MD
Brunswick, Georgia [130] Assistant Pro essor
Division o Hospital Medicine
Sukit Chaiyachati, MD Department o Internal Medicine
Assistant Pro essor o Medicine Emory University School o Medicine
Division o Hospital Medicine Emory University Hospital
Emory University School o Medicine Atlanta, Georgia [132]
xvi
Roger P. Clark, DO Frank E. Corrigan, III, MD
Assistant Pro essor o Medicine Cardiology Fellow
Division o Geographic Medicine and In ectious Diseases Emory University School o Medicine
Tu ts Medical Center Atlanta, Georgia [101]
Consultant, In ectious Diseases
Brigham and Women’s Faulkner Hospital Dominique L. Cosco, MD, FACP
Boston, Massachusetts [199] Assistant Pro essor o Medicine
Associate Program Director
John O. Clarke, MD Grady Memorial Hospital
Associate Pro essor o Medicine J. Willis Hurst Internal Medicine Residency Program
Division o Gastroenterology &Hepatology Emory University School o Medicine
C
Johns Hopkins University Atlanta, Georgia [230]
O
Baltimore, Maryland [97]
N
Douglas B. Coursin, MD, FCCP
T
Stephen D. Clements, Jr., MD Pro essor o Anesthesiology and Medicine
R
I
B
Pro essor o Medicine (Cardiology) University o Wisconsin School o Medicine and Public Health
U
R. Harold Harrison Chair in Cardiology Madison, Wisconsin [153]
T
O
Division o Cardiology
Lisa Criscione -Schreiber, MD, MEd
R
Department o Medicine
S
Emory University School o Medicine Associate Pro essor o Medicine
Atlanta, Georgia [130] Rheumatology Training Program Director
Duke University Medical Center
Steven L. Cohn, MD, FACP, SFHM Duke University School o Medicine
Pro essor o Clinical Medicine Durham, North Carolina [248, 249]
University o Miami Miller School o Medicine
Medical Director Yvette M. Cua, MD
UHealth Preoperative Assessment Center Associate Pro essor o Medicine
Director Department o Medicine
Medical Consultation Services Associate Vice Chair or Clinical A airs
University o Miami Hospital and Jackson Memorial Hospital Department o Medicine
Miami, Florida [49, 50] University o Louisville
Louisville, Kentucky [28-30, 32]
Lauren Colbert, MD
Fellow Randall Czajkowski, MS, RRA, RT(R)(CT)
Radiation Oncology Lead Clinical CT Technologist
MD Anderson Cancer Center Brigham and Women’s Hospital
Houston, Texas [182] Boston, Massachusetts [119]
Alexandra Columbus, MD Sonye K. Danof , MD, PhD

Resident Associate Pro essor o Medicine
General Surgery Department o Medicine
Brigham and Women’s Hospital Division o Pulmonary and Critical Care Medicine
Boston, Massachusetts [45] Johns Hopkins University School o Medicine
Jose F. Condado, MD, MS
Cardiology Research Fellow Jatin K. Dave, MD, MPH
Structural Heart and Valve Center Part-Time Instructor, Harvard Medical School
Division o Cardiology Division o Aging, Brigham and Women’s Hospital
Emory University School o Medicine Boston, Massachusetts
Atlanta, Georgia [131] Medical Director
Geriatrics and Senior Care Options
Nicholas J. Connors, MD Tu ts Health Plan
Assistant Pro essor o Medicine Watertown, Massachusetts [68, 73]
Division o Emergency Medicine David B. De Lurgio, MD
Section o Medical Toxicology Pro essor
Charleston, South Carolina [100] Clinical Cardiac Electrophysiology
Emory Saint Joseph’s Hospital
Darin J. Correll, MD Atlanta, Georgia [136]
Assistant Pro essor o Anesthesia
Harvard Medical School Steven B. Deitelzweig, MD, MMM, SFHM, FACP
Director Ochsner Health System
Postoperative Pain Management Service Medical Director o Regional Business Development
Department o Anesthesiology, Perioperative, and Pain Medicine System Chairman, Hospital Medicine
Chair, Acute Pain Committee Associate Pro essor o Medicine-Ochsner Clinical School [24]
xvii
Paul F. Dellaripa, MD Daniel D. Dressler, MD, MSc, SFHM, FACP
Associate Pro essor Pro essor o Medicine
Harvard Medical School Director
Division o Rheumatology Internal Medicine Teaching Services
Brigham and Women’s Hospital Emory University Hospital
Boston, Massachusetts [246] Associate Program Director
J. Willis Hurst Internal Medicine Residency Program
E. Patchen Dellinger, MD Co-Director
Pro essor and Vice Chair Semmelweis Society
Department o Surgery Emory University School o Medicine
University o Washington Atlanta, Georgia [101]
C
Seattle, Washington [55]
O
Jacob M. Drew, MD
N
Harry A. Demos, MD Assistant Pro essor
T
Associate Pro essor
R
Department o Orthopedics
I
B
Department o Orthopedics Medical University o South Carolina
U
T
O
R
Catherine E. DuBeau, MD
S
Rebecca Dezube, MD Pro essor o Medicine
Postdoctoral Fellow Family Medicine and Community Health, and Obstetrics and
Johns Hopkins University Gynecology
Pulmonary and Critical Care Medicine Clinical Chie o Geriatrics
Baltimore, Maryland [95] University o Massachusetts Medical School
Lorenzo Di Francesco, MD, FACP, FHM Worcester, Massachusetts [71]
Pro essor o Medicine Jenni er Duf , MD
Division o General Medicine &Geriatrics Assistant Pro essor o Medicine
Program Director Division o Hematology and Oncology
J. Willis Hurst Internal Medicine Residency Program Department o Medicine
Assistant Chie o Medicine, Grady Memorial Hospital University o Florida College o Medicine
Emory University School o Medicine Hematology and Oncology Section
Atlanta, Georgia [93] NF/SG Veterans A airs Medical Center
Shira Doron, MD, FIDSA Gainesville, Florida [175]
Antimicrobial Steward Liam Durcan, MD, FRCPC
Associate Hospital Epidemiologist Assistant Pro essor
Division o Geographic Medicine and In ectious Diseases Department o Neurology and Neurosurgery
Tu ts Medical Center McGill University
Boston, Massachusetts [191] Consultant Neurologist
Michael Dougan, MD, PhD McGill University Health Centre
Gastroenterology Fellow Montreal, Quebec, Canada [207]
Massachusetts General Hospital Kent Russell Edwards, Jr., MD
Boston, Massachusetts [183] Urology Research Assistant
James D. Douketis, MD, FRCP(C), FACP, FCCP University o South Carolina School o Medicine
Pro essor o Medicine Columbia, South Carolina [67]
McMaster University Mikhael F. El-Chami, MD, FACC, FHRS
Hospitalist Service Associate Pro essor o Medicine
St. Joseph’s Healthcare Hamilton Division o Cardiology-Section o Electrophysiology
Hamilton, Ontario, Canada [56, 252] Emory University
Aeron A. D. Doyle, MD, CM, FRCPC Atlanta, Georgia [134]
Assistant Pro essor o Anesthesiology Elwaleed A. Elhassan, MD, FACP, FASN
University o British Columbia Assistant Pro essor o Medicine
Department o Anesthesiology, Perioperative Medicine, and Pain Division o Nephrology and Hypertension
Management Wayne State University School o Medicine
Providence Health Care Detroit, Michigan [242]
Vancouver, British Columbia, Canada [47]
William J. Elliott, MD, PhD
Tracy J. Doyle, MD, MPH Chair
Instructor in Medicine Department o Biomedical Sciences
Harvard Medical School Chie , Division o Pharmacology
Pulmonary and Critical Care Medicine Pro essor o Preventative Medicine, Internal Medicine, Pharmacology
Brigham and Women’s Hospital Paci c Northwest University o Health Sciences
Boston, Massachusetts [85] Yakima, Washington [244]
xviii
John M. Embil, MD, FRCPC, FACP James C. Fang, MD, FACC, FAHA
Pro essor o Internal Medicine and Medical Microbiology Chie
University o Manitoba Cardiovascular Division
Director Pro essor o Medicine
In ection Prevention and Control Unit, Health Sciences Centre John and June B. Hartman Presidential Endowed Chair
Winnipeg, Manitoba, Canada [147, 202] Executive Director
Cardiovascular Service Line
Scott F. Enderby, DO, MMM, SFHM, FACP University o Utah Health Sciences Center
CEO/Medical Director Salt Lake City, Utah [129]
Bay Area Hospitalist Associates, Inc.
San Francisco, Cali ornia [25] John Fanikos, RPh, MBA
C
Department o Pharmacy Services
O
Jeannine Z. Engel, MD Brigham and Women’s Hospital
N
Associate Pro essor o Medicine Boston, Massachusetts [254]
T
Internal Medicine, Huntsman Cancer Hospital
R
I
Harrison W. Farber, MD
B
Physician Advisor, Billing Compliance
U
Compliance Services Pro essor o Medicine
T
O
University o Utah Health Sciences Boston University School o Medicine
R
Salt Lake City, Utah [31] Director
S
Pulmonary Hypertension Center
Joseph C. English, III, MD Boston Medical Center
Pro essor o Dermatology Boston, Massachusetts [237]
University o Pittsburgh Department o Dermatology
Pittsburgh, Pennsylvania [144] Claire E. Farel, MD, MPH
Clinical Assistant Pro essor o Medicine
Mary Eno, MD, MPH University o North Carolina School o Medicine
Regional Chie o Addiction Medicine Medical Director
Southern Cali ornia Permanente Medical Group UNC In ectious Diseases Clinic
Los Angeles, Cali ornia [251] UNC Institute or Global Health and In ectious Diseases
Andrew S. Epstein, MD Chapel Hill, North Carolina [203]
Assistant Attending Dimitrios Farmakiotis, MD
Memorial Sloan Kettering Cancer Center Assistant Pro essor o Medicine
Gastrointestinal Oncology Service New York University School o Medicine
New York, New York [182] Division o In ectious Diseases and Immunology, NYU Langone
Evert A. Eriksson, MD, FACS, FCCP Medical Center
Associate Pro essor o Surgery New York, New York [201]
Department o Surgery Jeanne M. Farnan, MD, MHPE
Medical University o South Carolina Associate Pro essor
Charleston, South Carolina [46] Section o Hospital Medicine
Samir M. Fakhry, MD, FACS Assistant Dean, Curricular Development and Evaluation
Charles F. Crews Pro essor o Surgery Pritzker School o Medicine
Chie , Division o General Surgery University o Chicago
Department o Surgery Chicago, Illinois [13]
Medical University o South Carolina Grace Farris, MD
Charleston, South Carolina [46] Beth Israel Deaconess Medical Center
Matthew E. Falagas, MD, MSc, DSc Boston, Massachusetts [168]
Adjunct Associate Pro essor o Medicine Kevin Felner, MD
Tu ts University School o Medicine Associate Pro essor
Boston, Massachusetts Division o Pulmonary and Critical Care
Director New York University School o Medicine
Al a Institute o Biomedical Sciences Harbor VA Medical Center
Director New York, New York [141]
Department o Internal Medicine and In ectious Diseases
Iaso General Hospital Andrew Z. Fenves, MD, FACP, FASN
Athens, Greece [184] Associate Pro essor o Medicine
Kenneth R. Falchuk, MD Clinician Educator Service
Associate Clinical Pro essor o Medicine Massachusetts General Hospital
Co-Director
Inf ammatory Bowel Disease Center
Department o Medicine, Division o Gastroenterology
Beth Israel Deaconess Medical Center
xix
Joseph D. Feuerstein, MD Michael Gardam, MD, MSc, FRCPC
Assistant Pro essor o Medicine Associate Pro essor o Medicine
Harvard Medical School University o Toronto
Attending in Gastroenterology Director
Center or Inf ammatory Bowel Disease In ection Prevention and Control
Beth Israel Deaconess Medical Center University Health Network
Boston, Massachusetts [164] Medical Director
Tuberculosis Clinic
Joseph J. Fins, MD, MACP Toronto Western Hospital
The E. William Davis, Jr., MD Pro essor o Medical Ethics and Toronto, Ontario, Canada [200]
C
Weill Cornell Medical College Brian T. Garibaldi, MD
O
Director Assistant Pro essor o Medicine
N
Medical Ethics and Attending Physician Johns Hopkins University School o Medicine
T
R
New York Presbyterian Hospital-Weill Cornell Medical Center Department o Medicine
I
B
New York, New York [33] Division o Pulmonary and Critical Care Medicine
U
T
O
Leslie A. Flores, MHA, SFHM
R
Nelson Flores Hospital Medicine Consultants Steven Garlow, MD, PhD
S
La Quinta, Cali ornia [26] Associate Pro essor
Chie o Psychiatry
John A. Flynn, MD, MBA, MEd, FACP, FACR Emory University Hospital
Medical Director Atlanta, Georgia [223]
Spondyloarthritis Program
Associate Dean and Executive Director Germán E. Giese, MD
Clinical Practice Association Assistant Pro essor o Medicine
Vice President, O ce o Johns Hopkins Physicians University o Miami Miller School o Medicine
Johns Hopkins University Attending, Division o Hospital Medicine
Baltimore, Maryland [76] University o Miami Hospital
Miami, Florida [37]
Ryan M. Ford, MD
Assistant Pro essor o Medicine Richard S. Gitomer, MD, MBA, FACP
Director o Viral Hepatitis Assistant Pro essor
Emory Transplant Center Emory University School o Medicine
Transplant Hepatologist President and Chie Quality O cer
Emory University Hospital Emory Healthcare Network
Atlanta, Georgia [159] Atlanta, Georgia [18]
Vance G. Fowler Jr., MD, MHS Jef rey J. Glasheen, MD, SFHM
Division o In ectious Diseases Chie Quality O cer
Duke University Medical Center University o Colorado Hospital
Durham, North Carolina [189] Associate Dean or Clinical A airs, Quality and Sa ety Education
Director
Gil Freitas, MD Institute or Healthcare Quality, Sa ety and E ciency
Division o Trauma, Burn, and Surgical Critical Care Pro essor
Harvard Medical School Division o General Internal Medicine
Fellow, Metabolic Support Service University o Colorado School o Medicine
Brigham and Women’s Hospital Aurora, Colorado [59]
Dragan Golijanin, MD
Joseph M. Furman, MD, PhD Associate Pro essor o Surgery
Pro essor Director
Departments o Otolaryngology Genitourinary Oncology
Neurology, Bioengineering and Physical Therapy The Warren Alpert Medical School o Brown University
University o Pittsburgh School o Medicine Providence, Rhode Island [180]
Director
Divisions o Balance Disorders Lucas Golub, MD
Pittsburgh, Pennsylvania [84] Emory University School o Medicine
Julia M. Gallagher, MD
Medical Director Steven M. Gorbatkin, MD, PhD
MGH Home Based Palliative Care Program Associate Pro essor
Division o Palliative Care Emory University School o Medicine
Massachusetts General Hospital Nephrologist, Atlanta VA Medical Center
Boston, Massachusetts [74] Decatur, Georgia [241]
xx
Norman D. Grace, MD Caroline N. Harada, MD
Lecturer on Medicine Associate Pro essor o Medicine
Harvard Medical School Assistant Dean or Community-Engaged Scholarship
Pro essor o Medicine University o Alabama School o Medicine
Tu ts University School o Medicine Division o Gerontology, Geriatrics, and Palliative Care
Sta Physician Birmingham, Alabama [166]
Division o Gastroenterology, Hepatology, and Endoscopy
Department o Medicine Nikroo Hashemi, MD, MPH
Brigham and Women’s Hospital Instructor o Medicine
Boston, Massachusetts [53, 160] Harvard Medical School
C
Yonatan H. Grad, MD, PhD Division o Gastroenterology and Hepatology
O
Assistant Pro essor Boston, Massachusetts [109]
N
Harvard T. H. Chan School o Public Health
T
Joaquim M. Havens, MD
R
Division o In ectious Diseases
I
B
Brigham and Women’s Hospital Assistant Pro essor o Surgery
U
T
O
Division o Trauma, Burn, and Surgical Critical Care
R
Charles S. Greenberg, MD Brigham and Women’s Hospital
S
Department o Medicine Boston, Massachusetts [45]
Division o Hematology/Oncology
Medical University o South Carolina Meghan Hayes, MD, FACP
Charleston, South Carolina [54] Department o Internal Medicine
Sutter Medical Group
Stephen B. Greenberg, MD, MACP Clinical Instructor
Distinguished Service Pro essor, Herman Brown Teaching University o Cali ornia
Pro essor Davis Medical Center
Baylor College o Medicine Sacramento, Cali ornia [218, 220]
Vice Chie o Sta , Chie o Medicine
Ben Taub Hospital Catherine P. M. Hayward, MD, PhD, FRCPC
Houston, Texas [198] Pro essor
Pathology and Molecular Medicine/Medicine McMaster University
Norton J. Greenberger, MD, MACP Head, Coagulation
Clinical Pro essor o Medicine Hamilton Regional Laboratory Medicine Program
Harvard Medical School West Hamilton, Ontario, Canada [172]
Senior Physician
Brigham and Women’s Hospital Galen V. Henderson, MD
Boston, Massachusetts [75] Director
Neurocritical Care and Neuroscience Intensive Care Unit
Anne F. Gross, MD Brigham and Women’s Hospital
Assistant Pro essor o Psychiatry Assistant Pro essor, Harvard Medical School
Associate Residency Training Director Boston, Massachusetts [209]
Oregon Health &Science University
Portland, Oregon [229] Kathie L. Hermayer, MD, MS, FACE, FACP
Angela S. Guarda, MD Medical Director or Diabetes Management Services at MUSC
Associate Pro essor o Psychiatry and Behavioral Sciences Division o Endocrinology, Diabetes, and Medical Genetics
Johns Hopkins School o Medicine Medical University o South Carolina
Director Chair, Diabetes Task Force
Eating Disorders Program Ralph H. Johnson Veteran’s A airs Medical Center
The Johns Hopkins Hospital Charleston, South Carolina [52]
Heather Herrington, MD
Navin R. Gupta, MD Associate Pro essor o Medicine
Department o Medicine Division o Gerontology, Geriatrics, and Palliative Care
Renal Division University o Alabama at Birmingham
Brigham and Women’s Hospital Birmingham, Alabama [166]
Stacy Higgins, MD, FACP
Sarah P. Hammond, MD Associate Pro essor o Medicine
Assistant Pro essor o Medicine Division o General Medicine and Geriatrics
Harvard Medical School Emory University School o Medicine
Division o In ectious Diseases Atlanta, Georgia [12]
xxi
Keiki Hinami, MD, MS Jef C. Huf man, MD
Collaborative Research Unit Associate Pro essor o Psychiatry
Cook County Health &Hospitals System Harvard Medical School
Chicago, Illinois [41] Medical Director
Inpatient Psychiatry
Ashley B. Hink, MD, MPH Massachusetts General Hospital
General Surgery Resident Boston, Massachusetts [229]
Department o General Surgery John T. Huggins, MD
Charleston, South Carolina [63] Associate Pro essor o Medicine
Division o Pulmonary and Critical Care
Kerstin Hogg, MD, MBChB, MSc
C
O
Assistant Pro essor Charleston, South Carolina [236]
N
T
McMaster University Daniel P. Hunt, MD
R
I
B
East Hamilton, Ontario, Canada [253] Pro essor o Medicine
U
Director
T
Fernando Holguin, MD, MPH
O
Emory Division o Hospital Medicine
Associate Pro essor o Medicine and Pediatrics
R
S
Asthma Institute Emory University School o Medicine
Division o Pulmonary Atlanta, Georgia [78]
Allergy and Critical Care Medicine
University o Pittsburgh William R. Hunt, MD
Pittsburgh, Pennsylvania [231] Assistant Pro essor o Medicine
Division o Pulmonary, Allergy, Critical Care, and Sleep Medicine
Anthony N. Hollenberg, MD Department o Medicine
Pro essor o Medicine Emory-Children’s Center or Cystic Fibrosis
Harvard Medical School McKelvey Lung Transplant Center
Chie , Division o Endocrinology, Diabetes, and Metabolism Emory University School o Medicine
Beth Israel Deaconess Medical Center Atlanta, Georgia [234]
Aubrey Ingraham, MD
Elizabeth H. Holt, MD, PhD Department o Internal Medicine
Associate Pro essor Kaiser Permanente
Yale School o Medicine Oakland, Cali ornia [39]
Yale Endocrinology
Yale Endocrine Oncology Program Bertrand L. Jaber, MD, MS
New Haven, Connecticut [240] Associate Pro essor o Medicine
Tu ts University School o Medicine
Michael H. Hoskins, MD Vice Chair or Clinical A airs
Assistant Pro essor Department o Medicine
Clinical Cardiac Electrophysiology Caritas St. Elizabeth’s Medical Center
Emory University Hospital Boston, Massachusetts [243]
Atlanta, Georgia [133, 136]
Claire S. Jacobs, MD, PhD
Susy Hota, MD, MSc, FRCPC Department o Neurology
Assistant Pro essor o Medicine Brigham and Women’s Hospital
University o Toronto Boston, Massachusetts [211]
Hospital Epidemiologist and In ectious Diseases Specialist
University Health Network Francine L. Jacobson, MD, MPH
Toronto, Ontario, Canada [200] Thoracic Radiologist at Brigham and Women’s Hospital
Assistant Pro essor
Liangge Hsu, MD Department o Radiology
Assistant Pro essor o Radiology Harvard Medical School
Division o Neuroradiology Boston, Massachusetts [107, 112-119]
Harvard Medical School Shilpa H. Jain, MD
Boston, Massachusetts [118] Clinical Assistant Pro essor (A liated)
Division o Endocrinology, Gerontology, and Metabolism
Margo S. Hudson, MD Stan ord University School o Medicine
Assistant Pro essor o Medicine Veterans A airs Palo Alto Health Care System
Harvard Medical School Palo Alto, Cali ornia [154]
Diabetes Management Service
xxii
Kunal Jajoo, MD Laurence Katznelson, MD
Assistant Pro essor Associate Dean o Graduate Medical Education
Harvard Medical School Pro essor o Neurosurgery and Medicine (Endocrinology and
Associate Physician Metabolism)
Brigham and Women’s Hospital Medical Director
Boston, Massachusetts [158] Pituitary Center
Stan ord University School o Medicine
Edward C. Jauch, MD, MS Stan ord, Cali ornia [154]
Pro essor
Director Clive Kearon, MB, MRCPI, FRCPC, PhD
Division o Emergency Medicine Jack Hirsh Pro essorship in Thromboembolism
C
Pro essor, Department o Neurosciences Department o Medicine
O
Vice Chair, Research, Department o Medicine McMaster University
N
Pro essor, Department o Bioengineering (Adjunct) Juravinski Hospital
T
R
Clemson University Hamilton, Ontario, Canada [253]
I
B
U
Robert T. Keenan, MD, MPH
T
O
R
Brent Jewett, MD Director
S
Clinical Instructor o Surgery Duke Gout and Crystal Arthropathy Clinic
Department o Surgery Duke University School o Medicine
Medical University o South Carolina Durham, North Carolina [247]
Corey D. Kershaw, MD
Danielle Jones, MD, FACP Associate Pro essor o Medicine
Associate Pro essor o Medicine Division o Pulmonary &Critical Care Medicine
Division o General Medicine and Geriatrics University o Texas Southwestern Medical Center
Emory University School o Medicine Medical Director, MICU
Atlanta, Georgia [93] William P. Clements Jr. University Hospital
Dallas, Texas [142]
J. Ryan Jordan, MD
Cardiovascular Disease and Clinical Cardiac Electrophysiology Adeel M. Khan, MD, MPH
South Denver Cardiology Associates, PC Taussig Cancer Institute
Littleton, Colorado [135] Cleveland Clinic Foundation
Cleveland, Ohio [181]
S. Andrew Josephson, MD
Vice President Claude Killu, MD
Neurohospitalist Society Intensive Care
Carmen Castro-Franceschi and Gladyne K. Mitchell Distinguished Los Angeles Medical Center
Neurohospitalist Pro essorship Los Angeles, Cali ornia [125]
Vice Chairman, Parnassus Programs Director, Neurohospitalist
Program Emmanuel S. King, MD, FHM, FACP
Department o Neurology Associate Pro essor o Clinical Medicine
University o Cali ornia, San Francisco Perelman School o Medicine
San Francisco, Cali ornia [104] University o Pennsylvania
Director o Clinical Operations
Brian W. Kaebnick, MD Section o Hospital Medicine
Structural Cardiology Fellow Division o General Internal Medicine
Department o Cardiology Hospital o the University o Pennsylvania
Emory University Hospital Philadelphia, Pennsylvania [21]
Joyce E. King, MD
Stephen P. Kalhorn, MD Assistant Pro essor
Assistant Pro essor Neurosurgery Georgetown University School o Medicine
Medical University o South Carolina Washington, DC
Charleston, South Carolina [64] Clinical Instructor
University o Maryland Medical School
Jameela Kari, MD Director Inpatient Medicine
Pediatric Nephrology Unit Family Medicine Residency
Department o Pediatrics Medstar Franklin Square Medical Center
King Abdulaziz University Baltimore, Maryland [228]
Jeddah, Kingdom o Saudi Arabia [239]
Emad Kishi, MD
Assistant Pro essor o Surgery
Division o Abdominal Transplant
Department o Surgery
xxiii
Joshua P. Klein, MD, PhD Mark S. Lachs, MD, MPH
Associate Pro essor o Neurology and Radiology Irene and Roy Psaty Distinguished Pro essor o Medicine
Harvard Medical School Co-Chie o Geriatrics and Palliative Medicine
Chie , Division o Hospital Neurology Weill Cornell Medicine
Department o Neurology Director o Geriatrics
Brigham and Women’s Hospital New York Presbyterian Health System
Boston, Massachusetts [106] New York, New York [167]
Michael Klompas, MD, MPH Victoria D. Lackey, MD

Associate Pro essor Duke University School o Medicine
Department o Population Medicine Division o Rheumatology and Immunology
C
Harvard Medical School Duke University Medical Center
O
Division o In ectious Diseases Durham, North Carolina [249]
N
T
Joshua R. Lakin, MD
R
I
B
U
Christopher Knudson, MD Dana-Farber Cancer Institute
T
O
Instructor o Medicine Boston, Massachusetts [105]
R
S
Emory University School o Medicine Albert Q. Lam, MD
Atlanta, Georgia [90] Associate Physician
Division o Renal Medicine
Serena Koenig, MD, MPH Brigham and Women’s Hospital
Assistant Pro essor Harvard Medical School
Division o Global Health Equity
Division o In ectious Diseases Lindy H. Landzaat, DO, FAAHPM
Brigham and Women’s Hospital Assistant Pro essor
Boston, Massachusetts [205] Division o Palliative Medicine
University o Kansas Medical Center
Sophia Koo, MD Kansas City, Kansas [217]
Harvard Medical School Vijay H. Lapsia, MD, MBBS
Division o In ectious Diseases Assistant Pro essor o Medicine
Brigham and Women’s Hospital Mount Sinai School o Medicine
Boston, Massachusetts [201] New York, New York [238]
Makeida B. Koyi, MD Lucia Larson, MD, FACP

Deputy Director o Adult Inpatient Consultation Service Associate Pro essor o Medicine
Clinical Instructor Director
Community Psychiatry Program Division o Obstetric Medicine
Department o Psychiatry and Behavioral Sciences Women’s Medicine Collaborative
Johns Hopkins Bayview Medical Center The Alpert Medical School o Brown University
Baltimore, Maryland [81] Providence, Rhode Island [218]
Svetlana Krasnokutsky, MD, MS Jodi Layton, MD

Assistant Pro essor o Medicine Assistant Pro essor o Medicine
Co-Director The Warren Alpert Medical School o Brown University
NYU Crystal Diseases Study Group Hematology and Oncology
New York University School o Medicine Rhode Island Hospital Comprehensive Cancer Center
New York, New York [247] Providence, Rhode Island [180]
Harold Kudler, MD Brian Leber, MDCM, FRCPC

Adjunct Associate Pro essor Pro essor o Medicine
Department o Psychiatry and Behavioral Sciences McMaster University
Duke University Medical Center Attending Physician
Durham, North Carolina [224] Hamilton Health Sciences/Juravinski Hospital and Cancer Centre
Hamilton, Ontario, Canada [170]
Carlos E. Kummer eldt, MD
Sta Pulmonologist Noah Lechtzin, MD, MHS, FCCP
TJ Samson Community Hospital Assistant Pro essor
Glasgow, Kentucky [236] Director
Johns Hopkins Adult Cystic Fibrosis Program
Pulmonary Director
Johns Hopkins Amyotrophic Lateral Sclerosis Clinic
Pulmonary and Critical Care Medicine
Johns Hopkins University
xxiv
Ji Yeon Lee, MD David J. Likosky, MD, SFHM, FAHA, FACP
Pulmonary/Critical Care Fellow President
Emory University School o Medicine Neurohospitalist Society
Division o Pulmonary, Allergy, Critical Care &Sleep Medicine Medical Director
Atlanta, Georgia [235] EvergreenHealth Neuroscience Institute
Clinical Assistant Pro essor
Linda A. Lee, MD University o Washington
Clinical Director EvergreenHealth
Division o Gastroenterology and Hepatology Kirkland, Washington [104]
Johns Hopkins University School o Medicine
Director o Endoscopy Ming Y. Lim, MB BChir
C
Johns Hopkins Hospital Department o Medicine
O
Director Division o Hematology/Oncology
N
Johns Hopkins Integrative Medicine &Digestive Center Medical University o South Carolina
T
R
Lutherville, Maryland [80] Charleston, South Carolina [54]
I
B
U
Linda S. Lee, MD Walter Limehouse, MD, MA, FACEP
T
O
Assistant Pro essor o Medicine Associate Pro essor o Emergency Medicine
R
Harvard Medical School Medical University o South Carolina
S
Director Charleston, South Carolina [33]
Endoscopic Education and Women’s Health in GI
Co-Director Lori-Ann Linkins, MD, MSc (Clin Epi), FRCPC
Pancreas Center Associate Pro essor
Brigham and Women’s Hospital Department o Medicine
Boston, Massachusetts [161] Division o Hematology &Thromboembolism
McMaster University
Blair J. N. Leonard, MD, PhD, FRCP Juravinski Thromboembolism Service
Senior Hematology Fellow MF1 Director/MF1 Hematology Subunit Planner
McMaster University Michael G. DeGroote School o Medicine
Hamilton, Ontario, Canada [170] Thrombosis &Atherosclerosis Research Institute
William I. Levin, MD
Associate Pro essor o Medicine Ra ael H. Llinas, MD
Division o General Internal Medicine Associate Pro essor o Medicine
University o Pittsburgh School o Medicine Associate Pro essor o Neurology
Pittsburgh, Pennsylvania [60] Chairman o Neurology
Johns Hopkins Bayview Medical Center
Katherine Lewis, MD, MSCR Baltimore, Maryland [89]
Assistant Pro essor o Medicine and Pediatrics
The Medical University o South Carolina Hermioni N. Lokko, MD, MPP
Charleston, South Carolina [52] Clinical Fellow o Psychiatry
Cindy Lien, MD Administrative Chie Resident
Assistant Pro essor o Medicine Massachusetts General Hospital
Palliative Care Physician
Internal Medicine Hospitalist Lenny López, MD, MPH, MDiv
Beth Israel Deaconess Medical Center Associate Pro essor o Medicine
Boston, Massachusetts [216] Chie o Hospital Medicine
San Francisco VA Medical Center
Elaine Chiewlin Liew, MD, FRCA University o Cali ornia
Assistant Pro essor o Anesthesiology San Francisco, Cali ornia [3]
Department o Anesthesiology and Perioperative Medicine
David Ge en School o Medicine David J. Lucier, Jr., MD, MBA, MPH
University o Cali ornia, Los Angeles (UCLA) Instructor o Medicine
Ronald Reagan UCLA Medical Center Harvard Medical School
Los Angeles, Cali ornia [153] Director o Quality and Patient Sa ety
Hospital Medicine Group
Assistant in Medicine
Division o General Internal Medicine
Massachusetts General Hospital
xxv
Courtney H. Lyder, ND, ScD(Hon), FAAN Merry Jenni er Markham, MD
Pro essor o Nursing, Geriatric Medicine, and Public Health Associate Pro essor o Medicine
Dean Emeritus Division o Hematology and Oncology
School o Nursing Department o Medicine
University o Cali ornia University o Florida College o Medicine
Los Angeles, Cali ornia [72] Gainesville, Florida [175]
William L. Lyons, MD Alayne D. Markland, DO, MSc

Pro essor Associate Pro essor o Medicine
Division o Geriatrics and Gerontology Division o Gerontology, Geriatrics, and Palliative Care
Department o Internal Medicine University o Alabama at Birmingham
C
University o Nebraska Medical Center Birmingham, Alabama [71]
O
Omaha, Nebraska [165]
N
Greg S. Martin, MD, MSc, FACP, FCCP, FCCM
T
Elizabeth H. Mack, MD, MS Pro essor and Associate Division Director or Critical Care
R
I
B
Pediatric Critical Care Division o Pulmonary, Allergy, and Critical Care
U
Medical University o South Carolina Emory University School o Medicine
T
O
Charleston, South Carolina [19] Director o Research
R
Emory Critical Care Center
S
James H. Maguire, MD, MPH Section Chie , Grady Memorial Hospital
Pro essor o Medicine Atlanta, Georgia [142]
Senior Physician R. Kirk Mathews, MBA
Division o In ectious Diseases Partner, Schmidt
Brigham and Women’s Hospital Mathews LLC
Boston, Massachusetts [205] Providing Executive Search and Leadership
Development Services [24]
Rahul Maheshwari, MD
Gastroenterology Fellow Melissa Mattison, MD, SFHM, FACP
Division o Digestive Diseases Assistant Pro essor o Medicine
Emory University School o Medicine Harvard Medical School
Atlanta, Georgia [159] Chie , Hospital Medicine Unit
Scott Manaker, MD, PhD Boston, Massachusetts [168]
Associate Pro essor o Medicine, Pulmonary, Allergy, and Critical
Care Division Saverio M. Maviglia, MD, MSc
Vice Chair or Regulatory A airs Assistant Pro essor o Medicine
Department o Medicine Harvard Medical School
Perelman School o Medicine, University o Pennsylvania Hospitalist Service
Philadelphia, Pennsylvania [27] Brigham and Women’s Hospital
E rén Manjarrez, MD, SFHM
Chie , Division o Hospital Medicine Laura K. Max, BA
Associate Pro essor o Clinical Medicine Clinical Research Assistant
Miller School o Medicine Johns Hopkins School o Medicine
University o Miami Anesthesiology and Critical Care Medicine
Miami, Florida [37] Johns Hopkins Hospital
Kimberly D. Manning, MD, FACP, FAAP
Associate Pro essor o Medicine Matthew W. McCarthy, MD
Director Assistant Pro essor o Medicine
Distinction in Teaching and Leadership Weill Cornell Medicine
J. Willis Hurst Internal Medicine Residency Program Assistant Attending Physician
Department o Medicine NewYork-Presbyterian Hospital
Division o General Medicine and Geriatrics New York, New York [33]
Atlanta, Georgia [5] Michael McDaniel, MD, FSCAI
Michael Manogue, MD Emory University School o Medicine
Fellow in Cardiovascular Disease Director
Emory University School o Medicine Cardiac Catheterization Lab
Atlanta, Georgia [133, 135] Grady Health Systems
Gary Margolias, MD
Assistant Pro essor o Anesthesiology
xxvi
Timothy B. McDonald, MD, JD Karina Meijer, MD, PhD
Pro essor, Anesthesiology and Pediatrics Division o Haemostasis and Thrombosis
Chie Sa ety and Risk O cer or Health A airs Department o Haematology
University o Illinois University Medical Centre Groningen
Chicago, Illinois [20] Groningen, The Netherlands [173]
Andrew McFarlane, MLT, ART, FCSMLS(D) David Meltzer, MD, PhD

Technical Specialist Fanny L. Pritzker Pro essor o Medicine
Molecular Hematology and Red Cell Disorders Economics and Public Policy Chie
Lecturer McMaster University Section o Hospital Medicine
Department o Medicine Director, Center or Health and The Social Science
C
McMaster University Medical Centre The University o Chicago
O
Hamilton, Ontario, Canada [169] Chicago, Illinois [e2]
N
T
John W. McGillicuddy, MD, FACS Peter A. Merkel, MD, MPH
R
I
B
Associate Pro essor o Surgery Chie , Division o Rheumatology
U
Medical University o South Carolina Pro essor o Medicine and Epidemiology
T
O
Charleston, South Carolina [66] University o Pennsylvania
R
Philadelphia, Pennsylvania [257]
S
Gerard Michael McGorisk, MD, FACC, MRCPI
Assistant Pro essor o Medicine Joseph J. Miaskiewicz, Jr., MD, FCCP, SFHM
Emory University Assistant Clinical Pro essor
Atlanta, Georgia [132] Tu ts Medical School
Chie o Utilization Review and Clinical Documentation
Sylvia C. McKean, MD, SFHM, FACP Hospitalist
Deputy Editor or Editorial Projects, UpToDate North Shore Medical Center
Boston, Massachusetts [107, 112, 116, 123, 124, e4] Salem, Massachusetts [110, 122]
Graham T. McMahon, MD, MMSc Chad S. Miller, MD, FACP, FHM
Adjunct Pro essor Associate Pro essor
Northwestern University Division Director
Feinberg School o Medicine General Internal Medicine
President and Chie Executive O cer Saint Louis University School o Medicine
Accreditation Council or Continuing Medical Education St. Louis, Missouri [92, 94]
Chicago, Illinois [150]
Tracey A. Milligan, MD, MS, FAAN
Julia McNabb-Baltar, MD Assistant Pro essor o Medicine
Instructor o Medicine Harvard Medical School
Harvard Medical School Vice Chair or Education
Center or Pancreatic Disease Department o Neurology
Division o Gastroenterology, Hepatology, and Endoscopy Brigham and Women’s Hospital
Department o Medicine Boston, Massachusetts [211]
Boston, Massachusetts [157] Elinor Mody, MD
Division o Rheumatology
Thomas E. McNalley, MD Brigham and Women’s Hospital
Associate Pro essor Boston, Massachusetts [127]
University o Washington School o Medicine
Department o Rehabilitation Medicine Daniel L. Molloy, Jr., MD
Seattle Children’s Hospital Division o Cardiology
University o Washington Medical Center Emory University
Seattle, Washington [69] Atlanta, Georgia [130]
Jakob I. McSparron, MD Paul A. Monach, MD, PhD

Instructor in Medicine Chie , Rheumatology Section
Harvard Medical School VA Boston Healthcare System
Division o Pulmonary, Critical Care, and Sleep Medicine Associate Pro essor
Beth Israel Deaconess Medical Center Section o Rheumatology
Boston, Massachusetts [85] Boston University School o Medicine
Niharika D. Mehta, MD
Assistant Pro essor o Medicine Carmen Monzon, MD
The Warren Alpert Medical School o Brown University Clinical Assistant Pro essor o Psychiatry and Human Behavior
Director o Ambulatory services The Warren Alpert Medical School o Brown University
Division o Obstetric and Consultative Medicine Women’s Behavioral Medicine, Women’s Medicine Collaborative
Women and In ants Hospital o Rhode Island Miriam Hospital
Providence, Rhode Island [221] Providence, Rhode Island [221]
xxvii
Luis Fernando Mora, MD Jenni er S. Myers, MD, FACP, FHM
The Arrhythmia Center o South Florida Associate Pro essor o Clinical Medicine
Delray Beach, Florida [134] Department o Medicine
Division o General Internal Medicine
CoLette Morgan, MD, FHM, CCDS, CDIP Section o Hospital Medicine
Assistant Pro essor Perelman School o Medicine
Division o Hospital Medicine University o Pennsylvania
Department o Medicine Philadelphia, Pennsylvania [21]
Atlanta, Georgia [32] Satish N. Nadig, MD, PhD, FACS
Assistant Pro essor
Christopher Moriates, MD
C
Surgery, Microbiology, and Immunology
O
Assistant Clinical Pro essor Medical University o South Carolina
N
Division o Hospital Medicine Transplant Surgery
T
University o Cali ornia at San Francisco
R
I
B
Director o Caring Wisely Program
U
UCSF Center or Healthcare Value Amulya Nagarur, MD
T
O
San Francisco, Cali ornia Instructor in Medicine
R
Director o Implementation Initiatives Harvard Medical School
S
Costs o Care, Inc. Hospital Medicine Group
Boston, Massachusetts [2] Massachusetts General Hospital
Ala Moshiri, MD, PhD
Assistant Pro essor o Ophthalmology Peter Najjar, MD
Eye Center Resident
University o Cali ornia, Davis Department o Surgery
Sacramento, Cali ornia [83] Brigham and Women’s Hospital
John E. Moss, MD Boston, Massachusetts [44]
Department o Critical Care Medicine Dale M. Needham, MD, PhD, FCPA
Mayo Clinic Florida Pro essor
Jacksonville, Florida [137] Division o Pulmonary &Critical Care Medicine
Department o Physical Medicine &Rehabilitation
Srinivasan Mukundan, MD, PhD School o Medicine, Johns Hopkins University
Associate Pro essor o Radiology Baltimore, Maryland [81]
Boston, Massachusetts [113] John Nelson, MD, MHM
Overlake Medical Center
L. Silvia Munoz-Price, MD, PhD Nelson Flores Hospital Medicine Consultants
Associate Pro essor o Clinical Medicine Bellevue, Washington [24]
Institute or Health and Society
Medical College o Wisconsin Karin J. Neu eld, MD, MPH
Enterprise Epidemiologist Clinical Director o Psychiatry
Froedtert Health System Johns Hopkins Bayview Medical Center
Milwaukee, Wisconsin [185] Associate Pro essor
Johns Hopkins University School o Medicine
Mandakolathur R. Murali, MD Department o Psychiatry and Behavioral Sciences
Director o Clinical Immunology Laboratory Baltimore, Maryland [81]
Harvard Medical School Tobenna Nwizu, MD
Boston, Massachusetts [230] Taussig Cancer Institute
Cleveland Clinic Foundation
Ernest Murray, MD Cleveland, Ohio [181]
Hospital Medicine Section
General Internal Medicine and Geriatrics Christopher D. Ochoa, MD
Medical University o South Carolina Fellow
Charleston, South Carolina [250] Pulmonary and Critical Care Medicine
Daniel M. Musher, MD Atlanta, Georgia [232]
Distinguished Service Pro essor o Medicine
Pro essor o Molecular Virology and Microbiology Victor M. Orellana, MD
Chie Emeritus, In ectious Disease Section Department o Medicine
Michael E. DeBakey VA Medical Center Rhode Island Hospital
Houston, Texas [186] Providence, Rhode Island [174]
xxviii
Karin Ouchida, MD Nicholas J. Pastis, MD
Assistant Pro essor o Medicine Assistant Pro essor o Medicine
Weill Cornell Medical College Division o Pulmonary and Critical Care
New York Presbyterian Hospital Medical University o South Carolina
New York, New York [167] Charleston, South Carolina [236]
Thomas A. Owens, MD Vihas Patel, MD

Vice President or Medical A airs Instructor
Chie Medical O cer Department o Surgery
Duke University Health System Director
Associate Pro essor o Medicine and Pediatrics Metabolic Support Service
C
Duke University School o Medicine Brigham and Women’s Hospital
O
Durham, North Carolina [189] Division o Trauma, Burn, and Surgical Critical Care
N
T
David A. Oxman, MD, FACP
R
I
B
U
Sidney Kimmel Medical College Timothy J. Patton, DO
T
O
Medical Intensive Care Unit Assistant Pro essor o Dermatology
R
Thomas Je erson University Hospital University o Pittsburgh
S
Philadelphia, Pennsylvania [206] Pittsburgh, Pennsylvania [144]
Menaka Pai, MD, MSc, FRCPC Jill M. Paulson, MD

Associate Pro essor Assistant Pro essor o Endocrinology
Department o Medicine George Washington University School o Medicine.
Associate Member, Department o Pathology and Molecular George Washington Medical Faculty Associates
Medicine Washington, DC [152]
McMaster University
Trans usion Medicine Quality Lead and Consultant Laboratory Allan B. Peetz, MD
Hematologist Instructor
Hamilton Regional Laboratory Medicine Program Department o Surgery
Hamilton, Ontario, Canada [56, 252] Brigham and Women’s Hospital
Division o Trauma, Burn, and Surgical Critical Care
Sumanta K. Pal, MD Harvard Medical School
Assistant Pro essor and Co-Director Boston, Massachusetts [42-45]
Kidney Cancer Program
Department o Medical Oncology &Experimental Therapeutics Vincent D. Pellegrini, Jr., MD
City o Hope Comprehensive Cancer Center John A. Siegling Pro essor and Chair
Duarte, Cali ornia [179] Department o Orthopedics
Robert M. Palmer, MD Charleston, South Carolina [65]
Director
Glennan Center or Geriatrics and Gerontology Jason Persof , MD, SFHM
Eastern Virginia Medical School University o Colorado School o Medicine
Nor olk, Virginia [87] Hospital Medicine Group
Aurora, Colorado [137]
Anand K. Pandurangi, MD, MBBS, DABPN
Pro essor o Psychiatry and Adjunct Pro essor o Radiology Brent G. Petty, MD
Vice Chair, Department o Psychiatry and Chair, Division o The Johns Hopkins Hospital
Inpatient Psychiatry Baltimore, Maryland [9]
Virginia Commonwealth University (VCU) Kurt P ei er, MD, FACP, FHM
Medical Director Pro essor o Medicine
Inpatient Psychiatry General Internal Medicine
VCU Health System Medical College o Wisconsin
Richmond, Virginia [225] Milwaukee, Wisconsin [51]
Jonathan B. Parr, MD, MPH Tania J. Phillips, MD
University o North Carolina School o Medicine Pro essor o Dermatology
Division o In ectious Disease Boston University School o Medicine
Chapel Hill, North Carolina [203] Boston, Massachusetts [148]
Jenni er C. Passini, MD Edward F. Pilkington, III, MD
Clinical Assistant Pro essor Instructor o Medicine
Hospital Medicine Hospitalist Service
University o Wisconsin Department o Medicine Brigham and Women’s Hospital
Madison, Wisconsin [153] Boston, Massachusetts [189]
xxix
Michael H. Pillinger, MD Amir A. Qamar, MD
Pro essor o Medicine and Biochemistry and Molecular Assistant Pro essor o Medicine
Pharmacology Tu ts University School o Medicine
Co-Director Boston, Massachusetts Senior Sta Hepatologist
NYU Crystal Diseases Study Group Lahey Hospital and Medical Center
New York University School o Medicine Burlington, Massachusetts [53]
Rheumatology Section Chie
VA New York Harbor Health Care System, New York Campus Susan Y. Quan, MD [97]
New York, New York [247] Clinical Assistant Pro essor (A liated)
Stan ord University School o Medicine
J. Richard Pittman, Jr., MD, FACP Division o Gastroenterology and Hepatology
C
Assistant Pro essor o Medicine Veterans A airs Palo Alto Healthcare System
O
Division o General Medicine and Geriatrics Palo Alto, Cali ornia
N
T
Timothy R. Quinn, MD, CM
R
I
B
Atlanta, Georgia [11, 101] Medical Director o Dermatopathology
U
Dermpath Diagnostics New England
T
O
Carol Pohlig, RN, BSN, CPC Marlborough, Massachusetts [146]
R
Senior Coding and Education Specialist
S
O ce o Clinical Documentation Talat H. Raja, MD
Department o Medicine Instructor
Hospital o the University o Pennsylvania Hospital Medicine
Philadelphia, Pennsylvania [27] Division o General Internal Medicine and Geriatrics
Timothy J. Poterucha, MD Charleston, South Carolina [82]
Resident Physician and Clinical Fellow
Harvard Medical School Graham W. Redgrave, MD
Department o Medicine Assistant Pro essor o Psychiatry and Behavioral Sciences
Brigham and Women’s Hospital Director or Residency Education
Boston, Massachusetts [108] Johns Hopkins School o Medicine
Assistant Director
Raymond O. Powrie, MD, FRCP, FACP Eating Disorders Program
Pro essor o Obstetrics, Gynecology, and Medicine The Johns Hopkins Hospital
Alpert School o Medicine o Brown University Baltimore, Maryland [226]
Chie Medical Quality O cer
Care New England John J. Reilly, Jr., MD
SVP or Population Health Vice Chancellor or Health A airs
Chie o Medicine Dean or School o Medicine
Women &In ants Hospital o Rhode Island University o Colorado School o Medicine
Providence, Rhode Island [219] Aurora, Colorado [60]
Michaella Maloney Prasad, MD Kerry Reynolds, MD

Assistant Pro essor o Urology and Pediatrics Instructor in Medicine
Medical University o South Carolina Harvard Medical School
Charleston, South Carolina [67] Attending Physician
Massachusetts General Hospital Cancer Center
Alicia Privette, MD Boston, Massachusetts [177, 183]
Assistant Pro essor o Surgery
Medical University o South Carolina Joseph Rhatigan, MD
Department o Surgery Associate Pro essor o Medicine
Charleston, South Carolina [46] Harvard Medical School
Division o Global Health Equity
Alberto Puig, MD, PhD, FACP Brigham and Women’s Hospital
Associate Pro essor o Medicine Boston, Massachusetts [e1]
Director Jessica Rimsans, PharmD, BCPS
Clinician Educator Service Senior Clinical Pharmacist
Massachusetts General Hospital Hemostatic and Antithrombotic Stewardship Pharmacist
Boston, Massachusetts [38] Brigham and Women’s Hospital
Rana C. Pullatt, MD, MS, MRCS, FACS, FASMBS
Diplomate in Obesity Medicine Tina Rizack, MD, MPH
Associate Pro essor o Surgery Assistant Pro essor o Medicine and Obstetrics and Gynecology
Director Robotic Surgery The Warren Alpert Medical School o Brown University
Director Bariatric Surgery VISN-7 Hematology/Oncology
Medical University o South Carolina Program in Women’s Oncology
Charleston, South Carolina [63] Women &In ants Hospital
xxx
Malcolm K. Robinson, MD Bradley T. Rosen, MD, MBA, FHM
Assistant Pro essor o Surgery Medical Director
Harvard Medical School ISP Hospitalist Service
Director Medical Director
Nutrition Support Service Supportive Care Medicine (Palliative Care)
Brigham and Women’s Hospital Director
Boston, Massachusetts [58] Care Transitions and Complex Medical Management
Associate Pro essor, Cedars-Sinai Medical Center
Katina Robison, MD Associate Pro essor, UCLA School o Medicine
Assistant Pro essor o Obstetrics and Gynecology Cedars-Sinai Health System
Program in Women’s Oncology Los Angeles, Cali ornia [120]
C
Department o Obstetrics and Gynecology, Women &
O
In ants’Hospital o Rhode Island Karen Rosene -Montella, MD, FACP
N
The Warren Alpert Medical School o Brown University Senior Vice President
T
R
Providence, Rhode Island [178] Vice Chair o Medicine or Quality/Outcomes
I
B
Division Chie Obstetric Medicine
U
Thomas P. Rocco, MD
T
Pro essor o Medicine and Obstetrics and Gynecology
O
Associate Pro essor o Medicine The Warren Alpert Medical School at Brown University
R
S
Providence, Rhode Island [219, 221, 222]
Clare Rock, MD, MS, MRCPI Assistant Pro essor o Medicine
Department o Medicine Harvard Medical School
Division o In ectious Diseases Hospitalist Service
Johns Hopkins University Brigham and Women’s Hospital
Baltimore, Maryland [195] Boston, Massachusetts [197, 201, 212]
Sarahi Rodríguez-Pérez, MD Stephen R. Rotman, MD
Director o Clinical Operations Gastroenterology Fellow
Assistant Pro essor o Clinical Medicine Brigham and Women’s Hospital
Miller School o Medicine Boston, Massachusetts [156]
University o Miami Hospital Joseph Rudolph, MD
Miami, Florida [37] Department o Neurology
Cleveland Clinic
Vinayak S. Rohan, MD Cleveland, Ohio [210]
Department o Surgery
Medical University o South Carolina Matthew L. Russell, MD, MSc
Charleston, South Carolina [66] Medical Director
Rehabilitation Service Units
Karen L. Roos, MD Hebrew Rehabilitation Center
John and Nancy Nelson Pro essor o Neurology Hebrew Senior Li e
Pro essor o Neurological Surgery Boston, Massachusetts [73]
Indiana University School o Medicine
Indianapolis, Indiana [192] Daniel F. Ruthven, MD
Clinical Associate o Psychiatry and Behavioral Sciences
Alexander E. Ropper, MD Johns Hopkins University School o Medicine
Assistant Pro essor Baltimore, Maryland [226]
Department o Neurosurgery
Baylor College o Medicine Arturo P. Saavedra, MD, PhD
Houston, Texas [208] Assistant Pro essor
Allan H. Ropper, MD, FRCP Medical Director
Pro essor o Neurology Medical Dermatology
Harvard Medical School Massachusetts General Hospital
Executive Vice Chair o Neurology Boston, Massachusetts [145]
Boston, Massachusetts [208, 212] Michel J. Sabbagh, MD
Assistant Pro essor
Anesthesia and Perioperative Medicine
Cheryl A. Sadow, MD
Assistant Pro essor o Radiology
Division o Abdominal Imaging and Intervention
xxxi
Bisan A. Salhi, MD, MA Jef rey L. Schnipper, MD, MPH, FHM
Assistant Pro essor o Emergency Medicine Associate Pro essor o Medicine
Emory University Harvard Medical School
Department o Emergency Medicine Director o Clinical Research, Hospitalist Service
Atlanta, Georgia [121] Brigham and Women’s Hospital
Boston, Massachusetts [40, 151]
John R. Saltzman, MD, FACP, FACG, FASGE, AGAF
Associate Pro essor o Medicine Robert W. Schrier, MD, MACP
Harvard Medical School Pro essor Emeritus o Medicine
Director o Endoscopy University o Colorado School o Medicine
Brigham and Women’s Hospital Division o Renal Diseases and Hypertension
C
Boston, Massachusetts [156, 162] University o Colorado Hospital
O
Aurora, Colorado [242]
N
Kenneth Sands, MD, MPH
T
Associate Pro essor o Medicine Allison R. Schulman, MD
R
I
B
Harvard Medical School Gastroenterology Fellow
U
Senior Vice President, Health Care Quality Brigham and Women’s Hospital
T
O
Beth Israel Deaconess Medical Center Boston, Massachusetts [158]
R
S
David A. Schulman, MD, MPH, FCCP
Milda Saunders, MD, MPH Associate Pro essor o Medicine
Assistant Pro essor Division o Pulmonary, Allergy Critical Care, and Sleep Medicine
Section o Hospital Medicine and MacLean Center or Clinical Emory University School o Medicine
Medical Ethics Atlanta, Georgia [235]
University o Chicago Sam Schulman, MD, PhD
Chicago, Illinois [e3] Thrombosis Service, McMaster Clinic
Hamilton Health Sciences-General Hospital
Marianne E. Savastano, MS, CCC-SLP Hamilton, Ontario, Canada [173]
Speech/Language Pathology Practice Leader
Stroke and Spinal Cord Injury Programs Richard M. Schwartzstein, MD
Spaulding Rehabilitation Hospital Gordon Pro essor o Medicine and Medical Education
Boston, Massachusetts [70] Harvard Medical School
Associate Chie , Division o Pulmonary, Critical Care, and
Paul E. Sax, MD Sleep Medicine
Pro essor o Medicine Beth Israel Deaconess Medical Center
Division o In ectious Diseases Julian L. Sei ter, MD
Boston, Massachusetts [203] Harvard Medical School
Senior Nephrologist and the James Haidas Family Master Clinician
Adam C. Schaf er, MD Brigham and Women’s Hospital
Instructor in Medicine Boston, Massachusetts [61]
Hospital Medicine Unit Samir K. Shah, MD
Brigham and Women’s Hospital Division o Vascular Surgery
Boston, Massachusetts [35, 111] Brigham and Women’s Hospital
Chie Quality O cer and Hospitalist Daniel S. Shapiro, MD
Associate Pro essor o Medicine Pro essor o Internal Medicine
Medical University o South Carolina University o Nevada School o Medicine
Charleston, South Carolina [82, 190] Reno, Nevada [e5]
Lynn Schlanger, MD Ann M. Sheehy, MD, MS

Associate Pro essor Associate Pro essor and Division Head, Hospital Medicine
Emory University School o Medicine Department o Medicine
Atlanta, Georgia [241] University o Wisconsin
Madison, Wisconsin [153]
Robert K. Schneider, MD, FACP
Associate Pro essor Eugenie Shieh, MD
Departments o Psychiatry, Internal Medicine, and Family Medicine Clinical Fellow
Virginia Commonwealth University Johns Hopkins Medicine
Director o Mental Health and Primary Care Integration Division o Gastroenterology
McGuire VA Medical Center Baltimore, Maryland [80]
Richmond, Virginia [227]
xxxii
Deborah M. Siegal, MD, MSc, FRCPC Scot T. Smith, MD
Division o Hematology and Thromboembolism Chie Medical O cer
Department o Medicine Sound Physicians
McMaster University Denver, Colorado [25]
Diana L. Snow, MA, CCS, CPC, CHC
Eric M. Siegal, MD, SFHM Director o Revenue Integrity &Quality
Clinical Associate Pro essor o Medicine (Adjunct) University o Utah Healthcare
University o Wisconsin School o Medicine and Public Health University o Utah School o Medicine
Medical Director Salt Lake City, Utah [31]
Aurora Critical Care Service
David R. Snydman, MD, FACP, FIDSA
C
Aurora St Luke’s Medical Center
O
Milwaukee, Wisconsin [138] Pro essor
N
Tu ts University School o Medicine
T
Mark Siegler, MD Chie , Division o Geographic Medicine and In ectious Diseases
R
I
B
Lindy Bergman Distinguished Service Pro essor Hospital Epidemiologist
U
Pro essor, Departments o Medicine and Surgery Tu ts Medical Center
T
O
Director Boston, Massachusetts [191]
R
MacLean Center or Clinical Medical Ethics
S
University o Chicago Medical Center Society o Hospital Medicine Key Characteristics
Chicago, Illinois [e3] Workgroup [24]
Members o the Society o Hospital Medicine Key Characteristics
Ross D. Silverman, JD, MPH Workgroup are Patrick Cawley, MD, Steven Deitelzweig, MD, Leslie
Pro essor o Health Policy and Management Flores, MHA, Joseph Miller, MS, John Nelson, MD, Scott Rissmiller,
Indiana University Richard M. Fairbanks School o Public Health MD, Laurence Wellikson, MD, and Winthrop Whitcomb, MD
Pro essor o Public Health and Law
Indiana University Robert H. McKinney School o Law Aaron Sodickson, MD, PhD
Indianapolis, Indiana [34] Associate Pro essor o Radiology
Christian T. Sinclair, MD, FAAHPM Section Chie o Emergency Radiology
Assistant Pro essor Medical Director o CT
Division o Palliative Medicine Brigham and Women’s Hospital
University o Kansas Medical Center Boston, Massachusetts [113]
Kansas City, Kansas [217]
Lauge Sokol-Hessner, MD
Ajay K. Singh, MBBS, MBA, FRCP Instructor in Medicine
Renal Division Harvard Medical School
Brigham and Women’s Hospital Associate Director o Inpatient Quality
Boston, Massachusetts [239] Attending Hospitalist
Beth Israel Deaconess Medical Center
Anika T. Singh Boston, Massachusetts [16]
Renal Division
Brigham and Women’s Hospital Margarita Sotelo, MD
Boston, Massachusetts [239] Associate Clinical Pro essor
Divisions o Geriatrics and Hospital Medicine
Mousumi Sircar, MD San Francisco General Hospital
Geriatrics Fellow
San Francisco, Cali ornia [165]
Department o Gerontology Geof rey L. Southmayd, MD
Beth Israel Deaconess Medical Center Instructor o Medicine
Boston, Massachusetts [73] Emory University School o Medicine
Chie Medical Resident, Emory University Hospital
Gerald W. Smetana, MD, FACP J. Willis Hurst Internal Medicine Residency Program
Division o General Medicine and Primary Care Nathan Spell, MD
Beth Israel Deaconess Medical Center Associate Pro essor o Medicine
Boston, Massachusetts [51] Emory University School o Medicine
Chie Quality O cer
Dustin T. Smith, MD Emory University Hospital
Atlanta, Georgia [8, 10]
Kelly Cunningham Sponsler, MD, SFHM
Robert L. Smith, MD Assistant Pro essor o Medicine
Associate Pro essor
Section o Hospital Medicine
Division o Pulmonary and Critical Care
Vanderbilt University Medical Center
New York University School o Medicine
Nashville, Tennessee [14]
Harbor VA Medical Center
New York, New York [141]
xxxiii
Jerry E. Squires, MD, PhD Ashley Stuckey, MD
Department o Pathology and Laboratory Medicine Assistant Pro essor o Obstetrics and Gynecology
Medical University o South Carolina Program in Women’s Oncology
Charleston, South Carolina [57] The Warren Alpert Medical School o Brown University
Department o Obstetrics and Gynecology
Christopher J. Standaert, MD Women &In ants’Hospital o Rhode Island
Clinical Associate Pro essor Providence, Rhode Island [178]
Rehabilitation Medicine, Neurological Surgery, and Orthopedics
and Sports Medicine Prem S. Subramanian, MD, PhD
University o Washington School o Medicine Pro essor o Ophthalmology, Neurology, and Neurosurgery
University o Washington Medicine Sports and Spine Physicians Vice Chair or Academic A airs, Department o Ophthalmology
C
Clinic University o Colorado School o Medicine
O
Harborview Medical Center Aurora, Colorado [83]
N
Seattle, Washington [69]
T
Ram M. Subramanian, MD
R
I
Gerald W. Staton, MD
B
Associate Pro essor o Medicine and Surgery
U
Pro essor o Medicine Hepatology and Critical Care
T
O
Division o Pulmonary and Emory University School o Medicine
R
Critical Care Medicine Atlanta, Georgia [159]
S
Emory University School o Medicine Katelyn W. Sylvester, PharmD, BCPS, CACP
Atlanta, Georgia [232] Pharmacy Manager
Clinical Pharmacy Specialist
Arlene Stecenko, MD Brigham and Women’s Hospital
Associate Pro essor o Pediatrics and Medicine Boston, Massachusetts [254]
Chie , Division o Pulmonary, Allergy and Immunology,
Cystic Fibrosis and Sleep Jef rey A. Tabas, MD
Department o Pediatrics Pro essor o Emergency Medicine
Director Director o Outcomes and Innovations
Emory-Children’s Center or Cystic Fibrosis O ce o Continuing Medical Education
Associate Director University o Cali ornia
Emory-Children’s Center or Cystic Fibrosis and Airways San Francisco School o Medicine
Disease Research San Francisco, Cali ornia [36]
Emory University School o Medicine Jenni er K. Tan, MD
Atlanta, Georgia [234] Associate Physician
Daniel I. Steinberg, MD Northstar Dermatology
Mount Sinai Beth Israel Fort Worth, Texas [149]
New York, New York [7] Todd A. Taylor, MD
Michael Sterling, MD Assistant Pro essor o Emergency Medicine
Assistant Pro essor o Medicine Emory University
Division o Pulmonary, Allergy and Critical Care Medicine Atlanta, Georgia [121]
Emory University School o Medicine Tracy J. Tipton, MD
Atlanta, Georgia [139] Urology Resident
Theodore A. Stern, MD Medical University o South Carolina
Ned H. Cassem Pro essor o Psychiatry in the eld o Charleston, South Carolina [67]
Psychosomatic Medicine/Consultation Catherine Dawson Tobin, MD
Harvard Medical School Assistant Pro essor
Chie , Avery D. Weisman Psychiatry Consultation Service Anesthesia and Perioperative Medicine
Massachusetts General Hospital Medical University o South Carolina
Director Charleston, South Carolina [62]
O ce or Clinical Careers
Massachusetts General Hospital Derrick J. Todd, MD, PhD
Boston, Massachusetts [229] Instructor in Medicine
Melissa B. Stevens, MD Division o Rheumatology
Assistant Pro essor o Medicine Brigham and Women’s Hospital
Division o Hospital Medicine Boston, Massachusetts [246]
Atlanta, Georgia David Tong, MD, MPH
Atlanta VA Medical Center Assistant Pro essor o Hospital Medicine
Decatur, Georgia [12] Department o Medicine
xxxiv
Anne C. Travis, MD, MSc Madeleine Verhovsek, MD, FRCPC
Department o Medicine Assistant Pro essor
Division o Gastroenterology, Hepatology and Endoscopy Division o Hematology and Thromboembolism
Brigham and Women’s Hospital McMaster University
Boston, Massachusetts [162] Consultant Laboratory Hematologist, Red Cell Disorders Laboratory,
Hamilton Regional
Glenn J. Treisman, MD, PhD Laboratory Medicine Program
Eugene Meyer, III Pro essor o Psychiatry and Medicine Hamilton, Ontario, Canada [77, 169]
Departments o Psychiatry and Behavioral Sciences and
Internal Medicine Donald C. Vinh, MD, FRCPC, FACP
Johns Hopkins University School o Medicine Assistant Pro essor
C
Baltimore, Maryland [228] Faculty o Medicine
O
McGill University
N
Elly Trepman, MD Division o In ectious Diseases, Department o Medicine
T
Department o Medical Microbiology
R
McGill University Health Centre
I
B
University o Manitoba Montreal, Quebec, Canada [202]
U
Winnipeg, Manitoba, Canada [147]
T
O
Kittane S. Vishnupriya, MBBS
Geof rey Tsaras, MB, ChB, MPH
R
S
Clinical Assistant Pro essor o Medicine Johns Hopkins University School o Medicine
University o Illinois College o Medicine at Rock ord Baltimore, Maryland [79]
Rock ord, Illinois [194]
Ruth Ann Vleugels, MD, MPH
Jef rey Turner, MD Associate Pro essor o Medicine
Assistant Pro essor o Medicine Harvard Medical School
Section o Nephrology Director
Yale University School o Medicine Autoimmune Skin Disease Program
New Haven, Connecticut [245] Department o Dermatology
Amit Uppal, MD Brigham and Women’s Hospital
Assistant Program Director Boston, Massachusetts [149]
Director o MICU Megan Ann Waldrop, MD
Bellevue Hospital Pediatric Neurology Fellow
Division o Pulmonary, Critical Care, and Sleep Medicine University o Cali ornia, Irvine
New York University Medical Center Children’s Hospital o Orange County
New York, New York [140] Orange, Cali ornia [213]
W. Alexander Vandergri t, III, MD Ruth H. Walker, MB, ChB, PhD
Associate Pro essor Neurosurgery Departments o Neurology
Medical University o South Carolina James J. Peters Veterans A airs Medical Center, Bronx, NY
Charleston, South Carolina [64] Mount Sinai School o Medicine
Joseph Varon, MD, FACP, FCCP, FCCM, FRSM New York, New York [210]
Chie o Critical Care Services Leon Walthall, MD
University General Hospital Hospital Medicine Section
Pro essor General Internal Medicine and Geriatrics
Department o Acute and Continuing Care Medical University o South Carolina
University o Texas Health Science Center at Houston Charleston, South Carolina [250]
Clinical Pro essor o Medicine
University o Texas Medical Branch at Galveston David A. Walton, MD, MPH
Pro essor o Medicine and Surgery UDEM, UNE, UABC, UAT, Division o Global Health
Anahuac, UACH, USON, UPAEP – Mexico [91] Brigham and Women’s Hospital
Boston, Massachusetts [e1]
Alvaro Velasquez, MD
Assistant Pro essor o Medicine John Scott Walton, MD
Division o Critical Care and Respiratory Medicine Associate Pro essor
Emory University School o Medicine Anesthesia and Perioperative Medicine
Atlanta, Georgia [90] Medical University o South Carolina
Nicole F. Velez, MD
Westmoreland Dermatology Associates
University o Pittsburgh Medical Center East
Pittsburgh, Pennsylvania [145]
xxxv
Annabel Kim Wang, MD Tosha B. Wetterneck, MD, MS
Associate Pro essor (Neurology) Department o Medicine
University o Cali ornia, Irvine School o Medicine and Public Health
Orange, Cali ornia Center or Quality and Productivity Improvement
Sta Neurologist University o Wisconsin Madison
VA Long Beach Healthcare System Madison, Wisconsin [41]
Long Beach, Cali ornia [213]
Omar Wever-Pinzon, MD
Sally Wang, MD Assistant Pro essor o Medicine
Instructor University o Utah School o Medicine
Harvard Medical School Department o Medicine, Cardiology Division,
C
Hospitalist Heart Failure Section
O
Brigham and Women’s Hospital University o Utah Health Science Center
N
Boston, Massachusetts [126] Salt Lake City, Utah [129]
T
R
I
Martha C. Ward, MD Christopher M. Whinney, MD, FACP, FHM
B
U
Assistant Pro essor Clinical Assistant Pro essor o Medicine
T
O
Department o Psychiatry and Behavioral Sciences Cleveland Clinic Lerner College o Medicine
R
Department o Medicine Chairman, Department o Hospital Medicine
S
Society Advisor, Osler Society Cleveland Clinic
Assistant Course Director Cleveland, Ohio [4]
Essentials o Patient Care Course
Emory University School o Medicine I. David Wiener, MD
Atlanta, Georgia [223] Pro essor o Medicine and Physiology and Functional Genomics
University o Florida College o Medicine
Theodore E. Warkentin, MD, FRCP(C), FACP, FRCP(Edin) Chie , Nephrology and Hypertension Section
Pro essor Gainesville VA Medical Center
Department o Pathology and Molecular Medicine and Department Gainesville, Florida [238]
o Medicine Michael G. DeGroote School o Medicine
McMaster University Jef rey G. Wiese, MD, FACP, FSM, SFHM
Trans usion Medicine, Hamilton Regional Laboratory Medicine Pro essor o Medicine
Program Tulane University
Service o Clinical Hematology, Associate Chairman, Department o Medicine
Hamilton General Hospital Senior Associate Dean or Graduate Medical Education
Hamilton, Ontario, Canada [171] Director
Tulane Internal Medicine Program
Kathryn Webert, MD, MSc, FRCPC Tulane University Health Sciences Center
Associate Pro essor New Orleans, Louisiana [92, 94]
Department o Pathology and Molecular Medicine
McMaster University B. Robinson Williams, III, MD
Medical Director, Utilization Assistant Pro essor o Medicine
Canadian Blood Services Program Director
Hamilton, Ontario, Canada [172] Cardiovascular Disease Fellowship
Steven Weinberger, MD, FACP Atlanta, Georgia [135]
Executive Vice President and CEO
American College o Physicians Neil H. Winawer, MD, SFHM
Adjunct Pro essor o Medicine Pro essor o Medicine
University o Pennsylvania Emory University School o Medicine
Philadelphia, Pennsylvania Director
Senior Lecturer on Medicine Hospital Medicine Unit
Harvard Medical School Grady Memorial Hospital
Boston, Massachusetts [6] Atlanta, Georgia [230]
Saul N. Weingart, MD, PhD Eric S. Winer, MD

Chie Medical O cer and Senior VP Medical A airs Division o Hematology/Oncology
Tu ts Medical Center Rhode Island Hospital
Pro essor o Medicine Providence, Rhode Island [174]
Tu ts University School o Medicine Kristin R. Wise, MD, FHM
Boston, Massachusetts [15] Assistant Pro essor o Medicine
Stacy Westerman, MD, MPH Hospital Medicine Section, General Internal Medicine and Geriatrics
Fellow, Cardiovascular Disease Medical University o South Carolina
Department o Medicine, Division o Cardiology Charleston, South Carolina [250]
xxxvi
Karl D. Wittnebel, MD, MPH Brian K. Yorkgitis, DO
Medical Director Assistant Pro essor
Pre-Operative Pain Program Department o Surgery
Department o Medicine University o Florida College o Medicine-Jacksonville
Cedars-Sinai Medical Center Jacksonville, Florida [43]
Hospitalist
Division o General Internal Medicine Bishoy Zakhary, MD
Cedars-Sinai Medical Center Pulmonary and Critical Care Fellow
Los Angeles, Cali ornia [120] Department o Pulmonary, Critical Care, and Sleep Medicine
New York University Medical Center
Brian D. Wol e, MD New York, New York [140]
C
O
Hospital Medicine Section Shanta M. Zimmer, MD
N
University o Colorado Denver Associate Pro essor o Medicine
T
University o Pittsburgh
R
University o Colorado Hospital
I
B
Aurora, Colorado [59] Director
U
Internal Medicine Residency Training Program
T
O
Kenneth E. Wood, DO, FCCP University o Pittsburgh Medical Center
R
Chie Medical O cer Pittsburgh, Pennsylvania [88]
S
Geisinger Medical Center
Director Camilla Zimmermann, MD, PhD, FRCPC
Center or Systems Re-engineering in Healthcare, Geisinger Head, Palliative Care Program
Health System University Health Network
Clinical Pro essor o Medicine Director o Research
Temple University School o Medicine Lederman Palliative Care Centre
Geisinger Medical Center Princess Margaret Hospital
Danville, Pennsylvania [153] Associate Pro essor o Medicine
University o Toronto
Rollin Wright, MD Toronto, Ontario, Canada [215]
Division o Geriatric Medicine and Gerontology
University o Pittsburgh
Pittsburgh, Pennsylvania [87]
Irene M. Yeh, MD, MPH

Division o Adult Palliative Care
Dana-Farber Cancer Institute
xxxvii
SECTION REVIEWERS
Joanna M. Bonsall, MD, PhD Don S. Dizon, MD, FACP
Assistant Pro essor o Medicine Associate Pro essor o Medicine
Division o Hospital Medicine Harvard Medical School
Emory University School o Medicine Clinical Co-Director, Gynecologic Oncology
S
E
Atlanta, Georgia Director
C
Part I, Section 3 The Oncology Sexual Health Clinic
T
I
Massachusetts General Hospital Cancer Center
O
Yvette M. Cua, MD
N
Associate Pro essor o Medicine Part VI, Section 7
R
E
V
Associate Vice Chair or Clinical A airs Francine L. Jacobson, MD, MPH
I
E
Department o Medicine Thoracic Radiologist at Brigham and Women’s Hospital
W
University o Louisville Assistant Pro essor
E
R
Louisville, Kentucky Department o Radiology
S
Part I, Section 6 Harvard Medical School
Jatin K. Dave, MD, MPH Part V, Section 2
Part-Time Instructor
Harvard Medical School Tina Rizack, MD, MPH
Division o Aging, Brigham and Women’s Hospital Assistant Pro essor o Medicine and Obstetrics and Gynecology
Boston, Massachusetts The Warren Alpert Medical School o Brown University
Medical Director Hematology/Oncology
Geriatrics and Senior Care Options Program in Women’s Oncology
Tu ts Health Plan Women &In ants Hospital
Watertown, Massachusetts Providence, Rhode Island
Part III Part VI, Section 7
Steven B. Deitelzweig, MD, MMM, SFHM, FACP Karen Rosene -Montella, MD, FACP
Ochsner Health System Senior Vice President
Medical Director o Regional Business Development Women’s Services and Clinical Integration, Li espan
System Chairman, Hospital Medicine Vice Chair o Medicine or Quality/Outcomes
Associate Pro essor o Medicine-Ochsner Clinical School Division Chie Obstetric Medicine
Part I, Section 5 Pro essor o Medicine and Obstetrics and Gynecology
The Warren Alpert Medical School at Brown University
Providence, Rhode Island
Part VI, Section 12
Dustin T. Smith, MD
Atlanta, Georgia
Part I, Section 2
xxxix
FOREWORD
I well remember reading the landmark article by Wachter and important topics in both clinical and nonclinical areas, ranging rom
Goldman entitled “The emerging role o ‘hospitalists’ in the value-based medicine to transplant surgery consultation. The sec-
American health care system,” published in the New England Journal tion on billing, coding, and clinical documentation has been greatly
of Medicine in 1996.1 In this article, the authors recognized the need expanded, as has coverage o a wide host o malignancies. Because
F
or “a new breed o physicians … specialists in inpatient medicine” o the importance o the recovery period and transitions to a variety
O
R
and coined the term “hospitalist” to re er to this new type o physi- o settings a ter hospital discharge, a welcome new section on reha-
E
cian specialist. Since then, the specialty o hospital medicine has bilitation and skilled nursing care has also been added.
W
become an increasingly popular and success ul career pathway, The editors and authors are to be congratulated on again having
O
R
and has expanded beyond its roots in internal medicine to other made a major contribution to the care o hospitalized patients and
D
disciplines, such as pediatrics, amily practice, and obstetrics. The to those physicians, whether or not they ormally identi y them-
Society o Hospital Medicine estimated there were approximately selves as hospitalists, who care or these patients. Given the breadth
44,000 hospitalists in the United States in 2014, and predicted that and the depth o this text, there are ew questions that clinicians
number will continue to grow.2 will not be able to answer or guidance that they will not be able to
When hospital medicine started, the expertise o hospitalists was obtain about how to provide the best care or the wide spectrum o
ocused on the clinical issues surrounding care o the hospitalized their hospitalized patients.
patient. More recently, there has been increasing emphasis on the Steven E. Weinberger, MD, MACP, FRCP
hospitalist’s role in designing and improving the systems o care Executive Vice President and Chie Executive O cer
in the hospital. These added responsibilities have necessitated an American College o Physicians
expansion o the hospitalist’s skills set beyond just a clinical knowl- Adjunct Pro essor o Medicine
edge base to an understanding o such topics as teamwork, transi- Perelman School o Medicine at the University o Pennsylvania
tions o care, quality metrics and improvement, and patient sa ety, Senior Lecturer on Medicine
among many others. A consequence o this proli eration o speci c Harvard Medical School
competencies has been the creation o an optional pathway or
internal medicine hospitalists to maintain their certi cation with the
American Board o Internal Medicine, o cially re erred to as Focused
Practice in Hospital Medicine. REFERENCES
In the rst edition o Principles and Practice of Hospital Medicine,
1. Wachter RM, Goldman L. The emerging role o “hospitalists” in
McKean and her co-editors took on the herculean task o assem-
the American health care system. N Engl J Med. 1996;335:514-517.
bling an outstanding group o contributing authors and putting
together a superb, comprehensive textbook o hospital medicine 2. Bureau o Labor Statistics. http://www.bls.gov/careeroutlook/
that was published in 2012. In this second edition, the editors 2015/youre-a-what/hospitalist.htm . Accessed April 12, 2016.
have not only updated content but have also added a number o
xli
PREFACE
Since its initial publication in 2012, Principles and Practice of Hospital concentrates on what the consulting hospitalist needs to know
Medicine has become established as a leading resource or the when consulting on patients undergoing bariatric surgery, neuro-
specialty o hospital medicine. More than 200 renowned generalists surgery, orthopedic surgery, transplant surgery and urologic proce-
and specialists contributed to make this book comprehensive and dures. All chapters ocus on problems commonly encountered in
P
authoritative, but as practical as possible. Clinical chapters presented the hospital setting, such as assessment and management o the
R
E
questions that commonly arise in the course o practice and empha- diabetic patient, risk assessment and risk reduction or patients with
F
A
sized core concepts with well-illustrated subject matter, radiology, end-stage liver disease, and preoperative assessment o patients
C
clinical images and quick-view decision trees. The scope o content with hematologic disorders.
E
de ned most o the eld o hospital medicine as it existed in 2012, Part III: Rehabilitation and Skilled Nursing Care. This new part,
and the ormat o the text itsel was enhanced both with an online written primarily by experts in rehabilitation medicine, provides
edition available through the widely used AccessMedicine.com, and key in ormation that hospitalists need to consider as they work to
an app version or use on iPad. ensure sa e transitions rom the inpatient setting to extended care
Since the publication o the rst edition, the eld o hospital acilities. Individual chapters address rehabilitation options, physi-
medicine has continued to evolve into areas beyond evidence- cal and occupation therapy, common issues such as bowel and
based general medical care into the practice o co-management bladder incontinence, dysphagia, pressure ulcers, care o surgical
o surgical and medical subspecialties, rehabilitation medicine, and wounds and pressure ulcers. The chapter on patient sa ety and quality
palliative care. Driven by quality improvement e orts, as well as improvement emphasizes core concepts embraced by hospitalists—
reimbursement models such as bundled payments, the last ew the multidisciplinary approach, prevention o complications, and
years have seen an increased emphasis on coordination o care patient-centered communication in the transition o patients to
between acute care hospitals and other settings, including skilled and rom the post-acute setting. The chapter on hospice ocuses on
nursing acilities, rehabilitation acilities, and long-term acute care common issues that clinicians need to address as they shi t toward
acilities. The rapid growth o the eld has been accompanied by an palliative care and consider the best setting or their patients.
emerging cadre o outstanding clinicians and leaders, both at the Part IV: Approach to the Patient at the Bedside. These chap-
local, national, and international level, and this book is the product ters provide detailed guidance or the initial inpatient evaluation,
o their collective e orts. diagnostic testing, and management o patients with common
The second edition o Principles and Practice of Hospital Medicine presenting complaints that may be encountered at the time o
provides tools to address the unique set o challenges hospitalists admission or in the middle o the night. Each disorder is addressed
ace in a healthcare system that ought to be sa er and more e ec- rom the perspective o hospital care, which in many cases di ers
tive. It comprehensively covers topics not included in any other signi cantly rom initial outpatient care or the same complaint.
print or online textbook. For example, this edition has new sections Even experienced clinicians will nd value in reviewing an initial,
and chapters on the value and values o hospital medicine; practi- sometimes algorithmic, approach to common problems such as
cal, specialty in ormation relating to what consulting hospitalists anemia, alls, delirium, dizziness and vertigo, insomnia, numbness,
need to know as they co-manage patients rom other services; key and weakness (how to localize the problem). Many o the chapters
in ormation in rehabilitation and skilled nursing care pertinent to re er to subsequent chapters in Part VI, which covers the diagnosis
patient sa ety and quality; expanded content on the approach to and management o speci c diagnoses.
the patient at the bedside and clinical conditions in the inpatient Part V: Diagnostic Testing and Procedures. E ciency o care,
setting. Using the basic ormat o the rst edition, all content has reduced cost, especially length o stay, coupled with high quality
been updated to incorporate new medical knowledge relevant to begins with clinical problem solving at the time o admission. This
the practice o hospital medicine. part explains how to interpret common admission tests, such as liver
The second edition has six major parts, covering issues o impor- biochemical tests or arterial blood gas reports, and how to avoid
tance to hospitalists everywhere: waste ul, unnecessary medical tests and treatments. The radiology
Part I: The Specialty of Hospital Medicine and Systems of Care. section reviews indications o radiology studies typically ordered in
The authors o this section represent some o the most knowledge- the hospital setting, a general approach to interpretation, patient
able and orward-thinking people in the areas o value based medi- sa ety issues in imaging and procedures per ormed by interven-
cine, critical decision making at the point o care, transitions o care, tional radiologists. A comprehensive textbook in hospital medicine
patient sa ety and quality improvement, practice management, would not be complete without a section on procedures. The
ethics and pro essional development. This part emphasizes the mul- procedures’ section provides some standardization o procedure
tidisciplinary approach, teamwork, prevention o hospital-acquired per ormance, highlights indications, initial assessment, prevention
complications, and patient-centered communication to ensure sa e o complications, and interpretation o results with links to video
and e cient care transitions and hando s. online resources that provide additional instruction, not possible
Part II: Medical Consultation. This part explains the traditional in a text ormat. This section includes the core set o procedures
role o the medical consultant and updates preoperative cardiac most likely to be per ormed or supervised by hospitalists and
and pulmonary risk assessment and risk reduction. Chapters that acknowledges local and regional variations in the role o hospitalists
ref ect the evolving role o hospitalists in co-management o surgi- per orming or supervising these procedures.
cal patients include general principles o surgery and anesthesia, Part VI: Clinical Conditions in the Inpatient Setting. Updated
perioperative pain management, and management o common clinical content across the breadth o hospital medicine includes major
complications in noncardiac surgery. The surgical specialties section disciplines in internal medicine such as cardiology, gastroenterology,
xliii
and in ectious diseases as well as sections with special relevance to hospital medicine has evolved and the skills required o hospitalists
hospital medicine, such as geriatrics, palliative care, psychiatry, toxi- so that they can provide exceptional patient care and clinical care
cology, and addiction. In response to the evolving role o hospitalists leadership. We thank the American College o Physicians or its col-
on oncology inpatient services, the section covering hematology laborative publishing arrangement with McGraw-Hill that included
and oncology has been substantially expanded. input into the editors, contributors, and overall scope or this new
Electronic chapters (available on AccessMedicine.com) cover edition. Through its engagement in this book, the college advances
hospital medicine aspects o global health and hospital medicine, it mission to enhance the quality and e ectiveness o health care.
the core competencies o hospital medicine, the economics o
health care, principles o medical ethics, and bioterrorism. Sylvia C. McKean, MD, FACP, SFHM
In summary, the second edition o Principles and Practice of
Hospital Medicine takes into account how the eld and practice o
P
R
E
F
A
C
E
xliv
ACKNOWLEDGMENTS
The editors o Principles and Practice of Hospital Medicine would like including amily and riends. Finally, we wish to recognize physicians
to acknowledge and thank our publisher McGraw-Hill, speci cally, who took the time out o their busy schedules to review chapters
James Shanahan, publisher; Amanda Fielding, editor; Kim Davis, and/or sections o the book that clearly bene ted rom their valu-
managing editor; Laura Libretti, administrative assistant; Richard able expertise.
A
Ruzycka, production manager; and the numerous people assisting
C
K
them to complete this e ort. We also express our gratitude to the
N
many contributors who worked diligently to create a comprehen- Sylvia C. McKean, MD, FACP, SFHM, FACP
O
sive resource or our readers and all the people who supported us,
W
L
E
D
G
M
E
N
T
S
xlv
PART I
The Specialty of Hospital Medicine
and Systems of Care
The Value and Values of 22 The Role of Information Technology in

SECTION 1 Hospital Medicine Hospital Quality and Safety . . . . . . . . . . . . . . . . . . . . . . . . . . 134
1 The Face of Health Care: Emerging Issues

SECTION 5 Practice Management
for Hospitalists . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2 Value-Based Health Care for Hospitalists . . . . . . . . . . . . . . . 10 23 Building, Growing and Managing a
3 Racial/Ethnic Disparities in Hospital Care . . . . . . . . . . . . . . . 18 Hospitalist Practice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
4 Comanagement of Orthopedic Patients ............... 23 24 Best Practices in Physician Recruitment and Retention . 150
5 Professionalism in Hospital Medicine . . . . . . . . . . . . . . . . . . 29 25 Teamwork in Leadership and Practice-Based

Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
6 Principles of Leadership . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
26 Negotiation and Conflict Resolution . . . . . . . . . . . . . . . . . . 164
SECTION 2 Critical Decision Making at the

Point of Care SECTION 6 Billing, Coding, and Clinical
Documentation
7 Principles of Evidence-Based Medicine
and Quality of Evidence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 27 Professional Coding and Billing Guidelines
for Clinical Documentation . . . . . . . . . . . . . . . . . . . . . . . . . . 173
8 Diagnostic Reasoning and Decision Making . . . . . . . . . . . . 45
28 Consultation, Comanagement, Time-Based,
9 Principles of Evidence-Based Prescribing . . . . . . . . . . . . . . . 56
and Palliative Care Billing . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182
10 Summary Literature: Practice Guidelines
29 Billing for Procedures and Use of Modifiers
and Systematic Reviews . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
in Inpatient Practice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192
11 Practical Considerations of Incorporating
30 Billing in the Teaching Setting and Billing
Evidence into Clinical Practice . . . . . . . . . . . . . . . . . . . . . . . . 70
with Advanced Practice Providers . . . . . . . . . . . . . . . . . . . . 198
31 Hospital-Driven Documentation ..................... 204
SECTION 3 Transitions of Care
32 Taming the ICD-10 Monster ......................... 210
12 Care Transitions into the Hospital:
Health Care Centers, Emergency Department,
SECTION 7 Principles of Medical Ethics
Outside Hospital Transfers . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
and Medical-Legal Concepts
13 Care Transitions within the Hospital: The Hand-Off ..... 84
33 Common Indications for Ethics Consultation .......... 217
14 Care Transitions at Hospital Discharge . . . . . . . . . . . . . . . . . 90
34 Medical-Legal Concepts: Advance Directives
and Surrogate Decision Making . . . . . . . . . . . . . . . . . . . . . . 224
SECTION 4 Patient Safety and Quality
Improvement 35 Medical Malpractice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
15 Principles of Patient Safety: Intentional

Design and Culture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99 SECTION 8 Professional Development
16 Patient-Centered Care . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104 36 Principles of Adult Learning and Continuing
17 Harnessing Data to Make Quality Improvement Medical Education . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243
Decisions: Measurement and Measures . . . . . . . . . . . . . . . 110 37 Cultural Competence ............................... 246
18 Standardization and Reliability . . . . . . . . . . . . . . . . . . . . . . . 113 38 Career Design and Development in Academic
19 Tools to Identify Problems and Reduce Risks . . . . . . . . . . 118 and Community Settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253
20 Preventing and Managing Adverse Patient 39 Mentorship of Peers and Trainees . . . . . . . . . . . . . . . . . . . . 257
Events: Patient Safety and the Hospitalist . . . . . . . . . . . . . 124 40 Research in the Hospital . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 264
21 Principles and Models of Quality Improvement: 41 For Individuals and Practices: Career
Plan-Do-Study-Act . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130 Sustainability and Avoiding Burnout ................. 273
1
SECTION 1
The Value and Values of
Hospital Medicine
3
CHAP TER
1 INTRODUCTION
Hospital medicine is entering its third decade since the term
“hospitalist” was rst created by Wachter and Goldman. There are
now approximately 40,000 hospitalists and the number is likely to
reach 50,000 in the next decade. The specialty has emerged rom
vigorous early debate about whether there was su cient evidence
to justi y the role o hospitalists to a point where hospital medicine
programs are hard-wired and indispensable in the majority o US hos-
pitals. Early leaders and pioneers in hospital medicine were requently
The Face of Health two to three physician programs with heavy call burdens. Current
hospital medicine groups, particularly in larger hospitals, average
Care: Emerging approximately 15 clinicians and have complicated alternating sched-

ules with clear time o . Early hospitalists o ten served in hospital
Issues for Hospitalists leadership roles with no dedicated time. They were o ten the only
source o quality and sa ety leadership across the hospital. Today, hos-
pitalists requently have committed time to serve in quality and sa ety
positions as well as other medical leadership roles. Most hospitals
now have numerous sta that oversee various quality unctions and
Patrick J. Cawley, MD, MHM ably assist these hospitalist leaders. While some aspects have greatly
advanced or hospital medicine, two things have not changed:
• Pressure to care or sick patients requiring hospitalization—In many
hospitals, the severity o illness across the patients cared or by
hospitalists continues to rise. The number o patients covered
by hospitalists grows in most hospitals and it is not unheard o
or all medical inpatients to be cared or by hospitalists.
• Juggling time between clinical care and hospital leadership—
While there are greater numbers o hospitalists and more
hospital sta ocused on quality and sa ety, the need or clini-
cian leadership has never been more necessary. The vacuum
caused by physicians no longer seeking hospital sta privileges
as well as the overall urgent need or a new paradigm in health
care to increase quality and decrease the cost o care has only
resulted in greater need or hospitalists and physician leader-
ship. Hospitalists are quickly expected to assume ormal leader
roles as well as clinical team leaders.
The need or inpatient care will not go away but how we care
or the acutely ill may change. The demand or physician leader-
ship does not diminish, but our ocus on certain issues will vary. So
what are the issues acing hospitalists in the next decade? Based on
care ul study o patients, amilies, hospitalists, hospitals, and health
systems, we can predict the likely emerging issues.
“VALUE”
Similar to history, certain issues in health care predominate dur-
ing a particular era because o timing related to unique discovery,
emerging evidence, or simply greater knowledge by the masses o
a problem. In the late 1990s, quality began its emergence as the
principal concern in health care. This started in the 1980s with early
studies examining inconsistency and errors as well as problems
associated with overuse, misuse, and underuse o health care ser-
vices. The Institute o Medicine released reports about the quality
o health care in the United States in 1994, 1999, and 2001. Each
o these had a slightly di erent ocus, and over the next decade,
improving quality became a common theme in health care. American
hospitals, health systems, and clinicians responded, albeit slowly and
deliberately, but by 2010, no area o medicine had ailed to grasp
the undamentals o the quality movement and make changes
or the better. During this same decade, another concept gained
5
broad consensus—improved quality is usually less expensive. This ve organizing characteristics that need to guide the thinking o
rst began as outcomes that were seen alongside quality improve- people in a high-reliable organization:
ment, but steadily it became clearer and today there is wide spread • Preoccupation with ailure—Everyone in the organization is
evidence that quality and cost are closely interconnected. The best
P
ocused on errors and near-misses in order to learn rom them
A
quality is o ten the least expensive. As the linkage grew, the ocus and gure out how to prevent recurrence. Finding and xing
R
on value in clinical care emerged to a signi cant degree among the problems is everyone’s responsibility. Leaders support and
T
large numbers o leaders in health care. The value equation (value encourage this approach.
I
= quality/cost) in health care is certainly not new. One can go back • Reluctance to simpli y interpretations—Everyone in the organi-
decades and read numerous re erences in the health care literature. zation is driven to ask why something happens and do not rely
There have been many proposed versions o the value equation on the rst or easiest explanation.
depending largely on how one de nes quality. The early concept • Sensitivity to operations—Everyone in the organization is ocused
T
h
o value was mainly in re erence to the health care system and the on ways their work processes might break down and are then
e
overall approach to strategically improving health care. Today, a ter encouraged to share potential ailures and create best practices.
S
p
almost two decades o quality improvement along with the more Situational awareness is part o the organizational culture.
e
recently acceptance o cost containment as an independent qual-
c
• Commitment to resilience—Everyone in the organization works
i
a
ity measure, the concept o value is rapidly approaching majority
l
quickly to contain errors that do occur in order to minimize
t
y
acceptance by physicians and health care leaders. The next decade potential harm. Additionally, errors do not disable the organi-
o
will see value broadly and deeply pushed into the health care sys-
f
zation. The organization responds robustly and looks or new
H
tem. Forward thinking hospitalists and hospital medicine groups will solutions to prevent catastrophes.
o
s
be ready or the emergence o value as an outcome by which they De erence to expertise—Everyone in the organization listens to
p
•
i
will be measured.
t
people who have the most developed knowledge o the task
a
l
at hand and empowers them to make decisions in order to
M
quickly mitigate harm. Sometimes, those individuals might not
e
PRACTICE POINT
d
have the most seniority, but they are still encouraged to voice
i
c
•
i
Hospitalists, by virtue o their unique role in the health care their concerns, ideas, and input—regardless o hierarchy.
n
e
system, will be expected to embrace the concept o value These organizing processes are essential to the development o
a
(value = quality/cost) and drive its use into everyday clinical
n
a highly reliable organization and all leaders and physicians in an
d
practice. Quality is de ned as sa e, e ective, e cient, equitable, organization attempting to achieve high reliability should under-
S
y
patient centered, and timely care. Cost is de ned as the unit stand the theory and application o each characteristic.
s
t
cost o care delivery by the hospitalists. Each o these should be
e
m
de ned, measured, and incrementally improved by the hospital
s
medicine team. PRACTICE POINT
o
f
•
C
Hospitalists are key to the development o an HRO because
a
o their central role in hospital operations. Hospitals on the
r
e
HIGH RELIABILITY
HRO journey will rely heavily on hospitalists to success ully
As quality and sa ety improvement has been embraced by health navigate the process. Hospitalists should study the science o
care, it has become abundantly clear that more radical trans orma- high reliability as it holds great promise to trans orm hospital
tion is necessary in order to accomplish a greater magnitude o medicine and health care in general.
error reduction and quality consistency. This need or trans orma-
tion led health care to an examination o how other industries
such as nuclear power and aviation have achieved a higher degree PATIENT-CENTERED CARE/CONSUMERISM
o reliability resulting in sa ety improvement and avoidance o
The word “consumer” turns o many physicians, but the movement
potentially catastrophic events. Subsequent research revealed
cannot be underestimated in how it is changing health care. The
that success ul organizations in high-risk industries achieve high
most strident opponents to the word requently become some o
reliability by maintaining a cultural mind ulness that allows them
the most vocal advocates or a more patient-centric approach when
to continually reinvent themselves in the ace o highly complex
they become users o the health care system.
environments.
The history o patient-centered care can be directly traced to
As high-reliability organizations (HRO) were studied, it became
the civil and human rights activism in the 1960s. In 1978, Angelica
clear that there are common challenges across organizations pursu-
Thieriot started the Planetree organization a ter a series o upset-
ing high reliability:
ting personal health care experiences. The organization “vowed to
• Hypercomplex environment reclaim or patients the holistic, patient-centered ocus that medi-
• Tight coupling teams where members depend on tasks per- cine had lost.” Harvey Picker ounded the Picker Institute in 1986 or
ormed across their team similar reasons and then teamed up with the Commonwealth Fund
• Extreme hierarchical di erentiation to research a patient-centric approach. The term “patient centered”
• Multiple decision makers in a complex communication was coined and then later changed to “patient and amily centered”
network by the Institute or Family Centered Care which was ounded in
• Need or requent, immediate eedback 1992, subsequently becoming the Institute or Patient and Family
• Compressed time constraints Centered Care in 2010. Through the Picker/Commonwealth Fund
Many o these challenges exist in hospitals, and as a result the research, we began to see di erences in what patients noted as
concept o high-reliability organizations in health care has been important versus physicians. Surveying patients about patient expe-
broadly embraced as potentially the trans orming approach needed rience with the health care system started. This led to the Institute
to vastly improve quality and sa ety. o Medicine emphasizing the concept o patient-centered care in its
As health care incorporates the methodology o HRO, it is impor- reports on quality care in the United States. Patient-centered care
tant to study the work o Weick and Sutcli e who have identi ed was de ned as providing care that is respect ul o and responsive to
6
individual patient pre erences, needs, and values and ensuring that improve in ormation trans er as patients move between providers.
patient values guide all clinical decisions. However, EHR issues such as how in ormation is displayed as well as
C
In addition to patient-centered care, two additional notions are the lack o linkage between di erent EHR systems limit the e ective-
H
important to understand. Both are subsets o patient-centered care, ness o EHRs in ully improving the knowledge “voltage drop.”
A
which as an expression has come to mean a lot o di erent things, In the last several years, the ocus on thirty-day hospital readmis-
P
and it is important to di erentiate. sion rates has become a signi cant driver o the emphasis on care
T
E
Patient experience is the sum o all interactions, shaped by an transitions. This is largely driven by the Centers or Medicare and
R
organization’s culture, that inf uence patient perceptions across a Medicaid penalties to hospitals with poor per ormance e ective
1
continuum o care. It is measured by patient surveys. For hospitalists, October 1, 2012. Readmission rates are seen as a potential outcome
the most requent measurement method is the HCAHPS (Hospital measure or poor care transitioning. There is increasing evidence
Consumer Assessment o Healthcare Providers and Systems) survey that socioeconomic actors play a large role in determining readmis-
T
h
which is used or adult inpatients and measures patients percep- sion rates, so one should be care in completely linking readmission
e
tions about communication with nurses, communication about rates to poor transition e orts by hospitals and hospitalists. As the
F
a
medicines, communication with doctors, pain management, cleanli- link to socioeconomic actors is urther understood, this will likely
c
e
ness and quietness o the hospital environment, responsiveness o require hospitals and health systems to work closely with commu-
o
hospital sta , discharge in ormation, and the overall rating o the nity programs and governments to address these actors i more
H
hospital. Patient experience and satis action surveys are common trans ormative improvement is to be made in readmission rates.
e
a
to all hospitals and are important as a method o understanding the
l
t
patient viewpoint. They are used as quality measure to be continu-
h
PRACTICE POINT
C
ously improved, but they are only one element o patient-centered
•
a
care and should be understood as that. Health care will remain ragmented, but the advent o
r
e
electronic health records and ormal transition o care programs
:
Patient engagement are actions taken by individuals to obtain
E
o er potential or lessening the dangers to patients during this
m
the greatest bene t rom the health care services available to
time. All hospitalists and hospital medicine groups should have
e
them. Patient engagement occurs when patients eel empowered
r
g
to move to a state o active participation and sel -e cacy in manag- an active transition o care program.
i
n
g
ing their health. This does not mean just obeying directives rom
I
health care providers, but rather moving to a higher plain o involve-
s
s
u
ment, interest, and sel -awareness. TELEHEALTH/TELEMEDICINE
e
s
Why is patient engagement so important? Engaged patients have
Telehealth is a broad term used to re er to the process o providing
o
better health outcomes, incur less costs, and enjoy greatest value
r
(quality/cost) rom health care system. A patient can conceivably health care services at a distance. It is increasingly per ormed via
H
synchronous or asynchronous video connection. Telemedicine is a
o
be satis ed with their health care experience while having minimal
s
component o telehealth and re ers to the delivery o clinical services
p
engagement.
i
to patients in other locations. Examples o telemedicine include
t
a
In order to optimize patient engagement, there are key patient
l
video consultations with physicians, digital transmission o medical
i
s
concepts to understand and incorporate into the health care experi-
t
imaging, and remote monitoring o intensive care unit patients.
s
ence such as literacy level, readiness to learn, readiness to change,
learning style, and patient activation. The delivery o clinical care via video is rapidly improving. Early
issues such as network availability, capability, and delivery cost
continue to progress on an annual basis. Open, secure so tware
PRACTICE POINT plat orms are replacing closed proprietary systems. This brings in an
era o ease o use as well as signi cant cost decrease. Patients are
• Hospitalists should embrace the concepts o patient-centered increasingly accepting telehealth visits. The growing use o smart-
care, patient satis action, and patient engagement. Each o phones and tablets by patients and amilies present a potential plat-
these concepts overlaps, yet di ers in meaning ul ways and orm or the provision o enhanced communication and clinical care.
should be used in the right context. Regulators o the practice o medicine such as state boards o
medicine have been slow to ully accept and grant unrestricted
licensure, but public demands or telehealth are likely to pressure
TRANSITIONS OF CARE regulators to relax constraints. Similarly, 2015 was a landmark year
Since the dawn o hospital medicine, there has been a ocus on or telemedicine in that over hal o the states now require health
care transitions, which are the movement patients make between plans to cover and reimburse or telemedicine in the same manner
health care practitioners during the course o a chronic or acute ill- and at the same rate as in-person services. Telehealth parity in the
ness. Early in the hospital medicine movement, this time period was remaining states or via ederal mandate is only a matter o time.
o ten re erred to as the “voltage drop.” Because o the discontinuity Provider acceptance has been limited at rst and mainly related
introduced when a patient’s outpatient physician is di erent than to early adopters and innovators trialing new methods and systems.
their inpatient physician, the importance o care transition in terms As telehealth visits have increased, the need or more dedicated
o patient quality and sa ety is paramount. One o the pro essional provider time and systems to schedule, record, and ollow-up visits
organizations o hospital medicine, the Society o Hospital Medicine, has increased.
was an early adopter in promulgating the study o transitions and
the determination o best practices by hospitalists and hospitals. PRACTICE POINT
Many potential models have been developed to improve hospital
• Telehealth is rapidly growing and likely to impact the practice
to home transitions including BOOST, Care Transition Intervention, o hospital medicine over the next decade in many potential
Project Red, and the Naylor Model. Health care payers are developing ways. Examples include the care o patients by hospitalists
nancial incentives or hospitalists as well primary care physicians to remotely, enhanced availability o specialists in many hospitals,
ocus on care transitions. The wider use o electronic health records and tele-critical care monitoring.
(EHR) by physicians and hospitals holds great promise as a tactic to
7
TRANSPARENCY relevance. In addition, genetic tests are now directly marketing
to the consumer who comes to the physician seeking additional
There are multiple major approaches to improving quality and
knowledge.
sa ety in health including adherence to a quality improvement
The combination o genomic technologies and the electronic
P
methodology, outcomes transparency or public reporting, nancial
A
health record holds great promise or medical care to become more
incentives, regulation or accreditation, and competition. Outcomes
R
precise and individualized, however many record systems are o ten
transparency is the one which most physicians and organizations
T
lacking core data to optimize this connection. To ully permeate
have the least expertise, but it is very high on the list o importance
I
these technologies into medical care, health care leaders and physi-
or patients and amilies. There is not a long history o transparency
cians must ensure that electronic health records are built to obtain
in health care and one o the earliest was the publication o hospital
appropriate in ormation, such as demographics, medications, diag-
mortality rates and outliers by the Healthcare Financing Administra-
nosis, vital signs, and other in ormation.
tion (HCFA) in the 1980s. This resulted in backlash rom hospital
T
h
Pioneers o learning health care systems have established stream-
executives who when surveyed overwhelmingly ound the data
e
lined consenting processes and data warehouses. However, systems
S
o little use. Since then, there has been persistent and increasing
p
are not consistently incentivized to build these capacities. Health
outcomes transparency including severity adjusted mortality rates
e
c
leaders want evidence be ore investing resources, yet without those
or coronary artery bypass gra ts survey in New York and Pennsyl-
i
a
investments such evidence is hard to gather.
l
vania in the late 1980s and early 1990s, and National Committee
t
y
or Quality Assurance reporting in 1993 o managed care plans and
o
f
Healthcare E ectiveness Data and In ormation Set (HEDIS) indica- PRACTICE POINT
H
o
tors. In 2002, the Joint Commission began reporting outcomes and
s
• Hospitalists should prepare or the coming clinical relevance in
p
the National Quality Foundation developed the serious reportable
i
t
events classi cation. By the mid-2000, many companies such as genetic testing, particularly in the patient’s individual potential
a
l
HealthGrades and Consumer Report as well as multiple states were response to drug therapy.
M
• Hospitalist leaders should work with their hospital in ormation
e
releasing a variety o health care outcome reports on a regular basis.
d
Price transparency in health care also began to rise. In 2006, Presi- technology team to ensure that the hospital electronic health
i
c
i
record is optimized to link genomic and patient in ormation.
n
dent Bush signed a Transparency Executive Order, which urther
e
opened avenues or pricing and quality transparency. In 2015, the
a
n
Robert Wood Johnson Foundation held its second transparency
d
summit with the ollowing conclusions:
S
TEAMWORK TRANSFORMATION
y
s
1. Transparency may not be ubiquitous, but it is now a perma- The acceptance and rapid growth o hospitalists, based on personal
t
e
nent eature o the health care landscape.
m
observation, in many hospitals and health systems is tied to the
2. For all the progress made in transparency, there is much more
s
team approach by early hospitalists. These hospitalists were avail-
o
work to be done. able to hospital personnel to a great degree than nonhospitalists
f
C
3. There is a paradox: The more transparency we have in health care, and this o ten resulted in close relationships and greater teamwork.
a
the more we expose how little we actually know or understand.
r
Studies o nursing satis action showed higher results in hospitals
e
Health care transparency is part o a larger transparency trend with hospitalists. Today, many hospitalist leaders have typically been
due to the rise in global market economics and its resultant demand trained in the mechanics o teamwork and advocate or greater
or data, an increase in communication technology which acilities interaction and voice across the hospital medicine team.
transparency, and an “internet culture” which demands a more The ocus on value in the coming decade will orce a reexamina-
open interaction. Health care transparency is clearly growing and tion o the hospital medicine team. The typical hospitalist is likely
not a ad. The OpenNotes project, in which patients are able to see to assume the care o much greater numbers o patients on a daily
their doctors’visit notes, ound that patients become more engaged basis and this will only be achieved through a trans ormation o the
with such access. This is likely to become routine in the near uture. hospital medicine team. To alter the present system will require rst
While transparency in health care is largely viewed positively, or and oremost a willingness o both hospitalists and hospital leaders
patients to achieve optimal engagement rom transparency will to modi y it. This is not an easy task as signi cant change manage-
require better design principles into public reports. ment skills are required as well as a commitment to a higher level
o team reliance, communication, and trust. There will need to be a
diligent ocus on patient satis action, each team member’s satis ac-
PRACTICE POINT tion, and clinical outcomes. It can be achieved aster i there is payer
• Hospitalist and hospital medicine group should prepare or f exibility in pro essional and hospital reimbursement.
greater transparency and expect probable ull access by
patients and amilies in the uture to the medical record as well PRACTICE POINT
as quality and pricing data.
• Hospitalists should continue to ocus on teamwork and
• We should not underestimate the patients and amilies ability continue to advocate or an interpro essional multidisciplinary
to assimilate the in ormation in public transparency reports! approach. Forward thinking hospitalists should trial the use o
nonphysician providers, such as nurses, pharmacists, advanced
practice providers, nutritionists, and others in innovative ways
GENOMICS/GENETICS in order to achieve greater value to the patients.
The past two decades has birthed a dizzying array o genomic tech-
nologies that ranges rom rapid genome sequencing to the ability
to evaluate genomes across large populations to the discovery o HOSPITALIST/PHYSICIAN STRESS
genetic basis o certain diseases as well drug responsiveness. The The health care system is undergoing tremendous change and
combination o these genomic technologies is resulting in new because o the central role in the health care system physicians will
types o genetic testing which will increasingly result in clinical experience signi cant change. In addition, there are pro essional
8
reimbursement challenges, increased scrutiny in terms o qual- PRACTICE POINT
ity outcomes, cultural transparency, and increasing demands or
• Hospitalists and hospital medicine groups should ocus on
C
physician time and presence. Incorporating new electronic health
H
ensuring high unctioning leadership and leadership succession
records into the daily work o physicians requires retraining and
A
planning within the hospital medicine group at all times.
optimization. Medical knowledge and techniques advance requiring
P
Hospital medicine groups can enhance their e ectiveness and
constant relearning. Administrative burdens related to reimburse-
T
should implement high-per ormance strategy.
E
ment continue to rise. There are simply many pressure points on the
R
individual physician. For hospitalists, who interact with patients and
1
amilies at a very stress ul time—acute illness requiring hospitaliza- SUGGESTED READINGS
tion, this can lead to tense and demanding interactions and com-
munications. On the leadership side, there is pressure or hospitals Berwick DM, Wald DL. Hospital leaders’ opinions o the HCFA mortal-
T
h
to deliver greater value and all hospital leaders, including hospital- ity data. JAMA. 1990;263(2):247-249.
e
ists, will experience great demands to develop and lead innovations.
F
Cawley PJ, Deitelzweig S, Flores L, et al. The key principles and char-
a
The list o stress points seems endless and overwhelming on the
c
acteristics o an e ective hospital medicine group: an assessment
e
individual hospitalist.
guide or hospitals and hospitalists. J Hosp Med. 2014;9:123-128.
o
However, the ability to positively impact patients and amilies has
Coleman EA. Falling through the cracks: challenges and opportuni-
H
never been greater in the history o medicine. Hospitalists should
e
ties or improving transitional care or persons with continuous
a
recognize that despite the numerous stressors as well as ambiguity
l
complex care needs. J Am Geriatr Soc. 2003;51(4):549-555.
t
o the uture, the ability to help patients and amilies is tremen-
h
C
dously satis ying and pro essionally ul lling. Hansen LO, Greenwald JL, Budnitz T, et al. Project BOOST: e ective-
a
Pro essional dissatis action and burnout is treatable i recognized ness o a multihospital e ort to reduce rehospitalization. J Hosp
r
e
:
and managed appropriately. There are personal and organizational Med. 2013;8:421-427.
E
m
approaches to prevent physician distress as well as optimize physi- Institute o Medicine. America’s Health in Transition: Protecting and
e
cal, emotional, and psychological wellness. Hospitalists and particu- Improving Quality. Washington, DC: National Academies Press; 1994.
r
g
larly hospital medicine leaders should ully understand the signs and
i
n
Institute o Medicine. Crossing the Quality Chasm: A New Health
g
symptoms o distress and have options or re erral.
System or the 21st Century. Washington, DC: National Academy
I
s
s
Press; 2001.
u
e
PRACTICE POINT Institute o Medicine. To Err Is Human: Building a Sa er Health System.
s
Washington, DC: National Academy Press; 1999.
o
• All hospitalists should be aware o high burnout and
r
Jack BW, Chetty VK, Anthony D, et al. A reengineered hospital dis-
H
pro essional dissatis action potential and understand the
o
signs and symptoms in order to appropriately intervene to charge program to decrease rehospitalization: a randomized trial.
s
p
treat physician distress. Hospital medicine group leaders need Ann Intern Med. 2009;150(3):178-187.
i
t
a
to activity manage high physician stress levels and institute Naylor MD, Aiken H, Kurtzman ET, Olds DM, Hirschman KB. The
l
i
s
physician wellness programs.
t
importance o transitional care in achieving health re orm. Health
s
Af (Millwood). 2011;30(4):746-754.
Roberts KH, Rousseau DM. Research in nearly ailure- ree, high-
LEADERSHIP
reliability organizations: having the bubble. IEEE Trans Eng
The need or hospital medicine leadership has never been greater. Manage. 1989;36(2):132-139.
It was there at the dawn o hospital medicine and it will not wane
Wachter RM, Goldman L. The emerging role o “hospitalists” in the
in the coming years. The demand spans rom clinical team leader-
American health care system. N Engl J Med. 1996;335:514-517.
ship to program leadership to hospital and health system senior
executive levels. I you have been a user o the health care system, Weick KE, Sutcli e KM. Managing the Unexpected: Assuring High Per or-
you understand the need immediately. We can do great things in mance in an Age o Complexity. San Francisco, CA: Jossey-Bass; 2001.
medicine, but it can be undone by some o the most trivial issues. Wol JA, Niederhauser V, Marshburn D, LaVela SL. De ning patient
To recti y this, particularly in the hospital setting, requires ocus and experience. Patient Exp J. 2014;1:7-19.
leadership as well as the right emotional commitment by hospital-
ists. William Welch, the rst dean o the Johns Hopkins School o ONLINE RESOURCES
Medicine, said it eloquently in his re erence to a new approach to
Center or Advancing Health. 2010. A New De nition o Patient
medical training—”I can think o but one motive which might in uence
Engagement: What Is Engagement and Why Is It Important.
you to come here with us and that is …our ideals and our uture…This
http :/ / www.c ah.org / p d s/ CFAH_En g ag e m e nt_Be h avior_
will not appeal to the great mass o the public, not even to the medical
Framework_current.pd . Accessed December 19, 2015.
public, or a considerable time. What we shall consider success, the mass
o doctors will not consider a success.” Weiss A, Dantzler S. 2015. Three Key Lessons rom the Healthcare
As hospitalists, we should keep the patient in our ocus at all times Transparency Summit. Robert Wood Johnson Foundation. http://
and continue to lead the trans ormation in quality and sa ety across www.rwj .org/ en/ culture-o -health/ 2015/ 04/ 3_key_lessons_
health care. romt.html. Accessed December 19, 2015.
9
CHAP TER
2 INTRODUCTION
The eld o hospital medicine was built on the premise that hospi-
talists would promote and deliver more e cient, sa er, and higher-
quality inpatient care. Indeed, over the past decade hospitalist
care has led to shorter lengths o stay and relatively lower hospital
costs. However, as national health care costs have continued to rise
unabated, on track to consume approximately 20% o the United
States gross domestic product by 2020, the government, payers,
and the public have all ocused renewed e orts on improving
Value-Based Health health care value—commonly de ned as
Care for Hospitalists Value =

Quality of care (including outcomes and patient experience)
Costs
Hospital costs represent the single largest segment o the nearly

$3 trillion annual US health care expenditure. Thus, hospitalists are
Christopher Moriates, MD vital to any e ort to rein in health care costs. This chapter reviews
concepts and strategies critical or hospitalists to understand in the
emerging world o value-based health care.
HOSPITAL COSTS IN THE NATIONAL SPOTLIGHT

In February 2013, Time magazine published an expose on health
care costs, “Bitter Pill: Why Medical Bills Are Killing Us,” which was
trumpeted across popular media and helped the hospital “charge-
master” become nearly a household term. The chargemaster (also
known as the charge description master or “CDM”) is the list o prices
or the tens o thousands o billable items at a given hospital. Shortly
ollowing the Time article, the Centers or Medicare and Medicaid
Services (CMS) publicly released a database o how hospitals billed
Medicare or the 100 most common inpatient procedures, reveal-
ing in stark relie the ba ing amount o variation in charges and
reimbursements or the same procedures between similar hospitals.
Later that same year, the New York Times published a ront-page arti-
cle with the headline, “As Hospital Prices Soar, a Stitch Tops $500,”
continuing to shine a bright national spotlight on the issue o hospi-
tal costs. As the “Bitter Pill” and the “$500 stitch” highlighted, charges
ound on hospital bills usually appear arbitrary and grossly inf ated.
Despite the pressures to increase transparency, health care
costs have largely remained hidden rom the public and medical
pro essionals. As a result, hospitalists are generally not aware o the
costs associated with their care. In addition, most clinicians nd the
concepts o “charge,” “price,” “cost,” and “reimbursement” con us-
ing (Table 2-1). In most medical centers, the majority o health
care transactions occur between the organization and large payer
organizations, such as insurance companies or Medicare. The price
or charge re ers to the amount reported on the bill to each o these
payers. The cost depends on perspective; providers, payers, and
patients each evaluate costs di erently:
• To providers, costs are the expense incurred to deliver health
care services to patients.
• To payers, costs are the amount payable to the provider or
services rendered.
• To patients, costs are the amount payable out-o -pocket or
health care services.
The chargemaster is theoretically meant to relate to both costs
and payments, but since there is tremendous inexplicable variation in
prices between similar organizations and because the prices are highly
inf ated, the chargemaster routinely ails at this unction. Instead, the
10
without prices, it is easy to unwittingly order the let mignon every
TABLE 2-1 Costs, Charges, Reimbursements, and time. So, why not just put the prices back on the menu?
C
Prices in Health Care Initial studies on this strategy showed mixed results, and the
H
conventional wisdom evolved that displaying price in ormation had
A
Term Definition
limited e ect, with prices o ten becoming “white noise” and being
P
Cost Account o the true cost o providing health
quickly disregarded. More recent studies, however, including a con-
T
services, de ined rom a speci ic stakeholder
E
perspective (provider, patient, payer, society) trolled trial at Johns Hopkins suggest that, perhaps due to the recent
R
global attention to the importance o health care costs, clinicians are
Charge Amount asked by a provider or delivering a
2
now more likely to react to price in ormation. Displaying the Medicare
service (typically more than reimbursement,
used as a starting point or negotiation) Allowable Rates o lab tests to hospital physicians led to substantial
decreases in certain higher-cost lab tests and resulted in a more than
V
Reimbursement Amount paid to the provider or delivering
$400,000 net cost reduction over the course o a 6-month interven-
a
a service
l
u
tion period. It is not clear i this e ect too will abate over time.
e
Price Amount paid by the patient or receiving a
-
Similarly, a study using dollar signs ($-$$$$) to translate relative costs
B
service
a
o antibiotics on culture and susceptibility testing reports resulted in a
s
e
signi cant decrease in prescriptions or high-cost antibiotics.
d
Source: Moriates C, Arora V, Shah N. Understanding Value-Based Healthcare.
Taken as a whole, a 2015 systematic review on the topic o provid-
H
New York: McGraw-Hill; 2015.
e
ing price in ormation or diagnostic testing concluded that “charge
a
l
t
in ormation changed ordering and prescribing behavior” in the
h
majority o studies.
C
chargemaster is generally used as a starting point or closed-door
a
bargaining with di erent payers. While insurance companies pay a Remaining challenges include determining which price to display
r
e
relatively small raction o the charge on the chargemaster, uninsured (the charge, the Medicare allowable ee, the estimated marginal
o
patients have o ten been stuck with ull chargemaster prices. cost, or some other measure), as well as whether prices should be
r
H
The A ordable Care Act (ACA) now ormally requires all provid- displayed or all orders or rather be limited to only speci c orders
o
that may be ordered requently or that are associated with high cost
s
ers to publish their chargemasters, and some states are requiring
p
or marginal bene t.
i
hospitals to also disclose the “allowed amount” (contractually
t
a
agreed amount paid by a private insurance company) to any patient
l
i
s
t
who asks. In Cali ornia, health care providers cannot bill uninsured
s
patients an amount greater than the reimbursement the hospital
PRACTICE POINT
would receive rom a government payer. • Recent research suggests that displaying price in ormation
Newer methods or determining more accurate measurements o at the point o care may help clinicians and patients make
actual costs are now increasingly being applied in health care. For high-value care decisions.
example, Michael Porter and Robert Kaplan rom Harvard Business
School have advocated or the use o time-driven activity-based
costing (TDABC). With TDABC, the costs o space, nonconsumable HOSPITAL PAYMENTS SHIFTING FROM VOLUME
equipment, and administrative overhead are all assigned minute- TO VALUE
to-minute cost rates that are relevant to speci c processes o care.
I there is one thing that most policymakers can agree on, it is that
The care that is delivered over an entire episode o care is broken
the payment system, which currently rewards volume o services
down into discrete activities or process steps, such as check-in,
delivered, should be realigned to compel the delivery o value. Not
vitals and intake, physician evaluation, nursing care, and so on. A
all policymakers, however, agree exactly how to best do that.
cost is assigned to each step by tracking who is doing the activity,
Medicare’s Value-Based Purchasing (VBP) program has already
what resources they use, which space they are in, and how long it
tied a percentage o hospital payments to metrics o quality, patient
takes them. Each item (personnel, resources, and space) is assigned
satis action, and cost. In addition, with the proli eration o account-
a per-minute cost rate by bundling together all costs ( xed and
able care organizations (ACOs) and other bundled payment models,
variable) and then dividing by the total amount available or patient
hospitals will continue to have an increasing share o reimburse-
care. For a more detailed explanation o TDABC, one can re er to
ment at risk related to the value o care that they deliver. According
“How to Solve the Cost Crisis in Health Care” by Kaplan and Porter
to the US Health and Human Services Secretary, Medicare aims to
in the Harvard Business Review (2011). Using TDABC, some progres-
have at least 50% o all payments tied to quality or value through
sive medical centers have begun to establish a true “cost-master” to
alternative payment models by the end o 2018.
replace the controversial charge-master.
■ MEDICARE’S HOSPITAL VALUE-BASED
PRACTICE POINT PURCHASING PROGRAM
The ederal government introduced their hospital VBP program
• Most hospitalists do not set the prices on the chargemaster
in 2012, initially with 1% o Medicare hospital payments based on
or on hospital bills, but hospitalists can advocate or a more
some measures o quality. This percentage will continue to rise. The
rational health care pricing system and or increased price and
rst quality indicators included process measures or pneumonia,
data transparency at their hospitals.
acute myocardial in arction, congestive heart ailure, health care-
associated in ections, and patient experience (largely based on
patient survey responses to the Hospital Consumer Assessment
■ THE EFFECT OF PRICE TRANSPARENCY ON o Healthcare Providers and Systems Hospital survey [HCAHPS]).
HOSPITALIST ORDERING PRACTICES Subsequently, risk-adjusted mortality, hospital-acquired conditions,
One seemingly obvious solution to hospitalists’ lack o knowledge and patient sa ety were added. The 2016 VBP metrics include eight
about costs is to provide diagnostic test prices at the point o order- clinical process o care measures, eight patient experience dimen-
ing. A ter all, many have remarked that when ordering o a menu sions, three 30-day mortality outcome measures, one Agency or
11
TABLE 2-2 Medicare’s Hospital Value-Based Purchasing Metrics for Fiscal Year 2016
Domain Weight Measures

P
Clinical process o care 10% Fibrinolytic therapy received within 30 min o hospital arrival in patients with acute
A
myocardial in arction
R
T
In luenza immunization
I
Initial antibiotic selection or community acquired pneumonia in immunocompetent
patient
Surgery patient on a beta blocker prior to arrival that received a beta blocker during
perioperative period
T
Prophylactic antibiotic selection or surgical patients
h
e
Prophylactic antibiotics discontinued within 24 h a ter surgery ends
S
p
Postoperative urinary catheter removal on postoperative day 1 or 2
e
c
Surgery patient who received appropriate venous thromboembolism prophylaxis
i
a
within 24 h prior to surgery to 24 h a ter surgery
l
t
y
Patient experience o care 25% Communication with nurses
o
f
Communication with doctors
H
o
Responsiveness o hospital sta
s
p
Pain management
i
t
a
Communication about medicines
l
M
Cleanliness and quietness o hospital environment
e
d
Discharge in ormation
i
c
i
Overall rating o hospital
n
e
Outcome 40% Acute myocardial in arction 30-d mortality rate
a
n
Heart ailure 30-d mortality rate
d
Pneumonia 30-d mortality rate
S
y
s
Complication/patient sa ety indicators (AHRQ PSI-90 composite score)
t
e
Catheter-associated urinary tract in ection
m
s
Central line-associated bloodstream in ection
o
f
Surgical site in ections in colon surgery and abdominal hysterectomy
C
a
E iciency 25% Medicare spending per bene iciary
r
e
Source: Data rom www.medicare.gov. Accessed May 8, 2015.
AHRQ: Agency or Healthcare Research and Quality. AHRQ PSI-90 is a composite score consisting o eight weighted component patient sa ety indicator mea-
sures: pressure ulcers, iatrogenic pneumothorax, central venous catheter-related bloodstream in ections, postoperative hip racture, postoperative pulmonary
embolism or deep vein thrombosis, postoperative sepsis, postoperative wound dehiscence, and accidental puncture or laceration.
Healthcare Research and Quality composite score, our health care- directly incentivize quality and e ciency. Strategies or payments
associated in ection rates, and one e ciency measure based on exist on a spectrum rom straight ee- or-service to xed global
Medicare spending per bene ciary (Table 2-2). budgets. I we consider reimbursements to a hospital, a payer may
Payment or achieving higher-quality metrics seems to be a pay a speci c amount or every service delivered ( ee- or-service),
step in the right direction or our health care system, but there are or each day in the hospital (Per Diem), or each episode o hospi-
criticisms that the current mechanism will un airly punish sa ety net talization (eg, Diagnosis Related Groups [DRGs]), or or each patient
hospitals and clinicians caring or the most vulnerable populations. in their community considered to be under their care (Capitation).
Alternatively, the hospital could be given a xed ee or all services
■ HOSPITAL COMPARE per ormed on every patient during a ull year (Global Budget).
Currently, the majority o payments are still primarily based on ee-
CMS hopes to also drive value through better public transparency
or-service, but this is projected to rapidly change (Figure 2-1).
o quality and cost data via their Hospital Compare website (www.
Bundled payments could theoretically encourage improved e -
medicare.gov/hospitalcompare). Hospital Compare provides data
ciency and reductions in hospital-acquired complications as these
on a large number o metrics and even allows the public to select
would lead to increased costs, length o stay, and spent resources.
up to three hospitals at a time to compare head to head. In an e ort
For example, a hospital could be paid one ee or pneumonia,
to make the website more user- riendly or public consumers, CMS
regardless o the number and type o interventions or resources
recently borrowed a strategy rom the vast majority o popular rat-
used. CMS has used a prospective f at ee per inpatient episode o
ing websites and added a “star rating.” The star rating, rom one to
care, based on a diagnosis-related group (DRG) system, since 1983.
ve stars, is initially based on validated patient experience metrics.
CMS sets the base payment amounts “ or the operating and capital
costs that e cient acilities would be expected to incur in urnish-
■ BUNDLED PAYMENTS AND ACCOUNTABLE ing covered inpatient services.” This rate is then weighted by DRG
CARE ORGANIZATIONS (which accounts or relative severity o a given condition), and then
Whereas the VBP program is based on annual rewards and penalties, adjusted according to an algorithm that accounts or a number o
other payment models including bundled payments aim to more actors such as the regional cost o labor, and whether the hospital
12
2014 Ho s pital pro je c tio ns fo r 2019
Othe r (e g, s ha re d Othe r (e g, s ha re d
C
s avings ), 3% s avings ), 3%
H
A
P
Globa l
T
payme nts,
E
Globa l
10%
R
payme nts,
2
14%
Epis ode of
ca re /bun dle d, 11% Fe e -for-s e rvic e, 34%
V
Pay-for-pe rform a nce , Epis ode of ca re /
a
Fe e -for-s e rvic e, 57%
l
bun dle d, 16%
u
9%
e
-
B
a
s
Ca pita tion ,
e
d
11% Pay-for-pe rform a nce ,
Ca pita tion ,
H
21% 13%
e
a
l
t
h
C
a
Figure 2 1 Hospital or hospital system reimbursements. (The State o Value-Based Reimbursement and the Transition rom Volume to Value
r
e
in 2014. McKesson Health Solutions, 2014.)
o
r
H
o
s
is a teaching acility. Medicare also provides higher payments or use beyond benchmarks, and unnecessary choice o higher-cost
p
i
patients with “complicating or comorbid conditions,” or with “major services (Table 2-3). Unnecessary services account or $210 billion o
t
a
l
complicating or comorbid conditions.” For particularly complex waste annually. This is the area o waste that individual hospitalists
i
s
t
patients, Medicare provides “outlier payments” that are calculated have the most direct control over.
s
based on an imprecise ratio o costs to charges.
Global or bundle payments could potentially combine pay-
■ OVERUSE IS A PATIENT SAFETY PROBLEM
ments across di erent providers and settings, encouraging better
coordination and communication between hospitals, postacute The 1998 IOM National Roundtable on Healthcare Quality classi ed
care acilities (eg, a skilled nursing acility [SNF]), and outpatient three types o health care quality problems: underuse, overuse,
providers. This is some o the logic behind the emergence o ACOs, and misuse. However, the ollowing decade o the patient sa ety
which were included as part o the ACA to provide an experiment movement ocused nearly exclusively on preventable complications
in global payments and shared risk. According to the CMS de ni- related to misuse. Only recently has overuse o medical care—which
tion, ACOs are “groups o doctors, hospitals, and other health care re ers to providing care in circumstances where the potential or
providers, who come together voluntarily to give coordinated high harm exceeds the potential bene ts—gained attention as an impor-
quality care to their Medicare patients.” When an ACO delivers high tant patient sa ety hazard.
quality care at low costs, the organization shares in the savings that Overuse o medical care is a widespread problem in the US
the ACO achieves. health care system. According to a 2011 study, nearly hal o primary
Interpretations o the early results o ACOs have been mixed. care physicians in the United States believe that their patients are
The pilot Pioneer ACO program, which included 32 medical care receiving too much care. Overuse o medical care can directly lead
organizations, was estimated to save 1.2% o health care spending, to patient harm as a result o the known risks or adverse e ects o
translating to about $400 million, over the rst 2 years. Critics point the provided test, procedure, or medication. There are numerous
out that the pilot programs were highly selected and unlikely to requently cited instances o overuse, including inappropriate imag-
represent the abilities o the rest o the health care system once ing, laboratory tests, antibiotics, and catheter usage (Table 2-4).
the model is more widely deployed. Moreover, 13 o the 32 pilot For example, despite evidence and clear guidelines that suggest
programs had dropped out due to not achieving savings in the imaging is unhelp ul or patients with acute low back pain who lack
rst year or because they elt that the program was too complex speci c clinical ndings, routine diagnostic imaging is requently
with too many quality metrics to track. While we await additional obtained or these patients. This places patients at risk or excessive
research evidence on the true potential impact o ACOs, their prem- radiation, costs, and substantial downstream e ects, including ine -
ise in improving e ciency and coordinated care is strong, and their ective spine operations and perceptions o lessened overall health
numbers will likely expand in the near uture. status.
Antibiotic prescribing is another area ri e with overuse, which has
led to the emergence o a number o antibiotic-resistant pathogens,
HEALTH CARE WASTE making in ections more di cult to treat. When prescribed incor-
The Institute o Medicine (IOM) estimated that over $750 billion rectly, antibiotics pose serious risks to both individual patients and
annually spent on health care does not make anyone healthier, and the public health at large. Antibiotic overuse can place patients at
thus is considered waste. This represents up to more than 30 cents risk or allergic reactions, antibiotic-associated diarrhea, and other
on every health care dollar spent. Although there are many con- dangerous adverse e ects.
tributors to health care waste including prices that are excessively More medical care may also lead to overdiagnosis and overtreat-
high, unwarranted administrative costs, raud, and ine ciencies due ment, which may result in a cascade e ect o potential harms,
to system errors and ailures o coordination, the largest compo- including adverse events, mistakes, anxiety and disability, and
nent is unnecessary services, which includes overuse, discretionary additional unnecessary treatments. With patients bearing more
13
TABLE 2-3 Areas of Health Care Waste
Areas of Waste in Annual

P
Healthcare Amount Examples Potential Strategies for Addressing
A
Unnecessary services $210 billion Antibiotics or nonbacterial in ections or Clinician decision making, “Choosing
R
MRIs or routine low back pain Wisely,” appropriateness criteria, high-
T
value care committees or programs
I
Ine iciently delivered $130 billion Lack o interoperability between electronic Lean, care pathway redesign,
services health records resulting in missing in ormation accountable care organizations,
electronic health record coordination
Prices that are too high $105 billion MRIs cost approximately $1080 in the United States Cost transparency, regulatory
T
h
and $280 in France, mostly due to di erences in measures
e
price setting
S
p
Excess administrative $190 billion American medical providers spend our times as Payment re orm, insurance orm
e
c
costs much interacting with insurance companies compared standardizations
i
a
l
to Canadians (who have a single payer system)
t
y
Fraud $75 billion Billing or services that were not rendered, or FBI and other law en orcement
o
f
“upcoding” or a procedure or diagnosis that is more
H
complex than the actual procedure or diagnosis
o
s
p
Missed prevention $55 billion Failing to provide appropriate immunizations Improved access to primary care,
i
t
opportunities or counseling decision support systems, accountable
a
l
care organizations
M
e
TOTAL $750 billion
d
i
c
i
FBI, Federal Bureau o Investigation; MRI, magnetic resonance imaging.
n
e
a
and more o the cost o care themselves, some have urther argued A survey study using clinical vignettes o common hospital clinical
n
d
that clinicians should also consider the potential nancial harm to situations revealed a large amount o overuse o testing among
S
individual patients due to excessive medical evaluations and subse- practicing hospitalists, with 52% to 65% o respondents requesting
y
s
t
quent overtreatments. unnecessary testing in a preoperative evaluation scenario, and 82%
e
m
The many drivers o overuse include medical culture, ee- or-service to 85% in a syncope work-up scenario. The majority o physicians
s
payments, patient expectations, and ear o malpractice litigation. reported that they knew the testing was not clinically indicated
o
f
based on evidence or guidelines, but were ordering the test due
C
to a desire to reassure the patients or themselves. This nding sug-
a
r
e
gests e orts to decrease overuse will need to engage clinicians and
TABLE 2-4 Common Examples of Overuse in Medical Care
patients in ways that help overcome the attitude that more testing
Some Common is required to provide reassurance.
Areas of Overuse Examples
Antibiotics or • 70% o patients with acute bronchitis STRATEGIES FOR HOSPITALISTS TO PROVIDE
nonbacterial are prescribed antibiotics, a rate that HIGH-VALUE CARE
illnesses has been increasing over time As payment systems and health care organizations shi t toward
• Antibiotic prescribing could potentially rewarding and supporting a ocus on value, individual hospital-
be improved in 37% o common ists can help deliver higher value care or their patients through:
inpatient prescription scenarios, (1) providing appropriate care, (2) ensuring care coordination,
according to the Centers or Disease (3) considering patient a ordability in customizing treatment plans,
Control and Prevention
and (4) leading local value improvement initiatives.
Diagnostic imaging • An estimated 3.8 million Americans
receive routine imaging or low back ■ PROVIDING APPROPRIATE CARE
pain each year
Hospitalists should address the problem o overuse by directly
• One-third o imaging per ormed in the
practicing appropriate care or their patients. Emerging resources
emergency department or suspected
pulmonary embolism may be avoidable or identi ying speci c targets o common overuse include the
Choosing Wisely lists, guidelines, and appropriateness criteria. The
Laboratory testing • Approximately 20% o lab testing may
Choosing Wisely campaign (www.choosingwisely.org) is an e ort
represent overutilization
organized by the ABIM Foundation to engage specialty societies in
• Daily blood tests are routinely drawn in identi ying lists o commonly overused medical services “that physi-
many patients in the hospital, which can cians and patients should question.” In 2013, the Society o Hospital
contribute to hospital-acquired anemia
Medicine published an initial Choosing Wisely list or both adult and
Urinary and central • Between 21% and 63% o urinary pediatric hospital medicine (Table 2-5), and many other pro es-
venous catheters catheters are placed in patients who do
sional organizations’ Choosing Wisely lists (eg, American College o
not have an appropriate indication
Physicians, American Academy o Neurology, etc) have components
• According to one study, hospitalists that apply directly or indirectly to hospital medicine practice.
elt 10%-25% o peripherally inserted One strategy or encouraging and communicating appropriate
central catheters (PICCs) placed at their
care is to create a cognitive orcing unction by explicitly document-
acility were inappropriate or avoidable
ing these types o decisions in daily progress notes. For example,
14
charge o their care. Hospital physician discontinuity may lead to
TABLE 2-5 Society of Hospital Medicine Choosing Wisely Lists increased resource utilization and lower patient satis action. Coor-
C
dinating structured hando s between inpatient providers and with
H
Pediatric Hospital
Adult Hospital Medicine Medicine outpatient providers during transitions in care is critical to delivering
A
Recommendations Recommendations high-value care.
P
Currently, about one- th o Medicare patients are readmitted
T
1. Do not place, or leave in 1. Do not order chest
E
place, urinary catheters or radiographs in children within 30 days o hospitalization, and more than hal o these
R
incontinence or convenience, with uncomplicated patients do not see any outpatient health care provider between
2
or monitoring o output asthma or bronchiolitis. these visits. This population o requently readmitted patients is
or noncritically ill patients particularly important or hospitalists. Some care coordination
(acceptable indications: critical programs have been experimenting with the use o hospitalists
V
illness, obstruction, hospice, that care or a subset o the highest risk patients both during hos-
a
l
u
perioperatively or <2 d or pitalization and ollowing discharge, either in a high-risk clinic or
e
urologic procedures; use weights
-
at postacute care acilities such as skilled nursing acilities. These
B
instead to monitor diuresis).
a
physicians are increasingly becoming known as “extensivists.” The
s
e
2. Do not prescribe medications 2. Do not routinely early data on the cost-e ectiveness o these types o programs
d
or stress ulcer prophylaxis to use bronchodilators have been mixed, but, much like ACOs, it may be too early to draw
H
medical inpatients unless at high in children with
e
conclusions.
a
risk or GI complication. bronchiolitis.
l
t
h
3. Avoid trans using red blood cells 3. Do not use systemic ■ CONSIDERING PATIENT AFFORDABILITY
C
just because hemoglobin levels corticosteroids in
a
More Americans than ever be ore are on high-deductible insurance
r
are below arbitrary thresholds children under
e
such as 10, 9, or even 8 mg/dL in 2 y o age with an plans, making them responsible or an increasing share o health
o
r
the absence o symptoms. uncomplicated lower care costs. As “ nancial harms” or individual patients become
H
respiratory tract increasingly recognized, and more patients orgo recommended
o
s
in ection. medical treatments due to out-o -pocket costs, hospitalists must
p
i
4. Avoid overuse/unnecessary use 4. Do not treat customize care plans to help patients a ord their care.
t
a
l
o telemetry monitoring in the gastroesophageal Hospitalists may be able to improve their prescribing practices,
i
s
t
hospital, particularly or patients re lux in in ants particularly at the time o discharge. Nearly one-quarter o hospital-
s
at low risk or adverse cardiac routinely with acid ized adults in a survey reported cost-related underuse in the year
outcomes. suppression therapy. prior to admission, and only 16% o patients knew how much their
5. Do not per orm repetitive CBC 5. Do not use continuous prescribed medications at discharge would cost them. Virtually
and chemistry testing in the ace pulse oximetry nobody had spoken to their inpatient providers about the cost o
o clinical and lab stability. routinely in children the newly prescribed drugs.
with acute respiratory
Discussing drug costs with patients has been shown to be
illness unless they
are on supplemental strongly associated with providing individualized lower-cost medi-
oxygen. cation options. Health care pro essionals and patients can rely on
an increasing number o reely available resources that provide
Source: Adapted rom the Society o Hospital Medicine’s adult and pediatric price in ormation and cheaper alternatives or most medications
hospital medicine Choosing Wisely recommendations. www.choosingwisely (Tab le 2-6). High-value prescribing has been de ned as “providing
.org. Accessed May 8, 2015. the simplest medication regimen that minimizes physical and nan-
cial risk to the patient while achieving the best outcome.”
hospitalists Drs Scott Flanders and Sanjay Saint recommend includ-
ing the indication, day o administration, and expected duration PRACTICE POINT
o therapy or all antimicrobial therapies in all progress notes and
sign-outs, as an approach or curbing inpatient antibiotic overuse. • Hospital sta can help screen patients or nancial concerns,
Likewise, hospitalists may eliminate use o routine labs, telem- particularly related to prescribed discharge medications. An
etry, continuous pulse oximetry, or other recurrent interventions or increasing number o tools are available to help determine the
monitoring by documenting daily the patient needs and reasons or most cost-e ective medication or a given patient’s condition
continued use or ordering. and insurance coverage (Table 2-6).
PRACTICE POINT ■ IMPLEMENTING VALUE-BASED INITIATIVES

• Avoiding overuse is the simplest way to simultaneously Hospitalists across the country have largely taken the lead on
enhance patient sa ety and decrease costs. Common areas o designing value improvement pilots, programs, and groups within
potential overuse in hospitalized patients include antibiotics, hospitals. Although value improvement projects may be built upon
telemetry and monitoring, imaging, and routine labs. the established structures and techniques or quality improvement
(see Section IV: Patient Sa ety and Quality Improvement, Chapter 21:
Quality Improvement Methodologies), importantly these programs
■ THE IMPORTANCE OF CARE COORDINATION should include expertise in cost analyses. Furthermore, some
The typical hospital patient is handed o rom one physician to traditional quality improvement programs have ailed to result in
another more than 15 times during a single 5-day hospital stay, actual cost savings; thus it is not enough to simply re-brand quality
a rate that has been increasing with new duty hour restrictions improvement with a banner o “value.” Value improvement e orts
and hospitalist sta ng models. Not surprisingly, studies show the must overcome the cultural hurdle o “more care as better care,” as
majority o hospital patients are unable to identi y the clinician in well as pay care ul attention to the diplomacy required with value
15
programs have led to remarkable improvements in hospital pro-
TABLE 2-6 Some Resources That Provide Price Information cesses and outcomes across the country rom sa ety net hospitals
for Medications like Denver Health to Veteran’s A airs hospitals to the University
o Michigan.
P
Websites Consumer http://www.consumerreports.
A
reports best org/cro/health/prescription- Similar in concept to lean are e orts to design and hone speci c
R
buy drugs drugs/best-buy-drugs/index.htm care pathways or certain patients and conditions. For example,
T
GoodRx www.goodrx.com many joint-replacement programs have created care pathways that
I
standardize when patients will have catheters removed, mobilize
Mobile Epocrates http://www.epocrates.com
applications with physical therapy, and be discharged to a speci c disposition.
Lowestmed http://www.lowestmed.com Hospitalists are increasingly creating similar models or patients
GoodRx http://www.goodrx.com with pneumonia, chronic obstructive pulmonary disease (COPD)
T
$4 Generic Walmart http://www.walmart.com/cp/4-
h
exacerbations, syncope, or other common clinical conditions. Inter-
e
drug lists Prescriptions/1078664 mountain Healthcare in Utah has applied evidence-based protocols
S
p
Target http://www.target.com/ to more than 60 clinical processes, re-engineering roughly 80% o all
e
pharmacy/generics
c
care that they deliver. Cincinnati Children’s Hospital partnered with
i
a
Sam’s Club http://resources.samsclub.com/ local physicians to create large-scale improvements in the care o
l
t
y
health-and-wellness/pharmacy- children with asthma, resulting in 92% adherence to best practices
o
and-health/Extra-Value-Drug-List. or asthma care, which has yielded many avoided hospital admis-
f
H
aspx sions and emergency department visits. These types o care rede-
o
s
Krogers http://www.kroger.com/topic/ signs and standardization promise to provide better, more e cient,
p
i
save-on-generic-prescriptions and o ten sa er care or more patients.
t
a
l
M
e
CONCLUSION
d
i
c
Hospitalists are now aced with the massive responsibility o provid-
i
n
improvement since reducing costs may result in decreased revenue ing better health care value. Based on the history o the hospitalist
e
a
or certain departments or even decreases in individuals’ wages. movement, we are up to the task. Value is de ned as providing the
n
d
The national nonpro t group Costs o Care has proposed a highest quality care at lower costs, and should include components
S
“COST” ramework to guide value improvement project design. o patient outcomes, sa ety, and experience. To achieve this goal,
y
s
COST stands or culture, oversight accountability, system support, measuring and understanding metrics related to quality and costs
t
e
and training. This approach leverages principles rom implementa- will be vitally important. There is an inexorable trend toward greater
m
s
tion science to ensure that value improvement projects success ully transparency in health care and it is likely soon that true health
o
provide multipronged tactics or overcoming the many barriers to care costs will be publicly accessible across the country. Evidence is
f
C
high-value care delivery. mounting that providing some cost in ormation at the point-o -care
a
r
may help support behavior changes and decrease unnecessarily
e
expensive test ordering.
PRACTICE POINT The most potent strategy or simultaneously improving care and
• Hospitalists are uniquely positioned to identi y potentially decreasing costs is to reduce waste and overuse, which accounts or
waste ul or ine cient practices within medical centers and more than $200 billion in health care costs annually in the United
to lead value improvement initiatives. Value improvement States and causes signi cant patient harms. There are many com-
work requires the inclusion o expertise in health care cost mon areas o overuse in hospital care, including the use o antibiot-
accounting, as well as thought ul diplomacy, and the design ics, telemetry monitoring, trans usions, imaging, catheter usage, and
o multipronged e orts that explicitly target culture, oversight, routine lab draws. Many o these are highlighted in Choosing Wisely
systems, and training (COST). lists and other emerging appropriateness criteria.
Hospitalists can deliver high-value care to their patients by
speci cally considering appropriateness o care, care coordination,
patient a ordability, and value-based initiatives and care pathways.
■ APPLYING LEAN AND REDESIGNING CARE PATHWAYS
On a health system level, methods or ensuring better value may
ocus on techniques to improve e ciency and decrease “de ects.” SUGGESTED READINGS
To achieve this goal, an increasing number o hospitals are now
adopting lean methodologies and systems. Lean principles stem Conway PH. Value-driven health care: implications or hospitals and
rom the Toyota Production System developed by the automaker in hospitalists. J Hosp Med. 2009;4(8):507-511.
Japan to ocus on improving quality while reducing waste. In 2002, Feldman LS, Shihab HM, Thiemann D, et al. Impact o providing ee
Virginia Mason Medical Center in Seattle amously began applying data on laboratory test ordering: a controlled clinical trial. JAMA
to health care the ve general principles o lean: (1) de ne value Intern Med. 2013;173(10):903-908.
rom the customer’s perspective, (2) identi y the value stream and
Flanders SA, Saint S. Why does antimicrobial overuse in hospitalized
remove any waste, (3) make value f ow without interruption, (4) help
patients persist? JAMAIntern Med. 2014;174(5):661-662.
customers pull value, and (5) pursue per ection.
Some lean tools have quickly been adapted to health care, Goetz C, Rotman SR, Hartoularos G, Bishop TF. The e ect o charge
including value stream maps that depict all o the individual steps display on cost o care and physician practice behaviors: a system-
in a process rom beginning to end, and provides a graphical tool atic review. J Gen Intern Med. 2015;30(6):835-842.
or identi ying any non–value-added steps, delays, waiting times, Institute o Medicine. Best Care at Lower Cost: The Path to Continu-
and ine ciencies, as well as the commitment to rapid improve- ously Learning Health Care in America. Washington, DC: National
ment cycles that are built around “small tests o change.” Lean Academies Press; 2012.
16
Kachalia A, Berg A, Fagerlin A, et al. Overuse o testing in preopera- Moriates C, Mourad M, Novelero M, Wachter RM. Development o a
tive evaluation and syncope: a survey o hospitalists. Ann Intern hospital-based program ocused on improving healthcare value.
C
Med. 2015;162(2):100-108. J Hosp Med. 2014;9(10):671-677.
H
Kaplan RS, Porter ME. How to solve the cost crisis in health care. Owens DK, Qaseem A, Chou R, Shekelle P. High-value, cost-
A
Harvard Business Review. 2011;89(9):46-52, 54, 56-61 passim. conscious health care: concepts or clinicians to evaluate the
P
T
Moriates C, Arora V, Shah N. Understanding Value-Based Healthcare. bene ts, harms, and costs o medical interventions. Ann Intern
E
New York: McGraw-Hill; 2015. Med. 2011;154(3):174-180.
R
2
V
a
l
u
e
-
B
a
s
e
d
H
e
a
l
t
h
C
a
r
e
o
r
H
o
s
p
i
t
a
l
i
s
t
s
17
CHAP TER
3 INTRODUCTION
Racial and ethnic disparities in care have been consistently docu-
mented in the treatment and outcomes o many common clinical
diseases. The 2003 Institute o Medicine (IOM) report, “Unequal
Treatment: Con ronting Racial and Ethnic Disparities in Health
Care,” de nes disparities as di erences in the treatment that are not
directly attributable to access-related actors, clinical needs, patient
pre erences, or appropriateness o intervention (Figure 3-1). The
elimination o health care disparities is a high priority or the ederal
Racial/Ethnic government and several academic organizations.
Documented disparities o disease prevention and treatment
Disparities in include rates o vaccination, cancer screening, secondary prevention

o myocardial in arction (MI), transplant surgery, curative surgery,
Hospital Care and angioplasty. Disparities in health outcomes include cardiovas-

cular disease, HIV/AIDS, diabetes, cancer, asthma, pregnancy out-
comes, mental health, and hospitalization.
Speci c examples include the ollowing (Table 3-1):
• A higher risk o stroke, heart ailure, and renal ailure associated
Lenny López, MD, MPH, MDiv
with hypertension (A rican Americans)
• A higher rate o complications rom diabetes (A rican Americans
and Native Americans)
• Later-stage colon, breast, and prostate cancer at presentation
(A rican Americans)
• Less aggressive evaluation and treatment: curative lung cancer
resection, cardiac catheterization, peripheral angioplasty, and
renal transplantation (A rican Americans)
• Diabetic more likely to receive amputations (A rican Americans)
• Higher death rates per 1000 hospital admissions in low-mortality
diagnosis-related groups (A rican Americans, Hispanics, and
the uninsured)
The observed racial/ethnic health care disparities have multi ac-
torial etiologies. Patients ace multiple barriers as they engage the
health care system: (1) personal and amily; (2) access to the health
care system (structural, nancial, types o services); and (3) the qual-
ity o the available providers (Figure 3-2). These barriers can occur
individually or in combination to have an additive e ect on health
outcomes.
PRACTICE POINT
• Disparities in health outcomes include the ollowing:
cardiovascular disease, HIV/AIDS, diabetes, cancer, asthma,
pregnancy outcomes, mental health, and hospitalization.
Hospitalists may signi cantly in uence the health status o
A rican American and Latino patients i they comprehend their
health care needs, communicate e ectively, and advocate or
additional local and institutional resources to ensure optimal
discharge back to the community.
Historically, disparities in hospital care originated in the policy o

hospital segregation during the rst 66 years o the 20th century.
Be ore the creation o Medicare, the Hospital Survey and Construc-
tion Act o 1946, commonly known as “Hill-Burton,” was the larg-
est ederal grant program in health care. This law was intended
to increase the number o hospital beds throughout the country.
However, this was the only ederal legislation in the 20th century
that explicitly permitted use o ederal unds to provide racially
18
C
H
Clinica l a ppropria te ne s s
A
e
a nd ne e d pa tie nt
r
a
P
pre fe re nce s
c
T
h
t
E
l
a
Diffe re nce The ope ra tion of
R
e
y
h
t
he a lth ca re sys te ms a nd
i
r
f
3
o
o
le ga l a nd re gula tory
n
y
i
m
t
i
clima te
y
l
Dis pa rity
n
a
t
i
u
r
o
o
R
N
Q
n
Dis crimina tion: bia s e s,
a
i
M
c
s te re otyping, a nd
i
a
l
unce rta inty
/
E
t
h
n
Figure 3 1 De ining di erences, disparities, and discrimination in populations with equal access to health care. (Reproduced, with permission,
i
c
rom Smedley BD, Stith AY, Nelson AR. Unequal Treatment. Con ronting Racial and Ethnic Disparities in Health Care. Washington, DC: National
D
i
Academies Press; 2002.)
s
p
a
r
i
t
i
e
exclusionary services (“separate but equal”) thus augmenting the quality o care or common medical and surgical conditions. One
s
wide divide in poor-quality hospital services and acilities or A rican study demonstrated that 90% o Hispanic and black Medicare ben-
i
n
Americans. Hospital segregation ended with President Johnson e ciaries receive their care at 25% o the 4500 acute care hospitals in
H
o
signing into law the Medicare bill on July 9, 1965. The Medicare bill the United States. Another study ound signi cant racial/ethnic dis-
s
p
was tied to Title VI o the Civil Rights Act o 1964, which banned parities in clinical processes or AMI, CHF, and pneumonia explained
i
t
discrimination in any activities that used ederal unds or training, primarily by hospital actors and not individual patient character-
a
l
employment, or construction. Because o this legal requirement, istics. They also ound that lower-per orming hospitals tended to
C
a
more than 95% o hospitals desegregated their acilities by the rst serve a larger proportion o minority patients. Similar ndings have
r
e
day Medicare was implemented on July 1, 1966, in order to receive been demonstrated using national Hospital Quality Alliance (HQA)
Medicare reimbursement. patient-level data. Other national studies demonstrate that A rican
The elimination o health care disparities is a high priority or the Americans go to hospitals that have lower rates o evidence-based
ederal government, and many academic organizations are begin- medical treatments and worse risk-adjusted mortality a ter AMI, are
ning to take steps to educate physicians about this problem. To less likely to receive optimal care or pneumonia as measured by
date, existing data suggest that the type o hospital acility and its national HQA measures, and have higher operative mortality risks or
location explain some o the observed racial and ethnic disparities eight di erent procedures because the hospitals they attend have
in health care services; less is known, however, about disparities in higher mortality rates or all patients.
hospital care. This chapter reviews the racial disparities in hospital In addition to di erences in quality o care based on institution,
care that can be impacted by hospitalists and proposes directions data suggest that among patients hospitalized in the same institu-
or uture research. tions, racial and ethnic disparities in care o ten exist. Several studies
have demonstrated signi cant racial and ethnic di erences in utili-
MECHANISMS AND ETIOLOGY zation o cardiovascular procedures or patients hospitalized within
the same institutions. For example, A rican Americans have a lower
■ INTERINSTITUTIONAL AND INTRAINSTITUTIONAL rate o coronary artery bypass procedures than whites, even with
VARIATIONS IN CARE similar presentation and clinical eatures.
Although where patients receive care likely explains some o the
observed racial and ethnic disparities in health care service, prior ■ DIFFERENTIAL UTILIZATION OF MEDICAL
studies suggest that hospital-level actors may play an important PROCEDURES AND TECHNOLOGY AVAILABILITY
role in creating disparities in care. In addition, minorities live dispro- Racial di erences in the utilization o medical procedures are well
portionately in parts o the country that have lower-quality hospitals documented, especially or “re erral-sensitive” procedures and inva-
and ewer primary care physicians. sive, costly procedures such as coronary revascularization. The rea-
Sa ety-net hospitals predominantly serve poor and underserved sons or these di erences are complex and may re ect di erences
patients and provide care or a disproportionate number o racial and in clinical presentation, medical decision making, di erential access
ethnic minorities in the United States. Multiple studies have shown that to providers and institutions providing procedures, and di erential
these hospitals o ten provide a lower quality o care. This decreased care at hospitals.
quality is likely due to shortages o resources, nurse sta , technical Studies have demonstrated that Caucasian patients more o ten
support such as health in ormation systems, and capital to make receive renal transplantation, cardiac surgical procedures, total joint
improvements. These hospitals have increased post-MI mortality rates replacement, and other procedures than do A rican Americans.
and decreased per ormance measure scores or acute MI, lower per- A rican Americans are less likely to receive coronary revasculariza-
ormance on national quality process indicators or acute myocardial tion compared to whites, even in hospitals with revascularization
in arction (AMI), congestive heart ailure (CHF), and pneumonia, and services. Additionally, A rican Americans are less likely to be trans-
higher postoperative colon cancer mortality rates. In addition, they erred rom hospitals without revascularization services to those
tend to have smaller gains over time on process measures or AMI, CHF, with these cardiac services, and, even when they are trans erred,
and pneumonia, and are less likely to achieve high-per orming status. they are still less likely to receive revascularization as compared
A large proportion o minority patients receive their care in a small to Caucasians. These di erences in procedure use have been
number o hospitals and these acilities seem to provide a lower associated with increased A rican American mortality rates. Finally,
19
may have lower rates o adoption o new technologies. In general,
TABLE 3-1 Racial/Ethnic Disparities in Disease Prevention, sa ety-net providers are slower to adopt new technologies than
Treatment, and Outcomes* non–sa ety-net providers. Providers and hospitals that invest in
technology score higher on standard quality measures. However,
P
Conditions where minorities documented with greater rates
A
compared with whites national studies have demonstrated that providers who cared or
R
uninsured and Medicaid A rican American and Hispanic patients
Cardiovascular disease:
T
are less likely to use electronic health records. These di erences
I
• Hypertension, stroke, congestive heart ailure underscore the need to provide higher unding o public and other
Incidence and mortality rom HIV sa ety-net hospitals in order to reduce disparities in health care by
Type 2 diabetes: ensuring the delivery o high-quality care or all patients.
• Prevalence
T
■ PATIENT EXPERIENCES
h
• Amputations
e
Perceived provider attitude (both physicians and nonphysicians),
S
• Hospitalizations
p
including perceptions o provider prejudice, by minority patients
e
Malignancy:
c
has been shown to have a direct relationship to patient decision
i
a
• Advanced-stage breast cancer
l
making and perceived quality and satis action. Reports o patient
t
y
• Advanced-stage colon cancer experiences with health care are there ore important correlates with
o
f
• Advanced-stage prostate cancer quality. There are signi cant di erences in hospitalized patients’ sel -
H
reported experiences among di erent groups. A rican American and
o
Pregnancy outcomes:
s
Latino patients are less satis ed with their hospital care, particularly
p
• In ant mortality
i
t
in the dimension o having their pre erences respected consistent
a
• Low birth weight
l
with prior studies demonstrating racial/ethnic di erences in satis ac-
M
• Maternal mortality tion with provider communication and management. Both A rican
e
d
End-stage renal disease American and Hispanic patients report that perceived attitudes o
i
c
i
Preventable hospitalizations social workers and nursing sta have an important direct relation-
n
e
Conditions where minorities documented with lower rates ship with their perceived satis action and quality. Hispanic patients
a
correlate high satis action with care when well-quali ed medical
n
compared with whites
d
Cardiovascular procedures: interpreters are available.
S
y
• Cardiac catheterization
These ndings have several important implications or prioritizing
s
t
quality improvement e orts in improving patient satis action with
e
• Peripheral artery angioplasty
m
care. First, physicians and hospital sta should strive to better under-
s
• Implantable cardiac de ibrillators stand and address the expectations o A rican American and Latino
o
f
Adult vaccinations patients. Second, hospital administrators should include allied health
C
pro essionals and social workers in addition to nurses and physicians
a
Solid organ transplantations
r
e
Secondary prevention or myocardial in arction in training on patient-centered and culturally appropriate counseling
Conditions where minorities documented with greater rates techniques and communication. All clinicians should use medical
compared with whites interpreters when English is not the rst language. Finally, hospitals
should collect satis action data strati ed by race and ethnicity in
Curative surgery or lung cancer
order to better tailor quality improvement (QI) e orts.
Inadequate hemodialysis
Renal transplantation
■ ROLE OF HOSPITALISTS
Receipt o recommended care or acute myocardial in arction
Hospitalists are specialists in the general medical care o hospital-
Receipt o recommended care or pneumonia
ized patients. Their activities include patient care, teaching, research,
Receipt o recommended care or congestive heart ailure and leadership related to inpatient care. The number o hospitalists
Receipt o recommended care or type 2 diabetes continues to grow signi cantly across the nation; as a result, this
Satis action with hospital care specialty will care or increasing numbers o hospitalized under-
Age-adjusted li e expectancy served patients. Existing literature has demonstrated that hospital-
ists are associated with lower inpatient costs and shorter lengths
Adapted rom Unequal Treatment and Agency or Healthcare Research and
*
o stay compared to general internists and amily physicians, and
Quality. National Healthcare Disparities Report 2009. Agency or Healthcare such savings did not have a detrimental e ect on rates o death or
Research and Quality, US Department o Health and Human Services. readmission. Importantly, hospitalists are associated with provid-
ing higher-quality inpatient care because o closer adherence to
procedure volume has been shown to be a proxy or quality o care. treatment guidelines and better post discharge ollow-up. A recent
A rican Americans and Hispanics tend to get care at low-procedure- national study has demonstrated that hospitals with hospitalists
volume hospitals with low-volume surgeons and cardiologists. A were associated with better per ormance on quality indicators or
study o coronary artery bypass surgery (CABG) surgery outcomes AMI, pneumonia, and the composite domains o disease treatment,
in New York State ound that A rican Americans and Asians were diagnosis, counseling, and prevention controlling or hospital char-
more likely to receive care rom surgeons with higher risk-adjusted acteristics such as size, location, ownership type, and staf ng avail-
mortality. A rican Americans at low-volume hospitals have greater ability. There are, however, no data reported on the per ormance
risk-adjusted mortality than Caucasian patients a ter elective aortic o hospitalists compared to other providers relating to their care o
abdominal aneurysm (AAA) repair, CABG, and carotid endarterec- di erent ethnic groups.
tomy (CEA). Previous research using national quality measures has ound sub-
Technology availability likely contributes to low per ormance. stantial variability and room or improvement in the care o hospital-
Hospitals with a high proportion o A rican American inpatients ized patients across medical conditions. With the continued growth
20
Ba rrie rs
C
Pe rs ona l/fa mily
H
• Acce pta bility
A
• Culture
P
• La ngua ge /lite ra cy Us e of s e rvic e s Me d ia tors Outc ome s
T
• Attitude s, be lie fs
E
Vis its Qua lity of provide rs He a lth s ta tus
R
• Pre fe re nce s • Prima ry ca re • Cultura l compe te nce • Morta lity
• Involve me nt in ca re
3
• S pe cia lty • Communica tion s kills • Morbidity
• He a lth be havior • Eme rge ncy • Me dica l knowle dge • We ll-be ing
• Educa tion/income • Te chnica l s kills
Proce dure s • Functioning
R
S tructura l • Bia s /s te re otyping
a
• Preve ntive Equity of s e rvice s
c
• Ava ila bility
i
• Dia gnos tic Appropria te ne s s of ca re
a
Pa tie nt views of ca re
l
• Appointme nts
/
• The ra pe utic • Expe rie nce s
E
Effica cy of tre a tme nt
• How orga nize d
t
h
Pa tie nt a dhe re nce • Sa tis fa ction
n
• Tra ns porta tion
• Effe ctive
i
c
Fina ncia l pa rtne rs hip
D
• Ins ura nce cove ra ge
i
s
p
• Re imburs e me nt
a
r
leve ls
i
t
i
• Public s upport
e
s
i
n
Figure 3 2 Barriers and mediators o racial/ethnic health care disparities. (Adapted, with permission, rom Cooper LA, Hill MN, Power NR.
H
o
Designing and evaluating interventions to eliminate racial and ethnic disparities in health care. J Gen Intern Med. 2002;17:477-486. Copyright
s
p
2002 Society o General Internal Medicine.)
i
t
a
l
C
a
o the hospitalist inpatient care model, urther research is needed clinical providers are not only an inconvenience to work ow but are
r
e
to delineate the speci c hospitalist model characteristics associated also a patient sa ety hazard.
with improved quality and outcomes o care. The hospitalist model Inpatient care is a continuum o clinical care encounters with
o inpatient care should be considered an essential component both clinical and nonclinical sta . All patient interactions with clini-
o quality improvement or hospitals seeking to improve inpatient cal providers should document whether an interpreter or a bilingual
care. This is especially true or public/municipal hospitals and clinician was used in order to allow quality assessment both in real
smaller hospitals, which have been shown to be consistently associ- time and retrospectively. It is likely not enough to have interpreter
ated with lower quality and provide care or the disproportionate use only at the time o admission or possibly only at discharge given
number o racial/ethnic minority patients. that the clinical care received by patients’ change may change
rapidly during admission. Increasing the availability and use o pro-
■ HOSPITALISTS AND THE USE OF INTERPRETERS essional interpreters throughout the hospital stay is essential. This
IN THE CLINICAL ENCOUNTER will also require educating physicians about the quality o available
In 2011, over 25 million Americans rated themselves as speaking interpretation modalities. Recent studies have demonstrated that
English less than “very well,” and as a result, health care providers remote telephone or video con erencing modes are as e ective
increasingly encounter patients with limited English pro ciency as in-person interpretation as measured by communication qual-
(LEP). The use o pro essional medical interpreters is associated ity and patient satis action. Use o these non–in-person methods
with increased patient satis action, quality o care, and improved should be encouraged in the hospital setting where in-person
disease-speci c process measures and outcomes. The Institute o interpretation may not be readily available throughout the 24-hour
Medicine report Crossing the Quality Chasm states that the use o an hospital continuum o care.
interpreter is not only a quality but also a patient sa ety imperative.
Patients with communication problems due to language barriers
PRACTICE POINT
are at high risk or preventable adverse events. Recent research has
documented that patient sa ety events a ecting hospitalized LEP • To ensure high quality and maximize patient sa ety, always use
patients are more severe and more requently due to communica- certi ed interpreters in any clinical encounter with patients
tion errors compared to English-speaking patients. Similarly, drug who have limited English pro ciency. Hospitals should strati y
complications in outpatients are more common in patients who their adverse events data by language in order to improve
speak a primary language other than English. Hospitals should patient sa ety.
strati y their adverse events data by language in order to improve
patient sa ety.
Although Federal law, the Institute o Medicine, and hospital
guidelines, including the Joint Commission standards, all recom- CONCLUSION
mend the routine use o pro essional interpreters during clinical Racial/ethnic disparities in hospital care and outcomes are based in
encounters, many LEP patients do not have access to interpreters a tradition o hospital segregation that has residual, signi cant, and
and some clinicians continue to “get by” without using interpret- persistent e ects. A growing body o literature has demonstrated
ers consistently. In addition, nonpro essional interpreters cause disparities in processes o care, utilization o procedures, clinical out-
an increase in interpretation errors that may potentially harm the comes, and patients’ experiences. These disparities are due in large
patient through the misinterpretation o in ormation and alteration part to di erences in access to higher-quality hospitals; however,
o key patient details. Language barriers between patients and their variation in intrainstitutional care has also been documented.
21
A small number o hospitals nationwide serve a high number o Hasnain-Wynia R, Baker DW, Nerenz D, et al. Disparities in health
minorities. Ameliorating health care disparities includes improving care are driven by where minority patients seek care. Examina-
hospital quality o care. The hospitalist model o inpatient care pro- tion o the Hospital Quality Alliance Measures. Arch Intern Med.
vides a vehicle or driving high quality by engaging hospitalists as 2007;167:1233-1239.
P
A
quality improvement and patient sa ety leaders. All clinicians should Hicks LS, Tovar DA, Orav EJ, Johnson PA. Experiences with hospital
R
strive to improve quality and patient sa ety through the consistent care: perspectives o black and Hispanic patients. J Gen Intern Med.
T
use o certi ed interpreters during clinical encounters involving 2008;23(8):1234-1240.
I
patients with limited English pro ciency.
Jha AK, Orav EJ, Li Z, Epstein AM. Concentration and quality o
It is incumbent among policy makers, hospital administrators,
hospitals that care or elderly black patients. Arch Intern Med.
and clinicians to develop strategies to achieve racial/ethnic equity
2007;167:1177-1182.
in care. These strategies should include (1) examining within-
T
Jha AK, Orav EJ, Li Z, Epstein AM. The characteristics and per or-
h
institution di erences in care, outcomes, and patient experience
e
based on patients’ race/ethnicity; (2) improving in rastructure and mance o hospitals that care or elderly Hispanic Americans.
S
Health Af . 2008;27:528-537.
p
quality within largely minority-servicing institutions; and (3) devel-
e
oping QI initiatives ocused on cultural competency and targeting
c
López L, Hicks L, Cohen AP, McKean S, Weissman JS. Hospitalists and
i
a
high-risk racial and ethnic groups such as those with limited English the quality o care in hospitals. Arch Intern Med. 2009;169:1-6.
l
t
y
pro ciency. López L, Huerta D, Soukup J, Rodriguez F, Hicks L. Use o interpret-
o
f
ers by physicians or hospitalized limited English pro ciency
H
SUGGESTED READINGS
o
patients and its impact on patient outcomes. J Gen Intern Med.
s
p
2015;30(6):783-789.
i
t
Agency or Healthcare Research and Quality. National Healthcare
a
Trivedi AN, Nsa W, Hausmann LR, et al. Quality and equity o care in
l
Disparities Report 2013. http://nhqrnet.ahrq.gov/inhqrdr/reports/
M
U.S. hospitals. N Engl J Med. 2014;371(24):2298-2308.
nhdr. Accessed April 30, 2015.
e
d
Werner RM, Goldman LE, Dudley RA. Comparison o change in qual-
i
Divi C, Koss RG, Schmaltz SP, Loeb JM. Language pro ciency and
c
ity o care between sa ety-net and non-sa ety-net hospitals. JAMA.
i
n
adverse events in US hospitals: a pilot study. Intl J Qual Health Care.
2008;299:2180-2187.
e
2006;18:383-388.
a
n
Groeneveld PW, Lau er SB, Garber AM. Technology di usion, hospital
d
S
variation, and racial disparities among elderly Medicare bene ciaries
y
s
1989–2000. Med Care. 2005;43:320-329.
t
e
m
s
o
f
C
a
r
e
22
CHAP TER
4 INTRODUCTION
From the beginning o the hospitalist movement, hospitalists have
lled a collaborative role in assuming care o primary care physicians’
patients in the hospital. Just as primary care physicians (PCPs) can-
not easibly be in two places at once (the o ce and the hospital),
surgeons and medical subspecialists cannot simultaneously man-
age complex inpatients and per orm procedures and other specialty
services. The limited surgical availability with restricted surgical
resident work hours creates added pressure on surgical residents
Comanagement of to maximize operating room time. Likewise, medical subspecialties
ace similar pressures with limited ellow work hours. The active
Orthopedic Patients involvement o a medical comanager may make practical and

economic sense i it is planned well and actively managed. In addi-
tion, co-management may improve the quality o care by having a
generalist on site to anticipate and address common problems that
Christopher M. Whinney, MD, FACP, FHM arise during hospitalization without the delays that may occur with
traditional medical consultation.
Comanagement is now a prominent practice pattern as an
integrated part o hospitalist practice. Comanagement practices
have now been described in collaboration with orthopedic surgery,
neurosurgery, vascular surgery, otolaryngology, hepatology, and
pediatrics. Thus, in all likelihood, the practice o comanagement
by hospitalists will not wane, and both surgeons and medical sub-
specialists will call on hospitalists in this collaborative spirit. Some
authors express concerns or exacerbating the work orce shortage
o internists by increasing overall workload with comanagement
and increasing the “silo” delineations in medicine. However, coman-
agement may provide value i there is a clear delineation o roles
and responsibilities and the value equation is articulated or all
parties—hospitalists, surgeons, patients, and hospital leadership.
Based on the author’s experience with orthopedic comanagement
and on other hospitalists’ success ul collaborations, this chapter
will suggest speci c steps that can be taken to initiate a potential
comanagement e ort and avoid common pit alls.
ANTICIPATE THE GROWTH OF THE SPECIALTY
IN THE HOSPITAL SETTING
Musculoskeletal disorders and diseases are the leading cause o dis-
ability in the United States and account or more than one-hal o
all chronic conditions in people over 50 years o age in developed
countries. One in two Americans has a musculoskeletal condition
requiring medical attention, twice the rate o chronic heart and
lung conditions. Annual direct and indirect costs or bone and joint
health are $874 billion, 5.7% o the gross domestic product.
Based on these data, it is little wonder that orthopedic surgeons
will have increasing volumes o patient visits and operative inter-
ventions in the coming years, especially in the setting o an aging
population with increasing expectations or unctional recovery and
quality o li e. The challenge associated with this growth will be the
increasing number o medical comorbidities in these older patients
and the need or systematic evaluation o these comorbidities to
optimize the perioperative course. It is estimated that surgery-
related costs will rise 50% and surgical complications 100% in the
United States in the next two decades.
REVIEW AVAILABLE LITERATURE ON COMANAGEMENT

Early literature on orthopedic comanagement ocused on geri-
atrician collaboration with surgeons. Despite inconsistent data on
length o hospital stay and mortality, these studies and more recent
23
ones demonstrate that systematic geriatric evaluation and manage- “comanagement” relationship? Some newer hospitalists or
ment can decrease the incidence o common postoperative medical hospitalist groups may accept this as part o their growth and
complications such as congestive heart ailure, arrhythmias, venous cultivation o their practice, whereas others ear the mis-
thromboembolism (VTE), and delirium, and improve compliance sion creep to becoming a “glori ed resident” as described
P
A
with antiosteoporotic therapy and VTE prophylaxis. More recent previously. In our institution, nurses channel most rst calls
R
literature has ocused on hospitalist collaboration with orthopedics or a variety o issues to one nurse practitioner (NP) and one
T
and has shown lower adjusted length o hospital stay and decreased physician’s assistant (PA), both stationed on the dedicated
I
complication rates in some studies, although mortality and readmis- orthopedic ward. I relevant medical issues arise that require
sion rates were not changed. In one study o hip racture patients, hospitalist involvement, they will nd us on the ward and relay
delirium was diagnosed more o ten in the comanagement group, the in ormation to us. This allows us as hospitalists to ocus on
and this was associated with an earlier discharge a ter surgery. This the more sophisticated medical issues that are more consistent
T
h
may ref ect greater attention to the presence o delirium, better with our scope o practice; in addition, we can also serve as the
e
documentation, and more prompt treatment. A single-center ret- gatekeeper to urther subspecialty consultation when needed,
S
p
rospective cohort study o 501 patients who experienced at least avoiding superf uous consultations and testing.
e
one postoperative complication ound that comanagement was
c
4. How would a comanagement service a ect other relation-
i
a
associated with a lower mortality rate and a shorter length o stay, ships? For example, consultants or other medical groups who
l
t
y
suggesting that this was a “rescue” phenomenon o the medically traditionally round on their own patients may lose revenue as
o
complex surgical patient. a result o this service.
f
H
5. Does the hospitalist service have adequate sta ng to expand
o
s
services? Initially, comanagement may be limited to certain
p
ASK KEY QUESTIONS SPECIFIC TO YOUR HOSPITAL
i
groups who admit to the hospitalist service, or to speci c diag-
t
a
Comanagement o surgical (and medical subspecialty) patients
l
noses, or time o day.
M
has rapidly evolved with much initial enthusiasm. However, when
e
proposals do not clearly delineate the nature o these relationships,
d
ANALYZE THE CURRENT STRUCTURE
i
c
great potential or con usion o roles, miscommunication, subopti-
i
n
mal patient care, and dissatis action o both parties can result as the Analysis o the current structure o care delivery in surgical or medi-
e
service expands and sta ng becomes more o an issue. Institutional cal subspecialty services serves a use ul measurement both as a
a
n
support or this activity is paramount and the medical administra- baseline and a ter the intervention. In our analysis at the Cleveland
d
S
tion should be involved in these initial meetings rom the outset. Clinic, we ound that there were signi cant di erences in the deliv-
y
s
Initially, ask the ollowing questions: ery o care on medical services and orthopedic services due to the
t
e
ollowing actors:
m
1. Why are we doing this?
s
To start out, it is best to explicitly clari y the motivation • Limited supervision o medical care provided by the NP and PA.
o
or starting such a program: Are the surgeons stretched thin • Competing responsibilities o orthopedic residents provid-
f
C
between operating room time and patient care responsibilities ing backup or other providers or assisting in the operating
a
r
on the f oors and in the o ce? Are orthopedic residents more room.
e
limited by duty hour restrictions and there ore less able to ocus • Lack o internal medicine training o orthopedic residents to
on patient care on the f oors? Are there concerns with care address complicated medical issues.
quality within the standard structure o medical subspecialty • Signi cant medical comorbidities o patients requiring routine
consultation? Are nursing and ancillary sta having issues with medical surveillance to prevent, detect, and intervene during
access to practitioners or patient medical needs and issues? their hospitalization.
Do they want someone to take on the role and responsibility • Limitations o general medical consultation service that typi-
o completing histories and physicals and discharge notes and cally “reacts” to consultation requests when problems have
summaries? These are just a ew questions that might help to already been identi ed and lacks the capacity to prospectively
ocus the expectations o the proposing surgeons. a rm or develop a medical plan o care or high-risk or compli-
2. Is the hospitalist service the best solution to this problem? cated medical patients.
Once it becomes clear what your surgical or medical col-
leagues desire, then clari y how hospitalist services and
skills can (or should) address these issues. Is a hospitalist the CLARIFY ROLES AND EXPECTATIONS
best solution to this problem? Certainly hospitalists are adept At this point clari ying and documenting roles and expectations
at addressing medical issues in hospitalized patients as medi- may avoid the potential or con usion and divisiveness. Table 4-1
cal consultants, but to what degree should they assume the outlines our expectations or our hospitalist roles and responsibilities
detailed minutia o patient care (ie, acetaminophen orders, and the ongoing orthopedic roles and responsibilities. Table 4-2
renewing intravenous f uids)? This presents the potential to lists the conditions that should trigger re erral to the hospitalist
become a “glori ed resident” in the care o these patients, comanagement service.
which many hospitalists abhor. In addition, expectations or These are not rm and inf exible rules; other comanagement
surgical wound care and drain management may be tasked relationships may opt to take on some o the responsibilities listed
to hospitalists despite concerns that this may extend out- above in the orthopedic section, such as blood and f uid manage-
side the scope o practice o their internal medicine training. ment, pain management, and communication with patients and
Similar concerns have been raised when hospitalists assume amilies regarding medical issues, all o which are in a reasonable
primary responsibility or patients with intracranial bleeds or scope o hospitalist practice. However, it is critical to delineate who
neurosurgeons. will take on these tasks and to have a mechanism or resolution o
3. What other options have been considered? disagreements. Some hospitalists may wish to “specialize” in coman-
I the rst call rom nursing or a problem traditionally has agement and become more amiliar with the specialty. See Part II
gone to orthopedic residents and/or sta who now are tied up (Medical Consultation), Section V (Specialty Consultation: What the
elsewhere, is the hospitalist the next logical call in an equitable Consulting Hospitalist Needs to Know).
24
TABLE 4-1 Delineation of Hospitalist versus Orthopedics Roles and Responsibilities
C
H
Hospitalist Orthopedics
A
Con er each morning with orthopedic providers and Retain the appropriate clinical support in rastructure including residents
P
review outpatient preoperative assessment (done in and current physician’s assistant and nurse practitioner positions
T
our hospitalist-run preoperative clinic) or the medical
E
comorbidities o nonelective admissions to identi y
R
suitable patients or comanagement
4
Promptly evaluate and document indings on comanaged Per orm daily rounding, assessments, and progress notes and orders or
patients, and enter orders on these patients routine and stable medical issues including postoperative orders
C
Follow-up on tests and studies ordered by the comanager Address the “ irst call” rom nurses or questions and patient assessments,
o
per orm ull and appropriate patient assessments prior to calling medical
m
physicians or urther support
a
n
Provide ormal and in ormal preoperative and postoperative Remain as primary service or patients without substantial medical
a
g
medical consultation as requested complexity
e
m
Provide teaching to orthopedic surgery service providers on Provide night, weekend, and holiday coverage o orthopedic patients with
e
medical issues support by the medicine consult resident
n
t
Participate in daily multidisciplinary rounds with nursing, Follow-up on studies and tests ordered by the orthopedic service
o
nonphysician providers, and case management to identi y
O
patients with ongoing medical needs not otherwise
r
t
captured by the mechanisms above
h
o
Address routine postoperative management orders including:
p
e
Initiate and comply with orthopedic protocols
d
i
c
Manage blood and luids
P
a
Manage pain and routine prn medications
t
i
e
Order DVT prophylaxis and medications
n
t
s
Assess and care or wounds and order perioperative antibiotics
Admit and plan discharge, prepare orms, and provide discharge
prescriptions
Communicate with amilies and acilitate discharge planning
Periodic meetings between hospitalist and orthopedic cham- o readiness or surgery and to acilitate optimization o key medical
pions as well as nursing, other members o the comanagement conditions. Instead o being seen in the hospital or by the patient’s
team, and administration should review program unctioning and primary care provider, clinicians with more concentrated expertise
processes o care; address speci c problems in direct patient care; in assessing and preparing patients or surgery see the patient in
and urther de ne roles and expectations. In addition, this group advance and then communicate their ndings and recommenda-
may assess the impact o the comanagement service on teaching tions to the surgeon and anesthesiologist. In our case, we also com-
o medical and surgical trainees and explore uture directions. Our municate with the orthopedic comanager via e-mail, page, or shared
medicine residents rotate with our hospitalist comanager as part o EMR(electronic medical record) patient list about planned admissions
their medical consultation experience, and our relationship provides o surgical patients with relevant comorbidities. The preoperative
opportunities or creating academic value such as the potential clinic model may reduce cancellation rates and may identi y decom-
or publication o outcomes data, collaboration in research, and pensated medical problems that might lead to increased periopera-
orthopedic quality improvement projects. Table 4-3 delineates the tive morbidity and mortality, as well as identi y conditions that could
development timeline o our comanagement service. easily decompensate i not scrutinized (eg, excessive intravenous
f uids postoperatively in a patient with systolic heart ailure).
PATIENT SELECTION AND TRIAGE
Decisions about which patients should be ollowed by the coman- MEASURING SUCCESS
ager should be based on patient needs and provider capacity. De ining what constitutes success o the program in measur-
Patients with minimal or no medical comorbidities rarely bene t able terms is an essential piece o the puzzle, as it may provide
rom hospitalist input and may only serve to direct clinician e orts in ormation about practice changes, variability o practice pat-
away rom the patients that need more intensive medical attention. terns, outcome changes, and inancial bene its or risks o the
Patients with acutely decompensated problems and/or multiple relationship.
chronic comorbidities will more likely bene t rom comanagement Tab le 4-4 lists some suggested metrics to consider at the
attention; in some circumstances, admission to a medical service with outset. It would also help to obtain data on these metrics prior
orthopedic consultation may be the most appropriate path. Figure 4-1 to the initiation o comanagement to determine the in luence o
delineates the decision process or triage at our institution. the program.
Especially challenging or the modern orthopedic surgeon is that
ROLE OF A PREOPERATIVE ASSESSMENT CLINIC since August 2008 the U.S. Centers or Medicare and Medicaid
Some institutions (including ours) have an internist- or hospitalist-run Services (CMS) has included deep venous thrombosis and pulmo-
preoperative assessment clinic that works in conjunction with the sur- nary embolism a ter total knee arthroplasty and total hip arthro-
geon and anesthesiologist to provide a broad systematic evaluation plasty on the list o nonreimbursed Hospital Acquired Conditions.
25
TABLE 4-2 Conditions Triggering Referral to Comanagement Service
Chronic Medical Conditions

P
• Stable or known coronary artery disease (chest pain, shortness o breath [SOB], electrocardiogram [ECG] changes)
A
R
• Congestive heart ailure (SOB, pulmonary edema, edema, oxygen desaturation)
T
• Hypertension (especially i blood pressure > 160 systolic blood pressure or > 100 diastolic blood pressure)
I
• History o stroke
• Moderate/severe peripheral vascular disease
• Mild-moderate chronic obstructive pulmonary disease (COPD) (SOB, wheezing, oxygen desaturation)
T
• Mild-moderate/stable asthma (SOB, wheezing, oxygen desaturation)
h
e
• Current antibiotic treatment or pneumonia/acute bronchitis
S
p
• History o upper/lower GI bleed in the last 3 mo (drop in Hgb/Hct, concern or active bleeding)
e
c
• Patients on chronic enteral tube eedings or hyperalimentation/total parenteral nutrition (TPN) (in addition to nutrition team/TPN consult)
i
a
l
t
• Diabetes mellitus type 1 or 2
y
o
• Stable psychiatric illnesses including a ective disorders, dementias, bipolar disorder, schizophrenia (with additional psychiatry
f
consultation or medication concerns or decompensation o psychiatric illness)
H
o
• Chronic anticoagulation (comanagement consultation on all patients)
s
p
• Recent anticoagulation or deep vein thrombosis (DVT) or pulmonary embolism (PE) within the last 6 mo (comanagement consultation
i
t
a
on all patients and possibly vascular medicine consultation)
l
M
• Chronic immunosuppression (prednisone, cyclosporine, methotrexate, FK506, azathioprine, TNF-alpha blockers, etc)
e
d
• Physiologic glucocorticoid treatment within the last year (≥ 7.5 mg/d o prednisone, or the equivalent, or ≥ 2 wks)
i
c
i
• Medical issues that require medical evaluation, monitoring, or treatment:
n
e
Atypical chest pain without evidence o an acute coronary syndrome
a
n
Shortness o breath
d
S
Acute DVT or PE
y
s
Baseline anemia or postoperative anemia
t
e
m
Urinary tract in ection with indwelling Foley catheter
s
Acute delirium
o
f
Electrolyte disorders
C
Hyperglycemia without evidence o diabetic ketoacidosis (DKA) or nonketotic/hyperosmolar state
a
r
e
Acute renal ailure
Others
A portion o the payment made by CMS to hospitals is withheld i a One must keep in mind that not all metrics can be expected to
patient experiences deep venous thrombosis or pulmonary embo- improve in a “positive” way; in the study described previously about
lism ollowing one o these procedures. While this decision has been hip ractures, the rates o delirium were increased, which tradition-
criticized because prophylaxis is neither per ect nor risk ree, it is a ally would be perceived as a negative result. However, this was due
reality o practice, and the hospitalist comanager must be aware o to increased recognition, documentation, and treatment o delirium
this and engage the orthopedist regarding appropriate evidence- by physicians, which many would agree is a bene cial intervention.
based prophylaxis methods. Also, these individual metrics are not in a vacuum; an increase in
TABLE 4-3 CCF Program Timeline
December 2007 Presentation o the concept o the “embedded consultant” with a mini white paper summarizing the
literature regarding bene its o comanagement by the department o orthopedics to the chairs o the
Medicine Institute and department o Hospital Medicine
December 2007–March 2008 Dra t proposal resulting rom outreach to existing programs and internal multidisciplinary planning
March 2008 Acceptance o pilot program by departments o Hospital Medicine and orthopedics
March 2008 Presentation o proposed pilot to hospital operations committee and approval o the hiring o two
additional ull-time equivalent (FTE) or the program
April 2008–July 2008 Recruitment and inalization o pilot protocols with NPs
August 2008 Kicko
Ongoing Oversight with orthopedic champion
Metrics collection
Creation o a link with IMPACT (preoperative clinic)
26
De te rmine
numbe r a nd
C
H
de gre e of
A
me dica l
P
comorbiditie s
T
(Ta ble 4-2)
E
R
4
Gro up 3:
C
Gro up 2: IMPACT
o
Gro up 1: Patie nt has g e ne rate d c o ns ults
m
Gro up 4:
Patie nt has multiple c o mplic ate d (Table 4-2):
a
Ac ute o r
n
no s ig nific ant ac ute o r c hro nic me dic al Ho s pitalis t in IMPACT
a
de c o mpe ns ate d
g
c hro nic me dic al is s ue s c o nditio n(s ) o r an ac ute ide ntifie s s ig nific ant
e
c hro nic me dic al
m
are ins ig nific ant and/o r is s ue that re quire s c o mo rbiditie s and
c o nditio n(s )
e
s table me dic ine c o ns ultatio n re c o mme nds
n
t
(Table 4-2) po s to pe rative fo llow-up
o
by the c o manag e r
O
r
t
h
o
p
e
Cons ide r a dmis s ion or
Coma na ge r s e e s a nd
d
tra ns fe r to a ge ne ra l or
i
Pa tie nt informa tion
c
No involve me nt eva lua te s pa tie nt,
s ubs pe cia lty me dica l
P
unle s s eva lua tion ma ke s e nte re d in s ha re d
a
s e rvice with orthope dic
t
for pe riope ra tive re comme nda tions EMR lis t, IMPACT
i
e
cons ulta tion for
n
ris k a s s e s s me nt is a nd pla ce s provide r notifie s
t
eve ntua l s urgica l
s
re que s te d a ppropria te orde rs, coma na ge r
inte rve ntion a s
follows pa tie nt da ily
indica te d
Coma na ge r s e e s a nd
eva lua te s pa tie nt,
ma ke s
re comme nda tions
a nd pla ce s
a ppropria te orde rs,
follows pa tie nt da ily
Figure 4 1 Cleveland Clinic orthopedic comanagement triage algorithm.
length o hospital stay coupled with decreased readmission rates

TABLE 4-4 Proposed Program Metrics might ref ect the hospitalist taking an extra hal to ull day to opti-
mize certain medical conditions, which results in ewer readmissions
Volume data
or decompensation.
Case mix Early results revealed that our program provided a net cost sav-
Patient satis action ings to the orthopedic department in terms o reduced surgical
Length o stay cancellations and improved patient satis action with care delivery.
OR cancellation rates
Hospital cost and ancillary utilization CONCLUSION
Productivity measures (RVUs and billing) Comanagement o specialty patients provides a novel diversity o
Provider satis action (hospitalist, orthopedist, nursing, residents) practice to hospitalists and may be pro essional rewarding with
improved collegiality with surgical specialties, opportunities or
Mortality
leadership, and quality improvement research. Engaging in a new
Unplanned ICU or medical service trans ers comanagement relationship requires orethought and planning,
Readmission rates and clari ying expectations, responsibilities, and metrics o success.
Quality/patient sa ety metrics However, when designed and managed well, a comanagement
JCAHO core measures service can bene t hospitalists, surgeons, and, most importantly, the
patients who trust us with their care.
27
SUGGESTED READINGS Zuckerman JD, Sakales SR, Fabian DR, et al. Hip ractures in geriatric
patients. Results o an interdisciplinary hospital care program. Clin
Fisher AA, Davis MW, Rubenach SE, et al. Outcomes or older Orthopaed Relat Res. 1992;274:213-225.
patients with hip ractures: the impact o orthopedic and geriatric
P
A
medicine cocare. J Orthopaed Trauma. 2006;20(3):172-178; discus-
ONLINE RESOURCES
R
sion 9-80.
T
Ja er A, Michota F. Why perioperative medicine matters more than American Academy o Orthopedic Surgeons. The Burden of
I
ever. Clev Clin J Med. 2006:73(Suppl 1);2006:S1. Musculoskeletal Diseases in the United States: Prevalence, Societal
Marcantonio ER, Flacker JM, Wright RJ, et al. Reducing delirium and Economic Cost; 2012. http://www.boneandjointburden.org/.
a ter hip racture: a randomized trial. J Am Geriatr Soc. 2001; Accessed September 13, 2015.
T
49(5):516-522. Hospital Acquired Conditions (Present on Admission Indicator).
h
e
Sharma G, Kuo Y, Freeman J, et al. Comanagement o hospitalized http://www.cms.gov/HospitalAcqCond. Accessed September 13,
S
surgical patients by medicine physicians in the United States. Arch 2015.
p
e
Intern Med. 2010;170(4):363-368.
c
Society o Hospital Medicine. Measuring Hospitalist Performance:
i
a
Strei MB, Haut ER. The CMS ruling on venous thromboembolism Metrics, Reports and Dashboards; 2006. http://www.hospitalmedi-
l
t
y
a ter total knee or hip arthroplasty: weighing risks and bene ts. cine.org/AM/Template.c m?Section=White_Papers&Template=/
o
f
JAMA. 2009;301(10):1063-1065. CM/HTMLDisplay.c m&ContentID=14632. Accessed March 10, 2010.
H
o
s
p
i
t
a
l
M
e
d
i
c
i
n
e
a
n
d
S
y
s
t
e
m
s
o
f
C
a
r
e
28
CHAP TER
5 PROFESSIONALISM
Pro essionalism in medicine has long ocused on the tenets o
patient wel are, patient autonomy and social justice. As health care
has evolved, our de nitions and value statements have broadened.
Evidence-based medicine, quality improvement, access to care,
cost-e ective practices, and conf icts o interest are now at the
ore ront o these discussions—all o which are salient in hospital
medicine.
The 2002 jointly published Medical Pro essionalism in the New
Professionalism in Millennium: A Physician Charter articulated undamental principles
and responsibilities to which all physicians should aim to maintain.
Hospital Medicine The ABIM Foundation, The European Federation o Internal Medi-
cine, and the ACP Foundation worked together to develop a charter
that reconciled the self ess expectations o the physician with the
ever-changing landscape o health care delivery. The American
Kimberly D. Manning, MD, FACP, FAAP Academy o Pediatrics’ Committee on Bioethics ollowed with their
policy statement on pro essionalism in pediatrics, which addressed
some o the unique considerations or those caring or children.
Hospitalists are trusted with the well being o some o the most
vulnerable patients. This distinctive contract with society calls or
some variation in our emphasis when discussing pro essionalism.
Here we will de ne pro essionalism in hospital medicine, clari y
concepts most applicable to those caring or hospitalized patients,
and translate this into clinical practice. These concepts apply to all
clinicians at all levels, team members, and hospital systems caring
or patients admitted to the hospital.
WHAT DOES “PROFESSIONALISM” MEAN IN

HOSPITAL MEDICINE?
Members o any pro ession are expected to have acquired a body o
knowledge and skills speci c to their chosen eld. Through a shared
commitment, there is sel -regulation and a contract with society to
judiciously apply skills and expertise. Hospitalists, like all physicians,
have made an agreement to heal. This pact requires an establish-
ment o trust and a willingness to place patient needs above all
other considerations. Sel -regulation is administered through state
medical boards, clinical leadership, and ethical codes. Pro essional-
ism is what binds this treaty to our pro ession.
The high stakes and lack o predictability in caring or inpatients
creates unique challenges or hospitalists. Beyond the standard de ni-
tions o medical pro essionalism, special emphasis on patient sa ety,
provider interdependency and communication, nancial reimburse-
ment and patient satis action reshape our understanding and must
always be considered. Finally, with a substantial component o
undergraduate, graduate and interpro essional training taking place
in the hospital setting, the impact o hidden curricula in medical edu-
cation is arguably greatest in hospital medicine. Given this, the ripple
e ects o our behaviors are ar reaching. Pro essionalism or the
hospitalist translates to more than outcomes or patients—it has the
potential to de ne generational culture in medicine.
PROFESSIONALISM AND PATIENT SAFETY

Sa ety or hospitalized patient relies upon a collaborative work
climate, teamwork, and e ective communication. Intimidating and
disruptive behaviors have been linked to increases in medical errors,
poor patient satis action, cost o care and higher rates o attrition
or sta . Behaviors that threaten the per ormance o the health care
team create signi cant barriers to quality.
29
In 2008, the Joint Commission released a sentinel event alert Additionally, rom this study we glean that certain job and provider
addressing behaviors that undermine a culture o sa ety. From this characteristics increased this likelihood.
came new requirements: Hospitalists with less clinical time were more likely to make un
o other colleagues. It is postulated that those with heavier clinical
P
1. Every hospital or patient care organization should have a code
A
o conduct de ning acceptable and disruptive behaviors. loads have more opportunities to orm relationships and are thus
R
2. Hospital leadership must have processes in place or managing more apt to avoid unf attering commentary o others. Moreover,
T
inappropriate or disruptive behaviors. more clinical time could also cause hospitalists to be less easily inf u-
I
enced by the opinions o others. Night work and age also seemed to
Disruptive behaviors have been described as a spectrum o
be actors in unpro essional behaviors. Those working at night elt
actions such as egregious verbal and physical actions to subtle
more pressure to wrap up work. Hasty handovers and celebration
re usals to cooperate within a system. Speci cally, this could mani-
o intercepted admissions are just a ew things that could happen
T
est as reluctance to answer phone calls or pages, condescending
h
as a result.
intonation with nurses, colleagues and patients and impatience
e
Given the understanding that unique characteristics increase the
S
with queries. These behaviors are not rare in health care organiza-
p
incidence o unpro essional behaviors, special attention should be
e
tions. In a survey conducted by the Institution o Sa e Medication
given to more junior hospitalists as well as those with night shi ts,
c
i
Practices on intimidation, as many as 40% o clinicians reported
a
limited clinical duties due to competing priorities, and at sites known
l
t
keeping quiet or remaining passive rather than con ront an intimi-
y
or heavier workloads. E orts to better understand the reasons or
o
dating person. Other surveys have revealed that such behaviors
participating in these behaviors should continue to be explored.
f
are not limited to one gender or a speci c discipline. In hospital
H
o
medicine, which relies upon requent transitions o care and heavy
s
p
collaboration with nurses and colleagues, this can pose signi cant PROFESSIONALISM AND PATIENT SATISFACTION
i
t
a
threats to patients. Patient satis action is linked to reimbursements and is a key actor
l
M
The standard in hospital organizations should also include (but in determining quality o care. Interestingly, the technical dimen-
e
are not limited to) these additional recommendations rom the Joint sions o quality o care are rarely called into question with regards
d
i
Commission: to hospitalized patients. Instead, perception o quality or patients is
c
i
n
1. Education o all team members de ned by the organization’s most connected to what is personally valued by patients and their
e
amilies. Among those things being treated with respect and dignity
a
code o conduct with particular emphasis on respect.
n
2. Accountability o all team members to model desirable behav- coupled with clear communication around treatment decisions are
d
areas that have been identi ed to be o paramount importance to
S
iors regardless o hierarchal rank with equitable consequences
y
patients. Pro essional behaviors are generally the de ning eatures
s
and rein orcement.
t
e
3. “Zero tolerance” or intimidating and/or disruptive behaviors o these perceptions.
m
with incorporation o speci c verbiage into by-laws and Surveys have been developed to best ascertain patient satis ac-
s
o
employee agreements. tion in hospitals. The Hospital Consumer Assessment o Healthcare
f
Provider and Systems (HCAHPS) is now the core metric against
C
4. Protection o those who report witnessing or experiencing
a
which health care systems are evaluated. As the rst national, stan-
r
disruptive behaviors through nonretaliation clauses in policy
e
statements. dardized and publicly reported survey o patients’ viewpoints on
5. Timely and thorough responses to all patients and/or amilies hospital care, HCAHPS now serves as a tool or valid comparisons
with emphasis on acknowledgement, empathy, and apology. between hospitals locally, regionally and nationally.
6. Development o organizational processes that solicit input A random sample o patients receives this survey between 48 hours
rom a broad representation o team members when address- and 6 weeks a ter discharge. Among the 32 questions about their
ing intimidating and disruptive behaviors including encour- hospitalization, patients are asked to recall their perceptions o phy-
agement o ongoing interpro essional dialogue. sician and nurse communication, responsiveness o hospital sta ,
7. O ering training or all leaders and managers o hospital communication about medications and care transition. Though
groups to build skills in team building, conf ict resolution, and HCAHPS data are not limited to hospitalists, as the “ ace” o the
eedback on unpro essional behavior. hospitalization or many patients their impact can be substantial.
8. Assessment o the working climate to assess perceptions o Special attention to communication skills training, strategic plan-
unpro essional behaviors and risk o harm to patients. ning around availability to patients and enhanced discharge paper-
9. Provision o anonymous and con dential surveillance and work and medication reconciliation are just a ew measures that
reporting systems or detecting unpro essional behavior. have been shown to improve potential areas o weakness. Other
10. Clear documentation o all attempts to address intimidating simple tips have been o ered to enhance patient satis action and
and disruptive behaviors with outcomes. experience include:
• Make a positive patient experience a part o the culture.
• Adopt patient-centered, multidisciplinary rounds with trans-
PROFESSIONALISM ISSUES AMONG HOSPITALISTS parent discussions that include patients.
Hospitalists strive to provide uninterrupted care to hospitalized • Avoid giving o cues o indi erence or uncaring (eg, rushing,
patients. Given this, transitions o care and responsibility are depen- lack o eye contact, standing instead o sitting).
dent upon respect ul interactions between providers. Provider inter- • Encourage every employee to think about purpose and people,
dependency also plays a role in patient satis action, burnout risk, not just tasks.
and quality. Healthy relationships and working climates are essential • Equip every team member with skills or handling patient or
though challenging to maintain. amily member complaints.
In a study published in the Journal o Hospital Medicine, Reddy • Recognize that system inadequacies can be responsible or
and colleagues looked at participation in unpro essional behav- individual lapses in pro essionalism. Unreasonable scheduling
iors among internal medicine hospitalists. They ound that actors algorithms and workloads are examples o system issues that
most likely to underlie such behaviors were: making un o others, can undermine pro essional behavior.
learning environment, workload management and time pressure. • Collect and act on data related to patient satis action.
30
Hospitalists are poised to have great impact in all o these areas certain prescribing and pro essional behavior. Institutional policies
through their continuum o contact throughout the hospitalization. are in place at many hospitals regarding disclosure o conf icts o
C
Intentional approaches to both individual behaviors and the sys- interest. This continues to be an important discussion at the level o
H
tems in which hospitalists work are essential. policy and education, as well as among pro essional organizations
A
such as Society o Hospital Medicine (SHM), American College o
P
PROFESSIONALISM AND THE ACADEMIC ENVIRONMENT Physicians (ACP), and Society o General Internal Medicine (SGIM).
T
E
The Accreditation Council o Graduate Medical Education (ACGME)
R
and the Liaison Committee on Medical Education (LCME) have ■ PROFESSIONALISM AND SOCIAL MEDIA
5
identi ed pro essionalism as a crucial competency or developing Social media has rede ned the inter ace between physicians,
physicians and have labored to build program requirements around patients, learners, and the public. The once passive viewing o
this expectation. For example, the LCME uses this verbiage in their
P
content on early iterations o the websites has exploded over the
r
accreditation data collection tools or medical schools:
o
last decade to include interactive exchanging, sharing and com-
e
A medical school ensures that the learning environment o munication in real time. Facebook, Twitter, Instagram, and LinkedIn
s
s
are among the most widely used social media plat orms, all made
i
its medical education program is conducive to the ongo-
o
n
ing development o explicit and appropriate pro essional even more accessible by recent advances in mobile smartphone
a
l
technology and their near universal use by the public. As o 2014,
i
behaviors in its medical students, aculty, and sta at all
s
m
locations and is one in which all individuals are treated with it was estimated that more than 1 billion people were on Facebook
i
and, o these individuals, nearly 400 million only log in using mobile
n
respect. The medical school and its clinical a liates share the
smartphones. Tablets, Internet ready televisions and even smart
H
responsibility or periodic evaluation o the learning environ-
o
ment in order to identi y positive and negative inf uences watches, all with capability to access these same social media plat-
s
p
on the maintenance o pro essional standards, develop and orms, have also emerged in very recent years. These developments,
i
t
a
conduct appropriate strategies to enhance positive and miti- along with more unrestricted Wi-Fi “hotspots” (even in most hospi-
l
M
gate negative inf uences, and identi y and promptly correct tals) and robust search engines using collective intelligence tools,
e
violations o pro essional standards. have closed the gaps that once stood between the pro essional and
d
i
private personas o physicians.
c
i
The hospital setting accounts or a substantial proportion o
n
Though social media presents new challenges to the landscape
e
learning environments in undergraduate and graduate medical o medical pro essionalism, it also presents opportunities. Speci c
education. Given this act, the importance o pro essionalism must health in ormation can be quickly disseminated to large groups
be emphasized or hospitalists through institutional e orts and o individuals in moments. Plat orms aimed toward health pro es-
programs. sionals such as Doximity o er encrypted, HIPAA compliant privacy
Medical education literature in orms us that learners report that settings or idea sharing, consultation, and collaboration in ways
values exhibited by their teachers and institutions directly impact previously un oreseen. Moreover, Twitter use at pro essional meet-
their own pro essionalism. Regrettably, a majority o medical students ings keeps attendees and allows those unable to leave the hospital
endorse witnessing peers and supervising physicians speaking to stay abreast o ongoing activities as they are happening. In less
o patients and/or other health care pro essionals in unf attering, than hal o a decade this culture o “tweeting” pro essional meet-
cynical or derogatory ways. Multiple studies have shown that the ings in real time has become not only a welcomed eature but an
humanistic attitudes essential to patient-centered care unravel dur- expected one as well.
ing medical school and residency, increasing the risk o burnout and A growing number o academic institutions and organizations
depression. have begun to develop policies about judicious use o social media,
Hospital medicine provides much o the backdrop or this “hid- most o which emphasize con dentiality. Pro essional development
den curriculum,” originally described by Ha erty as “a set o inf u- centered on social media is becoming more requent on meeting
ences that unction at the level o organizational structure and culture.” agendas. From concrete tips on privacy settings to abstract ways to
There can be pro ound dichotomy between the negative behaviors seize collaborative opportunities, hospitalists and other health care
exhibited by some physician role models and the desired attributes providers should all gain some pro ciency with social media and its
described in ormal curricula. The insidious nature o a hidden cur- potential impact on the care o patients.
riculum is power ul enough to become institutional culture. Newer
mandates through accreditation organizations call or more honest CONCLUSION
appraisals o the in ormal inf uences on learners. They also charge
Hospitalists, with their large imprint in medical education and con-
institutions with building ongoing strategies to reshape and miti-
tract to care or a diverse and vulnerable patient population, must
gate their untoward e ects.
approach pro essionalism with intentionality. New considerations
NEW CONSIDERATIONS such as direct pharmaceutical marketing and social media have
shi ted prior understandings o what it means to be a medical pro-
■ PROFESSIONALISM AND DIRECT PHARMACEUTICAL essional. Thought ul consideration o pro essional behaviors along
MARKETING with their impact on patient sa ety, patient satis action, and the hid-
The interaction between physicians and the pharmaceutical indus- den curriculum should guide ongoing individual, institutional and
try and its sales representatives remains controversial. Though organizational development.
lessening in requency, direct marketing (usually trade name drugs)
promoted to providers and groups through gi ts, sponsored meals SUGGESTED READINGS
and lectures, travel and symposia are not oreign to most physicians.
Though clinicians do not typically report that such interactions Gholami-Kordkheili F, Wild V, Strech D. The impact o social media
inf uence their prescribing practices, evidence supports otherwise. on medical pro essionalism: asystematic qualitative review o
A large meta-analysis o over 500 studies revealed that attending challenges and opportunities. J Med Internet Res. 2013;15(8):e184.
sponsored CME events, accepting unding or travel, and partici- Ha erty FW. Beyond curriculum re orm: con ronting medicine’s
pating in pharmaceutical sponsored lectures were associated with hidden curriculum. Acad Med. 1998;73(4):403-407.
31
Joint Commission: Behaviors that undermine a culture o sa ety. in the new millennium: a physician charter. Ann Intern Med.
Sentinel Event Alert. vol. 40. http://www.jointcommission.org/ 2002;136:243-246.
sentinel_event_alert_issue_40_behaviors_that_undermine_a_ Rosenstein AH, O’Daniel M. A survey o the impact o disruptive
culture_o _sa ety/. Accessed July 9, 2008.
P
behavior and communication de ects on patient sa ety. Joint
A
Kirk LM. Pro essionalism in medicine: de nitions and considerations Comm J Qual Patient Sa . 2008;34(8):464-471.
R
or teaching. Proc (Bayl Univ Med Cent). 2007;20(1):13-16.
T
Swick HM. Toward a normative de nition o medical pro essional-
Project o the ABIM Foundation, ACP–ASIM Foundation, and ism. Acad Med. 2000;75(6):612-616.
I
European Federation o Internal Medicine. Medical pro essionalism
T
h
e
S
p
e
c
i
a
l
t
y
o
f
H
o
s
p
i
t
a
l
M
e
d
i
c
i
n
e
a
n
d
S
y
s
t
e
m
s
o
f
C
a
r
e
32
CHAP TER
6 Physicians are requently called upon to take on leadership roles.
These roles can come in various orms, ranging rom academic
leadership roles (eg, division or department chie or chair) to educa-
tional leadership roles (eg, clerkship or residency program director)
to leadership roles in a practice setting (eg, director o a practice
group). Although some o the desired skills and competencies or
the leader may be speci c to particular roles and responsibilities,
others are more generic and applicable to any o these leadership
positions. This chapter concentrates initially on the generic aspects
Principles o o leadership and concludes by discussing some o the challenges

that are more speci c to hospitalists and to the hospital environ-
Leadership
ment. In addition, instead o trying to review the voluminous leader-
ship literature, presented here will be a personal perspective, based
upon experiences in a variety o leadership positions over many
years. Discussion o leadership will be divided into our primary
components: the personal attributes that a leader should demon-
Steven Weinberger, MD, FACP strate, the skills that should be acquired, a suggested approach to
reach a goal, and leadership challenges or hospitalists in the hospital
environment.
LEADERS VERSUS MANAGERS

Be ore considering the important attributes o a leader, it is worth-
while to understand the distinction between a leader and a man-
ager. Much has been written about these di erences, which can be
readily summarized and understood by any o several descriptions
or aphorisms:
• Leaders have ollowers; managers have subordinates. Individu-
als voluntarily ollow a leader because o the qualities o the
leader; subordinates work or managers because o the report-
ing relationship and the organizational authority vested in the
manager.
• Leaders lead people; managers manage tasks.
• Leadership is doing the right thing; management is doing
things right.
• Managers ocus on tactics and tasks; leaders ocus on strategy
and direction.
In act, however, these distinctions o ten blur in the setting o
actual roles and responsibilities in the workplace. The individuals
who are most success ul in assuming roles with greater authority
and responsibility are those who are both e ective leaders and
e ective managers. A leader who does not have good management
skills can generate visionary ideas but will be unable to implement
or operationalize them. A manager who does not have good lead-
ership skills will be unable to mobilize and motivate a supportive
team.
Some activities and responsibilities o a physician leading a group
o hospitalists can readily illustrate the di erences between lead-
ership and management. “Managing” the group means assuring
that the patients are covered, that transitions o care are e ectively
handled, that chart and billing documentation is complete and
accurate, and that teaching responsibilities are assigned and well
integrated with patient care responsibilities. In contrast, “leading”
the group means exploring and developing ideas or improving the
system and its productivity, improving quality o care, developing
the skills o the team, and acilitating the pro essional development
o the team members.
For the purposes o this chapter, I will primarily use the terms
“leadership” and “leadership skills,” recognizing, however, that we
33
are really considering both leadership and management skills. In respect and share the values o their colleagues and the leader, is
the medical leadership positions that are likely to be assumed by not an enjoyable workplace. It is also one that will never reach its
readers o this chapter, success will hinge upon both leadership and true potential.
managerial qualities and the importance o each in rein orcing the
P
A
other. There ore, the approach here to discussing leadership and
LEADERSHIP SKILLS
R
management qualities will be one o lumping rather than splitting.
T
An important prerequisite or a success ul leader is a high level o
I
competence in the particular eld related to the leadership role. For
DESIRABLE ATTRIBUTES OF THE LEADER example, an educational leader must be accomplished as a teacher
A leader must demonstrate pro essional integrity, including hon- and educator in order to command the necessary respect o trainees
esty. High standards o integrity and honesty are a prerequisite or and other educational colleagues. Similarly, a clinical leader must be
T
highly regarded as an excellent clinician in order to have credibility
h
obtaining the respect o colleagues, superiors, and subordinates.
e
The leader sets the model o behavior or the rest o the team, and with other clinicians.
S
Another important skill that leaders must acquire is the ability
p
lack o pro essional integrity exhibited by the leader will soon be
e
mirrored by cracks in pro essionalism among others. to negotiate e ectively. This topic is covered extensively in a later
c
i
a
The leader needs to be an e ective communicator; openness chapter (Chapter 26, Negotiation and Con ict Resolution), but it
l
t
y
and transparency in communicating to all constituencies assure is important to stress that leaders must exercise their negotiation
o
that everyone is on the same page. A commonly held perception skills in many settings—when dealing with superiors, dealing with
f
H
o a talented leader is o ten someone who can communicate both subordinates, or dealing with third parties with whom there is no
o
reporting relationship. Physicians typically have not been trained
s
values and vision, including a set o goals and how those goals
p
in negotiation, but ortunately a number o excellent and readily
i
might be achieved. However, that is only part o the communica-
t
a
accessible resources can provide valuable guidance to the previ-
l
tion equation, which also involves establishing and transmitting
M
expectations or others. It is critical that subordinates, trainees, and ously untrained leader.
e
Many leaders have responsibility or budgets, and a working
d
team members understand the expectations being placed on them,
i
c
including how and on what basis they are being judged. amiliarity with balance sheets and with revenue and expense state-
i
n
ments is there ore use ul. Although providing such nancial training
e
Communication must also occur on a two-way street, that is, the
a
leader must be an excellent listener as well. I the leader is unable or is beyond the scope o this chapter, a valuable resource that can
n
provide basic training in the principles o accounting is a short, easy
d
unwilling to hear what others have to say, he or she will be doomed
S
to ailure. Without ideas and eedback rom others, the leader will to read, programmed text used in many business schools.
y
s
invariably make mistakes o both commission and omission that can Finally, the responsibility o running meetings o ten accompanies
t
e
leadership roles. Everyone has participated in meetings that run
m
be avoided by hearing the ideas and opinions o others and consid-
e ectively, where the participants eel they have not wasted their
s
ering all perspectives when making important decisions.
o
Another important aspect o communication relates to eed- time, and they leave the meeting with a well-de ned action plan.
f
C
back. The e ective leader provides eedback in a constructive, On the other hand, everyone has also participated in meetings that
a
r
pro essional manner. This eedback must be based on established are poorly organized, do not make best use o participants’time, and
e
expectations, not on the subordinate’s or trainee’s ability to read the do not have a well-de ned purpose or outcome. Several important
leader’s mind and guess what s/he wants. In addition, the eedback aspects o well-designed meetings include:
should be done in a way that is ormative, that is, giving the receiver • Sending an advance agenda to the participants, so that
o the eedback an opportunity to, and advice on how to improve they can prepare appropriately.
per ormance. Providing only summative eedback at the end o a • Starting on time and f nishing on time (or early). Time is
responsibility or task allows no room or improvement and o ten an incredibly valuable commodity, and meeting participants
proves rustrating to the person receiving the eedback. become distracted and resent ul when time is wasted at the
An e ective leader remembers that success is determined by the beginning o a meeting or when a meeting runs overtime and
contributions or development o the people s/he leads or trains. potentially a ects their subsequent commitments.
The leader must acknowledge the contributions o others and not • Assuring that the meeting is interactive and makes good
always take credit or success. Nothing is as demoralizing to a team use o the participants’ time and expertise. A success ul
member as the eeling that his contributions are not being recog- meeting is not a monologue provided by the meeting coordi-
nized and that someone else is taking credit or his work or accom- nator. Rather, it involves active engagement and participation
plishments. At the same time, a success ul leader is cognizant o , and by all attendees, so that the participants eel they have contrib-
aims to promote the pro essional development o the individuals or uted to the meeting and have not just been wasting their time.
which she is responsible. For example, true success or an academic • Wrapping up the meeting with a summary and a well-
leader is o ten determined as much by the ultimate careers o the def ned action plan. Everyone should understand the out-
individuals trained by the leader, as it is by the academic contribu- come o the meeting and the expected action plan, including
tions o the leader himsel . Supporting the pro essional develop- the assignments that have been meted out to individuals and
ment o one’s trainees is one o the most important and enduring the timeline according to which they should be completed.
legacies that a leader can leave to the pro ession.
Finally, an intangible but critical quality o a success ul leader is
the ability to create and maintain a positive work environment. The REACHING A GOAL
leader needs to establish a workplace tone that is positive, in which In trying to generate ideas and complete desired tasks necessary to
people eel they are supported and a “can do” attitude prevails. reach a goal, the success ul leader may nd it help ul to use a set o
Productivity o the individuals or whom a leader is responsible is principles: (1) establish the goal, (2) identi y and include stakehold-
dependent upon their interest in, and commitment to the team ers, (3) assemble a team, (4) engage all team members to gener-
and to the shared goals and vision that have been de ned by the ate ideas, (5) accept ideas rom outside the team or organization,
leader. A setting in which individuals are competitive with each (6) delegate responsibility, and (7) assess interim outcomes and
other, where back-biting is common, and people eel they do not reassess the plan.
34
At the outset, the leader needs to establish the goal and communi- physicians, hospitalists are constantly working with nonphysician
cate it to others. In other words, the leader must de ne the destina- personnel and administrators whose management and reporting
C
tion be ore anyone can plot the route o how to get there. Success structure is quite independent o the physicians at the institution.
H
in achieving a goal depends critically upon the process o engaging A hospitalist leader who is trying to e ect change but is not part o
A
others at least as much as the goal itsel . This process involves iden- the hospital’s administrative hierarchy may have dif culty shaping
P
tifying stakeholders who are either impacted by the goal, or who will opinions and getting buy-in rom a group o nursing leaders or rom
T
E
eel disen ranchised i they have not been included in developing nonphysician hospital administrators. Even among physician leaders
R
or working to achieve the goal. For example, i a residency program at the hospital, challenges arise o ten centered around a competi-
6
director needs to restructure a schedule or the residents based tion or resources, so that the interpersonal relationships become
upon changing regulatory requirements, s/he is much more likely adversarial rather than cooperative.
to be success ul when including residents in developing the new An additional challenge con ronted by hospitalists stems rom
P
model, rather than imposing it upon them without their input. their basic demographic characteristics. Age, gender and even race
r
i
n
No matter how sound an idea or plan, its acceptance and imple- could impact abilities to break into a hospital hierarchy that tends to
c
i
p
mentation can be obstructed by those who were not included or be older, male, and Caucasian. A young physician who goes on sta
l
e
engaged—but eel they should have been—in its generation and as a hospitalist at the institution where s/he completed residency
s
o
development. may nd it dif cult to shake the image o being a resident rather
In making plans or a project or reaching a goal, the leader o ten than a sta member and colleague. On the other hand, a young
L
e
needs to establish a team o individuals who can work together. hospitalist who takes a position at an institution where s/he did not
a
d
Assembling the right people is critical or a success ul outcome. train may nd it dif cult to parachute in as a newcomer un amiliar
e
r
s
Team members need to be chosen based upon their skills, their with a particular hospital’s culture and personalities.
h
i
interest in and enthusiasm or the project, their ability to work well Although there are no proven methods to overcome such chal-
p
with others on the team, and their openness in providing ideas and lenges, some strategies may be help ul. First, it is extremely valuable
eedback about how things are going. to obtain the trust and support o a more senior, well-respected
Once the goal is established, stakeholders are identi ed, and the person, ideally a current physician leader at the institution. Such
team is assembled, the leader must recognize that s/he does not an individual not only can serve as a mentor and advisor to guide
need to generate all the ideas. Great work is typically done in teams the hospitalist in charting a path through un amiliar territory, but
in which everyone is encouraged to share ideas, no matter how crazy s/he can also smooth the way or the young hospitalist to become
or ar- etched they may initially seem. In addition, not all good ideas accepted by the more established hospital hierarchy. For example,
need to come rom within the team. One can adopt and build upon the support and trust rom a well-respected division chie or depart-
ideas and success ul initiatives that have been developed by indi- ment chair can be invaluable in easing the way or a hospitalist to
viduals outside the organization. The “not invented here” attitude deal with an older, potentially intimidating chie o surgery or
o ten precludes an open mind to accepting ideas that have worked hospital’s chie operating of cer.
success ully in other settings. Second, as mentioned earlier in this chapter, it is critical or any
The ideas necessary to reach a goal are o ten not grand and clinician leader, particularly a hospitalist leader, to be viewed by
sweeping ones. A series o small steps, each o which can be judged both physician and nonphysician sta as an outstanding clinician.
and modi ed as necessary based upon the outcome achieved, is It is very dif cult to have credibility in the hospital environment,
o ten more success ul than a single, revolutionary idea that does particularly rom physician colleagues and rom nursing sta , i one
not allow or opportunities to provide mid-course assessment and is viewed as a “clinical lightweight.” The hospitalist’s conscientious-
correction. ness, clinical skills, decision-making ability, communications skills,
In making the best use o the team members, the leader must and pro essionalism all contribute to the individual’s reputation and
be willing to delegate responsibility appropriately. From the time o ability to command respect rom others at the institution.
their training, physicians are o ten used to eeling that they need to Third, when trying to e ect change and garner support rom
take ull responsibility or a patient, and this attitude o individual both physician and nonphysician sta , the hospitalist leaders must
responsibility and accountability should o ten be modi ed when initially establish, promote, and ocus on the principles underlying
one assumes leadership responsibilities. Members o a team work any proposed plan. Although it is easy or a hospital administrator
best when they eel that responsibility has been bestowed upon to argue with a speci c proposal, it is much more dif cult to take
them. Delegating responsibility is not a sign o weakness; rather, a position against an ultimate goal o improving the quality and
appropriate delegation demonstrates an understanding o how sa ety o patient care, improving hospital systems and ef ciency, or
to share responsibility, engage others, make best use o available improving the nancial per ormance o the hospital.
resources, and capitalize on each person’s strengths. Finally, it is important or the hospitalist to seek out de ned
As a project progresses, the leader must critically assess interim leadership roles. Such roles can obviously be within the hospital
outcomes. Based upon these outcomes, the leader must be will- community, or example, by serving on committees. However,
ing to reassess the plan and adjust accordingly. The leader and establishing a presence and reputation outside the institution, or
the team members should also recognize that not all plans will example, through involvement at regional and national levels, can
be success ul. A plan that does not succeed should not neces- only help the hospitalist’s reputation and credibility within her own
sarily be viewed as a ailure. Important lessons are o ten learned hospital setting.
and new ideas generated based upon unanticipated problems or
unexpected results.
FINAL WORDS
The concept o being a “born leader” has clearly given way to a
LEADERSHIP CHALLENGES IN THE HOSPITAL philosophy that leadership skills can be learned. It is perhaps based
ENVIRONMENT on this premise that so many books and articles have been writ-
Hospitalists who are in leadership positions or who are expected to ten about every possible aspect o leadership. Yet, it is air to say
e ect change are o ten con ronted with challenges that arise speci - that many personal qualities and aspects o personality do have an
cally rom working in the hospital environment. Besides dealing with impact on potential success as a leader. When placed in a position
35
with leadership responsibilities, it is valuable to take some time to REFERENCES
sel -re ect upon personality traits and how they will likely in uence
leadership style. In addition, one should try to assess what additional Breitner LK, Anthony RN. Essentials of Accounting. 11th ed. Upper
skills s/he needs to acquire to be an e ective leader. Even though Saddle River, NJ, Prentice Hall, 2012.
P
A
physicians are o ten placed in either clinical or academic leadership Fisher R, Ury W, Patton B. Getting to Yes: Negotiating Agreement
R
positions, they have not typically received leadership training. Rec- without Giving. New York, NY: Penguin Books; 2011. Updated and
T
ognizing the interplay between personal style and leadership skills, revised ed.
I
and acknowledging the importance o sel -re ection on successes
Shell GR. Bargaining for Advantage: Negotiation Strategies for Reason-
and ailures as a leader will serve to make one an increasingly e ec-
able People. 2nd ed. New York, NY: Penguin Books; 2006.
tive leader over time.
T
h
e
S
p
e
c
i
a
l
t
y
o
f
H
o
s
p
i
t
a
l
M
e
d
i
c
i
n
e
a
n
d
S
y
s
t
e
m
s
o
f
C
a
r
e
36
SECTION 2
Critical Decision Making at
the Point of Care
37
CHAP TER
7 INTRODUCTION
■ A BRIEF HISTORY
The March 1, 1981 issue o the Canadian Medical Association Journal
included a landmark article titled “How to read clinical journals: I.
Why to read them and how to start reading them critically.” Writ-
ten by David Sackett, MD (1934–2015) o McMaster University, it
introduced a series o articles that highlighted the importance o
critical appraisal o the literature. Starting in 1993, a set o articles in
Principles o the Journal of the American Medical Association titled “Users’ guides
to the medical literature” reprised and expanded on the earlier
Evidence-Based series. These works, and other e orts by their authors, made criti-
cal appraisal o the literature accessible to the masses and laid the
groundwork or evidence-based medicine (EBM).
Medicine and Quality Gordon Guyatt, MD, coined the term “evidence-based medi-
cine” in the early 1990s, while he served as the internal medicine
o Evidence residency program director at McMaster University. Dr. Guyatt and
colleagues had incorporated critical appraisal o the literature into
the residency program curriculum, and Dr. Guyatt wanted a term to
describe and advertise their e orts.
Daniel I. Steinberg, MD EBM caught on quickly over subsequent years as practicing physi-
cians and training programs embraced and taught its methods, with
dissemination greatly ueled by the rise o the Internet.
■ ROLE OF CLINICAL JUDGMENT AND PATIENT

PREFERENCES IN EBM
An early criticism o EBM, which some still harbor, was that it did
not properly acknowledge the importance o clinical judgment or
patient pre erences. In an updated ramework or evidence-based
practice by R. Brian Haynes, P.J. Devereaux, and Gordon Guyatt in
2002, evidence-based decisions are based on our cardinal ele-
ments: (1) the research evidence, (2) the patient’s clinical state and
circumstances, (3) the patient’s pre erences, and (4) the clinician’s
judgment and expertise.
PRACTICE POINT
• Clinical judgment and expertise are essential to the practice o
EBM. These skills acilitate optimal decision making by allowing
the clinician to properly weigh the research evidence in the
context o the patient’s individual clinical circumstances and
pre erences. Decisions should never be based on the evidence
alone.
Practicing EBM may appear to be a straight orward a air with

its methodical approaches to clinical question construction and
to searching and critically appraising the literature. However,
hospitalists should not con use process with content, and they
will o ten nd that EBM tends to highlight clinical uncertainty and
gaps in the medical literature. High-quality evidence does not exist
to guide all clinical decisions, and extrapolation rom lower quality
evidence is o ten necessary. Bayesian diagnostic decision making
o ten relies on clinical judgment to ormulate pretest probabilities
or to deal with the uncertainty that accompanies inconclusive
post-test probabilities. Learning to deal with uncertainty is a core
competency o EBM, which draws heavily on clinical judgment
and experience.
39
STAYING UP TO DATE WITH THE LITERATURE PRACTICE POINT
■ PUSH INFORMATION RESOURCES • Hospitalists should strongly consider using an e-mail-based
alerting service or RSS (Rich Site Summary) eed aggregator
P
Few clinicians have the time to consistently read medical jour-
rom a high-quality evidence-based summary push resource to
A
nals, identi y relevant new research and critically appraise new
R
e ectively stay up to date on the literature.
studies to determine i they should be incorporated into one’s
T
practice. “Push” in ormation resources are resources that send
• Hospitalists can pair a virtual le cabinet with these resources
I
to orm an e ective in ormation management system that will
content out to their users on a regular basis. “Pull” in ormation
make evidence readily available at the point o care.
resources are databases that clinicians search in order to answer
a clinical question. Pull resources are discussed later in this chap-
ter. Tab le 7-1 lists selected high-quality push and pull in orma-
T
h
tion resources. ■ KEEPING INFORMATION AT HAND: THE
e
McMaster PLUS (Premium Literature Service) continuously VIRTUAL FILE CABINET
S
p
searches over 120 medical journals and selects evidence or critical Although the traditional way o storing articles or later re erence
e
c
appraisal. Articles that pass the critical appraisal process and are also is to use a physical le cabinet, this approach has a number o
i
a
l
rated as clinically relevant and newsworthy by their team o review- disadvantages. File cabinets are not mobile, they cannot be quickly
t
y
ers are then trans erred to the PLUS database. The PLUS database searched or updated, and determining how to best le something
o
f
contributes content to evidence-based summary resources such as or easy retrieval can be con using. A clinician might ask himsel /
H
EvidenceUpdates, ACP JournalWise, DynaMed, and ClinicalEvidence.
o
hersel in rustration: “Did I le that great article on pulmonary mani-
s
These resources all o er e-mail alerts to users. ACP Journal Club
p
estations o HIV under ‘HIV,’ or ‘in ectious disease,’ or ‘pulmonary’?”
i
t
and NEJM Journal Watch critically appraise and produce synopses They do not o er a way to electronically add content to them or
a
l
o high-quality evidence accompanied by expert commentary. electronically share content with others, making them incompatible
M
PubMed, through its ree account service “My NCBI,” allows users to
e
with modern communication methods such as e-mail.
d
receive the results o literature search strategies they either design The virtual le cabinet (VFC) is an Internet cloud-based electronic
i
c
i
or select (via the “Clinical Queries” eature) by e-mail on a regular document storage system that synchronizes across multiple elec-
n
e
basis. tronic devices (eg, smartphone, tablet, laptop computer). A VFC is an
a
n
e ective way or hospitalists to electronically le articles they receive
d
rom a push in ormation resource as described above or easy
S
y
retrieval at the point o care. Box, Dropbox, Evernote, and Google
s
t
Drive are some examples o the commercial products that currently
e
m
TABLE 7-1 High-Quality Push and Pull Resources exist that can be used as a virtual le cabinet. Products such as these
s
also o er easy options or electronically sharing content with others.
o
f
Push Resources Pull Resources
C
THE EBM PROCESS: ASKING AND ANSWERING
a
Resources that automatically Databases that are searched
r
CLINICAL QUESTIONS
e
send new, high-quality as needed to answer clinical
evidence to users via e-mail or questions Practicing EBM o ten involves asking and answering questions that
RSS eed aggregator arise during the care o patients. There are our steps in this process:
ACP JournalWise ACP Journal Club (1) asking a ocused clinical question, (2) searching the literature or
the best available evidence, (3) critically appraising the literature,
http://journalwise.acponline. http://annals.org/journalclub.
and (4) applying the literature to an individual patient. This chapter
org aspx
explores the basic principles o EBM as they relate to these our steps.
BMJ Clinical Evidence BMJ Clinical Evidence
http://clinicalevidence.bmj. http://clinicalevidence.bmj. ■ STEP 1: ASKING A FOCUSED CLINICAL QUESTION
com com Clinical questions all into two general groups: background or
DynaMed Cochrane Collaboration oreground questions. Background questions ask about general
https://www.dynamed.com http://www.cochrane.org knowledge, pathophysiology, epidemiology, and broad aspects o
diagnosis and treatment. “What are the treatments or epilepsy?”
Evidence Updates DynaMed is an example o a background question. Junior learners o ten ask
https://plus.mcmaster.ca/ https://www.dynamed.com background questions, and answers can o ten be ound in text-
evidenceupdates books. Foreground questions are more ocused, address speci c
NEJM Journal Watch NEJM Journal Watch clinical situations, and acilitate the delivery o the most up-to-date,
http://www.jwatch.org http://www.jwatch.org evidence-based care. Experienced clinicians ask oreground ques-
tions, with answers residing more in the medical literature. Hospital-
PubMed (using an My NCBI Practice Guidelines rom ists should always aim to construct ocused oreground questions.
account and search strategies pro essional societies, eg,
These are urther discussed below.
created by the user, see text) AHRQ National Guideline
Clearing House http://www. Most hospitalists would recognize the question, “Should patients
http://www.ncbi.nlm.nih.gov/ with heart disease receive regular vaccinations?” as one that is overly
pubmed guideline.gov
broad. Not all heart diseases are the same, nor are all vaccinations,
PubMed
and the speci c bene ts patients might reap rom vaccination are
http://www.ncbi.nlm.nih.gov/ not speci ed by the question. Clinical questions need to be ocused
pubmed
in order to be answerable. In addition to clinical questions about
Trip Database therapy, clinicians can ask ocused clinical questions about diag-
https://www.tripdatabase.com nostic tests, about the harm an intervention might cause, about
prognosis, or about di erential diagnosis.
40
Clinical questions should be constructed using the “P-I-C-O” or- can be best. JAMAEvidence catalogs evidence on the accuracy o
mat. “P” stands or “population” and describes the patient the ques- history and physical exam ndings. The Cochrane Collaboration
C
tion is about in proper detail. “I” stands or “intervention” and re ers ocuses on systematic reviews. No single resource is per ect and
H
to the therapy or diagnostic test in question. “C” stands or “com- clinicians should adopt a “toolbox” approach by becoming amiliar
A
parison” and describes either an alternative treatment or standard with a ew resources.
P
o care ( or questions about therapy) or the gold standard test ( or
T
E
questions about diagnostic tests). “O” stands or “outcome,” which
PRACTICE POINT
R
should be clinically important and patient-centered. Surrogate
7
markers o clinically important outcomes are acceptable. • Pull resources are databases that are searched in an on-demand
An example o a well-built clinical question about therapy is: “In way to answer a clinical question. Pull resources have di erent
patients admitted to the hospital with non-ST elevation myocardial and o ten complementary roles. None are per ect, and
P
in arction (P), what is the e ect o in uenza vaccination at discharge
r
hospitalists should adopt a “toolbox”approach in which they
i
n
(I) as compared to no vaccination (C) on recurrent acute coronary become amiliar with a ew resources. The type o question and
c
i
p
syndrome or mortality (O)?” the amount o time available to answer the question should
l
e
An example o a properly designed clinical question about a diag- help determine which resource the hospitalist consults.
s
o
nostic test is: “In patients presenting to the emergency department
with suspected in ection (P), how accurate is a history o shaking
E
v
chills (I), as compared a gold standard o blood cultures (C) in diag-
i
■ THE HIERARCHY OF EVIDENCE
d
e
nosing bacteremia (O)?”
n
Hospitalists should know which types o clinical trials will best
c
When clinical questions do not per ectly t into the P-I-C-O
e
answer di erent types o clinical questions, and which study designs
-
ormat, clinicians should ollow as many o the above principles as
B
will provide the most power ul results. The randomized controlled
a
possible.
s
trial (RCT) is the gold standard or determining the e ect o a thera-
e
d
peutic intervention.
M
PRACTICE POINT Determining the accuracy o a diagnostic test requires a prospec-
e
tive design in which the test is studied in the same clinical setting it
d
•
i
Clinical questions must be ocused to be answerable.
c
will be used, and is compared against an acceptable gold standard.
i
n
Hospitalists should use the widely accepted “Population– The e ect o a diagnostic test on clinical outcomes can be deter-
e
Intervention–Comparison–Outcome”(P-I-C-O) ormat to
a
mined by a randomized controlled trial, in which the test in ques-
n
construct ocused clinical questions. tion is treated as the intervention and another diagnostic approach
d
Q
(pre erably a gold standard i available) is considered the comparison.
u
A systematic review is a summary o the evidence on a topic in
a
l
i
■ STEP 2: SEARCHING THE LITERATURE which the literature search and selection o evidence has been per-
t
y
ormed in a rigorous, transparent, and reproducible way. The most
o
Pull information resources
valuable systematic reviews will also include a meta-analysis. In a
E
With a properly constructed clinical question in hand, the hospital-
v
meta-analysis, the results o multiple similar types o studies (RCTs,
i
d
ist can now search the literature to nd the answer. The rst step is observational studies, or studies o diagnostic tests) are statistically
e
to select an in ormation resource that is appropriate or the clinical
n
combined to o er more power ul results. What a meta-analysis
c
question and the amount o time available. Databases that are
e
gains in power, it can sometimes lose in applicability and ocus i
searched in an on-demand way in order to answer a clinical ques- too much clinical heterogeneity exists among the patients included
tion are called pull resources. rom individual studies. With that caveat, a high-quality systematic
In many cases, and especially when time is limited, one should rst review that includes a meta-analysis is considered to be the highest
consult a high-quality summary pull resource. Summary resources level o evidence. Table 7-2 describes the hierarchy o evidence or
that are requently updated assess the quality o the evidence pre- di erent types o clinical questions.
sented and are user- riendly and pre erable. Examples include BMJ
Clinical Evidence, ACP Journal Club, DynaMed, the Cochrane Collabo- ■ STEP 3: CRITICALLY APPRAISING THE LITERATURE
ration, NEJM Journal Watch, UpToDate, and practice guidelines rom
Although summary resources that appraise the medical literature
pro essional societies. All are highly use ul. Each has its strengths and
have risen in quality and are an essential resource or clinicians, they
weaknesses. UpToDate is ast to use, comprehensive, and provides
will not always provide the answer to a clinical question. In addi-
expert guidance in an easy to digest, narrative ormat, but it is not
tion, hospitalists may participate in discussions around particular
as rigorously constructed as the others. ACP Journal Club provides
studies, attend “journal club” con erences, or teach junior learners
excellent summaries o highly selected literature deemed valid and
about evidence-based medicine. Hospitalists must have solid critical
relevant to clinical practice, but as a result its database is not compre-
appraisal skills. The Users’ Guides to the Medical Literature (McGraw-Hill,
hensive. DynaMed is rigorously constructed and presents a lot o pri-
2014) is the benchmark textbook or learning how to practice EBM.
mary data rom clinical trials, o ten in an outline ormat. The Cochrane
It proposes an e ective method or critical appraisal that has been
Collaboration produces high-quality systematic reviews o the evi-
widely adopted. The principles and approach it endorses are dis-
dence. Practice guidelines are excellent resources that o er clear
cussed urther in this chapter.
recommendations, but their quality can vary, update intervals can be
In appraising any type o study, three broad questions must be
long, and users must pay close attention to the level o evidence and
answered:
strength o recommendations in published practice guidelines.
I one has more time, or i a deeper dive is needed a ter consult- 1. Are the results valid?
ing a summary resource, the primary literature can be searched via 2. What are the results?
PubMed (pre erably using “Clinical Queries” option or clinical ques- 3. How can I apply the results to patient care?
tions) or Trip Database (pre erably using “PICO search” option or The critical appraisal process asks these three questions o
clinical questions). For certain questions, a content-speci c resource each type o study, including those about therapy, diagnosis,
41
TABLE 7-2 Hierarchy o Evidence or Di erent Types o TABLE 7-3 Critical Appraisal Questions or an Article
Clinical Questions About Therapy
P
Type of Clinical Best Types of Articles (Listed in Decreasing Main Question Supplemental Questions
A
Question Level of Evidence) 1. Is the study a. Were patients randomized?
R
Therapy or harm 1. Systematic review/meta-analysis o valid?
T
b. Was group allocation concealed?
randomized controlled trials
I
c. Were patients in the study groups similar
2. Randomized controlled trial with respect to prognostic variables?
3. Cohort study d. To what extent was the study blinded?
4. Case-control study e. Was ollow-up complete?
T
h
5. Case series . Were patients analyzed in the groups
e
6. Case reports to which they were irst assigned
S
p
7. Expert opinion (ie, intention to treat)?
e
c
g. Was the trial stopped early?
i
Accuracy o a 1. Systematic review/meta-analysis
a
l
diagnostic test 2. What are the a. How large was the treatment e ect?
t
2. Prospective comparison against gold
y
results? (What were the RRR and the ARR?)
o
standard conducted in setting diagnostic
f
test will be used in practice b. How precise were the results? (What
H
o
Prognosis 1. Systematic review/meta-analysis were the con idence intervals?)
s
p
2. Prospective cohort study o a 3. How can I apply a. Were the study patients similar to my
i
t
a
representative, homologous patient the results to patients?
l
patient care?
M
group with appropriate ollow-up and b. Were all clinically important outcomes
e
objective outcomes. considered?
d
i
c
3. Retrospective case-controlled study c. Are the likely treatment bene its worth
i
n
Di erential 1. Systematic review/meta-analysis the potential harm and costs? (eg, what
e
is the number needed to treat? What is
a
diagnosis o a 2. Prospective evaluation o a representative
n
condition the number needed to harm?)
d
sample that includes de initive diagnostic
S
evaluation, per ormed in a setting similar
y
Adapted rom Guyatt G, et al., eds. Users’Guides to the Medical Literature: A
s
to actual practice
t
Manual for Evidence-Based Clinical Practice, 3rd ed. New York, NY: McGraw-Hill
e
m
Education; 2014.
s
o
harm, prognosis, and systematic reviews. Each o the three major
f
C
questions is answered through a subset o critical appraisal ques-
a
■ CRITICAL APPRAISAL OF AN ARTICLE
r
tions that are speci ic to each study type. The critical appraisal
e
ABOUT THERAPY
questions help determine i a study used proper methods to
prevent bias, i the results are large enough to be meaning ul, To critically appraise an article about therapy, the clinician should
and whether the results can be applied to a particular patient or answer the ollowing set o questions.
population.
Are the results valid
Were patients randomized? Randomization will best ensure
PRACTICE POINT that the intervention and control groups are equal at the start o
the trial, except or the intervention being tested. In observational
• Critical appraisal ocuses on answering three broad questions: studies, investigators must take special steps to ensure experi-
Are the results valid? What are the results? How can I apply the mental and comparison cohorts are evenly matched. Randomiza-
results to patient care? The Users’Guides to the Medical Literature tion does much o this automatically.
o ers a methodical approach to answering these questions or Was group allocation concealed? When allocation conceal-
studies about therapy, diagnosis, harm, and prognosis, and or ment is present, those enrolling patients into the study during
systematic reviews. randomization are blinded to what group (ie, intervention or
control) the patients are being assigned. Without allocation
concealment, or example, a patient being enrolled but who is
In recent years, a new type o evidence, the results o quality viewed as likely having a bad outcome might be steered into
improvement studies, has risen in prominence. As hospitalists the comparison group, potentially improving the results in the
o ten are involved in quality improvement e orts, they should intervention group.
have a working knowledge o how to critically appraise this type Were patients in the study groups similar with respect to
o evidence. The Users’ Guides to the Medical Literature o ers urther known prognostic variables? This is necessary to isolate the
instruction in this area. e ect o the intervention and minimize con ounders. Proper ran-
This chapter will illustrate the critical appraisal process through domization will ensure this. In the absence o randomization, cli-
analysis o an article about therapy, as randomized controlled trials nicians should look to see that the intervention and comparison
and prospective cohort studies are among the most common types groups were care ully matched so as to be equal or all possible
o evidence encountered in practice. Table 7-3 outlines the critical con ounders. This is o ten dif cult to do, which is why randomiza-
appraisal questions, which are discussed in detail below. Clinicians tion is pre erred.
should re er to the Users’ Guides to the Medical Literature or a com- To what extent was the study blinded? The term “double-
plete list o critical appraisal questions or di erent types o research blind” does not describe all parties that should be blinded in an
studies. RCT. For maximum validity, multiple groups should be blinded,
42
including those selecting patients or randomization (ie, alloca- PRACTICE POINT
tion concealment), the patients, those administering the inter-
•
C
vention, the data collectors/analysts, and the outcome assessors. A randomized controlled trial describes the average e ect o
H
When patients or those administering the intervention cannot be an intervention across the group o patients studied. The e ect
A
blinded (as in trials o certain surgeries or procedures), allocation an intervention will have on any individual patient can be
P
concealment, as well as blinding o data analysts and outcome determined by combining that patient’s baseline risk with the
T
relative risk reduction (RRR) reported in the trial. Clinicians can
E
assessors, is essential.
R
Was the ollow-up complete? Studies should track the out- estimate their patient’s baseline risk by comparing them to the
clinical characteristics and comorbidities o patients in a trial,
7
comes o all participants. Patients may be lost to ollow-up i they
su er a negative outcome or nd the intervention too dif cult to and by using their clinical judgment and expertise.
comply with. Both o these reasons would be highly relevant to
P
the results o a study.
r
i
n
Were patients analyzed in the groups to which they were ■ HOW PRECISE WAS THE ESTIMATE OF THE
c
i
p
rst assigned (ie, intention to treat)? The principle o “inten- TREATMENT EFFECT? (WHAT WERE THE
l
e
tion to treat” highlights that in a clinical trial, the o ering o an CONFIDENCE INTERVALS?)
s
o
intervention to participants is being tested as much as the other Con dence intervals provide more in ormation than P-values alone,
e ects o the intervention. I or instance participants do not like
E
giving an estimate o the range o possible results. Some high-
v
the taste o a pill or nd a study protocol too hard to comply
i
d
quality evidence-based summary resources, such as ACP Journal
e
with and drop out o a trial or asked to be switched to the other Club, emphasize con dence intervals and the help ul picture they
n
c
arm as a result, these consequences must be recorded as part paint o the results.
e
-
o the results o the study. Outcomes must be attributed to the In the study o ri aximin described above, the RRR = 51% (95% CI,
B
a
group to which participants were initially assigned. A trial that 20-71). In “plain English,” this 95% con dence interval tells us that
s
e
ollows the intention to treat principle will give the best esti- ri aximin most likely reduces in-hospital death by 51% (the “point
d
mate o what will happen i a therapy is o ered to a population.
M
estimate”) but it may reduce in death by as little as 20%, or by as
In a “per protocol analysis,” the study results represent only what
e
much as 71%. There is a 95% chance that the true e ect is between
d
happened to those who actually accept the intervention and
i
20% and 71%, a 2.5% chance the true e ect is below 20%, and a
c
i
complete the trial. This type o analysis can in orm what e ect
n
2.5% chance it is above 71%.
e
a therapy would have i taken properly by a highly compliant In order to determine whether a trial has ound two therapies to
a
n
patient. be equivalent, clinicians should examine the upper and lower limits
d
Was the trial stopped early? Follow-up must be an appropriate o the 95% con dence interval. I either would be clinically signi cant
Q
length or the outcome measured. For example, 3 days might be
u
i true, the two therapies studied cannot be called equivalent, and
a
an appropriate ollow-up period or an intervention to reduce
l
urther research is needed. A 2014 study by Regimbeau et al. pub-
i
t
y
acute pain, but it would likely be too short or an intervention lished in the Journal of the American Medical Association ound that in
o
designed to reduce LDL cholesterol or to improve unctional sta- patients undergoing cholecystectomy or acute calculous cholecys-
E
tus. Randomized controlled trials that are stopped early because titis, postoperative antibiotics reduced in ection by an absolute risk
v
i
o bene t may overestimate the e ect o an intervention. A large
d
reduction o 1.9% (95% CI, –9.0-5.1, P < 0.05). The con dence interval
e
bene t observed early in a trial may be due chance, and may be
n
indicates that antibiotics most likely reduce in ection by 1.9%, but
c
greater that what would be observed i the trial were allowed to
e
may reduce in ection by as much as 5.1% (in which case most clini-
run to completion. cians would prescribe them), or may increase in ection by as much as
9% (in which case most clinicians would not prescribe them). In this
What are the results study, the true e ect o antibiotics on postoperative in ection could
not be determined as they could be either bene cial or harm ul,
How large was the treatment ef ect (ie, what were the relative and urther study is needed. A common misinterpretation o these
risk reduction and absolute risk reduction?)? Clinicians should results, which could occur i the con dence intervals are not noted,
consider results o a study using the absolute risk reduction (ARR), would be: “the P-value is greater than 0.05 so there is no di erence
where the ARR% = event rate in comparison group – event rate in between antibiotics and placebo and the two are equivalent.”
experimental group. The relative risk reduction (RRR) is calculated as Two actors a ect the width o a con dence interval: the number
RRR% = event rate in comparison group – event rate in experimen- o patients and the requency o the outcome in a study. In our
tal group/event rate in comparison group. The RRR allows one to example o the Regimbeau trial, urther studies that enroll larger
determine the e ect o a therapy on an individual patient according numbers o patients or measure more postoperative in ections
to their baseline risk. could result in a narrower con dence interval as well as a di erent
Consider the study by Sharma et al., published in the American point estimate.
Journal of Gastroenterology in 2013, that randomized 120 hospital-
ized patients with cirrhosis and overt hepatic encephalopathy to
ri aximin versus placebo. In-hospital death occurred in 24% o the PRACTICE POINT
ri aximin group and in 49% o the placebo group. Here the ARR is
25% (49% – 24%) and the RRR is 51% (49% – 24%/49%). • Con dence intervals are pre erable to P-values when
Clinicians can use the RRR to estimate the e ect a therapy will considering the results o a clinical trial, as they give more
have on individual patients they treat that may be more or less sick in ormation about the range o possible results, including the
than the average patient in a study. For example, i a patient is esti- best and worst case scenarios.
mated to have a baseline risk o dying o 60%, ri aximin will reduce
this patient’s risk o dying to 30.6% (60% × 0.51). In this case the ARR
will be 60%-30.6% = 29.4% which is higher than what the ri aximin How can I apply the results to patient care
group as a whole experienced in the trial. In a similar way, lower Were the study patients similar to my patients? The more a
baseline risk will result in lower absolute risk reduction. patient meets the inclusion criteria, and the less they meet the
43
exclusion criteria, the more con dently the results o a study can three broad questions discussed above: a study must be valid, it
be applied to them. Clinicians should consider the setting o a must report important results, and it must be applicable to the
study as well as whether those who administered the interven- patient at hand. I any o these three elements is missing, the study
tion had specialized expertise that is not available locally. ndings may not be appropriate or implementation into practice.
P
A
Were all clinically important outcomes considered? The When a study has used valid methods, has reported highly impor-
R
“grandmother test” can help determine i an outcome is clinically tant results, and has enrolled patients clearly similar to the patient
T
relevant. Outcomes that would be valued by the average person in question, the hospitalist can con dently apply its ndings. But
I
(eg, someone’s grandmother) are clinically important; outcomes conducting clinical studies is o ten dif cult work, and ew studies
that would not be valued are not clinically relevant and should are per ect in every way. Clinicians need to learn which validity or
not be measured by clinical trials. For example, outcomes such applicability issues represent atal aws, and which ones still allow
as a reduction in pain, an increase in survival, or a reduction the results o a study to be considered. This is a skill that comes with
T
h
hospital admission are likely to be meaning ul to patients, while experience.
e
biochemical, laboratory, or purely hemodynamic outcomes are The hospitalist must remember that best evidence-based deci-
S
p
not. An exception is when nonclinical outcomes are established sions incorporate not only the evidence, but also the individual clini-
e
surrogate markers or clinically important outcomes. Composite
c
cal circumstances and pre erences o patients. In most cases, patient
i
a
outcomes o clinical endpoints are valid, but i possible studies values and pre erences are more important than the other actors.
l
t
y
should make clear how much each individual endpoint is driving
o
the composite result. CONCLUSION
f
H
This chapter has ocused on skills such as the construction o
o
s
ocused clinical questions and how to search and critically appraise
p
i
t
PRACTICE POINT the literature. These skills are necessary but not suf cient or the
a
l
practice o EBM. The hospitalist’s knowledge o the patient is at the
M
• Hospitalists should value studies that measure clinically
heart o evidence-based practice. The right clinical questions can-
e
important endpoints (or surrogate markers o these) over those
d
not be asked unless the hospitalist rst has a clear understanding
i
c
that measure physiologic or biochemical endpoints.
i
o the patient’s clinical issues, and the literature cannot be applied
n
e
to a patient without knowledge o their values and pre erences.
a
n
Communication skills, history and physical examination skills, illness
d
■ ARE THE LIKELY TREATMENT BENEFITS WORTH scripts, problem representation, and clinical reasoning skills are
S
THE POTENTIAL HARM AND COSTS? (WHAT IS THE
y
some o the ways hospitalists come to know their patients. To be a
s
NUMBER NEEDED TO TREAT? WHAT IS THE NUMBER
t
top-notch practitioner o EBM really starts and ends with being an
e
m
NEEDED TO HARM?) outstanding clinical doctor.
s
The number needed to treat (NNT) describes how many patients
o
f
must be treated with an intervention to produce one positive out- SUGGESTED READINGS
C
a
come or prevent one negative outcome. The NNT allows clinicians
r
e
to compare the e ects o di erent therapies, and is calculated as Devereaux PJ, et al. Double blind, you are the weakest link—
NNT = 100/ARR%. In the study by Sharma et al. discussed above, in- goodbye! Evidence-Based Med. 2002;7:14-15.
hospital death occurred in 24% o the ri aximin group and in 49% o
Guyatt G, et al., eds. Users’ Guides to the Medical Literature: A Manual
the placebo group. Here the ARR is 25% (49% – 24%) and the NNT is
for Evidence-Based Clinical Practice, 3rd ed. New York, NY: McGraw-Hill
4 (100/25). In other words, we need to give our patients ri aximin to
Education; 2014.
prevent one patient rom dying in the hospital.
In order to best in orm risk/bene t discussions about a therapy, Haynes RB, et al. Clinical expertise in the era o evidence-based
studies should measure important adverse e ects. The number medicine and patient choice. ACP Journal Club. 2002; March/April:
needed to harm (NNH) describes how many patients must be A11-A14.
treated or one to experience a particular adverse e ect. These two Regimbeau JM, et al. FRENCH Study Group. E ect o postoperative
numbers can be compared or an intervention and a particular antibiotic administration on postoperative in ection ollowing
adverse e ect to determine the net bene t or harm. In addition to cholecystectomy or acute calculous cholecystitis: a randomized
the likelihood o adverse events and their morbidity, the level o clinical trial. JAMA. 2014;312:145-154.
concern a patient has about particular side e ects must be con- Sharma BC, et al. A randomized, double-blind, controlled trial
sidered. Many studies do not assess cost, and those that do o ten comparing ri aximin plus lactulose with lactulose alone in treat-
determine cost-e ectiveness at the population level, which is less ment o overt hepatic encephalopathy. Am J Gastroenterol.
relevant to the individual patient. The extent to which a therapy 2013;108:1458-1463.
is covered by insurance is highly relevant to patients and should
always be considered.
■ STEP 4: APPLYING THE LITERATURE TO AN

INDIVIDUAL PATIENT
For the ndings o a study to be use ul in clinical care, the critical
appraisal process must yield a satis actory answer to each o the
44
CHAP TER
8 INTRODUCTION
Diagnosis is the art o identi ying a disease by the signs, symptoms,
and test results o a patient. Diagnosis stems rom the Greek word,
diagignoskein, which means to distinguish or discern. Indeed, the
ability to distinguish or discern a patient’s underlying illness is criti-
cal to being an e ective clinician as a hospital medicine provider. In
many cases, hospitalized patients may be quite complicated with
multiple competing possible reasons to explain their underlying
signs or symptoms. Patients do not always read textbook (ie, they
Diagnostic Reasoning may not always describe their symptoms or have ndings on exam
that are pathognomonic or as classically described). There ore, diag-
and Decision Making nostic reasoning and diagnostic decision making are crucial skills or
hospital medicine providers. In addition, cognitive biases exist and
diagnostic errors occur when there is any mistake or ailure in the
diagnostic process that leads to a misdiagnosis, a missed diagnosis,
Laurence Beer, MD, SFHM or a delayed diagnosis. This chapter will discuss diagnostic reason-
ing and diagnostic decision making.
Lucas Golub, MD
Dustin T. Smith, MD
DIAGNOSTIC REASONING
■ CLINICAL REASONING
Clinical reasoning is the process where a clinician applies reasoning
in combination with the clinician’s knowledge and skills (Figure 8-1).
Clinical reasoning is a constant process that does not end when the
diagnosis has been made. It may be considered complete upon
autopsy or when a gold standard has con rmed a diagnosis, but it is
important to acknowledge that in many instances the gold standard
is not 100% accurate. In the hospital setting clinicians are operating
under a running diagnosis until the diagnosis has been con rmed
and/or until the patient has improved both subjectively and objec-
tively. Sometimes patients do not improve when a treatment strat-
egy is implemented; thus, it is important that providers continuously
use clinical reasoning skills as in ormation is collected in an attempt
to veri y the diagnosis. Once a treatment or workup plan has been
implemented, it is crucial to reassess the patient’s response to this
to urther con rm i the correct diagnosis or treatment strategy has
been made (Figure 8-2). I the diagnosis does not appear correct
or the treatment strategy is ailing, the clinician must review the
in ormation and data collected and reconsider the other possible
diagnoses to explain a patient’s presenting signs and symptoms.
Hence, a clinician’s ability to success ully reason and diagnose is in
some ways anchored by that clinician’s ability to create an adequate
di erential diagnosis.
■ DIFFERENTIAL DIAGNOSIS
A di erential diagnosis is more than just a list o illnesses that poten-
tially explain why a particular sign, symptom, and/or diagnostic test
result exists. Rather, the clinician uses the di erential diagnosis to
distinguish a disease or condition rom others that present with
similar signs, symptoms, or diagnostic test results. Initially start with
a broad list o diagnoses until urther in ormation or data is obtained.
A clinician must consider the prevalence o disease and other ac-
tors, which make up a provider’s patient population, when ormulat-
ing a di erential. It is more o ten the case where a common disease
presents in an atypical ashion rather than a rare or exotic disease
actually being present.
45
is excluded. Once a diagnosis has been con rmed, the problem list
should be diagnosis-oriented rather than problem-oriented. Should
a patient not respond to a treatment plan or a diagnosis where
a clear and obvious bene t is expected, the clinician should then
P
A
Re as o ning re-evaluate the patient, review the data obtained, and nally provide
R
a summary o key data to ensure that the diagnosis is indeed correct.
T
I
PRACTICE POINT
• List the top diagnosis rst but keep it problem oriented until
the diagnosis has been made.
T
h
• Aggressively prioritize most likely then most harm ul in clinical
e
S kill Knowle dg e
workups.
S
p
• Exclude what diseases can easily be excluded and then remove
e
c
those rom di erential diagnosis.
i
a
l
•
t
Keep a broad di erential until the top diagnosis has been
y
con rmed and always plan or workup o other alternate
o
f
Figure 8 1 Clinical reasoning. diagnoses i the top diagnosis is excluded.
H
•
o
When presented with a clinical conundrum, begin by de ning
s
p
the problem, be deliberate and gather key in ormation,
i
PRACTICE POINT
t
a
and then nally summarize the case when aced with large
l
•
M
A di erential diagnosis is the ability to distinguish a disease amounts o in ormation.
e
or condition rom others that present with similar signs,
d
i
symptoms, or diagnostic test results.
c
i
n
■ ILLNESS SCRIPTS
e
a
n
List your top diagnosis rst ollowed by other potential diagnoses Illness scripts are made up o key risk actors, pathophysiology, and
d
or a speci c problem but keep it problem oriented until you have clinical presentations that summarize a speci c clinical problem or
S
y
the actual diagnosis (eg, chest pain di erential diagnosis: most likely syndrome. An illness script can be a very power ul tool or clinicians.
s
t
The number o illness scripts an individual clinician has increases
e
acute coronary syndrome but consider heart ailure, pneumonia,
m
pneumothorax, pulmonary embolism, and musculoskeletal pain). with time and experience. Illness scripts may also diminish i a
s
Aggressively prioritize workup o the most likely and most harm ul provider’s site o practice, specialty, or patient population becomes
o
f
(ie, li e threatening) diagnoses under consideration. Prioritize the limited (eg, a clinician whose only site o practice is in a private
C
hospital setting or adult hospitalized patients). An example o some
a
workup o acute and reversible diseases ollowed by chronic and
r
e
irreversible diseases (eg, delirium due to a medical cause versus commonly agreed upon illness scripts may be ound in Table 8-1.
chronic, progressive dementia). As in ormation or data that e ec-
tively rules out a particular diagnosis or a chie complaint becomes
PRACTICE POINT
available, remove that diagnosis rom your list and ocus your atten-
tion on remaining possibilities. Clinicians should always keep a broad • Utilize illness scripts, which can be very power ul tools or
di erential until the top diagnosis has been con rmed but also have clinicians and aid in diagnosis.
a plan or workup o other alternate diagnosis i the top diagnosis
■ REASONING PROCESSES
Imple me nt (Re )As s e s s
pa tie nt The dual-process theory o reasoning consists o two systems o rea-
pla n
soning, which clinicians can utilize to correctly arrive at a diagnosis
(Figure 8-3). The rst system is intuitive while the second system
is analytical. A care ul balance o these two systems is necessary to
ensure e cient and e ective clinical reasoning.
Ve rify Colle ct info
dia gnos is
PRACTICE POINT
• Per orm the dual-process theory o reasoning and make sure
to balance the intuitive and analytical components to maintain
highly e cient and e ective clinical reasoning.
Ge ne ra te
Colle ct more diffe re ntia l
info
System 1 (intuitive)
The rst system o the dual-process theory is intuitive, implicit or
automatic. O ten an unconscious or subconscious process involv-
Re fine Colle ct more ing pattern recognition or matching against illness scripts markedly
dia gnos is info a ects the clinician’s judgment; the clinician’s past clinical experi-
ence and clinical expertise weighs heavily on this system. Heuristics,
Figure 8 2 The process of clinical reasoning. or mental shortcuts, are also utilized in this system.
46
TABLE 8-1 Examples of Illness Scripts
C
Key Risk Factors Pathophysiology Clinical Presentation Diagnosis
H
A
• Usually 1-3 mo old in ants • Pyloric muscular hypertrophy Nonbilious projective vomiting Pyloric stenosis
P
• Rare a ter 12 wks o age • Pyloric canal narrows In ant demands to be re ed soon
T
a terwards
E
• 4:1-6:1 male: emale • Near-complete obstruction
R
• First born son No diarrhea
8
• Maternal smoking increases Severe weight loss
the risk Emaciated with “olive-like” mass on
abdominal exam
D
i
Hypochloremic metabolic alkalosis
a
g
Hypokalemia
n
o
• Usually idiopathic • Repetitive actions o the hand Pain and parasthesia in the medial Carpal tunnel
s
t
i
or wrist nerve distribution syndrome
c
• Common disorder among
R
adults • Increased pressure in the Worse at night
e
a
More requent in women intracarpal canal Changes in hand posture or shaking the
s
•
o
• Median nerve compression hand mitigates symptoms
n
• Involved in activities that
i
n
involve lexing or extending • Ischemia and mechanical + Phalen maneuver and Tinel test
g
o wrist repeatedly (eg, typing disruption o nerve
a
n
a book chapter)
d
• Uncommon event • Tear in the aortic intima Severe, “tearing” chest pain radiating to Aortic
D
e
• Most common predisposing • Degeneration o the aortic the back dissection
c
i
actor is hypertension media Blood pressure not equal in both arms
s
i
o
Typically occurs in 60- to Hemorrhage into the media Widened mediastinum on chest
n
• •
80-y-old men radiography
M
• Creation o a alse lumen
a
Acute hemodynamic compromise
k
• Propagation o dissection
i
n
both distal and proximal to
g
initial tear involving branch
vessels, aortic valve, and/or
pericardial space
• Ischemia, aoritic regurgitation,
and/or cardiac tamponade
• Most common indication or • Obstruction o the vestigial Abdominal pain, typically beginning Acute
emergent abdominal surgery vermi orm appendix in the periumbilical region and migrating appendicitis in
in childhood • In lammation o the to the right lower quadrant children
• More common in older appendiceal wall Anorexia
children and adolescents • Localized ischemia, Nausea or vomiting
• More common in boys per oration, and the Low-grade ever
development o a contained
abscess or generalized Peritoneal signs on abdominal
appendicitis examination
+McBurney’s point tenderness,
Rovsing’s sign, ilopsoas sign, and/or
obturator sign
Mild leukocytosis
• Most common cause o acute • Obstruction o the vestigial Abdominal pain, typically right lower Acute
abdomen vermi orm appendix quadrant appendicitis in
• Occurs most requently in • In lammation o the Anorexia adults
second and third decades appendiceal wall Nausea or vomiting
o li e • Localized ischemia, Low-grade ever
• 1.4:1 male: emale per oration, and the
development o a contained Peritoneal signs on abdominal
abscess or generalized examination
appendicitis +McBurney’s point tenderness,
Rovsing’s sign, psoas sign, and/or
obturator sign
47
The dual-pro c e s s the o ry Strategies to avoid pit alls exist include:
Sys te m 1: a utoma tic, • Always begin by de ning the problem a patient aces or reports.
s ubcons cious proce s s ing
Ye s
EXP ERT HEURISTIC • Deliberate and gather key in ormation, as opposed to gather-
?
P
ing large amounts o in ormation.
A
Re cognize d ? Dia gnos is
• Per orm a summary o key elements in abstract terms to pro-
R
vide a discriminatory and use ul overview o the patient’s clini-
T
No Sys te m 2: de libe ra te ,
cal case when aced with large amounts o clinical in ormation.
I
cons cious thought
This is o ten re erred to as problem representation.
Figure 8 3 The dual-process theory of reasoning. • Utilize illness scripts which summarize a speci c clinical problem.
• Pay particular mind to not prioritize exotic or rare diseases
initially or aggressive work-up (see base-rate neglect below).
T
h
• Always consider the diagnostic testing and treatment thresh-
e
System 2 (analytic)
olds or the diagnoses he/she is considering in the di erential.
S
p
The second system o the dual-process theory is the analytic, con- • Nonanalytic and analytic reasoning must be balanced, espe-
e
c
scious and deliberate type o reasoning. The clinician uses rational cially when a clinical pattern does not t a known illness script.
i
a
and care ul analysis when evaluating a di erential diagnosis or a
l
• Be cognizant o biases in clinical reasoning and avoid them at
t
y
speci c problem. This system involves hypothesis testing as a tool all costs.
o
or clinical reasoning (Figure 8-3).
f
• Always keep in mind that patients “do not always read the
H
textbook.”
o
■ PITFALLS AND STRATEGIES TO AVOID PITFALLS
s
p
i
t
Numerous pit alls exist in clinical reasoning and practice. Clinicians
a
■ COGNITIVE HEURISTICS AND BIASES
l
may overly rely on heuristics, or mental shortcuts. Cognitive biases
M
may alter a clinician’s judgment and clinical reasoning. In the mod- The worse heuristic comes rom the Greek word, heuriskein, meaning
e
d
ern era o clinical practice, in ormation overload and overreliance on to nd or discover. Diagnostic decision making is ultimately a path
i
c
i
diagnostic tests may lead to excessive testing or “overdiagnosis,” or toward discovery, requiring deliberate, systematic thought. There
n
e
the diagnosis o a “disease” that will never cause a patient symptoms are many avenues o systematic approach to a common question,
a
or death. Certain diagnoses prompt treatments and interventions each with unique issues and pit alls. A summary o cognitive heu-
n
d
which may cause harm, especially i the alternative is best managed ristics and biases with examples given can be ound in Table 8-2.
S
y
by watch ul waiting regarding the identi cation o an incidental
s
■ ANCHORING
t
nding or very early stage disease.
e
m
Anchoring bias arises when undue importance is placed on in or-
s
PRACTICE POINT mation that is received early in the diagnostic evaluation. The
o
f
• In the modern era o clinical practice, in ormation overload clinician does not adjust the di erential diagnosis commiserate
C
a
is another pit all as is overreliance on diagnostic tests, which with the totality o in ormation as new, possibly contradictory data,
r
e
may lead to overtesting or “overdiagnosis”(ie, the diagnosis o a emerges. The inverse o anchoring is known as the “order e ects”
“disease”that will never cause a patient symptoms or death). bias, where recent in ormation is given more weight in the deci-
sion process.
TABLE 8-2 Description of Cognitive Biases or Heuristics with Examples
Bias or Heuristic Description Example

Anchoring Undue ocus on a case’s starting point, without Patient with chronic obstructive pulmonary disease presents
adequate adjustment or new in ormation or dyspnea, does not improve with medical therapy, and is
later ound to have a pulmonary embolism
Availability (recall) Overestimation o diagnostic likelihood based Suspicion o pheochromocytoma in a patient with
on vivid or easily recalled events chronic, resistant hypertension
Base-rate neglect Impression that two diseases which present Overestimation o the probability o aortic dissection in
similarly are equally likely chest pain
Commission Pre erence or intervention over inaction Thrombolysis or submassive pulmonary embolus
Con irmation Attempting to con irm a diagnosis rather than Leading (non open-ended) questions aimed at eliciting
testing other possibilities history which con irms a suspected diagnosis
Hindsight Clinical decisions judged by the result rather Use o bilevel ventilation is judged to be bene icial
than the logic o the choices because the patient survived severe hypoxia
Omission Iatrogenesis is perceived to be worse than Surgery not pursued or appendicitis and overwhelming
naturally occurring adverse events sepsis because operative mortality is judged to be too high
Premature closure Early cessation o investigative thought once Altered mental status and ever are attributed to alcohol
a presumed diagnosis has been made withdrawal, but aspiration pneumonia is also present
Representativeness Patients do not read textbooks and pathology Cholecystitis dismissed because Murphy’s sign is
will not always present as expected negative
Search satis ying An end to in ormation gathering once Rib ractures are not identi ied on a chest radiograph
something has been ound with in iltrate
48
■ AVAILABILITY PRACTICE POINT
Availability bias, also re erred to as recall bias, involves the overesti-
•
C
A clinician should be cognizant o cognitive biases in clinical
mation o the likelihood o disease, creating a alse sense o preva-
H
reasoning and avoid them at all costs.
lence. Largely driven by the clinician’s personal experience, diseases
A
that have been encountered either recently or vividly may be given
• Remember that patients “do not always read the textbook.”
P
T
too much credence.
E
R
■ BASE-RATE NEGLECT ■ IMPROVING DIAGNOSTIC REASONING
8
Base-rate neglect occurs when true prevalence is distorted. The Both individual skills can be developed and systems-based practices
similar presentation o two di erent diseases creates a alse sense can be instituted to minimize diagnostic error and improve diagnos-
D
that the rare disease is equally likely compared to its true prevalence. tic reasoning. Individual skills that may be taught include situational
i
a
awareness, deliberate practice, and metacognition. Systems-based
g
■ COMMISSION BIAS
n
solutions include creating a process or providing individual eed-
o
back, developing a nonpunitive error reporting culture, the use o
s
This popular saying provides a good representation o commission
t
i
electronic medical records, and urnishing computer-based decision
c
bias: “when you are a hammer, everything looks like a nail.” Commis-
R
sion bias speaks to the clinician’s pre erence or active intervention support tools. A summary o cognitive debiasing strategies can be
e
a
over passive inaction. ound in Table 8-3.
s
o
n
■ SITUATIONAL AWARENESS
i
■ CONFIRMATION BIAS
n
g
Con rmation bias is the inclination to seek in ormation that con- Health care sa ety and quality researchers have borrowed concepts
a
n
rms the current clinical assessment rather than delving into alter- rom other high-risk industries like the military, aviation, and nuclear
d
native hypotheses. Con rmation bias may aggravate anchoring and power generation that have proven success ul at minimizing
D
adverse outcomes. Situational awareness re ers to having a keen
e
thereby cause a vicious cycle that hampers diagnostic momentum.
c
awareness o the multitude o environmental actors that inf uence
i
s
i
o
■ HINDSIGHT BIAS decision making in complex, dynamic situations. Having broad situ-
n
ational awareness acilitates understanding o what is happening,
M
Hindsight bias or outcome bias may produce misguided positive
why it is happening, and anticipating what is likely to happen next,
a
eedback i inappropriate testing occurred but a positive clinical out-
k
making it easier to identi y problems at an early stage, be ore they
i
n
come resulted or negative eedback despite appropriate manage-
g
become catastrophic.
ment i an adverse or unexpected outcome transpired. Knowledge
The concept o situational awareness may be used to mitigate
o a case’s outcome may inf uence perception o preceding events.
the types o errors engendered by both the limitations and com-
plexity o hospital medicine and by the complex organizational
■ OMISSION BIAS and systems structures that de ne modern hospitals. Sources o
Primum non nocere ( rst do no harm). Omission bias occurs when error due to the nature o hospitalist work include lack o long term
clinicians do not therapeutically intervene because an adverse event relationships with their patients and knowledge o their prior level
resulting rom treatment is perceived to be worse than one due to o unctioning, incomplete medical histories, large volumes o rap-
inaction, even i the likelihood is exactly the same in both scenarios. idly changing and sometimes conf icting data, polypharmacy, and
Excessive emphasis is placed on avoiding iatrogenic harm. misperception o disease prevalence due to a particular hospital’s
demographics. Systems-based sources o error, which in one study
■ PREMATURE CLOSURE are present in two-thirds o all errors, include discontinuities in care,
Deliberate diagnostic consideration may be curtailed hastily i a lapses in communication, and lack o coordination between depart-
presumed diagnosis is accepted be ore it has been completely ments. Explicit acknowledgement o and extra attention to both
established. Premature closure may occur either when a patient has sources o error can elucidate ways in which the individual clinician
multiple comorbid processes and the clinician xes on one or when and the institution as a whole may decrease error.
the wrong diagnosis is made in the rst place. Increasingly, hospitalists are working on wards organized as
Accountable Care Units (ACUs). ACUs by de nition are structured
■ REPRESENTATIVENESS around shared missions and goals, and are ideal organizations or
the implementation o group-based situational awareness tech-
Many diseases have a prototypical or “classic” presentation, and the niques. In this ormulation, team members are expected to oper-
probability o disease is o ten assessed according to how closely a ate rom the same script, share a common knowledge base and
speci c patient’s case mirrors that ingrained portrait o the disease language, anticipate the needs o other team members, and be
in question. This search or representative presentations will miss
atypical variants o common disease. It also predisposes clinicians to
dismiss diagnostic possibilities because a cardinal eature is absent,
even i that eature is not necessarily present in all cases. TABLE 8-3 Cognitive Debiasing Strategies to Improve
Clinical Reasoning
■ SEARCH SATISFYING
Situational awareness
Search satis ying describes the cessation o in ormation acquisition
Error reporting and accountability
once something has been ound. This heuristic is di erent rom
premature closure in that it applies to eatures beyond the primary Deliberate practice
diagnosis. With clinical ocus on this, other subtle eatures may Feedback
be missed including additional comorbidities, subtle lab abnor- Decision support tools and other electronic-based systems
malities, and radiologic ndings that do not t with the reason or Metacognition
presentation.
49
able to adapt to tasks as they arise. All team members, regardless o Web-based morbidity and mortality con erences, interactive
role on the team, are taught to recognize “red f ags” such as eeling medical cases, clinical conundrums, and other real-time tools have
con used, having a gut instinct that something is wrong, sensing largely replaced eedback rom autopsies and morbidity and mortal-
that other team members are overlooking important in ormation, or ity con erences. Case simulations and standardized patients, a now
P
A
observing improper procedures. They are then empowered to bring common eature o medical training programs, also present oppor-
R
these concerns to the attention o the rest o the team without ear tunities or trainees to receive real-time eedback.
T
o reprimand.
I
■ DECISION SUPPORT TOOLS AND OTHER
■ ERROR REPORTING AND ACCOUNTABILITY ELECTRONIC-BASED SYSTEMS
In order to learn rom medical errors clinicians must rst and ore- Electronic decision support tools (DSTs) include order sets, electronic
health record (EHR) embedded alerts and reminders, quality metric
T
most be encouraged to report them. Accomplishing this requires
h
the care ul cultivation o a nonpunitive culture in which team mem- dashboards, web-based di erential diagnosis engines (eg, ISABEL
e
and DXplain), and clinically oriented literature review resources such
S
bers are reed rom ear o reprisal. In order to urther de-stigmatize
p
as UpToDate and Epocrates. All o these have the potential o guid-
e
medical error and create a climate o transparency clinicians are now
c
ing physician behavior toward evidence-based practice (and away
i
being encouraged and in many institutions required to report medi-
a
l
rom anecdotal, bias-driven practice). Data regarding the e ective-
t
cal errors to their patients. These progressive views o error do not
y
ness o DSTs, except in very speci c domains such as medication
o
mean that clinicians should not be held accountable or their errors,
f
but rather that examination o errors should be or the purposes o dosing, or in institution-speci c applications, is mixed. One meta-
H
analysis ound that DSTs modestly improve morbidity but have no
o
quality improvement rather than punishment. Since many errors are
s
e ect on adverse events or mortality. Conversely, DSTs did seem to
p
at least in part due to f awed systems or a breakdown o multiple
i
t
improve adherence to preventative care guidelines and ordering
a
processes, an appropriate apportionment o responsibility between
l
clinical studies. There are no rigorous examinations o the potential
M
the hospital, clinician, and ancillary sta is warranted. In act, due
negative e ects o using DSTs. Nonetheless, DSTs provide a resource
e
to their presence throughout the hospital, hospitalists are uniquely
d
that is more ree rom the cognitive biases that inf uence much o
i
positioned to encounter, understand, and report on systems-based
c
i
physician clinical practice and may provide the occasion or a clinical
n
errors.
e
To get the most in ormation about medical errors hospitals “time-out” (Table 8-2).
a
EHRs themselves are an o t-overlooked orm o DST. By collating
n
should not only employ anonymous error-reporting systems, but
d
should actively solicit cases o medical error rom their sta . I these and organizing large volumes o data, well-designed EHRs acilitate
S
y
errors are not obvious because no substantial harm was caused or diagnostic ref ection, enhance collaboration, and provide a means
s
t
or receiving eedback. Increasingly, EHRs are also including ele-
e
because the patient nonetheless recovered, they might not other-
m
wise be reported. Even the reporting o “near-misses,” in which a ments o data analysis such as the ability to graph lab values over
s
time.
o
potential error is identi ed be ore it is actually committed, should
f
also be encouraged.
C
■ METACOGNITION
a
r
e
Metacognition is de ned as “thinking about thinking.” It involves
■ DELIBERATE PRACTICE
sel -questioning and ref ecting on personal biases and assump-
Rote practice (ie, expertise in any ield requires a minimum o tions. It implies an active, rather than an automatic, control o the
10,000 hours o practice) is not su cient. Instead, goal-directed, thinking process. Examples o metacognition include planning how
care ully measured, deliberate practice (ie, ref ective practice) is also to approach a cognitive task (eg, diagnosis) and how to evaluate
required. Without deliberate practice the mere accumulation o progress in that task. Several authors have suggested the use o a
experience leads to a plateauing o clinical skills and in some cases diagnostic checklist (Figure 8-7) to prompt metacognition.
even a decline as previously acquired knowledge is orgotten.
Deliberate practice is a training method that ollows the iterative DIAGNOSTIC DECISION MAKING
cycle o planning, per ormance, and ref ection. The planning phase
consists o goal setting, sel -motivation, and orientation toward ■ BEDSIDE DIAGNOSIS
learning. The per ormance stage consists o time-management, Bedside diagnosis with a comprehensive history and physical exam
care ul record-keeping, benchmarking, and seeking out expert assis- continue to orm the bedrock o clinical medicine despite the
tance and direction. The ref ective phase consists o sel -evaluation astronomical technical innovations o modern medicine. It is impos-
and studying shortcomings and ailures. sible to ully interpret even the best diagnostic testing without a
One review o research on deliberate practice ound that train- thorough prior knowledge o the patient. The adage that “ninety
ees should be given tasks with clearly de ned goals, motivated to percent o the diagnosis is in the history” remains true, even in the
improve, provided eedback as soon as practically possible, and 21st century. For example, the di erentiation o stable and unstable
given numerous opportunities to re ne their skills. Experts engaged angina (not to mention cardiac rom noncardiac causes o chest
in deliberate practice consciously seek out clinical challenges just pain) requires a skilled clinician at the bedside and carries dramatic
beyond the current skill level o trainees. Properly modeling such therapeutic implications.
habits or trainees allows them to become more sel -motivated, sel -
critical, and able to engage in li e-long improvement. ■ DIAGNOSTIC TESTS AND INTERVENTIONS
Some diagnostic tests are better screening tests or disease whereas
■ FEEDBACK others are better con rmatory tests. Some diagnostic tests are
Without eedback, it is impossible or clinicians to properly cali- power ul enough to screen and con rm disease. Many laboratory
brate their sense o diagnostic accuracy leading to both over- and tests do not simply report a positive (presence o disease) and nega-
undercon dence. Overcon dence has been identi ed as a requent tive (absence o disease) but rather a range requiring application
cause o diagnostic error. Feedback should be real time, speci c, and o the result to speci c patients. For example, a very low brain or
coupled with corrective instruction. B-type natriuretic peptide (BNP) level (eg, <50 pg/mL) is excellent at
50
excluding heart ailure in a patient presenting with a chie complaint accurate as i you had prevalence data, but more accurate than your
o dyspnea. Raising the cuto to either 100 pg/mL or 200 pg/mL is starting assumption. As you obtain more in ormation, you may con-
C
still good or the exclusion o heart ailure but not as power ul as tinue to incorporate this in ormation and re ne your posttest prob-
H
when using a lower cuto . Additionally, research has shown that ability using your knowledge to that point as the pretest probability.
A
patients with a history o le t ventricular dys unction who present
P
with dyspnea due to noncardiac causes o ten have a higher baseline
T
PRACTICE POINT
E
BNP level than those patients without a history o le t ventricular
R
dys unction. • Recognize that bedside diagnosis (ie, history and physical),
8
Consider the laboratory threshold cuto s used to report a test as diagnostic tests, and interventions all carry diagnostic power
abnormal. For example, a D-dimer assay cuto level 500 ng/mL per- and o ten times have been studied with published data on
orms very well as a rule out test in patients unlikely to have acute measures o diagnosis (eg, sensitivity and speci city) that can
D
venous thrombosis and/or pulmonary embolism. Lowering the be utilized when per orming Bayesian reasoning in clinical
i
a
g
cuto level to 250 ng/mL improves the ability o the D-dimer test to practice
n
exclude acute venous thromboembolism (VTE) with more certainty;
o
s
however, a lower cuto will increase alse positive results compared
t
i
c
with the traditional assay cuto level o 500 ng/mL. ■ PRETEST PROBABILITY
R
e
Pretest probability percentage represents the probability o a spe-
a
■ BAYES’ THEOREM
s
ci c pathology or a patient prior to initiating urther diagnostic
o
n
Bayes’ theorem quanti es how new evidence changes the probabil- testing. Pretest probability may be the prevalence o the disease in
i
n
ity that an existing belie is correct. Bayes’ theorem may be applied question in the population. At any point in the diagnostic workup,
g
the clinician may pause and estimate the prevalence or pretest
a
to any new clinical in ormation such as a historical nding, physical
n
exam nding, test result, or outcome ollowing an intervention. probability o disease in light o what in ormation is known about
d
the patient. The pretest probability o clinical gestalt is requently
D
Bayesian reasoning or analysis re ers to the pretest probability or
e
odds combined with the test result to the posttest probability or delineated as low, intermediate, and high probability or simplicity.
c
i
s
odds. Bayesian analysis o ers a simple construct or incorporating
i
o
new lab, radiology, and other clinical in ormation, as it arises, into ■ SENSITIVITY
n
M
our existing diagnostic hypotheses. For interested clinicians many A sensitive test correctly identi es patients who have the disease in
a
ree online calculators exist or doing quite sophisticated Bayesian question. To calculate sensitivity the number o patients who have
k
i
n
analysis. A simpli ed version o Bayes’ theorem states that the pre- the disease and test positive (ie, true positives or “TP”) is divided by
g
test odds multiplied times likelihood ratio equals the posttest odds. all who have the disease, including those who alsely test nega-
tive (ie, alse negatives or “FN”). Tests with high sensitivity are ideal
screening tests to discover as many patients as possible with the
Side bar 8-1 disease, requently with a tradeo o increased alse positive results.
Con rmatory testing may require a more speci c test.
The simpli ed Bayes’ theorem and likelihood ratios deal with
odds. Disease prevalence is quoted as a probability. To apply
TP
Bayes’ theorem one must be able to convert between the two.
Odds = Probability/1 – Probability and Probability = Odds/1 + TP+ FN
Odds. Consider a 43-year-old male with f ank pain presenting to
the emergency department. For such patients the prevalence ■ SPECIFICITY
o kidney stones is 20%. He undergoes renal ultrasound, which Speci city re ers to the ability o testing to recognize patients who
identi es a stone. Renal ultrasound has a likelihood ratio (LR) o 12 do not have the disease. To calculate speci city, the number o
i a stone is identi ed. patients who test negative and do not have the disease (ie, true neg-
atives or “TN”) is divided by the total number without the disease,
Step 1: Find the pretest probability (prevalence) o the disease: including those who alsely test positive (ie, alse positives or “FP”).
20%
Step 2: Find the LR o the Test: 12 TN
Step 3: Convert pretest probability to pretest odds:
TN+ FP
odds = probability/1 – probability = .2/.8 = .25
For example, D-dimer has high sensitivity and will be positive in
Step 4: Multiply pretest odds by the LR to obtain the posttest most cases o pulmonary embolism; however this comes at the cost
odds: o a high alse positive rate due to low speci city. A con rmatory,
.25 × 12 = 3 more speci c, imaging test may be required to make the diagnosis.
Step 5: Convert the posttest odds back to probability:
■ “SPIN” AND “SNOUT”
Probability = odds/1 + odds = 3/4 = .75 or 75% SPeci c tests rule IN disease (“SPIN”) while SeNsitive tests rule OUT
This method may be cumbersome, especially without a disease. (“SNOUT”).
calculator. As an alternative, some have suggested memorizing the
“Rule o 15’s”(Table 8-5) or keeping a nomogram (Figure 8-5) handy ■ POSITIVE PREDICTIVE VALUE
to estimate the change in the posttest probability rom the pretest Positive predictive value reports the probability that a patient has
probability based on the LR or a test. the disease a ter testing positive or it. Positive predictive value is
markedly dependent on the prevalence o the disease in question.
I data on prevalence is unknown and a clinician’s best guess is TP

the only starting point, Bayesian analysis may still be applied. Not as TP+ FP
51
■ NEGATIVE PREDICTIVE VALUE
TABLE 8-4 Using Likelihood Ratios to Calculate Posttest
Negative predictive value describes the probability that a patient Probability
does not have the disease ollowing a negative test result. Both
P
positive and negative predictive values are dependent on the preva- Step 1 Convert pretest probability to pretest odds
A
lence or pretest probability prior to testing.
R
pretest probability
T
= pretest odds
TN 1− pretest probability
I
TN+ FN Step 2 Multiply pretest odds by the likelihood ratio to
obtain posttest odds
■ ACCURACY pretest odds × likelihood ratio = posttest odds
T
h
Accuracy is the proportion o all test results, whether positive or
e
Step 3 Convert posttest odds to posttest probability
negative, which are correct. It assesses whether a test actually mea-
S
p
sures what it claims to measure (Figure 8-4). posttest odds
e
= posttest probability
c
i
1+ posttest odds
a
■ LIKELIHOOD RATIOS
l
t
y
o
Likelihood ratios are more help ul to make clinical decisions than
f
sensitivity and speci city. For a given diagnostic test, positive likeli-
H
o
hood ratios apply to positive results; negative likelihood ratios apply ■ POSTTEST PROBABILITY
s
p
to negative results.
i
Diagnosis depends on achieving or accepting a high posttest prob-
t
a
l
• POSITIVE LIKELIHOOD RATIO ability. Each diagnostic test may bring you closer to a diagnosis, but
M
The positive likelihood ratio is the likelihood that a patient with the there can rarely be 100% certainty. A posttest probability may be
e
d
disease tests positive compared to the likelihood that a patient with- calculated or any disease i two things are known: the pretest prob-
i
c
out the disease tests positive. I a test result is positive and the posi- ability o disease or the patient and the pertinent likelihood ratio
i
n
e
tive likelihood is greater than 1, then it is more likely than chance or the diagnostic test per ormed depending on whether the test is
a
that the patient has the disease. However i the likelihood ratio or positive or negative (Table 8-4).
n
d
that test is less than 1, then it is less likely the patient has this disease. The likelihood nomogram acilitates application o these calcula-
S
tions to clinical care, but it may not be help ul in clinical practice
y
s
to approach clinical problem-solving with such clinical precision
t
TP
e
m
sensivitiy %TP total with disease (Figure 8-5).
s
= = The “Rule o 15’s” is an easily remembered rule o thumb or
o
1− specificity %FP FP
adjusting pretest probability based on likelihood ratio. The sequence
f
C
total without disease o numbers 1, 2, 5, and 10 corresponds to positive likelihood ratio
a
r
values. For each progression in the sequence, the pretest probability
e
• NEGATIVE LIKELIHOOD RATIO increases by an additional 15% to arrive at a posttest probability. A
positive likelihood ratio o 1 or a positive test does not change the
The negative likelihood ratio is the likelihood that a patient with
the disease tests negative compared to the likelihood that a patient
without the disease tests negative.
.1 99
FN .2
1− sensitivity %FN total with disease
= = .5 95
specificity %TN TN
1 1000 90
total without disease 500
2 200 80
100
5 50 70
20 60
10 10 50
5 40
Patie nt has Patie nt do e s no t 2 30
(%) 20 1 (%)
dis e as e have dis e as e 30 20
.50
40 .20
50 .10 10
60 .05
True Pos itive Fa ls e Pos itive 5
Po s itive te s t 70 .02
re s ult .01
(TP ) (FP ) 80 .005 2
90 .002
.001 1
95 .5
Fa ls e Ne ga tive True Ne ga tive .2

Ne g ative te s t
re s ult 99 .1
(FN) (TN)
Pre te s t Like lihood Pos tte s t
proba bility ra tio proba bility
Figure 8 4 The 2 × 2 table. Figure 8 5 Nomogram.
52
eel com ortable not ordering an abdominal CT or questionable
TABLE 8-5 The Rule of 15’s appendicitis: 5%, 1%, or less than 1%? Lower testing and treatment
C
thresholds will nd and treat more disease, but at the cost o more
H
Likelihood Ratio Change in Probability
alse positives, adverse outcomes rom some ultimately unneces-
A
0.1 –45% sary interventions, radiation and contrast exposure, and increased
P
0.2 –30% cost burden or our already overburdened health care system, to
T
E
0.5 –15% name a ew. Ideally, any such threshold would be the point at which
R
1 No change bene ts o either testing or treatment outweigh the risks; however
8
2 +15% this remains a judgment call, and the real numbers behind these
decisions remain unknowable.
5 +30%
D
10 +45%
i
a
PRACTICE POINT
g
n
•
o
It is important to know the diagnostic testing and treatment
s
t
thresholds or certain patients and speci c illnesses being
i
c
probability o disease (ie, pretest = posttest). Similar adjustments or
considered.
R
negative likelihood ratios may be conducted by taking the inverse
e
a
o the corresponding positive likelihood ratios. The inverse o 2 is
s
o
0.5, which decreases probability by 15% (Table 8-5). Estimating
n
■ CLINICAL EXPERIENCE AND EXPERTISE
i
the change in probability or values that all between these notable
n
g
numbers is easy or positive likelihood ratios between 2 and 6, as Expertise in medicine may be de ned as reliably superior per or-
a
mance through highly developed perceptual and cognitive skills.
n
there is a 5% change or each single digit (ie, 2 = 15%, 3 = 20%,
d
4 = 25%, 5 = 30%, 6 = 35%). This simple sequence breaks down at In medicine this is mani est as an organized and e cient approach
D
higher values. to hypothesis generation and testing. Experts at medical diagnosis
e
c
share the ollowing attributes:
i
s
i
o
• Accurate calibration: match between observed and objectively
n
PRACTICE POINT
measured diagnostic skills.
M
• Utilize the “rule o 15’s”to quickly calculate posttest probability • Use o intuitive type I reasoning or common problems or
a
k
or a disease i the likelihood ratios are known or the diagnostic limited time and analytic type II reasoning or less common
i
n
g
test per ormed. problems, double-checking type I conclusions, or when time
more readily available.
• Extensive internal database o “illness scripts,” previously
■ DIAGNOSTIC TESTING AND TREATMENT THRESHOLDS acquired collections o signs, symptoms, and patient character-
Answering the question o whether to pursue urther diagnostic istics, to which a new patient presentation can be compared.
testing requires an understanding o testing and treatment thresh-
olds. The testing threshold is a pretest probability which is high ■ DIAGNOSTIC ERROR
enough to warrant urther diagnostic testing, while the treatment In 1999, the landmark Institute o Medicine report To Err is Human:
threshold is a pretest probability which is high enough orgo any Building a Safer Health System identi ed medical error as the th
more testing and warrants intervention or the presumed pathology leading cause o death in the United States. Diagnostic error includes
(Figure 8-6). incorrect diagnosis, missed diagnosis, and materially delayed diag-
These thresholds will vary depending on any number o actors nosis. It may proceed rom any cognitive ailure in the diagnostic
including the patient, the disease in question, and the clinician. process including ailure take an adequate history, per orm an ade-
For example, highly morbid disease processes, especially so-called quate exam, ormulate or prioritize a di erential diagnosis, create
cannot miss diagnoses, will have lower diagnostic and treat- and enact a diagnostic plan, correctly interpret test results, or seek
ment thresholds. At what pretest probability should the clinician specialist opinion. While many diagnostic errors are inconsequential,
in several studies in the United States and abroad, diagnostic errors
accounted or roughly 10% to 20% o all adverse events.
100%
Tre atme nt thre s ho ld
PRACTICE POINT
75% • A diagnostic error occurs when there is any mistake or ailure
in the diagnostic process that leads to a misdiagnosis, a missed
diagnosis, or a delayed diagnosis.
50%
Across many studies the diagnostic error rate, at least or clinical
specialties like internal medicine, has consistently been around 15%.
This consistency, however, belies the inherent di culty in studying
25%
diagnostic error. Diagnostic error is a process measure rather than
an easily measured endpoint. Diagnostic error is a concept built
Te s ting thre s ho ld on re erence to a gold standard such as a con rmatory test, clinical
0% outcome, second opinion, consensus review, or autopsy that not
in requently do not exist or do exist but are not readily available.
Figure 8 6 Arbitrary diagnostic testing and treatment thresholds. Error reporting systems have historically been underutilized and
53
A c he cklis t fo r diag no s is
Obta in your own, comple te me dica l his tory.

P
A
Pe rform a focus e d a nd purpos e ful phys ica l exa mina tion.
R
T
Ge ne ra te s ome initia l hypothe s e s a nd diffe re ntia te the s e with a ppropria te a dditiona l
I
que s tions, phys ica l exa mina tion, or dia gnos tic te s ts.
Pa us e to re fle ct—Ta ke a dia gnos tic time -out:
• Wa s I compre he ns ive ?
T
h
• Did I cons ide r the inhe re nt flaws of he uris tic thinking?
e
S
• Wa s my judgme nt a ffe cte d by a ny othe r bia s e s ?
p
e
c
• Do I ne e d to ma ke the dia gnos is NOW, or ca n I wa it?
i
a
l
• Wha t’s the wors t-ca s e s ce na rio? Wha t a re the “do not mis s ” e ntitie s ?
t
y
o
f
Emba rk on a pla n, but a cknowle dge unce rta inty a nd e ns ure a pa thway for
H
follow-up.
o
s
p
i
t
Figure 8 7 Achecklist for diagnosis.
a
l
M
e
d
i
c
i
even when employed o ten lack, within their schema, a category or
n
and o ten times have been studied with published data on
e
diagnostic error. There is no evidence base on which to rest conclu- measures o diagnosis (eg, sensitivity and speci city) that can
a
sions about when a diagnosis should or could have been made and
n
be utilized when per orming Bayesian reasoning in clinical
d
whether a more accurate or timely diagnosis would have a ected practice.
S
y
outcomes. • In the modern era o clinical practice, in ormation overload
s
t
e
is another pit all as is overreliance on diagnostic tests, which
m
CONCLUSION may lead to overtesting or “overdiagnosis”(ie, the diagnosis o a
s
Numerous authors have suggested clinicians should use a diagnos-
o
“disease”that will never cause a patient symptoms or death).
f
tic checklist to help avoid or mitigate diagnostic error. An example •
C
It is important to know the diagnostic testing and treatment
a
o a checklist or diagnosis can be ound in Figure 8-7. thresholds or certain patients and speci c illnesses being
r
e
considered.
PRACTICE POINT • Utilize the “rule o 15’s”to quickly calculate posttest probability
or a disease i the likelihood ratios are known or the diagnostic
• A di erential diagnosis is the ability to distinguish a disease test per ormed.
or condition rom others that present with similar signs, • A diagnostic error occurs when there is any mistake or ailure
symptoms, or diagnostic test results. in the diagnostic process that leads to a misdiagnosis, a missed
• List the top diagnosis rst but keep it problem oriented until diagnosis, or a delayed diagnosis.
the diagnosis has been made.
• Aggressively prioritize most likely then most harm ul in clinical
workups. SUGGESTED READINGS
• Exclude what diseases can easily be excluded and then remove
those rom di erential diagnosis. Bowen JL. Educational strategies to promote clinical diagnostic
• Keep a broad di erential until the top diagnosis has been reasoning. N Engl J Med. 2006;255:2217-2225.
con rmed and always plan or workup o other alternate Charlin B, et al. Scripts and clinical reasoning. Med Educ.
diagnoses i the top diagnosis is excluded. 2007;41:1178-1184.
• When presented with a clinical conundrum, begin by de ning Croskerry P. The importance o cognitive errors in diagnosis and
the problem, be deliberate and gather key in ormation, strategies to minimize them. Acad Med. 2003;78:775-780.
and then nally summarize the case when aced with large
amounts o in ormation. Croskerry P. A universal model or diagnostic reasoning. Acad Med.
2009;84:1022-1028.
• Utilize illness scripts, which can be very power ul tools or
clinicians and aid in diagnosis. Elstein AS, Schwartz A. Clinical problem solving and diagnostic
• Per orm the dual-process theory o reasoning and make sure decision-making: selective review o the cognitive literature. BMJ
to balance the intuitive and analytical components to maintain 2002;324:729-732.
highly e cient and e ective clinical reasoning. Elstein AS. Heuristics and biases: selected errors in clinical reasoning.
• A clinician should be cognizant o cognitive biases in clinical Acad Med. 1999;74:791-794.
reasoning and avoid them at all costs. Ericsson KA. Deliberate practice and acquisition o expert per or-
• Remember that patients do not always read the textbook.” mance: a general overview. Acad Emerg Med. 2008;15:988-994.
• Recognize that bedside diagnosis (ie, history and physical), Graber ML, et al. Cognitive interventions to reduce diagnostic error:
diagnostic tests, and interventions all carry diagnostic power a narrative review. BMJ Qual Saf. 2011;21:535-557.
54
Marcum JA. An integrated model o clinical reasoning: dual- ONLINE RESOURCE
process theory o cognition and metacognition. J Eval Clin Pract.
C
2012;18:954-961. Journal o the American Medical Association Collection. The Rational
H
McGhee S. Evidence-Based Physical Diagnosis, 3rd ed. St. Louis, MO: Clinical Exam. http://jama.jamanetwork.com
A
Saunders; 2012.
P
T
Norman GR, Eva KW. Diagnostic error and clinical reasoning. Med
E
Educ. 2010;44:94-100.
R
8
D
i
a
g
n
o
s
t
i
c
R
e
a
s
o
n
i
n
g
a
n
d
D
e
c
i
s
i
o
n
M
a
k
i
n
g
55
CHAP TER
9 INTRODUCTION
Medications are the principal tools doctors use to maintain health,
reverse illness, and extend patients’ survival, hope ully with good
quality o li e. Yet medications may also cause serious illness and
ail to have the desired e ect i they are used improperly. Medica-
tions may be extraordinarily expensive, and the cost to individual
patients, to hospitals, and to our health system has become almost
prohibitive. Thus, the proper use o medications and prescribing o
medications is critically important.
Principles of PRINCIPLES OF RATIONAL THERAPEUTICS
Evidence-Based “It’s not likely to be harm ul” is no justi cation or trying something
without demonstrated e cacy or the patient’s problem, unless
Prescribing the intervention is in the setting o a clinical trial or the patient is

in ormed o o -label use without evidence o bene t.
PRACTICE POINT
Brent G. Petty, MD
• Be ore any medication is ordered in a hospital or prescribed or
an outpatient, the prescriber needs to consider the
(1) e cacy, (2) sa ety, and (3) cost o the medication, in that
order o importance. Without e cacy or the condition being
treated, no medication should be given. There is a risk o toxicity
with virtually all medications, so there must be a consideration
o risk and bene t be ore starting or continuing medications.
EVIDENCE FOR EFFICACY

The quality o medical studies supporting the use o medications
varies widely. In recent years, the quality o data has been graded by
the groups reviewing the literature and making recommendations,
such as the Chest guidelines or anticoagulation (Table 9-1). These
grading systems consider the methodologies o the studies as well
as the strength o the results.
Among the di cult issues with clinical trials is whether they can
be extrapolated or all drugs in the same class. In general, extrapo-
lation across a class is somewhat hazardous, as drug ormulation,
absorption, duration o e ect, and sometimes drug interactions di -
er among drugs in the same class. Even with HMG-CoA reductase
inhibitors, whose e ects on LDL cholesterol are mostly a ected by
drug potency and can o ten be equated through adjustment o
dose, the e cacy related to clinical outcomes and adverse e ects
may vary. What is true or one drug in a certain class may not be true
or other drugs in the same class.
Another issue regarding the validity o clinical trials is the use
o “surrogate markers” in place o “hard clinical end points.” An
example is a reduction o human immunode ciency virus (HIV) RNA
levels as a surrogate or medication e cacy instead o extended sur-
vival in patients with acquired immune de ciency syndrome (AIDS).
Some surrogate markers have been demonstrated through rigorous
clinical studies to be closely associated with hard clinical end points,
providing assurance that they may be trusted as substitutes. Other
surrogate markers have less data to justi y their use as substitutes.
The hazard o using surrogate markers is exempli ed in a study o
interleukin-2 therapy in patients with HIVin ection, which showed a
substantial and sustained elevation o CD4+ cell count over a period
o 7 to 8 years average ollow-up, but demonstrated no improve-
ment in survival or the incidence o opportunistic in ections.
56
TABLE 9-1 Strength of Recommendations Grading System
C
H
Grade of Benefits vs Risks and Methodological Strength of
A
Recommendation Burdens Supporting Evidence Implications
P
Strong Bene its clearly Consistent evidence rom randomized Recommendation can apply to most
T
recommendation, high- outweigh risk and controlled trials without important patients in certain circumstances. Further
E
quality evidence (1A) burdens or vice versa limitations or exceptionally strong research is very unlikely to change our
R
evidence orm observational studies. con idence in the estimate o e ect.
9
Strong Bene its clearly Evidence rom randomized controlled Recommendation can apply to most
recommendation, outweigh risk and trials with important limitations patients in certain circumstances. Higher-
moderate-quality burdens or vice versa (inconsistent results, methodological quality research may well have an
P
r
evidence (1B) laws, indirect or imprecise) or very important impact on our con idence in
i
n
strong evidence rom observational the estimate o e ect and may change the
c
i
p
studies. estimate.
l
e
Strong Bene its clearly Evidence or at least one critical Recommendation can apply to most
s
o
recommendation, low outweigh risk and outcome rom observational studies, patients in certain circumstances. Higher-
or very low-quality burdens or vice versa case series, or randomized controlled quality research is likely to have an
E
v
evidence (1C) trials, with serious laws or indirect important impact on our con idence in the
i
d
evidence. estimate o e ect and may well change
e
n
the estimate.
c
e
Weak recommendation, Bene its closely Consistent evidence rom randomized The best action may di er depending on
-
B
high-quality evidence balanced with risk controlled trials without important circumstances or patient or societal values.
a
s
(2A) and burdens limitations or exceptionally strong Further research is very unlikely to change
e
d
evidence orm observational studies. our con idence in the estimate o e ect.
P
r
Weak recommendation, Bene its closely Evidence rom randomized controlled Best action may di er depending on
e
s
moderate-quality balanced with risk trials with important limitations circumstances or patient or societal values.
c
r
evidence (2B) and burdens (inconsistent results, methodological Higher-quality research may well have an
i
b
laws, indirect or imprecise) or very important impact on our con idence in
i
n
strong evidence orm observational the estimate o e ect and may change the
g
studies. estimate.
Weak recommendation, Uncertainty in the Evidence or at least one critical Other alternatives may be equally
low- or very low-quality estimates o bene its, outcome rom observational studies, reasonable. Higher-quality research is
(2C) risks and burden; case series, or randomized controlled likely to have an important impact on our
bene its, risk and trials, with serious laws or indirect con idence in the estimate o e ect and
burden may be evidence. may well change the estimate.
closely balanced
Another common outcome strategy in clinical trials is the use o (around 1-10 per thousand) will be identi ed. More rare (and o ten
“composite end points,” combining as an “event” any one o several more serious) side e ects may only become recognized with much
conditions, such as cardiac death, non atal myocardial in arction, more extensive use, involving tens o thousands o people. The expe-
and admission to a hospital or unstable angina. Obviously, all o rience with drugs such as troglitazone emphasizes the importance o
these conditions are de ensible as outcomes in patients with coro- postmarketing reporting o toxicities associated with newly approved
nary artery disease, but they are decreasingly reliable as “hard clinical medications to MedWatch (the FDA Sa ety In ormation and Adverse
end points” or an intervention intended to inf uence the course o Event Reporting Program) and/or to the manu acturer.
coronary artery disease. Especially when one o the three conditions There is a risk o toxicity with virtually all medications, so there must
contributing to the composite end point is the result o variable cli- be a consideration o risk and bene t be ore starting or continuing
nician judgment (eg, when to admit a patient or unstable angina), medications. In many cases, the toxicity emerges without warning
the reliability o the composite end point decreases. (“idiosyncratic”), such as rashes in response to sul a drugs. These
All three o these problems with assessing e cacy o medications “adverse drug events” are usually unpredictable and are not consid-
relate to how physicians evaluate and utilize randomized clinical tri- ered “medication errors.” In other cases, the possible toxicities o med-
als (RCTs) or making clinical decisions in their practices. The applica- ications can be identi ed and treated be ore they become clinically
tion o results rom an RCT to a speci c patient ideally depends on dangerous (eg, hypokalemia with loop diuretics or hyperkalemia with
assuring that the patient is similar to those described as subjects or angiotensin-converting-enzyme [ACE] inhibitors). These adverse drug
the study and that the intervention being tested will be provided in events are not medication errors either, unless the patient is not moni-
the same way described in the RCT. Both o these elements (similar- tored appropriately with occasional serum potassium measurements.
ity to the subjects and ollowing the intervention described in the
study) require a care ul review o the “Methods” section o the paper, COST
not just the “Abstract” and/or “Conclusions” section. The cost o medical care is staggering, partly ueled by the cost o
medications. The contribution o medication cost to overall health
SAFETY care expenses more than doubled rom 4.7% in 1982 to 10.5% in 2002.
Throughout all phases o drug development be ore drug approval Interestingly, while drug costs continued going up therea ter, the rate
(phases I, II, and III), sa ety is assessed, but at best these studies involve o increase in the cost o prescription drugs generally decreased over
only a ew thousand study subjects or the vast majority o drugs. the years rom 1998 to 2008 and leveled o to around 4% in the years
With this number o patients, only side e ects o moderate requency rom 2007 to 2011. In 2012, there was actually a 1% decline in total
57
dollars spent on medications in the United States, then the expen- Mobitz type I (aka Wenkebach), the occasional missed beat is o
ditures rebounded to a 3.2% increase in 2013 and then to a whop- no clinical consequence, creates no risk or the patient, and almost
ping 13% increase in 2014, the largest increase since 2001. This large always resolves without intervention. Treating such a patient with
increase was driven by the largest number o new molecular entities atropine or a pacemaker would be a mistake, introducing some risk
P
A
launched in more than a decade (numbering 42), by reduced patent o toxicity or complication or no clinical bene t, so no treatment is
R
expirations, and by increased price and volume o drugs ordered to the best approach in this case.
T
treat patients with hepatitis C virus (HCV). The high cost o medica- The prescriber should also consider coexisting medical condi-
I
tions a ects patients as individuals, who sometimes nd that they are tions that might likewise bene t rom the same therapy, as this may
unable to a ord their medications, and as a result these patients o ten magni y the bene t o the medication without adding additional
go without them. This “economic noncompliance” increases during risk o toxicity. For example, in a patient with hypertension who
di cult economic periods or when people have xed incomes and also su ers rom requent migraine headaches, a beta blocker or
T
h
must choose between paying or these medications or or their ood calcium channel blocker such as verapamil might be avored over
e
or mortgage or instance. The high cost o medicine also a ects hos- other medications because they may reduce the requency and/or
S
p
pitals and health systems. I hospitals and health systems would pay severity o the migraine episodes at the same time the blood pres-
e
less or their medications, they would have more unds available or
c
sure is being reduced.
i
a
capital improvements or expanded personnel services. Patients with potentially li e-threatening conditions (eg, meta-
l
t
y
Clinical trials have increasingly been including assessment o the static cancer) are o ten treated with potent medications with the
o
quality o li e saved, not just the survival rate. The measure o quality- potential o side e ects that are not only miserable but may also be
f
H
adjusted li e years (QALYs) is a standard and internationally recog- li e-threatening. When treatments are similar in e cacy but di er in
o
s
nized method to assess the relative bene t o medical interventions. types o toxicities, the patients’ pre erences are important, since hair
p
i
It combines duration o survival and the quality o li e during each loss may be more adverse or some patients than risk o in ection or
t
a
l
year o li e. Although one treatment might help patients live longer, incidence o diarrhea. Tailoring the medications used in such cases
M
it might also have serious side e ects (eg, it might make them eel preserves the patient’s autonomy and properly respects his or her
e
d
sick or put them at risk or other illnesses). Another treatment might right to choose among reasonable options. This is an example o the
i
c
not extend survival but it may improve quality o li e (eg, by reduc- important principle o shared decision making.
i
n
e
ing pain or other symptoms o disease). The quality o li e rating
a
can range rom 0 (ie, worst possible health) to 1 (ie, best possible PRACTICE POINT
n
d
health). Having the QALYmeasurement allows one to consider cost
• When treatments are similar in e cacy but di er in types o
S
e ectiveness (ie, how much the drug or treatment costs per QALY).
y
toxicities, the patients’pre erences are important. Tailoring
s
This is the cost o providing a year o the best quality o li e available,
t
e
the medications used in such cases preserves the patient’s
m
which could be one person receiving one QALY, but is more likely
autonomy and properly respects his or her right to choose
s
to be a number o people receiving a portion o a QALY (eg, our
o
people receiving 0.25 QALY). In this example, cost e ectiveness is among reasonable options.
f
C
expressed as dollars per QALY.
a
r
Cost e ectiveness analysis is another increasingly popular
e
INDIVIDUAL RISK
approach to assess the impact o interventions that may have
Sometimes patient characteristics create special susceptibility to
nancial bene t. For example, aspirin’s cost is much lower than the
adverse events. For example, the toxicity seen with the nucleoside
cost o caring or the number o heart attacks and strokes it pre-
reverse-transcriptase inhibitor abacavir causes a hypersensitivity
vents. Sometimes the bene t is secondary or indirect. For example,
reaction in about 5% to 10% o patients. This reaction usually occurs
acetylcholinesterase inhibitors are reported to cause a temporary
in the rst 2 months o treatment and is su ciently severe enough
delay in the cognitive decline o patients with dementia. I this
that it requires discontinuation o the drug. The mani estations
delay in cognitive decline can prevent a patient rom requiring
include ever, rash, and respiratory, gastrointestinal, and constitu-
institutionalization or ull-time care at home or a period o months
tional symptoms. The reaction was ound to be associated with the
or years, the costs o such care may be much more than the cost
HLA-B*5701 gene variant. Investigators in Australia have demon-
o the medication and so the medication would then be deemed
strated that screening with genotyping be ore instituting abacavir
cost e ective. Policymakers, including governmental bodies, pay-
therapy is e ective in reducing the number o such reactions. In act,
ers, and inf uential oundations, are interested in maximizing cost
there were no hypersensitivity reactions in the HLA-B*5701-negative
e ectiveness. They are convinced, with some justi cation that many
patients. This exempli es the value o pharmacogenomic biomark-
practices and interventions might well be replaced with less costly
ers to enhance the treatment o patients with medications.
approaches, without diminishing the quality o the care and the
Patients age 65 or over constitute about 15% o the US popula-
bene t our patients derive. “Choosing Wisely,” an initiative o the
tion, but they consume around 30% o the medications prescribed.
American Board o Internal Medicine (ABIM) Foundation and Con-
The natural deterioration o both renal and hepatic unction with
sumer Reports, is a good example.
age makes older patients more susceptible to toxicity rom the
OTHER FACTORS INFLUENCING MEDICATION SELECTION regular use o medications. Figure 9-1 demonstrates the altered
pharmacokinetics a ter an intravenous dose o verapamil in an
■ PATIENT PREFERENCES AND VALUES 82-year-old man as compared to a 23-year-old man. Elimination is
With rare exception, prescribers have a number o possible medications delayed and the peak blood concentration o verapamil is higher
or managing diseases, and each may cause likely responses, either good in the older man, perhaps related to a modest change in volume
or bad, in addition to the intended response. In all cases, the patient’s o distribution. Figure 9-2 shows the relationship o age and
inclination to accept the proposed therapy should be considered. intravenous diazepam dose needed to achieve adequate sedation
The very choice o initiating medication treatment or not or a procedure. As predicted by pharmacokinetics, the dose
should be weighed. It is always an option in medicine to do noth- needed or an older adult is less than that needed or a younger
ing (ie, o er no treatment), and sometimes no treatment is the best patient. Figure 9-3 shows the relationship o age and serum
option. For example, in a patient with an acute in erior wall myocar- concentration o diazepam needed to achieve adequate sedation
dial in arction who develops second-degree atrioventricular block, or a procedure. Note that the concentration required diminishes
58
100
70
C
H
50
Ve ra pa mil 10 mg IV
A
40
)
82-ye a r-old ma le
P
L
n
o
30
m
23-ye a r-old ma le
T
i
t
/
a
g
E
r
n
t
20
(
n
R
e
n
c
o
n
i
9
t
o
a
c
r
t
n
m
10
e
u
c
r
n
e
P
7
o
S
c
r
i
l
n
i
5
m
c
a
i
4
p
p
a
l
r
e
3
e
s
V
o
2 0 Age
E
v
i
Figure 9 3 Relationship of age and serum concentration required to
d
e
1 achieve a desired effect.
n
0 4 8 12 16 20 24
c
e
Time (hours )
-
B
a
to nd. In act, most investigators begin with the expectation o
Figure 9 1 Altered pharmacokinetics in the elderly. (Reproduced, with
s
e
showing a di erence between the test compound and either stan-
d
permission, rom Schwartz JB. Clinical pharmacology. In: Hazzard
dard treatment or inactive treatment (ie, placebo). For example, i
P
WR, et al., eds. Principles of Geriatric Medicine and Geronotology,
r
one were comparing the e ect o two HMG-CoA reductase inhibi-
e
3rd ed. New York, NY: McGraw-Hill; 1994.)
s
tors (aka statins) on serum cholesterol, a null hypothesis could be:
c
r
i
“There is no di erence between atorvastatin and rosuvastatin in
b
i
n
steadily with advancing age, demonstrating that older adults are patients with hypercholesterolemia and symptomatic coronary
g
more sensitive to diazepam than younger patients. So or both o artery disease.” The study requires a su cient sample size to attempt
these reasons (delayed elimination and increased sensitivity to the to disprove the null hypothesis with a certainty o at least 95% that
medication), dosing diazepam is best accomplished with particular the degree o di erence between the response to the two drugs is a
caution in older adults. While this relationship between medication true di erence and not just the result o chance (alpha or type I error
concentration and age is not seen with all benzodiazepines, starting = 0.05). Investigators may expect that there will not be a signi cant
at low doses and increasing slowly, according to individual patient di erence between the two arms and employ what is called a non-
response, is an especially good principle when prescribing medica- in eriority study. Con rming that two medications are equivalent or
tions or older adults. that one is nonin erior requires a larger sample size than needed or
con rming that medications are di erent. The beta, or type II, error is
ASSESSING THE EVIDENCE normally set at 0.2, but when investigators want more certainty that
the observed similarity is more likely to be true than just the result
Well-designed, randomized, controlled, “blinded” or “masked,” and
o chance, the beta error may be reduced to 0.1.
prospective clinical trials provide the strongest evidence to direct
Once the study is completed with the intended sample size, the
medical practice. Each o these elements increases the likelihood
results are analyzed. The most balanced approach is to assume
that the results o the study can be accepted rather than be the
that either o the two groups could be superior to the other, which
result o chance. The study question is ideally ramed as a null
leads to a two-tailed statistical test. It is especially interesting to see
hypothesis, which is o ten not what the investigators actually expect
how each arm per ormed compared to the predicted response.
The analysis can determine whether one group had a more avor-
able outcome than the other, and by how much they di ered. The
di erence is statistically signi cant i it is less likely than 5% to have
reached that di erence through chance alone. The 5% threshold is,
o course, arbitrary as a level to embrace an observation with abso-
lute conviction versus 6% to discount the di erence as nothing very
meaning ul. In act, when the di erence reaches a 6% degree o cer-
tainty or being beyond a chance nding, it seems inappropriate to
e
say that the per ormance o the groups was “not di erent.” The truth
s
o
is that the groups’per ormance was di erent, but that di erence did
D
not reach the accepted level o statistical signi cance. In such cases,
one o ten hears the term “trend” used to describe the di erence,
and the di erence would likely have reached statistical signi cance
i the study’s sample size were larger and the proportional responses
continued to occur with the same requency in additional subjects.
Too o ten readers ignore the “Methods” section o published
0 papers, giving their limited time and attention instead to the “Abstract,”
Age a ew gures or tables in the “Results” section, and the highlights o
Figure 9 2 Relationship of age and dose required to achieve a desired the “Discussion” section. This approach may save time, but it ignores
effect. the critical in ormation about characteristics o the study population
59
recruited, what kinds o patients were excluded, how other medi- upon trans er or at discharge should take into account not only
cations were managed, and many other aspects that ultimately the medications the patient was taking in the hospital just be ore
determine whether the results o the study are valid and whether trans er or discharge, but also the patient’s “home medications.”
they can be applied to any other population/patient group besides The purpose o this dual consideration is to avoid costly and poten-
P
A
those enrolled in the study. The paper rises or alls on its methods, so tially hazardous duplication o medications. As already explained, a
R
results or conclusions are not valid i the procedures involved with patient may receive one proton pump inhibitor while in the hospital
T
conducting the study are seriously f awed. (eg, pantoprazole), which is di erent than the one taken at home
I
(eg, omeprazole) or one that could be prescribed at discharge (eg,
THE HOSPITAL FORMULARY lansoprazole). Patients have been known to be taking supplies o
both war arin and “Coumadin” ollowing hospital discharge because
Decisions about which speci c drugs are included in a hospital’s or-
one had been provided by prescription rom the amily doctor and
mulary are usually made by a multidisciplinary ormulary committee,
T
h
the other was prescribed by the hospital doctor(s). Since generic
which includes physicians and other prescribers, nurses, pharmacists,
e
and brand products may look di erent, it is not hard to understand
S
and others. This group reviews the available data on e cacy, sa ety,
p
how patients may not recognize a hazardous duplication.
and cost o products proposed to be added to or removed rom the
e
c
hospital ormulary. To be an e ective member o such a committee,
i
a
THERAPEUTIC DRUG MONITORING
l
t
an individual would need to understand the importance o e cacy,
y
sa ety, and cost as they relate to the population o patients served by Treatment o any patient should ollow the “ideal therapeutic algo-
o
f
the hospital. Clear thinking is especially important as the committee rithm” (Table 9-2). First, the prescriber should have a therapeutic
H
goal in mind, whether it is to lower the blood pressure to a certain
o
members consider the hospital’s wel are in addition to advocacy or
s
point, reduce the hemoglobin A1c below a certain threshold, or
p
individual patients or a small group o patients. Workf ow issues or all
i
t
members o the care team are important. Likewise, the incremental drive the LDL cholesterol down under 100 mg/dL. With the goal
a
l
cost o one medication compared to another may ultimately inf u- in mind, an appropriate agent is selected, and then an appropriate
M
dose o the agent is chosen. When relevant patient characteristics
e
ence whether su cient money is available or other hospital priori-
d
ties, such as hiring or maintaining sta members. It is the balance o or concomitant medications are known, the dose may be individu-
i
c
i
alized somewhat. A ter allowing a su cient period o time or the
n
these multiple issues that makes the work o the ormulary commit-
e
tee interesting and important. intervention to reach a substantial or peak e ect, which may be days
a
or weeks, a repeat measurement is per ormed and is compared to
n
During a hospitalization, the patient may receive a di erent drug
d
than what he or she was taking at home be ore admission. This may the pretreatment reading and the therapeutic goal. Then, whatever
S
y
be the result o provider pre erence or ormulary restriction. Hospital the starting dose may have been, adjustments in the dose may well
s
t
be needed to achieve the therapeutic goal. A ter the response to
e
ormularies are either open or closed, and may have additional
m
restrictions. Open ormularies allow prescribers to order any mar- the new dose is observed, another adjustment in dose, or adding
s
keted product, and the patient will get whatever speci c product or substituting another medication, can be considered. All the while
o
f
was ordered. Closed ormularies limit the selection o medications there is monitoring or evidence o adverse e ects.
C
“Therapeutic drug monitoring” is a term that usually implies
a
to a small number o products within either a chemical class or an
r
e
indication class. For example, rather than having all H2 receptor the measurement in some body f uid o a substance that is either
antagonists and proton pump inhibitors on the hospital’s ormulary, the medication that is being monitored or a related substance.
the hospital may restrict the choice to amotidine and pantoprazole. Therapeutic drug monitoring is best employed when certain cri-
These determinations are generally made based on the assumptions teria can be met (Table 9-3). I measuring a physiological result
o (1) equal, or at least adequate, e cacy, (2) no worse toxicity pro le (eg, prothrombin time) or i a drug concentration is part o the
or the selected product, and (3) substantial cost savings. monitoring (eg, phenytoin serum level), one must be con dent that
the laboratory to be used can measure the item accurately and in
a timely ashion. Then, it must be known that the e cacy o the
MEDICATION RECONCILIATION
drug is enhanced or the toxicity o the drug is reduced by adjust-
When the history, physical examination, and relevant laboratory data ing the dose o the medication. We should avoid the temptation to
have been obtained, treatment begins. The treatment may be either measure drug concentrations just because we can. Achieving and
speci c (ie, based on the establishment o a speci c diagnosis) or maintaining results in a “therapeutic range” should reduce the risk o
empiric (ie, based on the best guess o diagnosis using the available toxicity or improve e cacy, or ideally both. I the e cacy or toxicity
evidence and considering the usual etiology responsible or the con- o a medication cannot reliably be improved by adjusting the dose
dition, such as the most likely bacterial pathogens or a community- to achieve a result in the therapeutic range, then therapeutic drug
acquired pneumonia). The initiation o new medication in the hospital monitoring is not o value.
or in the outpatient setting must be ramed on the background o
the medication(s) that the patient previously had taken. Medication
reconciliation re ers to the cognitive process o considering the TABLE 9-2 Ideal Therapeutic Algorithm
immediate previous medications (eg, the patient’s “home medica-
tions”) when ordering new treatment. Medication reconciliation is not 1. Determine the therapeutic goal.
simply duplicating the patient’s “home medications” (eg, outpatient 2. Choose an appropriate agent.
medications) into the hospital’s order system, but rather a thought ul 3. Choose the appropriate dose, individualizing or each patient
consideration o the value and appropriateness o providing each when possible.
medication in light o the patient’s new medical situation. There 4. Know when/how to monitor or e ectiveness and sa ety,
should be a conscious decision, or each and every medication, including the essential criteria or appropriate therapeutic
whether to stop, continue, or modi y administration o the drug. drug monitoring.
At the time o trans er to a new service or level o care and at the 5. Know how to adjust the therapy (eg, increase the dose, add
end o the patient’s hospitalization, another medication reconcilia- another medication, switch to another agent etc) to attain
tion should occur. This process di ers rom the one at the time o the therapeutic goal and avoid toxicity.
admission because the consideration o medications to prescribe
60
marketing motivation o the suppliers o the samples. Samples may
TABLE 9-3 Criteria for Appropriate Therapeutic Drug be allowed in the o ces o individual practitioners or in the clinical
Monitoring
C
space o a multidisciplinary group. Some hospitals have centralized
H
1. Medication concentration or e ects can be measured reliably samples into the pharmacy to be dispensed to the medically indi-
A
and accurately. gent with special prescriptions, while other hospitals have orbidden
P
drug samples altogether. There are justi cations on all sides o this
T
AND
E
issue, but i samples are allowed in an o ce, practice, or hospital,
2. The e icacy o medication treatment can be enhanced by
R
their use should be documented in each case they are dispensed,
achieving a certain concentration or e ect range.
9
the patient should be supplied with product in ormation (eg, rom a
AND/OR pharmacy), and expiration and recalls should be monitored.
3. The toxicity o medication treatment can be reduced by Samples are not the only strategy used by industry to inf uence
P
maintaining a certain concentration or e ect range. prescribing and medication-ordering habits. Direct-to-consumer
r
i
n
advertising, gi ts, grants, support o clinical investigation, journal
c
i
p
advertising, and even unrestricted donations to hospitals and medi-
l
e
cal schools have the potential to introduce a sense o obligation and
s
Measuring drug concentrations in plasma or serum establishes
o
indebtedness in those who inf uence the speci c medications that
individual patient pharmacokinetics. One well-done drug concen- patients receive. The issue is complicated, but we should institute
E
v
tration is more valuable than any algorithm that seeks to predict measures to minimize our biases or their e ects. We must act as
i
d
concentration or e ect using patient characteristics, comorbidities,
e
objectively as possible or the bene t o our patients.
n
or other actors. Poorly done therapeutic monitoring may produce
c
“Academic detailing” is a concept that has been proposed to help
e
results that are misleading, and in this way are worse than having no
-
reduce undue inf uence rom industry. Academic detailing involves
B
testing at all. The duration o an in usion and the correct timing o
a
the distribution o knowledge rom trusted medical personnel, o ten
s
e
the sample a ter the in usion are critical to having results that can be the leaders o academic departments or divisions, government
d
assessed in light o published data and guidelines. Especially hazard- advisers (eg, rom the U.S. Food and Drug Administration, Centers
P
r
ous is drawing blood samples or drug concentrations too soon a ter or Disease Control and Prevention, or National Institutes o Health),
e
s
an intravenous dose o a drug, which may put the sample in the
c
or other external experts who can be invited into hospitals to edu-
r
i
period o the alpha hal -li e or distribution phase rather than in the cate the hospitals’ medical sta s on the pros and cons o various
b
i
beta hal -li e or elimination phase, which is a more predictable and
n
medications. Leaders and other members o the hospital’s ormulary
g
interpretable portion o the drug elimination curve. On the other committee, as well as leaders and members o the hospitalist group
hand, especially with oral medications, checking drug concentra- at the hospital, may be considered or academic detailing i they
tions a ter too ew doses are given to have the concentrations at or have appropriate knowledge about the medications under discus-
near steady state may lead to an underestimation o the adequacy sion and are ree rom conf icts o interest that would potentially
o a dose, and a premature increase to a higher dose may lead to a ect their opinions.
serious toxicity when the drug concentration does achieve steady
state at a level too high or sa ety. SUGGESTED READINGS
ROLE OF PHARMACISTS IN ASSISTING WITH Guyatt GH, Norris SL, Schullman S, et al. Methodology or the devel-
MEDICATION ORDERING
opment o antithrombotic therapy and prevention o thrombosis
More hospitals are employing pharmacists to assist prescribers in guidelines: antithrombotic therapy and prevention o thrombosis.
their management o patients. These trained pro essionals are espe- Chest. 2012;141:53S-70S.
cially knowledgeable about medication issues, including indications
IMS Institute or Healthcare In ormatics, Press Releases, IMS Health
or medications, their doses (in both normal and physiologically
Study. 2014 A Record-Setting Year or U.S. Medicines. http://www.
impaired patients), drug-drug and drug- ood interactions, and other
us.imshealth.com. Accessed May 9, 2015.
important matters in medication use. They are amiliar with resources
that help identi y medications, including oreign and generic prod- INSIGHT-ESPRIT Study Group and SILCAAT Scienti c Committee.
ucts. Some pharmacists serve on hospital policy-making committees Interleukin-2 therapy in patients with HIV in ection. N Engl J Med.
because o their perspectives related to drug dispensing and moni- 2009;361:1548-1559.
toring. In some hospitals, they see patients with conditions such as Mallal S, Phillips E, Carcosi G, et al. HLA-B*5701 screening or hyper-
hypertension and evaluate the propriety o the patients’medications, sensitivity to abacavir. N Engl J Med. 2008;358:568-579.
the patients’ knowledge o their medications and how to use them, Petty BG. Trends in medication use: implications or medication
and the most likely adverse events that the patients may encounter. errors. J Pharmacist Fin Econ Pol. 2006;15:137-174.
Pharmacists round with care teams and provide in ormation on
medications during the discussions about the patients. In some Reidenberg MM, Levy M, Warner H, et al. Relationship between
states, pharmacists may be granted ordering authority by the hos- diazepam dose, plasma level, age, and central nervous system
pital, and pharmacists may sta anticoagulation monitoring clinics. depression. Clin Pharm Ther. 1978;23:371-374.
Pharmacists, like nurses, are the physicians’ compatriots and can help Schumock GT, Li EC, Suda KJ, et al. National trends in prescription
physicians avoid making serious errors. As important team members, drug expenditures and projections or 2015. Am J Health Syst
they should be heeded, respected, and appreciated. Pharm. 2015;72:717-736.
MANAGEMENT OF DRUG SAMPLES ONLINE RESOURCES

The topic o drug samples is only a portion o the larger topic o
the relationship o prescribers and hospitals to the pharmaceuti- • Choosing Wisely. An initiative of the ABIM Foundation. http://
cal industry. Whether to allow samples in a practice or hospital at www.choosingwisely.org/
all can be very controversial. At balance is the advantage o “ ree” • MedWatch. The FDA Safety Information and Adverse Event Report-
medication or those patients who cannot a ord it versus the clear ing Program. http://www. da.gov/Sa tey/MedWatch/
61
CHAP TER
10 INTRODUCTION
Summary literature represents the highest quality o evidence and
the top portion o the evidence-based medicine pyramid. In this
chapter, summary literature re ers speci cally to practice guide-
lines and systematic reviews. I done properly, authors will have
per ormed a comprehensive systematic review o the literature
available as well as summarized ndings rom the available litera-
ture or both o these types o manuscripts. For practice guidelines,
the summary typically exists in the orm o graded recommenda-
Summary Literature: tions. For systematic reviews, the summary comes in the orm o a
meta-analysis where a quantitative statistical analysis is conducted
Practice Guidelines o pooled data rom several separate but similar studies. Although
articles eaturing literature reviews (ie, narrative review, aka “review
and Systematic articles”) by authors exist or many clinical topics in medicine, the
methods used or reviewing the literature in these articles are not
necessarily systematic or comprehensive and it is typically le t
Reviews to the authors’ discretion or which data and studies to include.
Table 10-1 o ers a comparison o narrative reviews, systematic
reviews, meta-analysis, and guidelines.
Dustin T. Smith, MD PRACTICE GUIDELINES

Bhavin Adhyaru, MD ■ INTRODUCTION
Hospitalists ace the challenging task o managing patients with
a diverse array o important and sometimes complicated medical
conditions. Clinical practice guidelines exist and are designed to
assist both health care providers and patients in making appropriate
decisions regarding clinical care or the prevention, diagnosis, and
treatment o health care conditions. Practice guidelines are system-
atically developed statements summarizing available medical litera-
ture or a speci c clinical circumstance. The summary or practice
guidelines usually comes in the orm o graded recommendations
based on the amount o data and quality o evidence that exists or a
speci c clinical circumstance. Ideally, practice guidelines should be
presented in a user- riendly ormat that allows hospitalists the ability
to deliver the best medical care possible.
■ BENEFITS
Medical practice guidelines acilitate consistency, e ciency, and
e ectiveness to improve health care outcomes. Good clinical
practice guidelines have been compared to receiving good recom-
mendations rom a consultant. Practice guidelines should provide
clear recommendations, discuss alternatives, acknowledge biases,
and consider extenuating circumstances or a speci c clinical case.
■ BARRIERS
Physician adherence to guidelines has been demonstrated to be
surprisingly low at times. Guidelines have been shown to have had
a limited e ect on changing physicians’ practices and behaviors
despite widespread implementation. Incorporating guidelines into
the daily practice ace numerous barriers, including both internal
(eg, provider awareness or amiliarity) and external (eg, environmen-
tal-related or systems-related actors). Some physician perceived
barriers to using guidelines include concerns or reduced autonomy
and/or loss o the patient-physician relationship when ollowing an
oversimpli cation or “cookbook” approach to medicine or that they
may not be practical or easy to use. Practice guidelines may not
be applicable to a clinician’s patient and/or practice population
62
TABLE 10-1 Comparison and Contrast of Overview Article Types
C
H
Narrative Review Systematic Review Meta-Analysis Guideline
A
Focused Clinical Question Not generally Yes Yes Yes
P
(Multiple questions addressed)
T
E
Structured Search Strategy No Yes Yes Yes
R
Selection o Included Articles Dependent on author Systematic Systematic Systematic
1
Quality Assessment o included articles No Yes Yes Yes
0
Qualitative or Quantitative result Qualitative Qualitative Quantitative Qualitative and Quantitative
Sources/Example UptoDate or Dynamed Cochrane Review Cochrane Review http://www.guideline.gov
S
u
m
m
a
r
■ GROUP COMPOSITION
y
(eg, outpatient management guidelines when a providers’sole orm
L
o practice is the inpatient setting). Concerns regarding the credibil-
i
The validity o clinical practice guidelines is markedly impacted
t
e
ity o a guideline may have existed in the past; however, numerous by the composition o the developers. Selection o a multidisci-
r
a
pro essional groups have worked to improve the standards and
t
plinary team to review data and generate the guidelines should
u
credibility o practice guidelines.
r
decrease the chance or bias. A guideline that is developed by only
e
:
one group o medical subspecialists may be biased to likely recom-
P
r
■ STANDARDS TO IMPROVE CREDIBILITY
a
mend an intervention that alls within their realm o practice. The
c
t
In 2012, the Guidelines International Network (G-I-N) proposed stan- multidisciplinary team should include both generalists and special-
i
c
e
dards or clinical practice guidelines. This global network comprises ists, providers rom both medicine and surgery i applicable, as well
G
numerous organizations and individual members rom countries as experts rom other allied health pro essions such as nursing, phar-
u
represented on all continents o the world to support evidence- macy, and therapy elds (eg, respiratory therapy i applicable). Mem-
i
d
e
based health care and improved health outcomes by reducing inap- bers should have expertise in the scope o practice the guideline is
l
i
n
propriate variation throughout the world. Their published mission is trying to address and may have extensive research backgrounds
e
s
to lead, strengthen, and support collaboration and work within the in the general or speci c eld o medicine or which the guideline
a
guideline development, adaptation, and implementation commu- is being developed. Many pro essional medical societies sponsor
n
d
nity. The G-I-N addressed the key components o high quality and guideline development so o ten times the group is comprised o
S
trustworthy guidelines, that is, composition o guideline develop- members rom the actual organization sponsoring the guideline.
y
s
t
ment group, decision-making process, conf icts o interests, scope Additionally, many well- ormed groups who develop clinical prac-
e
m
o a guideline, methods, evidence reviews, guideline recommen- tice guidelines will include an expert in statistics, evidence-based
a
dations, rating o evidence o and recommendations, peer review medicine, and/or grading recommendations.
t
i
c
and stakeholder consultations, guideline expiration and updating,
R
and nancial support and sponsoring organization. The G-I-N also ■ CONFLICTS OF INTERESTS
e
v
houses the world’s largest international guideline library.
i
As or any publication, authors must report any conf icts o interests
e
w
The Appraisal o Guidelines or Research and Evaluation Enter- somewhere in the text or each individual. Any and all relevant
s
prise (AGREE) is another group that has worked to address the nancial relationships with any commercial interests related, directly
variability in practice guideline quality by creating the AGREE Instru- or indirectly, to the subject o the clinical practice guideline must
ment. The AGREE Instrument is a tool that assesses methodological be addressed in the guideline. I members o the group comprising
rigor and transparency in guideline development and use. The the guideline do have any signi cant conf icts o interests, this may
AGREE II is comprised o 23 items and organized into six quality a ect the validity o the guideline in a negative way. The role o the
domains or practice guidelines: scope and purpose, stakeholder sponsoring organization in ormulating the guidelines should also
involvement, rigor o development, clarity o presentation, applica- be made clear in the text o the guidelines.
bility, and editorial independence. A modi ed ramework in both
interpretation and assessment o the quality o clinical practice ■ LITERATURE REVIEW
guidelines is divided into three general domains: validity, results,
The validity o a clinical practice guideline is markedly a ected by
and applicability.
the completeness o the search strategy and review o literature
used to ormulate recommendations. The search strategy should be
■ VALIDITY
reported a ter the introduction but be ore the recommendations
The credibility o clinical practice guidelines is markedly swayed in the guideline similar to a “methods” section in a research manu-
by the validity in which guideline developers gathered, appraised, script. The review typically ollows the protocol o a valid systematic
and combined evidence to ormulate recommendations. Clinicians review and meta-analysis. The search or relevant studies includes
should consider the ollowing components that determine the electronic databases (eg, National Center or Biotechnology In or-
validity o a guideline: scope and purpose, group composition, con- mation PubMed), con erence abstracts/symposia, and re erences
f icts o interests, literature review, group processes, external review provided by the identi ed studies. In some instances, researchers o
and endorsements, editorial independence, and current guidelines. previous studies may need to be contacted in order to gather nal
or complete results i they have not been published and in order to
■ SCOPE AND PURPOSE (AKA PRIORITY SETTING) avoid publication bias. The search terms, synonyms, di erent spell-
Well-written clinical practice guidelines should begin by describing ings (eg, anemia and anaemia), and Boolean Operators used are also
the scope and purpose o the guidelines. The scope and purpose important to ensure an exhaustive search and provide evidence o
o ten include the target patient population including but not limited validity or the guidelines. The criteria or including and excluding
to parameters or age, gender, ethnicity, and disease type/subtypes. studies used to ormulate the guidelines should be de ned and
63
reported within the text. The process o reviewing included stud- o the guideline. I this is the case, guideline developers should
ies or quality should be elucidated within the text. Additionally, explain which evidence was considered or not considered in mak-
other limits or the search strategy and literature review such as by ing the nal recommendations.
language (eg, articles published in English only), date o publication,
P
A
demographics, or subgroups should be reported and considered ■ RESULTS
R
when assessing the validity o guidelines. Some o this may be clear The results o clinical practice guidelines di er rom that o an
T
in the scope and purpose o the clinical practice guideline but an observational study, clinical trial, or meta-analysis. Whereas these
I
explicit description o the literature review used to collect data in publications typically report results in statistical and numerical
ormulating the clinical practice guideline is crucial in determining terms, practice guidelines provide results in the orm o recommen-
the validity o the guideline. dations as both the results o the systematic review and conclusions
made by the guideline developers. Practice guidelines that are
T
■ GROUP PROCESSES
h
summarized appropriately should involve graded recommenda-
e
The process used by the group developing the guidelines should tions (ie, the recommendation receives an assigned grade by the
S
p
be made clear in the text o the guidelines. Numerous group pro- experts involved in developing the guidelines). There are numerous
e
c
cesses exist which have been validated (eg, Delphi method) but a grading systems in place or recommendations and many guideline
i
a
l
particular clinical practice guideline may be distinctive and require developers adopt a system that is endorsed or created by either
t
y
a unique process. The process should be de ned, systematic, and air the sponsoring organization or pro essional society representing
o
f
while still allowing members o the group to equally contribute to the guideline. Some have called or the adoption o a standard
H
the decision making and the recommendations developed therein
o
approach to grading recommendations or all guidelines. The uni y-
s
the clinical practice guideline.
p
ing principle or most grading systems is that both the strength o a
i
t
recommendation (eg, e ect size and/or risk-bene t ratio) and level
a
l
■ EXTERNAL REVIEW AND ENDORSEMENTS o evidence supporting a recommendation are considered when
M
e
Valid clinical practice guidelines must undergo an external review by assigning a grade.
d
The Grading o Recommendations Assessment, Development
i
c
another group o experts or a separate medical pro ession society or
i
and Evaluation (GRADE) Working Group came together in 2000 with
n
organization that did not sponsor or participate in the initial devel-
e
opment o the guidelines. One or more societies or organizations the goal o developing a common, sensible, and transient approach
a
n
may have reviewed the guidelines prior to publication and also pro- or grading quality o evidence and strength o recommendations.
d
The GRADE system classi es quality o evidence into one o our
S
vided their endorsement or the guidelines. Two or more organiza-
y
levels: high, moderate, low, or very low. With the GRADE system,
s
tions may ormulate joint guidelines, a process which increases the
t
evidence that is based on randomized controlled trials begins as
e
validity o the guidelines as long as they adhered to proper group
m
processes. For example, societies representing both interventional high-quality evidence but may be lowered i any o the ollowing
s
are present: study limitations, inconsistency o results, indirectness
o
cardiologists and cardiothoracic surgeons may jointly ormulate a
f
o evidence, imprecision, or reporting bias. Observational studies
C
clinical practice guideline on coronary revascularization.
a
start as low quality but may be graded up i the size o the treat-
r
e
ment a ect is very large, i a dose-response relation is present, or i
■ EDITORIAL INDEPENDENCE
all possible biases would decrease the size o the treatment a ect. In
The authors developing clinical practice guidelines should have the GRADE system, high-quality evidence is where urther research
editorial independence, both rom any relevant conf icts o interests is very unlikely to change the con dence in the estimate o e ect or
as well as the sponsoring organization. This may markedly a ect the an intervention. Moderate quality is when urther research is likely to
validity o the guidelines i complete independence is not main- have an impact and may change the estimate. Low quality is when
tained. For many sponsoring organizations, there may exist a guide- urther research is very likely to have an impact and likely change
line development committee to ensure accurate and consistent the estimate. Very low quality is when any estimate o e ect is very
guidelines, both in content, development, and structure. It should uncertain. The GRADE system o ers two grades or the strength o
be explicitly stated that the actual group o authors who comprised a recommendation: strong or weak. A strong recommendation is
the guidelines had editorial independence throughout the process. assigned i the bene cial e ects o an intervention clearly outweigh
the risks or clearly do not. A weak recommendation is assigned i
■ CURRENT GUIDELINES there is a concern either because o the quality o the evidence,
Valid clinical practice guidelines should ref ect current research uncertainty in bene t-risk ratio or an intervention, uncertainty or
to accurately address the scope and purpose o the intended variability in values and pre erences, or uncertainty about whether
guideline. The time it takes to develop a guideline is considerable an intervention represents a wise use o resources. An example o a
and puts the guideline at risk o being out o date at the time o grading system or Clinical Practice Guideline recommendations is
publication i current research is not included. Approximately 7% o provided in Table 10-2.
guidelines are out o date at the time o publication and the median Clinicians should amiliarize themselves with the grading sys-
survival o guidelines o is approximately 5.5 years. There are three tem used to grade the recommendations when reviewing prac-
important timelines to consider in the assessment o whether a tice guidelines. The grading process should be explicit and well
guideline ref ects current practice: de ned. The concepts o strength o recommendation and level o
1. The date o publication or the guideline or which the nal evidence supporting a recommendation will be urther elucidated
recommendations were given. below.
2. The dates o the literature reviewed in ormulating the guide-
lines to determine i recent evidence was utilized. ■ STRENGTH OF RECOMMENDATIONS
3. The date and the procedure or updating the guideline. The strength o a recommendation, sometimes re erred to as the
In some instances there may ongoing studies in progress during class o a recommendation (eg, Class I), designates whether a treat-
the guideline development that may change the recommendations ment or intervention is recommended or not as well as the level o
64
TABLE 10-2 A Grading System for Evaluating Evidence
C
H
Quality of Evidence Strength of Recommendation
A
1 (Strong) 2 (Weak)
P
T
A (High) 1A—Strong recommendation, high-quality 2A—Weak recommendation, high-quality
E
evidence. evidence.
R
Consistent evidence or RCTs without important Consistent evidence rom RCTs without important
1
limitations or exceptionally strong evidence rom limitations or exceptionally strong evidence rom
0
observation studies. observational studies.
B (Moderate) 1B—Strong recommendation, moderate-quality 2B—Weak recommendation, moderate-quality
S
evidence. evidence.
u
m
Evidence rom RCTs without important limitations Evidence rom RCTs without important limitations
m
(inconsistent results, methodological laws, indirect or (inconsistent results, methodological laws, indirect or
a
imprecise), or very strong evidence rom observational imprecise), or very strong evidence rom observational
r
y
studies. studies.
L
i
t
C (Low) 1C—Strong recommendation, low- or very low-quality 1C—Weak recommendation, low- or very low-quality
e
r
evidence. evidence.
a
t
u
Evidence or at least one critical outcome rom Evidence or at least one critical outcome rom
r
e
observational studies, case series, or rom RCTs with observational studies, case series, or rom RCTs with
:
P
serious laws or indirect evidence. serious laws or indirect evidence.
r
a
c
t
RCT, randomized controlled trial.
i
c
e
G
u
i
d
certainty behind the recommendation. The strength o a recom- evaluation o very limited patient populations resulted in only a “C”
e
l
mendation may be a ected by numerous actors but generally is a grade or level o evidence. Likewise, a recommendation may be
i
n
e
representation or the size o treatment e ect and degree o bene t considered o low strength i the treatment e ect or bene t-risk
s
versus risk or a speci c intervention. Some interventions may have ratio is very small and still considered o high level o evidence i
a
n
an un avorable risk-bene t pro le and thus would receive a recom- multiple populations have been studied and results have proven
d
S
mendation strength denoting the risk is greater than the expected to be both certain and precise. Repeated studies are less likely
y
s
bene t. Some interventions shown to be o no bene t may be clas- when the ollowing occurs:
t
e
si ed as such.
m
• Expected bene ts clearly outweigh any undesirable e ects o
a
t
the intervention as highly recommended.
i
c
■ LEVEL OF EVIDENCE • Desired bene ts more closely balanced with any undesirable
R
e
The level o evidence surrounding a recommendation is an esti- e ects, given only marginal bene t upon initial study.
v
i
e
mate o the certainty or precision or the treatment e ect o a
w
speci c intervention. The level o evidence is graded based on the
s
number o di erent populations studied, the actual number o ■ APPLICABILITY
study participants, and also what types o studies exist or a spe- Clinical practice guidelines that are use ul must be applicable to a
ci c intervention. A higher level o evidence would be assigned to clinical provider’s patient population and clinical scope o practice.
an intervention that has been validated in multiple study popula- The practices and recommendations described in guidelines must
tions and with a high number o study participants. A higher level be able to be replicated by clinical providers in their own practice in
o evidence would also be assigned to an intervention where data order or them to be relevant and worthwhile. Additionally, actors
has been acquired by randomized controlled trials rather than such as disease prevalence, risk actors, comorbidities, and individual
observational studies. The lowest level o evidence designation patient pre erences that di er rom the target population rom
is typically assigned to the recommendation when very limited which the guidelines were developed may a ect the applicability o
populations were evaluated or a speci c intervention and thus the guidelines to di erent patient populations.
only expert opinion, case reports, or previously accepted standard Conf icting guidelines or discordant recommendations may be
o care exist to support a recommendation. Although there is made by di erent organizations such a venous thromboembolism
some relation and overlap between the concepts o strength o prophylaxis in subpopulations (eg, orthopedic surgery). Clinicians
recommendation and level o evidence or a recommendation, may not be able to reconcile discordant recommendations or every
they may exist exclusively. The strength o a recommendation or occasion.
treatment based on a study demonstrating a very large treatment Tools are available to acilitate guidelines comparisons so that
e ect would be considered to be high; however, the level o evi- clinicians may determine the quality and applicability o guideline to
dence would be considered low i based on a single observational speci c patient populations. One example o a tool that exists to aid
study until additional studies involving either more populations clinicians is the “Compare Guidelines” tab that exists at www.guide-
and/or higher quality studies (eg, randomized, controlled trial) line.gov sponsored by Agency or Healthcare Research and Quality.
are completed. For example, the American Heart Association and The simplest and most appropriate approach should always take
American College o Cardiology in 2013 grades the use o diuret- into account individual patient risk actors, health care pre erences
ics or symptom relie in patients with heart ailure and volume and goals o care. Be ore o ering interventions (including screening)
overload as a “Class I” recommendation, indicating the bene t o and therapies to patients, always ask the question: “Do these clinical
this therapy is much greater than the risk (ie, bene t >>> risk); the practice guidelines apply to my patient?”
65
PRACTICE POINT besides English based on prevalence o diseases in certain countries.
Consideration o these actors can reduce the potential or publica-
• Practice guidelines provide three important bene ts or tion bias, which may produce misleading summary e ects.
hospital medicine providers by acilitating consistency,
P
e ciency, and e ectiveness to improve health care outcomes.
A
■ INCLUSION AND EXCLUSION CRITERIA
R
Similar to high-quality recommendations rom a consultant,
In systematic reviews, inclusion and exclusion criteria or articles
T
practice guidelines should give clear recommendations, discuss
should be stated and there should be a discussion o the patients
I
alternatives, acknowledge biases, and consider extenuating
included, the exposures or interventions, the outcomes o interest
circumstances or a speci c clinical case.
and the methodological standards or study selection (eg, random-
• A modi ed ramework or hospital medicine providers to
ized controlled trials, diagnosis cohort, or observational studies).
use in both interpreting and assessing the quality o clinical
To improve the quality o studies, i randomized controlled trials
T
practice guidelines is divided into three general domains with
h
exist, then these should be used pre erentially to minimize bias.
e
subdomains:
There should be speci c inclusion and exclusion criteria based on
S
p
M Validity: scope and purpose, group composition, conf icts o
the clinical question such as patient demographics (eg, age, sex,
e
interests, literature review, group processes, external review
c
ethnicity, etc).
i
a
and endorsements, editorial independence, and current
l
t
y
guidelines ■ QUALITY OF STUDIES
o
M Results: strength o recommendations and level o evidence
f
There should be an assessment o individual studies that are
H
M Applicability
o
included in the systematic review. For therapy studies, there should
s
p
be evaluation or randomization (ie, allocation concealment), blind-
i
t
a
ing (o investigators, patients, data collectors), and loss to ollow-up
l
SYSTEMATIC REVIEWS
M
as these can introduce systematic error in the meta-analysis. Ideally,
e
the studies should report intention-to-treat (ie, all patients who
d
■ INTRODUCTION
i
underwent allocation are analyzed regardless o how long they
c
i
n
There are our major overviews, and they include narrative reviews, stayed in the study) or per protocol (ie, only patients who remained
e
systematic reviews, meta-analysis, and guidelines. They represent within the protocol or a predetermined period are analyzed). For
a
n
the highest quality o evidence on the quality pyramid as they diagnostic studies, there should be evaluation o the appropriate
d
synthesize evidence rom many sources. Narrative reviews elucidate re erence standard.
S
y
a health context, condition, or intervention. They tend to answer Typically in systematic reviews, there should be quality assess-
s
t
e
background questions, biologic, or social issues. However, the meth- ment by more than one individual and investigator other than the
m
ods or article inclusion and quality assessment are not structured principal investigator o the review. There are several ways to assess
s
and the credibility o the review is tied to the reviewer expertise. Sys-
o
or the quality and di erences amongst the individuals making
f
tematic reviews and meta-analyses may ocus on therapy, diagnosis, the quality assessment. The kappa statistic (κ) represents the inter-
C
a
or prognosis depending on the clinical question at hand. Table 10-1 reviewer reliability and is a measurement o the agreement between
r
e
highlights the similarities and major di erences between the various observers beyond chance. The κ-value ranges rom –1.0, which is no
types o overview articles. agreement, to 1.0, which is per ect agreement. For most studies, a
The goals o systematic reviews include answering questions κ-value o 0.8 or higher is acceptable and represents excellent agree-
where conf icting data exist and/or where sample sizes o individual ment and a κ-value o less than 0.4 represents poor agreement. The
trials are small. The ultimate goal o systematic reviews may be to Jadad score is another structured way o assessing the quality o
generate hypotheses. Meta-analyses are a type o systematic review individual studies. The score is based on upon points or random-
that use quantitative methods to combine results rom multiple ization, double blinding, and withdrawals. Table 10-3 shows the
studies to yield a summary estimate o e ect and a con dence inter- modi ed Jadad scoring template or randomized controlled trials.
val around the estimate. A common source to search or systematic The Newcastle-Ottawa score can be used or nonrandomized con-
reviews and meta-analyses is the Cochrane Database (http://www. trolled trials.
cochranelibrary.com/).
■ META-ANALYSIS
■ DEFINE THE QUESTION I it seems reasonable and suitable, data rom several studies can
Whereas in narrative reviews the clinical question tends to be be combined to give an overall estimate result. The overall results
more general, the clinical question in systematic reviews should produced rom the results o a larger number o patients than indi-
be ocused in a ormat with a clearly de ned patient population, vidual studies are more accurate and reduce type 2 error (ie, ailing
intervention, control, and clear outcomes. An example would be: to detect a di erence that exists between the two groups), which
“What is the risk o bleeding in adult patients with recurrent venous increases the power o the analysis. It is important that the results
thromboembolism on rivaroxaban as compared to war arin?” can be appropriately combined (ie, individual studies have similar
interventions, patients, outcome measures); otherwise, it would
■ LITERATURE SEARCH be like the popular idiom: “combining apples and oranges.” Data
Because the goal o systematic reviews is to synthesize in ormation in meta-analyses are presented as Forest plots. There are various
rom many sources, it is important that a thorough review o the lit- statistical methods used to combine di erent data types, which will
erature is done to prevent missing relevant studies. There should be be discussed later in this chapter.
a discussion o the bibliographic databases searched with the search
terms included. The search strategy and included databases will ■ POOLED ESTIMATESAND FOREST PLOT
depend on the question being asked. There should be an attempt Figure 10-1 shows an example o a orest plot produced or a meta-
to locate unpublished studies such as con erence abstracts, experts analysis. On the le t side, the names o the individual studies are
in the eld, pharmaceutical companies, and investigators o stud- listed. The green box represents the individual data rom the studies
ies. There should also be consideration to searching in languages and the risk ratio (with 95% con dence interval) is calculated taking
66
sample size. The x-axis represents the net e ect or risk ratio either
TABLE 10-3 The Modified Jadad Scoring System for avoring control or experimental with the baseline at 1.0, which
C
Randomized Controlled Trials would indicate no e ect. I the line or diamond width cross 1.0, then
H
A
Score the result is not statistically signi cant. I the result is on the le t side
P
o the plot, it avors experimental. I the result is on the right side o
1. Was the study described as randomized? I yes,
T
the plot, then it avors the control.
score 1 point.
E
R
2. I yes to question 1, was an appropriate ■ HETEROGENEITY
1
randomization sequence described and used
0
(eg, table o random numbers, computer When the results are pooled in a meta-analysis, determine that
generated, etc)? I yes, score 1 point. the data being pooled are appropriate without signi cant di er-
ences between study results (ie, “we are not comparing apples and
S
3. I yes to question 1, was an inappropriate
u
method to generate the sequence o oranges”). This is done using various tests or heterogeneity. The null
m
randomization used (patients were allocated hypothesis o the test or heterogeneity is that the underlying e ect
m
alternately, or according to date o birth, is the same in each study so that all o the variability between stud-
a
r
hospital number, etc)? I yes, subtract 1 point.
y
ies is just due to chance alone. A visual inspection or heterogeneity
L
4. Was the study described as double blinded? I can be done. Random error (ie, chance) is a plausible explanation
i
t
e
yes, score 1 point. or the di erences in the point estimates i the con dence intervals
r
a
overlap widely (Figure 10-2A). Random chance cannot explain the
t
5. I yes to question 4, was an appropriate method
u
r
o blinding used (eg, identical placebo, active di erences in the apparent treatment e ect i the con dence inter-
e
:
placebo, dummy, etc)? I yes, score 1 point. vals do not overlap (Figure 10-2B). The P-value is a statistical test o
P
r
6. I yes to question 4, was an inappropriate signi cance or heterogeneity or di erence between study results.
a
c
method or blinding used (eg, comparison o In most instances, there is no heterogeneity i the P-value is greater
t
i
c
tablet vs injection with no double dummy)? I than 0.05 (this is set based on type I error), as this represents the null
e
yes, subtract 1 point.
G
hypothesis holding true. Random chance alone cannot explain the
u
7. Were the withdrawals and dropouts described? di erences between studies and there is heterogeneity between
i
d
I yes, score 1 point. the study results i the P-value is less than 0.05.
e
l
i
Another test or the magnitude o heterogeneity is the I2 statistic,
n
e
a percentage rom 0% to 100%. Zero percent represents no hetero-
s
a
geneity and 100% represents signi cant heterogeneity. Ideally or a
n
into account the weight o the study. The weight (ie, importance in
d
meta-analysis, there should be minimal heterogeneity, typically with
the overall result) is based on the sample size and precision o the
S
I2 values o less than 40%. The higher the I2 statistic, the more cau-
y
results. More weight is given to studies with a larger sample size and
s
tious we should be in interpreting the treatment e ects. Signi cant
t
e
a smaller standard deviation.
heterogeneity may be related to signi cant di erences in patient
m
The orange box shows the pooled data and risk ratio. The abso-
a
populations, interventions, and/or outcomes being measured.
t
lute risk di erence and number needed to treat or harm can also
i
c
be calculated rom this gure. In this example meta-analysis, the
R
■ PUBLICATION BIAS AND FUNNEL PLOTS
e
total experiment event rate is 0.25 and the control event rate is 0.16.
v
i
Thus, the approximate risk ratio is 1.57 ( avors control), absolute risk In meta-analyses, publication bias may o ten be present. Results
e
w
increase is 0.09, and the number needed to harm in this case is 11. with signi cant positive ndings are more likely to be submitted
s
Whereas many o the individual studies are small and not statistically and accepted or publication. This phenomenon leads to an over-
signi cant, when pooled together, the overall e ect becomes statis- estimate o treatment e ects. Publication bias may be evident rom
tically signi cant with a narrower con dence interval. comparing results rom published versus registered trials, published
The right side o the graph shows the individual mean e ect as studies versus doctoral dissertations, and randomized trials versus
a dot with the line representing the 95% con dence interval. The observational studies. Studies with signi cant positive results are
larger the dot, the larger the weight it is given to the overall result. likely to generate multiple publications, hence propagating the bias.
The diamond represents the overall e ect size, and the size o the Researchers rom countries where English is not the major spoken
diamond is dependent on the 95% con dence intervals. A larger language may tend to submit nonsigni cant results to their own
and/or wider diamond implies larger con dence intervals and a low respective domestic journals. Searching or articles published only
Expe rime ntal Co ntro l Ris k ratio Ris k ratio

S tudy o r s ubg ro up Eve nts To tal Eve nts To tal We ig ht M–H, Fixe d 95%CI M–H, Fixe d 95%CI
Bla ck 2004 3 25 2 23 4.1% 1.38 [0.25, 7.53]
Dolittle 1999 6 26 5 22 10.8% 1.02 [0.36, 2.88]
Finlay 1999 15 56 10 54 20.2% 1.45 [0.71, 2.93]
Higgins 2001 4 44 5 42 10.2% 0.76 [0.22, 2.65]
Le ctor 2000 24 75 12 71 24.5% 1.89 [1.03, 3.49]
S tra nge love 2005 17 54 10 52 20.2% 1.64 [0.83, 3.24]
Wa ts on 2005 12 43 5 42 10.0% 2.34 [0.90, 6.04]
To tal (95% CI) 323 306 100.0% 1.57 [1.14, 2.15]
Tota l eve nts 81 49
He te roge ne ity: Chi2 = 3.09, df = 6 (P = 0.80); I2 = 0%

0.01 0.1 10 100
Te s t for ove ra ll e ffe ct: Z = 2.78 (P = 0.005) Favours expe rime nta l Favours control
Figure 10 1 Example of a forest plot in a meta-analysis.
67
P
A
R
T
I
A 1.0 Re la tive ris k B 1.0 Re la tive ris k
T
Figure 10 2 Two plots demonstrating heterogeneity in meta-analysis. Plot (A) by visual inspection shows minimal heterogeneity (I2 0, P-value > 0.05)
h
e
and plot (B) demonstrates significant heterogeneity (I2 > 40%, P-value < 0.05).
S
p
e
c
i
a
l
in English adds to publication bias. A unnel plot is used to assess included or a meta-analysis or when one large study may be more
t
y
or the presence o publication bias in a meta-analysis. As shown in trustworthy than smaller studies with di ering results.
o
f
Figure 10-3A, a plot o the sample size (or standard error) versus the A random-e ects model assumes that the studies include a
H
e ect size (relative risk or risk ratio) is made. Each dot represents random sample o the population o studies that address the ques-
o
s
the overall e ect rom one randomized controlled trial. The larger tion in the meta-analysis. Because there are di erences in patients,
p
i
t
the sample size is (or smaller the error), then the higher the dot interventions, and outcomes in the studies, each study estimates
a
l
on the graph is placed. The dotted red line shows the overall esti- a di erence underlying true e ect that has a normal distribution.
M
mate and i there are no missing studies, then the results should The pooled e ect in the random-e ects model is not a single e ect
e
d
be located in a symmetric, triangular area centered on the overall e ect but a mean e ect across populations, interventions, and outcomes.
i
c
i
o all studies. I there are missing studies as in seen Figure 10-3B, This model takes into account both within-study variance as well as
n
e
then they will appear as a gap in the portion o the unnel plot between-study variance (ie, heterogeneity). In the random-e ects
a
where one would expect to nd negative studies. The unopposed model, large studies with wider con dence intervals have more
n
d
positive studies will shi t the apparent treatment e ect (blue line) weight. The random-e ects model is generally considered the more
S
y
toward a larger size than it really is. reliable approach o the two models.
s
t
Figure 10-4 highlights some o the di erences in these two mod-
e
m
■ FIXED-EFFECTS MODEL VERSUS els. In Figure 10-4A, both models yield a similar net estimate when
s
RANDOM-EFFECTS MODEL there is no heterogeneity. However, the xed-e ects model yields
o
f
There are two statistical models used to combine the results rom tighter con dence intervals compared to the random-e ects model,
C
which is a more conservative estimate o the overall e ect, when
a
multiple studies in a meta-analysis: the xed-e ects model and the
r
e
random-e ects model. Each model has advantages that inf uence there is similar sample size in each study but variability between
interpretation o results. studies as in Figure 10-4B.
A xed-e ects model assumes there is a single true value underly-
ing all o the study results. All the studies should give identical esti- ■ SUBGROUP AND SENSITIVITY ANALYSIS
mates o the e ect i all o the studies addressing the same question Subgroup analyses are o ten presented in systematic reviews to
are large and ree o bias. Variance in this model is derived only rom determine i certain groups o patients may respond to treatment
within-study variance and does not incorporate between-study vari- as compared to other groups (eg, patients more or certain comor-
ance (ie, heterogeneity). A xed-e ects model generally gives more bidities versus patients with less or no comorbidities). While these
weight to larger studies with narrower con dence intervals. This subgroup analyses may help individualize treatment, they must
model may be reasonable when there are a small number o studies be interpreted with caution. A ew principles or critical questions
Actua l Appa re nt
tre a tme nt tre a tme nt
Ove ra ll e s tima te e ffe ct e ffe ct
e
e
z
z
i
i
s
s
e
e
l
l
p
p
m
m
a
a
S
S
A Effe ct s ize (re la tive ris k) B Es tima te s ize (re la tive ris k)
Figure 10 3 Funnel plot for assessment of publication bias. Plot (A) shows a normal funnel plot and plot (B) shows an example of how missing
studies shift the treatment effect.
68
C
H
A
P
T
Fixe d-e ffe ct Fixe d-e ffe ct
E
R
Ra ndom-e ffe ct Ra ndom-e ffe ct
1
0
A 1.0 Re la tive ris k B 1.0 Re la tive ris k
S
Figure 10 4 Example of a random-effects model and a fixed-effects model on overall effect with (A) minimal variability between study results and
u
m
(B) significant variability between study results.
m
a
r
y
SUGGESTED READINGS
L
to help guide in determining the credibility o subgroup analysis
i
t
e
include:
r
a
Atkins D, Best D, Briss PA, et al. Grading quality o evidence and
t
(1) Does chance explain the subgroup di erence?
u
strength o recommendations. BMJ. 2004;328:1490.
r
(2) Is the subgroup consistent across studies?
e
Brouwers MC, Kho ME, Browman GP, et al. Development o AGREE II,
:
(3) Is there a biologic plausibility supporting the subgroup e ect?
P
part 1: per ormance, use ulness and areas or improvement. CMAJ.
r
(4) Is the subgroup di erence suggested by meta-analysis com-
a
c
parisons within rather than between studies? 2010;182:1045-1052.
t
i
c
Brouwers MC, Kho ME, Browman GP, et al. Development o AGREE II,
e
A sensitivity study may also be per ormed in a meta-analysis i part 2: assessment o validity o items and tools to support appli-
G
there are relatively lower quality data used. It may re er to a com-
u
cation. CMAJ. 2010;182:E472-E478.
i
d
parison o analyses using the random-e ects model versus xed-
e
Guyatt GH, Oxman AD, Kunz R, et al. GRADE: an emerging consensus
l
e ects model. Alternatively, it may re er to the use o quality scores
i
n
on rating quality o evidence and strength o recommendations.
e
o included trials to modi y the impact o lower quality studies to the
BMJ. 2008;336:924-926.
s
overall result o the analysis. Although this may reduce the contribu-
a
Irwig L, Tosteson AN, Gastonis C, et al. Guidelines or meta-analyses
n
tion o lower quality studies, there may be bias because it allows
d
authors to somewhat arbitrarily reduce the e ects or dismiss the evaluating diagnostic tests. Ann Intern Med. 1994;120:667-676.
S
y
e ects o studies they may not like. Jadad AR, Moore RA, Carroll D, et al. Assessing the quality o reports
s
t
e
o randomized clinical trials: is blinding necessary? Controlled Clin
m
Trials. 1996;17:1-12.
a
■ APPLICABILITY
t
Moher D, Cook DJ, Eastwood S, et al. Improving the quality o
i
c
The next question asks i the results may be applied to a speci c reports o meta-analyses o randomized controlled trials: the
R
e
patient and/or patient population now that it has been determined QUOROM statement. Quality o Reporting Meta-analyses. Lancet.
v
i
that the systematic review and meta-analysis answers the original
e
1999;354:1896-1900.
w
clinical question, that there was an appropriate search or the evi-
s
Murad M, Montori VM, Ioannidis JA, et al. Understanding and apply-
dence and an assessment o the quality o the evidence, and that ing the results o a systematic review and meta-analysis. In: Guyatt
the study results were appropriate to combine. This step requires G, Rennie D, Meade MO, Cook DJ, eds. Users’Guides to the Medical
consideration o all clinical outcomes as well as the bene ts versus Literature. 3rd ed. New York, NY: McGraw-Hill; 2014;471-490.
the harms and costs. Evidence-based medicine relies not only on
Neumann I, Aki EA, Vandvik PO, et al. How to use a patient manage-
the best evidence available but also clinical expertise, clinical expe-
ment recommendation: clinical practice guidelines and decision
rience, and patient pre erences. Clinical experience improves the
analysis. In: Guyatt G, Rennie D, Meade MO, Cook DJ, eds. Users’
e ciency o diagnosis and treatment, expands the ability to deter-
Guides to the Medical Literature. 3rd ed. New York, NY: McGraw-Hill;
mine the applicability o research data to the individual patients,
2014;547-560.
and allows or the consideration o patient pre erences. First ask the
question: “Do these results apply to my patient?” whenever consid- Qaseem A, Forland F, Macbeth F, et al. Guidelines International Net-
ering o ering interventions and therapies to patients based on data work: toward international standards or clinical practice guide-
rom systematic reviews and meta-analyses. lines. Ann Intern Med. 2012;156:525-531.
ONLINE RESOURCES
PRACTICE POINT
• Agency for Healthcare Research and Quality (AHRQ) National
• Meta-analysis provides a quantitative, pooled estimate o a Guideline Clearing House. http://www.guideline.gov.
treatment e ect. Key advantages o a meta-analysis are that it • American Medical Association, JAMAevidence. http://jamaevi-
can pool smaller studies yielding a larger sample size as well as dence.mhmedical.com.
summarize disparate data.
• Cochrane Collaboration, Cochrane Library. http://www.
• Systematic reviews and meta-analyses are subject to
cochranelibrary.com.
publication bias, which can lead to some studies being missed
and thus not included in the analysis. Bias in individual studies • Guidelines International Network (G-I-N). http://www.g-i-n.net.
will lead to bias in the systematic review and meta-analysis. • The Appraisal of Guidelines for Research and Evaluation Enter-
• The nal conclusion should be based on the evidence, rather prise. http://www.agreetrust.org.
than personal opinion. • The Grading of Recommendations Assessment, Development and
Evaluation (GRADE) Working Group. www.gradeworkinggroup.com.
69
CHAP TER
11 INTRODUCTION
With over 21.5 million unique articles and more than 1 million ran-
domized controlled trials indexed in MEDLINE as o 2014 and more
than 1 million new publications published and indexed annually,
clinicians now must process a vast volume o medical literature.
Many clinicians eel like they are drowning in in ormation. Hospital-
ists must balance the need to nd relevant and accurate answers to
their clinical questions with the need or e ciency in nding those
answers to immediately guide high-quality care to multiple acutely
Practical ill patients. Formulating and answering questions e ciently and
e ectively will improve care and reduce the rates o consultation,
Considerations testing, and potential errors.

The volume o data and limited time or searching or answers
o Incorporating compound each other. A recent systematic review examining clini-

cal questions raised by clinicians at the point o care ound approxi-
mately one clinical question arises or every two patient encounters.
Evidence into Clinical Clinicians only pursued 51% o the questions raised and ound
answers to only 78% o those questions pursued. Studies reported
Practice the main barriers to seeking in ormation included a clinician’s lack o

time and a doubt that a use ul answer existed. The state o relying on
in ormation already known prevails when the energy required to get
a new answer outweighs the perceived bene t. Clinical inertia may
be illustrated by considering the gap between the potential bene ts
J. Richard Pittman, Jr., MD
o evidenced-based care and actual rates o implementation as in
Mikhail Akbashev, MD the treatment o heart ailure with reduced ejection raction (HFrEF).
Optimal implementation o strong evidence-based therapies or
HFrEF could save an estimated 35,000 to 117,000 thousand lives per
year. See Chapter 129 (Heart Failure).
In an ideal practice setting, the majority o clinical questions
would have a readily accessible, evidence-based answer. Clinicians
would have current knowledge o guideline-based therapy and
could apply pertinent point-o -care reminders rom the electronic
medical record or best practice or every patient under their care.
As a result, patients would yield maximal bene t rom clinical trials
and guideline-driven in ormation. This ideal state may not be attain-
able; however, clinicians may take steps toward better utilizing the
evidence-based answers currently available to meaning ully impact
clinical practice.
CURRENT CONSEQUENCES AND MOTIVATORS

FOR CHANGE
■ PRACTICING WITH OUTDATED INFORMATION

Relying on outdated in ormation or patient care (ie, clinical inertia)
may limit potential bene ts o current therapies and expose patients
to risks o disproven therapies. An example would be practicing
based o outdated guidelines or the treatment o blood choles-
terol to reduce atherosclerotic cardiovascular disease (ASCVD) risk
in adults where a clinician tailors treatment and statin dosing based
solely on the low-density lipoprotein cholesterol (LDL-C) levels
rather than selecting a moderate- or high-intensity statin strategy
based on ASCVD risk, an LDL-C ≥ 190 mg/dL, and/or the presence
o diabetes as recommended in the current updated guidelines.
■ UTILIZING UNVERIFIED INFORMATION SOURCES

Clinicians may utilize unveri ed in ormation sources. Examples o
possible unveri ed in ormation sources include seemingly more
in ormed colleagues, television programming/advertisements,
70
popular media, drug company representatives, and searching or a con erence or reading a publication on mobile resources
“evidence” or answers to clinical questions using search engines that that may be tempting to clinicians to sign up or too many
C
may provide in ormation rom unveri ed sources (eg, “Googling”). resources at one time.
H
Heavy reliance on unveri ed in ormation sources can signi cantly
A
• Selection methodology or choosing the articles presented
P
limit the quality o in ormation with which a clinician practices and may not always be clear.
T
should be avoided i possible.
E
Suggestions
R
■ SEARCH SATISFACTION BIAS 1. Setting up digital resources. It may be help ul or a clinician
1
to initially take consider what in ormation he or she needs
1
Depending on a clinician’s com ort level with evidence-based medi-
and the areas o knowledge de ciencies (eg, original research
cine and available search tools and time, a clinician may only read
manuscripts, clinical updates, and guidelines). Any individual
article summaries or conclusions rom abstracts. This represents a
P
clinician may not be aware o each o her or his personal
r
a
orm o search satis action bias. When an answer is ound to a clini-
c
de ciencies, especially since “blind spots” are by de nition not
t
cal question that is requently encountered in a clinician’s scope o
i
c
seen. Thought ul attention to one’s knowledge base, de cien-
a
practice, the clinician might jump to apply this answer to all sub-
l
cies, and practice gaps may lead to crucial rst digital steps
C
sequent patients without ully assessing the validity and quality o
o
or change. In addition, a comparison o pro essional needs
the source or considering the applicability o this in ormation to the
n
and knowledge gaps should be made with regards to any
s
individual patient.
i
d
resources a clinician is actively subscribed. A busy clinician
e
r
should pay care ul attention to cutting down on in requently
a
■ EVIDENCE-BASED PRACTICE IMPROVES PATIENT
t
used resources prior to subscribing to new ones.
i
o
OUTCOMES
n
Consideration should be made or an individual clinician’s
s
Perhaps the strongest and most important motivator or change pre erred method o receiving in ormation. Some like in orma-
o
should be the act that evidence- or guideline-based practice has tion to come straight to their e-mail with a link to a website
I
n
been shown to improve patient outcomes across a wide spectrum while others pre er opening a designated application on their
c
o
o illnesses. device(s). E-mails clutter e-mail inboxes, but most people
r
p
regularly check e-mail. Mobile applications clutter your devices
o
r
a
SOLUTIONS TO INCORPORATE EVIDENCE INTO while also continuing to collect new material unnoticed to the
t
i
CLINICAL PRACTICE clinician. This can be corrected by removing unused applica-
n
g
Solutions should take into consideration how a clinician might tions that clutter mobile devices while also taking steps to
E
v
encounter in ormation during the f ow o practice. “Keeping up” become amiliar with how applications work and aggregate
i
d
with the literature is increasingly di cult, and staying abreast o in ormation. Most tools discussed in this chapter strive to
e
n
important developments that may not directly relate to patients deliver in ormation in a way that best ts the workf ow o
c
e
that the clinician is actively managing. “Keeping up” single-source di erent clinicians (Table 11-1). The key is or any individual
i
n
publications rom an area o clinical ocus are use ul, but they may practitioner to identi y resources most convenient and reliable
t
o
narrow a clinician’s awareness o broader advances across di erent or that individual clinician.
C
disciplines. There are numerous methods to “keep up” with the Once an account is setup or the any o the resources, the
l
i
n
literature that transcend traditional postal mail by utilizing e-mail or next step is to select specialty and topic pre erences. Once
i
c
these steps are completed, the in ormation will then be sent
a
designated applications or mobile devices. Users sign up or these
l
either via e-mail or to a mobile application depending on
P
“keeping up” resources as reputable and reliable evidence sent to
r
a
them rom trusted sources. These tools may help clinicians keep which resource is selected (Figure 11-1).
c
t
up with a particular eld or topic by organizing in ormation by dis- A critical step is to then monitor the use ulness o in orma-
i
c
e
cipline, relevance to that discipline, and newsworthiness or impact tion and then a ter a period o time (eg, a couple o weeks or a
level. This in ormation may be delivered to an e-mail inbox, a web- month) go back into the tool and adjust the settings or unsub-
site, or to a discrete mobile application. The ocus below will be on scribe rom the resource based on its help ulness or not.
ree resources available or those included with pro essional society 2. Schedule time or new reading. A seemingly simple step,
memberships common or most hospital medicine providers. but one that can make a large di erence in “keeping up” is to
Advantages o push resources available or most hospital medi- schedule time or new reading. Carrying a paper or virtual le
cine providers include: with articles, presentations, and/or books may make a hal an
hour be ore rounds, between appointments, or while sitting in
• Preset, reliable stream o in ormation lowers the energy
carpool lines a lot more productive.
required to incorporate new in ormation into clinical practice,
3. Attend institutional con erences. Attending institutional
leaving a clinicianless subject to clinical inertia or anchoring.
con erences (eg, Grand Rounds, noon con erences) is an excel-
• Many o the tools are ree and can ocus on a speci c discipline
lent way or a clinician to acquire new in ormation while also
or a topic.
providing the opportunity to spark conversations and clinical
• The volume o in ormation presented can be adjusted based
questions with colleagues. Volunteering to present a topic
on clinician desire or tolerance.
at con erence promotes expertise acquisition o new and
• The requency o the delivery o this in ormation can be
updated clinical knowledge.
customized.
4. Participate in pro essional societies and their scientif c
• Article summaries by colleagues trained in evidence-based
meetings. Scienti c meetings hosted by medical pro essional
analysis can be accessed and can raise awareness o potential
societies are designed to help clinicians stay abreast o the
inclusion biases.
ever-changing in ormation and also to o er credit or Continu-
Disadvantages include: ing Medical Education.
• These streams may become overwhelming i the amount 5. Encourage a culture o learning. Activities such as journal
and requency o content is not properly customized to the club, case con erence, and clinical updates in hospital medi-
individual clinician pre erences. There is a risk o attending cine at hospitalists’ sites o practice can be an e ective way to
71
pertinent to the patient, with speci c, actionable in ormation that is
TABLE 11-1 Example Digital Resources or “Keeping Up” quickly and reliably accessed. Unlike the “keeping up” resources pre-
viously discussed, these “quick questions” need to be answered or a
E-mail/Website-Based Website Link
real patient right now. That is why using these types o resources is
P
BMJ Evidence Updates https://plus.mcmaster.ca/
A
essential in day-to-day patient care. Fortunately, multiple resources
evidenceupdates/
R
exist including current textbooks, database search tools such as
T
ACP Journal Club http://annals.org/journalclub.aspx Trip Database, popular portable resources such as UpToDate or
I
NEJM Journal Watch www.jwatch.org Dynamed, and large databases like PubMed. These resources have
PubMed—My NCBI http://www.ncbi.nlm.nih.gov/ been developed to answer speci c, patient-related questions. Hos-
pubmed/ pitalists should check with their institutional subscriptions rst and
Application-Based Website Link utilize these be ore personally enrolling in subscriptions that may
T
(App)
h
be expensive.
e
Docphin https://www.docphin.com/ Developing “quick questions” require clinicians to stop their f ow
S
p
Read by QXMD http://www.qxmd.com/apps/read-by- o work and look or an answer. Cumbersome retrieval o in orma-
e
tion may lead to clinical inertia. Initial searching using summary
c
qxmd-app
i
a
resources, such as textbooks or review articles, can be e cient and
l
Doximity https://www.doximity.com/
t
y
more likely to yield pertinent answers. However, these resources
o
are subject to authorship biases. Important articles could poten-
f
H
tially be omitted while smaller studies maybe overly emphasized
o
s
promote a culture o learning, which can help providers urther in recommendations. Summary resources can also be delayed in
p
i
keep up with the f ow o new in ormation. incorporating the most recent study results, even sometimes those
t
a
6. Utilize the electronic health record. Perhaps the most perti-
l
o large pivotal studies. Many clinicians start with a Google search
M
nent time to receive updated in ormation on any treatment or with Wikipedia as the top listed initial resource. This may o ten be a
e
d
diagnostic test is at the time o ordering the treatment or test. quick and satis ying way to answer simple medical de nition ques-
i
c
Recent advances and expansions in electronic health records tions such as “what is an antimitochondrial antibody.” Mainstream
i
n
(EHR) o er a limited opportunity or “just-in time” learning
e
publicly available search engines such as Google or Yahoo can
a
about speci c medical concepts. Many EHRs incorporate o er ast and ree searches but o ten are less reliable and may not
n
d
evidence-based recommendations into order sets, as well as provide a robust and accurate answer to clinical questions as would
S
linking users to literature regarding updated recommenda- more rigorously vetted medical resources, such as AccessMedi-
y
s
tions. A recent meta-analysis showed improvement in both cine, ClincialKey ( ormerly MD Consult), DynaMed, Medscape (a.k.a.
t
e
e ciency and adherence to guidelines through use o EHR.
m
eMedicine), UpToDate and a growing list o other compiled medical
s
Similarly, a study o an older computer support tool or anti- summary resources and reviews (Table 11-2). These resources have
o
biotic management ound a dramatic decrease in pharmacy made incorporation o new in ormation a priority.
f
C
antibiotic expenditures and mortality rates over a 6-year period.
a
r
A systematic review o 68 controlled trials o computer decision
e
support systems ound the majority o the trials demonstrated Suggestions
bene t in physician per ormance and patient outcomes. Mul- Despite the apparent limitations, actively and e ciently searching
tiple health systems have implemented quality improvement or answers to clinical queries is an essential part o patient care.
projects to adhere to core measures using order sets within the Overcoming clinical inertia to identi y and explore questions is the
EHRs with great success. rst step. This may be as simple as an index card with questions
or the day or lunches with colleagues to discuss cases. Having
■ QUICK QUESTION identi ed questions, a clinician needs to ormulate the question
The context is essential or any clinical question. The ideal “quick as speci cally as possible and identi y an e cient search strategy
question” resource would be readily available at the point o care, (Figure 11-2).
*Ale rt cut off s core s Re leva nce 6 Newsworthine s s 5

To s e e how ma ny a rticle s you a re like ly to re ce ive a t va rious leve ls of re leva nce a nd
newsworthine s s, try cha nging the leve ls a bove, the n look a t the gra ph be low. At the curre nt
cutoffs s hown a bove, on ave ra ge you would ha ve re ce ive d 9 a rticle s pe r month in the la s t 12
months
14
12
11 11
9
t
n
8 8 8 8
u
o
7 7
c
e
5
l
c
i
t
r
A
Aug S e p Oct Nov De c Ja n Fe b Ma r Apr May Jun Jul
14 14 14 14 14 15 15 15 15 15 15 15
Month/Ye a r
*Ale rt Fre que ncy Da ily
Figure 11 1 Screenshot of BMJ evidence updates. This is an example of a free online “keeping up” resource which allows users filter paper alerts by
topic, relevance and newsworthiness, to narrow to a manageable amount.
72
TABLE 11-2 Example Digital Resources or Answering Clinical Questions
C
H
Resource (Publisher) Description CME
A
AccessMedicine Compendium o rapidly searchable medical textbooks Yes
P
T
E
R
McGraw-Hill
1
Clinical Key (a.k.a. MDConsult) Automated evidence search engine, allow users to search the primary Yes
1
medical literature, guidelines and standard textbooks
P
r
a
c
Elsevier
t
i
c
Dynamed Collection o evidence-based summary reviews on common clinical topics; Yes
a
l
presented in bulleted ormat with speci ic citations and assessment o
C
evidence quality or recommendations. Systematic rapid inclusion o newly
o
n
EBSCO Health published high-quality evidence, including systematic reviews
s
i
d
Essential Search engine to browse guidelines and systematic reviews; integrated with Yes
e
r
other resources such as calculators, coding helper, or decision support tools
a
Evidence Plus
t
i
o
n
s
o
I
Wiley
n
c
o
Medscape Compilation o summaries by specialists in the ield; includes new article Yes
r
p
(a.k.a. eMedicine) ormats
o
r
a
t
i
n
g
WebMD
E
v
PubMed Clinical Queries Indexed database o articles rom a multitude o medical journals No
i
d
e
n
c
e
NIH
i
n
t
TRIP Database Automated evidence search engine o primary studies, guidelines, textbooks No
o
and other clinical resources with ilters available; scores articles based on how
C
(Translating Research Into Practice)
l
recent, pertinence, and publication source
i
n
i
c
a
l
P
TRIP
r
a
c
Up To Date Evidence-based online textbook. Compendium o expert written Yes
t
i
c
invited reviews on a broad array o topics; evidenced based and graded
e
recommendations, primarily prose and directive in nature
Wolters Kluwer
Many hold to dogma and use PubMed as an initial search. This These questions should be very concrete and speci c. A standard
o ten yields a time-consuming search, with many published articles “PICO” (Patient, Intervention, Comparison, Outcome) ormat iden-
to si t through, and challenges nding relevant studies e ciently. ti es the pertinent parts o a question (eg, “What is the e ect o
This also has been shown to take 41% longer (29 vs 17 minutes) than 23-valent pneumococcal polysaccharide vaccine versus placebo
UpToDate searches in a cohort o resident physicians. Furthermore, on mortality in patients with systolic heart ailure exacerbations?” or
a study o 54 medical students, residents, and aculty searching “What is the relative risk o bleeding with rivaroxaban versus war arin
multiple methodologies to answer critical-care-related questions in patients with venous thromboembolism?”). For questions too
showed that users most requently searched Google rst (45% o speci c or narrow to be ound in summarized resources, a primary
the time), with Google and UpToDate providing aster answers than literature search may be pre erred. PubMed o ers ree access to
PubMed (3.8 vs 3.3 vs 4.4 minutes, respectively). Importantly, Google users, and its “Clinical Queries” search option quickly lters results or
and UpToDate were more likely to lead to a correct answer than clinical and systematic review articles. Other electronic resources like
PubMed (60% vs 70% vs 36%, respectively). The highest value rst TRIP (Translating Research into Practice) database can quickly lter
resource is o ten a compiled resource such as Dynamed or UpTo- studies indexed in MEDLINE, guidelines and other resources, and
Date, but multiple resources are available rom various publishing it also ranks them on quality and type o study. Guideline searches
organizations (Table 11-2). may be completed through the National Guideline Clearinghouse
Regardless o the initial search strategy, there are o ten clinical (http://www.guideline.gov), which is a resource maintained by
situations that require more nuanced answers and speci c expertise. the U.S. Department o Health and Human Services. This resource
73
STEP 1 STEP 2
Whe re Can I Find The
Hig he s t Quality Ans we r
P
Ide ntify a Clinic al Que s tio n
A
• As s pe cific a s pos s ible the Fas te s t?
R
• Focus on pa tie nt, Is The re a De dic ate d Re s o urc e ?
T
inte rve ntion, compa ris on, (ie, De finitions in dictiona rie s )
I
outcome • Ele ctronic me dica l re cord
• Re na l dos ing of me dica tions
Ye s • Antibiotic re comme nda tions ) No
• Drug re fe re nce
• Antibiotic dos ing guide
T
h
• Me dica l ca lcula tor?
e
• Othe r
S
p
e
c
Is This a Co mmo n To pic ?
i
a
S ummary Re s o urc e Ye s • Would the re be g uide line s for this ?
l
t
y
• Dyna me d • Am I unfamiliar with this topic?
o
• UpToDa te
f
H
• Othe rs (ba s e d on
o
ins titutiona l No
s
p
ava ila bility)
i
t
a
l
M
e
d
Primary Lite rature S e arc h:
i
c
i
• Pubme d clinica l que rie s -
n
e
review a rticle s
a
• Trip da ta ba s e
n
• Cochra ne da ta ba s e
d
STEP 4
S
Did This Ans we r • Othe rs
y
My Que s tio n?
s
Try Ano the r Re s o urc e
t
e
If s o, S top He re • Conta ct a libra ria n
m
• As k a cons ulta nt/
s
STEP 3 No s pe cia lis t
o
f
C
a
Figure 11 2 An effective search strategy for answering clinical questions.
r
e
acilitates comparison o guidelines on topics published rom di er- is connected to a clinician’s personal computer or mobile device
ent societies (eg, breast cancer screening guidelines rom American via the Internet. The advantages o this include an online backup
College o Physicians versus American College o Obstetrics and and the ability to synchronize les across multiple devices with very
Gynecology). Although somewhat cumbersome to use initially, it is little e ort. Many o these services allow a clinician to keep a copy
readily updated. Actively recruiting clinical trials are indexed in the o the les on a computer or mobile devices so these les can be
registry ClinicalTrials.gov (https://clinicaltrials.gov), provided by the accessed even when not connected to the Internet. These systems
U.S. National Institutes o Health. retrieve in ormation by storing text documents that can be searched
Although there are advantages and disadvantages to the di er- using terms that obtain results based not just on the title o the le
ent resources, the least use ul resource is the unavailable resource. but also rom the text o the document. This allows use o these
For this reason, clinicians should nd the resources available at resources with almost no organization necessary.
their home institutions. Clinicians should identi y clinical questions With a digital storage solution, the important tactile experience
and explore the available resources to nd answers. Ideally, a query o touching and annotating the original article is lost. The clinician
should obtain a high-quality answer in less than 5 minutes. must convert rom paper to digital les, name olders, and move
documents to the olders that have been created. A similar e ort
■ STORAGE FOR FUTURE REFERENCE would be required to set up a new le cabinet, but many struggle to
A challenge o the digital and increasingly paperless work environ- invest the time into a similar digital system. Theoretically, this digital
ment is how to keep up with the materials that an individual clini- storage solution is still a mechanical system that could ail.
cian would like to save or uture re erence. Mobile technology can
Suggestions
really help take busy clinicians ar beyond the traditional physical le
cabinet into a system that can travel on mobile devices with certain 1. Understand the cloud Many clinicians already have user
ones that can even help create bibliographies i needed. accounts or cloud-based storage solutions but are simply
Many digital storage solutions ollow the “ reemium” model. unaware that they do. Programs such as Dropbox, Google
“Freemium” is a pricing strategy that re ers to an initial ree price or Drive, Box, Evernote, OneDrive, and iCloud all meet the de ni-
the limited use o an application that then tiers to paid subscriptions tion or cloud-based storage solutions. I a clinician already has
or heavier use. The real advantage o the digital storage solutions one o these accounts, then it may be best or that clinician to
is their use o storing les virtually in the “cloud.” The “cloud” re ers rst explore all the eatures o that solution, be ore considering
to the storage o les in a server (ie, high-capacity computer) that others (Table 11-3).
74
TABLE 11-3 Example Cloud-Based Storage Solutions
C
H
Platform Cost/Memory Pros Cons
A
Box.com Free up to 10 Gb, then $10/mo or Only solution that advertises HIPAA Web editing,
P
100 Gb complaint or enterprise clients; great commenting, and
T
amount ree; can edit documents on tagging have some laws
E
R
web, make comments, and assign tasks to that limit their use ulness
documents
1
1
Dropbox Free up to 2 Gb, $10/mo Established company; cross plat orm with Small amount o data ree
1 Terabyte o storage easy sharing and well-integrated into mobile
devices
P
r
a
c
Evernote Free up to 60 Mb data trans er per Accepts multiple types o inputs including Best use requires some
t
i
mo, then $45/per y or heavier use audio, photos, scanned documents, and organization
c
a
web clips; advanced organization with
l
C
notebooks and tagging
o
n
Google drive 15 Gb ree, paid plans or 100 Gb Best or working on iles with other users, Navigation o the drive is
s
i
per mo even simultaneously; great i you use somewhat nonintuitive
d
e
Android mobile device(s)
r
a
t
i
Microso t Onedrive Free up to 7 Gb Ideal i heavy user o Microso t ® O ice,
o
n
especially on a windows plat orm
s
computer
o
I
n
c
o
r
p
o
While the solutions can be accessed through websites using but it should be noted the “personal service” o Box is not
r
a
a web browser, most also o er applications that can be down- HIPAA compliant.
t
i
n
loaded and placed on personal computers and mobile devices
g
4. Annotate articles Some clinicians pre er to read paper articles,
that allow or more convenient access o their stored les. The
E
which then allows or highlighting or annotating. Digital ver-
v
applications work by turning les and olders meant or storage
i
d
sions o these articles o ten allow o highlighting or annotating
e
into ones that are automatically synchronized and immediately as well depending on the viewer used on a computer or mobile
n
c
uploaded to the cloud, which allows them to thereby be avail- device. This latter way o reviewing articles allows the clinician
e
able rom anywhere.
i
to more easily save notes taken right inside the cloud-based
n
t
2. Use cloud-based storage A ter gaining an understanding o storage solution being used. For those who still pre er the paper
o
C
the “cloud,” clinicians should strongly consider using cloud- version to initially review an article, clinicians should consider
l
i
based storage in their practice. document scanner to convert the annotated paper version o
n
i
c
In general, any o these cloud-based storage solutions will the article into a digital one which can then be stored in these
a
l
work to get started, so once a choice is made regarding which cloud-based storage plat orms. The scanner that uses “Optical
P
r
one to use then time must be invested by users in order to Character Recognition” or OCR to turn the documents scanned
a
c
learn how to use all o the eatures. Many have “getting started” into a text searchable document a ter storage provides optimal
t
i
c
videos which can be viewed on their respective websites. Once unctionality. At the time o this chapter being dra ted, Evernote
e
a clinician is amiliar with the tool, time should be set aside or is the only one o the cloud-based solutions that has annotation
at least an hour or two to allow or some les or articles to be (eg, highlighting) or drawing built into its system. Many o these
moved over to the cloud olders. Investing time up ront will other plat orms can integrate easily with so tware/application
allow a busy clinician to get the most o out o the system viewers or readers (eg, “Portable Document Format” or PDF)
and generally be better organized moving orward in clinical which are available on most mobile devices.
practice. Choosing a system-based organization (eg, cardiol-
ogy, gastroenterology, in ectious diseases, and so on.) versus by ■ CULTURAL CHANGE
more speci c topic areas (eg, heart ailure, cirrhosis, pneumonia, Fostering a culture o education, accountability or actions, and
and so on.) depends on the le contents and personal pre er- support or evidence-based, high-quality care can encourage all
ences o the individual clinician. clinicians to keep learning or better patient care. Although intan-
The next time a clinical question comes up about the topic gible personal attributes contribute greatly to culture, leadership
previously searched and saved in this storage system, the clini- can adjust policy to achieve desired outcomes by nurturing a work-
cian can simply check the system rather than doing another place o collaboration, learning, open eedback, and accountability.
search. A supportive, collaborative environment encourages clinicians to
3. Comp ly with p atient conf dentiality As with any new discuss individual studies or publications, treatment decisions on
developments or advances in technology that a ects or complicated cases, and even to work together on quality improve-
modi es a clinician’s practice, care should be taken to assure ment and system-based projects. Impediments to collaboration
the sa ety and protection o a patient’s con dentiality. Most among hospitalist groups include provider scheduling and work-
cloud resources are not compliant with the Health Insurance load. Seven-day-cycled schedules may isolate providers to those
Portability and Accountability Act (HIPAA). As such, patient other providers on or the same shi ts only. Potential interventions to
or personal health in ormation should not be stored on improve collaboration could include standardized work schedules,
these digital and cloud-based drives. The notable exception physician lounges, physician ca eterias, grouped o ces, and consid-
is “enterprise services” provided by the digital plat orm Box, eration o patient census caps.
75
Open eedback also plays an important role in incorporating evi- SUGGESTED READINGS
dence and improving patient care. Increasing isolation o physicians
and high workload limits the available time to discuss the manage- Banzi R, et al. Speed o updating online evidence based point o care
ment o individual patients and/or to ollow up on subsequent care summaries: prospective cohort analysis. BMJ. 2011;343:d5856.
P
A
while o -service. Hando s could be an opportunity or colleagues Campanella P, et al. The impact o electronic health records on
R
to review care and potentially noti y providers o updates in hospital healthcare quality: a systematic review and meta-analysis. Eur J
T
medicine or new literature. In order to be success ul, this should be Public Health. pii:ckv122 [Epub]; Jun 30, 2015.
I
done in a collegial manner with adequate time allotted. Hospitalists
Del Fiol G, et al. Clinical questions raised by clinicians at the point o
should be encouraged to both give and request eedback regularly.
care: a systematic review. JAMAIntern Med. 2014;74:710-718.
Morbidity and mortality con erences may also o er an opportunity
to discuss updates in new literature using a pertinent clinical case. Elliott DJ, et al. E ect o hospitalist workload on quality and e -
T
ciency o care. JAMAIntern Med. 2014;174:786-793.
h
“Word o mouth” spread o in ormation is a pertinent and power ul
e
way to share medical in ormation and encourage a culture that Hunt DL, et al. E ects o computer-based clinical decision support
S
p
strives to keep up with the literature. systems on physician per ormance and patient outcomes: a sys-
e
c
tematic review. JAMA. 1998;280:1339-1346.
i
a
CONCLUSION
l
Kronen eld MR, et al. Survey o user pre erences rom a com-
t
y
The volumes o literature and increasing demands on physician parative trial o UpToDate and ClinicalKey. J Med Libr Assoc.
o
f
time have complicated the process o nding and incorporat- 2013;101:151-154.
H
ing evidence into clinical practice. Multiple new methods exist to
o
Michtalik HJ, et al. Impact o attending physician workload on patient
s
receive new evidence on pertinent topics in manageable volumes.
p
care: a survey o hospitalists. JAMAIntern Med. 2013;173:375-377.
i
t
Compiled review resources are signi cantly quicker and more reli-
a
Rangachari P, et al. Awareness o evidence-based practices alone
l
able initial resources than primary literature searches or most clini-
M
cal questions. Cloud-based storage options o er the advantages o does not translate to implementation: insights rom implementa-
e
d
better search-ability and portability over ling cabinets and stacks o tion research. Qual Manag Health Care. 2013;22:117-125.
i
c
i
printed manuscripts and journals; most cloud storage options are Sayyah EL, et al. To compare PubMed Clinical Queries and UpToDate
n
e
not HIPAA compliant but are good or storing medical re erences. in teaching in ormation mastery to clinical residents: a crossover
a
Simple steps can be taken to oster a positive learning culture at randomized controlled trial. PLoS One. 2011;6:e23487.
n
d
a hospitalist’s respective institution (eg, starting a journal club or Thiele RH, et al. Speed, accuracy, and con dence in Google, Ovid,
S
y
scheduling lunchtime discussions o current topics). Some hospi- PubMed, and UpToDate: results o a randomised trial. Postgrad
s
t
talists may pursue resources to better keep up, while others may
e
Med J. 2010;86:459-465.
m
inquire at their institution or “quick question” resources. Some clini-
s
cians may already be inundated with what they have collected and
o
ONLINE RESOURCES
f
need to invest in a cloud-based resource to decrease the stress every
C
time they ask, “Where is that article?” Encouraging a work culture
a
r
• Society of Hospital Medicine’s learning portal at https://
e
shi t toward addressing gaps may help clinician scope with in or-
mation overload. Even the most current or seemingly up-to-date shmlearningportal.org
clinicians started one step at a time and invested continued e ort. • Table 11-1 Digital Resources
76
SECTION 3
Transitions of Care
77
CHAP TER
12 INTRODUCTION
For patients being admitted into the hospital, hospital admission
may be the rst and most signi cant care transition that they will
experience in their medical care. The number o inpatients being
cared or by their primary care physicians has decreased signi cantly
in the last several years. In a study reported in JAMA in 2009, outpa-
tient to inpatient continuity with a primary care physician decreased
rom 44.3% in 1996 to 31.9% in 2006, correlating with the growth o
hospital medicine during the same time period. As a result, patients
Care Transitions are requently cared or by physicians who are meeting them or
or the rst time in the hospital, and who are un amiliar with their
into the Hospital: medical history, past hospitalizations, or amily and social support
network. In addition, due to the “shi t” structuring o many hospital
Health Care medicine groups, patients are likely to be cared or by multiple phy-
sicians during a single hospital stay, each one having to learn anew
the subtleties o their history.
Centers, Emergency Patients may be admitted to the hospital through the Emergency
Department (ED), directly rom an outpatient o ce, or trans erred
Department, Outside rom an outside acility such as another hospital, a Skilled Nurs-
ing Facility (SNF) or a Subacute Rehabilitation Facility (SAR). While
Hospital Transfers
these transitions have much in common, they also have unique
challenges in care transitions that are speci c to their sites. We will
discuss these challenges as well as potential solutions here.
ADMISSIONS FROM THE EMERGENCY DEPARTMENT

Joanna M. Bonsall, MD, PhD The majority o unscheduled admissions to the hospital come
Stacy Higgins, MD, FACP through the Emergency Department. The American College o
Emergency Physicians estimates that over the past decade the
Melissa B. Stevens, MD
percent o admissions through the ED has increased rom 64% to
over 80% while at the same time there was a decline in the per-
centage o unscheduled admissions rom clinics or doctor’s o ces.
Patients admitted through the Emergency Department experience
two transitions—the transition into the ED and the transition rom
the ED into the hospital. As patients’ transition rom the ED to
the hospital the need to not only trans er in ormation but also to
clari y who is primarily responsible or the patients’ care is critical
as there is o ten a delay between the exchange o in ormation
and the physical relocation o patients. Admissions that occur dur-
ing shi t changes may be particularly problematic as they result
in multiple trans ers o in ormation and responsibility. In the ED,
patients may be admitted at a shi t change with data pending
rather than be signed out to a new ED provider. These transitions
can result in ambiguity and con icting expectations between
ED providers and hospitalists speci cally regarding patient care
responsibilities and can contribute to dropped in ormation, delays
in treatment, and other errors that threaten the sa ety o patients.
Strategies to improve the transitions are discussed below and
outlined in Table 12-1.
■ TRANSITION OF INFORMATION
Hospitalists and ED physicians have varied expectations o the ED
to hospital handof both in terms o the in ormation that should be
communicated and the data that should be available at the time
o admission. ED physicians and hospitalists have dif erent roles
in patient care and dif erent in ormation needs, which may lead
to erroneous assumptions and misunderstandings. Understand-
ing the actors that af ect care o patients and communication on
both sides is crucial. Fatigue and increased workload can impact
79
is clear that the admitting provider is responsible or the patient
TABLE 12-1 Strategies to Improve ED to Hospital Transitions while they remain at their acility and once the patient leaves the
admitting acility they become the responsibility o the accepting
1. Vital signs should be rechecked at regular intervals while
provider. However, patients do not typically leave the ED immedi-
P
patients are boarded, consistent with protocol or their
A
intended loor location. ately a ter they have been accepted or admission by the hospitalist,
R
leading to con usion on the part o patients, nurses and other staf
2. Orders written in the emergency department (ED) should be
T
readily available or hospitalists to review in the electronic as to who is responsible or the patient’s care. The ED provider and
I
medical record (EMR). hospitalist should identi y pending labs and studies and who will
be responsible or ollow up and communication o results. In the
3. Formalized structured eedback should immediately
occur when there is an adverse event or hando event that patients remain in the ED or any length o time a ter
miscommunication. the communication o in ormation between the ED provider and
T
h
hospitalist, there should be an established protocol or identi ying
4. Standardized hando tools including an admission check-list
e
the responsible provider to be contacted with any new test results,
S
should be used to set expectations.
p
question/concerns, or any change in the patient’s clinical condition.
e
5. Clearly delineate the responsibility or ollow-up o pending
c
tests results.
i
a
■ PROLONGED ED BOARDING
l
t
6. Minimize asking EM physicians to order additional tests
y
ED overcrowding and inadequate inpatient capacity lead to pro-
o
be ore patients are accepted or admission or trans er to the
f
loor unless the tests will alter bed acuity level. longed ED boarding and increased risk o harm not only or admit-
H
o
7. Clearly identi y the clinician o record. ted patients but or ED patients as well. Admitted patients boarded
s
p
8. Hospitalist changes o shi t or boarded patients should be in the ED do not receive the same level o care as they would on an
i
t
inpatient unit and studies have shown that prolonged ED board-
a
communicated to ED sta expediently.
l
ing is associated with an increase in preventable adverse events,
M
9. Responsibilities or boarded patients should be speci ied at
e
the institutional level with written policy. length o stay and mortality. Institutions should develop a clear plan
d
or standardized communication and trans er o responsibility o
i
c
10. Prioritize getting boarders out o the ED to their inpatient
i
n
patients with prolonged ED boarding that should include a plan or
units.
e
shi t changes and may include standardized order sets to be initi-
a
n
ated in the ED prior to trans er to an inpatient unit.
d
Reducing ED boarding is a challenge at most acilities that
S
y
communication and updating o patient in ormation and contribute requires the input o a multidisciplinary team. Prolonged ED board-
s
t
e
to unintended outcomes. ing is usually the result o hospital overcrowding and there are
m
In order to improve sa ety and quality o care and avoid ambigu- many possible strategies to improve patient ow including: using
s
ity hospitalists and ED providers at any acility should clearly identi y a “ ull capacity protocol” where the burden o boarding patients is
o
f
and mutually agree upon a minimum set o data elements that shared between the emergency medicine and hospital medicine
C
a
should be part o the transition. The use o checklists and standard- department; coordinating discharges as early in the day as possible;
r
e
ized communications tools are recommended in assisting this moving toward a 24/7 operational culture, active bed management/
process. In keeping with existing guidelines, the minimum data set hiring a “Bed Czar” to match bed needs to resources and utilizing
should include a principle diagnosis and problem list, medication observation units.
list, patient’s cognitive status, and test results/pending tests. In or-
mation should be communicated between providers in a secure, ■ COMMUNICATION BETWEEN THE PCP
private and HIPAA compliant manor. Communication o in orma- AND THE HOSPITALIST
tion should be done in person or by phone with an opportunity or
Barriers in communication between the primary care provider
both the admitting and receiving provider to ask questions and get
(PCP) and the ED physician or admitting hospitalist are similar on
clari cation and eedback. This last step, while recommended or all
admission and discharge. On admission, patients may be in pain,
handovers, is especially crucial in this type o transition, given the
con used, and distressed, providing limited or unreliable histories.
dif erent backgrounds and roles o the providers.
The communication with the PCP can provide in ormation on past
While hospitalists requently ask or additional in ormation to
medical history, past hospitalizations, previous testing done, medi-
help with patient management decisions, such requests may lead
cation reconciliation, social history, and the baseline medical status
to a delay in patient trans er. Requests or additional workup prior
o the patient. However, conversations require identi cation o the
to admission may be appropriate i this in ormation will af ect one
PCP by the patient, ability to reach the PCP in a timely ashion, easy
o the ollowing: (1) the level o inpatient care (ie, ward vs telemetry
access by the PCP to the patient’s updated medical record, and a
unit vs ICU), (2) the admitting service (medicine vs surgery vs other),
mutually available time or the conversation. System barriers inter-
or (3) time to obtain critical in ormation (ie, the patient may be able
rupting this communication include inadequate reimbursement
to get a CT more rapidly in the ED than once they are admitted to
to the PCP or the time involved in reviewing the record and the
the oor). In the event that additional workup is warranted prior
conversation; time shortage in a busy outpatient schedule; and
to a patient being accepted or admission, the ED provider and
ragmented patient care with an incomplete record. There is also
hospitalists should have a protocol or communicating in ormation
the transition process barrier o the lack o a standardized commu-
about the patient at shi t change with their colleagues who will be
nication system—success ul completion o a single phone call may
responsible or ollowing up these tests and ultimately admitting
require multiple calls in each direction as neither party is available
these patients.
at the time o the phone call, and lengthy waits between calls; text
pages may not be responded to until records are reviewed and
■ TRANSITION OF RESPONSIBILITY there is time in the schedule; and e-mails may not be secure and do
Unlike other transitions, the exact moment o the trans er o respon- not allow or interactive dialogue.
sibility or patient care may be ambiguous. When patients are admit- In cases where the hospitalist and primary care physician share a
ted rom the outpatient clinic setting or skilled nursing acilities, it common electronic health record, communication through the EHR
80
can be secure and timely, with ongoing real time access by both signout, they can advise against a direct admission. Familiarity and
parties. However, this is likely the exception rather than the rule, and trust between the outpatient and hospitalist group can help acili-
C
it assumes that patients obtain all o their care within a single health tate direct admissions as each learns the other’s practice style and
H
system. Where this does not exist, methods such as e-mail and ax
A
com ort.
P
are acceptable. With the growth o the Patient Centered Medical While a patient may initially be assessed as stable or direct admis-
T
Home, it is possible that care coordinators in the health care system sion, an extensive wait time in patient registration where the patient
E
can help to acilitate in ormation trans er rom the medical home to is unmonitored or may miss scheduled medications may lead to
R
the admitting physician. Employment o nurse case managers by the patient becoming unstable. As the number o inpatient beds is
1
practices can assist with the barrier o communication coordination, reduced, and the ll rate goes up, patients may have to wait several
2
have access to the patient’s medical record, and can standardize the hours be ore receiving a bed assignment. Solutions include having a
trans er o in ormation. Care Initiation Unit that serves as a transition zone where the patient
C
can check in, have their vital signs checked, and be evaluated by the
a
r
e
admitting hospitalist there. This decreases time spent in the ED and
PRACTICE POINT
T
catches patients who may need stabilization prior to going to the
r
a
• •The ED-to-hospital admission process can be a high-risk event oor bed. It also allows or continuous monitoring and initiation o
n
s
orders while awaiting a bed assignment.
i
due to dif ering backgrounds/roles o providers and ambiguity
t
i
o
at the time o admission about trans er o responsibility or
n
■ TRANSFER OF INFORMATION
s
patient care. To minimize the risk, ED and hospital medicine
i
n
groups should develop mutually agreed upon in ormation As with ED-to-hospitalist communications, primary care providers
t
o
checklists or standardized templates to aide trans er o and hospitalists have dif erent patient care roles and perspectives,
t
h
in ormation; and should develop clear guidelines around who which may contribute to misunderstandings and missed in orma-
e
will assume responsibility or pending test results and changes tion. Opportunities to improve care at this transition can likely
H
o
in patient status in the period a ter the patient has been improve the experience o the hospitalized patient, reduce unnec-
s
p
accepted by the hospitalist but still remains in the ED. essary testing, reduce length o stay, and potentially reduce cost to
i
t
a
the system. On admission, as with all care transitions, there should
l
:
be a minimum shared data set that should include the rationale
H
e
■ DIRECT ADMISSIONS FROM AMBULATORY or admission, the working diagnosis, the problem list, key history
a
l
components and recent changes, relevant laboratory and radiologic
t
HEALTH CARE CENTERS
h
results, medication list and allergies, and patient/ amily pre erences
C
While ewer patients are being admitted directly rom a physician’s
a
and support system. Ideally this in ormation should be shared both
r
o ce to the hospital, there are signi cant advantages to continuing
e
verbally and in a written ormat in a HIPAA compliant manor. Verbal
C
this practice or appropriate patients. A direct admission means the conversations allow the hospitalist opportunities to ask questions
e
patient can avoid the potentially long wait in the ED and prolonged ED
n
and clari y in ormation. As the “accepting” hospitalist may not be
t
e
boarding while also reducing ED overcrowding. The direct admission the hospitalist who ultimately cares or the patient, the in ormation
r
s
process also allows the opportunity or the PCP to speak directly to the
,
should also be in an accessible paper or electronic ormat or uture
E
hospitalist, sharing the patient’s medical history as well as providing
m
providers to re erence. This may be sent with the patient or sent by
some anticipatory guidance to the hospitalist, who is not as amiliar
e
ax. Alternatively, the receiving hospitalist may ll out an electronic
r
g
with the patient. However, as with all transitions, the direct admission template during or a ter a phone conversation; however, this may
e
has some speci c risks that need to be anticipated and managed.
n
be cumbersome. On initial contact, agreeing upon how, when, and
c
First, the patients appropriate or direct admission should be care ully
y
under what circumstances urther communication should occur
D
selected, to ensure that they will be admitted to the correct care loca- between in and outpatient physicians should be established.
e
tion and that they are not at risk or deterioration be ore being seen by
p
a
the hospitalist. Second, there should be clear communication between
r
t
m
the PCP and the hospitalist regarding the patient to be admitted. PRACTICE POINT
e
n
• •A direct admission rom the PCP’s o ce to the hospital can
t
,
■ SELECTION OF APPROPRIATE PATIENTS save the patient hours o waiting in the ED and can help
O
FOR DIRECT ADMISSION
u
reduce ED overcrowding. However, patients appropriate
t
s
Advantages to admitting a patient directly to the hospital include or direct admission should be selected care ully, and clear
i
d
e
convenience to the patient, improved patient satis action, reduced communication between the PCP and hospitalist should occur.
H
crowding o the Emergency Department, and decreased cost to the Institutions should partner with admitting PCPs to develop o
s
system. In evaluating who is appropriate or direct admission, there a checklist to determine whether a patient is appropriate or
p
i
are a ew recommendations in the literature: direct admission, and standardize what in ormation should be
t
a
transmitted and how that in ormation should be transmitted.
l
• The admitting diagnosis is airly certain and no additional triage
T
r
is needed.
a
n
• The patient is clinically stable in their vital signs and does not
s
e
require supplemental oxygen, immediate IV uids, antibiotics ■ ADMISSIONS/TRANSFERS FROM OUTSIDE HOSPITALS
r
s
or urgent imaging. While trans ers rom outside hospitals account or a minority o hos-
• The patient has been seen and evaluated on the day o admis- pital admissions (3%-5%), trans erred patients typically have a higher
sion by their primary provider. severity o illness and are more medically complex. Additional risks
• The patient arrives at the hospital early in the day (be ore 4 pm) present at time o trans er o ten include a medical record system
to acilitate communication between the admitting physician that is not shared with the trans erring hospital and delays in trans er
and the hospital team and be ore shi t changes. due to transportation, distance rom receiving hospital, bed avail-
I the admitting hospital team is concerned about the patient’s ability and physician acceptance, resulting in more than two-thirds
condition or need or more extensive initial workup a ter getting o patients arriving at the accepting acility at night. While studies
81
Another risk inherent to interhospital trans ers is in delaying the
TABLE 12-2 Risks of Interhospital Transfer appropriate care o the patient. Delays in patient care are multi-
actorial and can be caused by lack o expertise at the trans erring
Prior to Transfer During Transfer After Transfer
acility as well as delays in acceptance at an appropriate acility and
P
Delays in care Decompensation Discontinuity o
A
arrangement o transportation. Minimizing these delays requires a
initiation due to during trans er care plan
R
coordinated ef ort among administrators, physicians, and medical
lack o expertise
T
Arrival at night Unnecessary and/or staf . Trans er centers and rapid trans er protocols at large medical
I
Delays in care Arrival to duplicative testing centers have helped acilitate trans ers by (a) ensuring that a bed is
due to delays inappropriate level Medication errors available at the appropriate level o care; (b) coordinating conversa-
in inding an o care
accepting acility Back-end tions between the sending and receiving physicians; and (c) in some
discontinuity cases, “triaging” trans ers to ensure that those with the most urgent
T
Inappropriate
h
needs have priority. Delays in care initiation can be minimized by
trans er
e
the receiving physician making treatment recommendations to the
S
p
sending physician when applicable. Implementation o the pro-
e
posed therapy should be up to the discretion o the sending physi-
c
i
a
cian. In addition, coordinated ef orts can also determine optimal
l
t
have shown improved outcomes or disease-speci c trans ers such
y
timing and method o transport.
o
as myocardial in arctions and trauma, in general, trans erred patients I trans er is delayed by more than a ew hours, a repeat verbal
f
H
have overall higher levels o morbidity and mortality; these dif er- conversation should occur between the current sending and receiv-
o
ences cannot always be accounted or by severity o illness alone.
s
ing physicians to ensure that no changes have occurred and that
p
There ore, although only a minority o patients go through a trans-
i
the physician who is receiving the patient at time o trans er has the
t
a
er, special attention should be paid to this care transition.
l
opportunity to ask clari ying questions. I at any point a patient has
M
Reasons or trans er typically include disease-speci c interven- a change o status, this should be communicated to the receiving
e
tions not available at the trans erring institution and availability o
d
physician as well.
i
c
subspecialty expertise at the receiving institution. However, reasons
i
n
or trans er can also be less well-de ned and include lack o bed ■ RISKS AFTER TRANSFER
e
availability at the sending institution, patient request, and insurance
a
n
Once the patient has trans erred, the main risks are with discontinu-
coverage reasons. Medicaid patients continue to be trans erred at a
d
ity o care plans—both discontinuity o care between the sending
S
higher rate than those covered by private insurance. Most studies
y
and receiving acilities and discontinuity once the patient has been
s
on interhospital trans ers have ocused on disease-speci c reasons
t
discharged or sent back to the receiving acility. In addition, trans-
e
or trans er such as patients requiring percutaneous angioplasty or
m
erred patients o ten receive unnecessary repeat imaging or testing,
myocardial in arctions, trauma patients, or stroke patients requir-
s
leading to increased charges and radiation doses. To minimize
o
ing management at a comprehensive stroke center. Other patient
f
discontinuity, best practice suggests the inclusion o the ollow-
C
populations that have been studied include critically ill patients
ing in the documentation sent to the receiving hospital: re erring
a
being trans erred to a hospital with greater intensive care capacity.
r
e
physician and contact number, reason or trans er and nature o
Patients trans erred or reasons that all outside these parameters
illness or injury, current and outpatient medications, vital signs and
have been not been studied; thus overall best practices must be
relevant physical ndings at time o trans er, results o pertinent
in erred rom the existing literature.
diagnostic studies, treatment up until time o trans er, and pending
Challenges that are speci c to interhospital trans ers can be
tests (including cultures). Use o a standardized accept orm—either
divided up into risks prior to trans er, during trans er, and a ter
by the sending or receiving acility—have been shown to reduce
trans er (Table 12-2). Prevention o decompensation during travel
missing in ormation. Another method to minimize discontinuity,
requires specialized transport and personnel and will not be covered
particularly or patients being trans erred or a single procedure, is
in this chapter.
to trans er the patient back to the sending acility when appropriate.
At time o discharge, care should be taken by the receiving acility
■ RISKS PRIOR TO TRANSFER
to include the relevant hospital course and diagnostic ndings rom
Prior to trans er, the risks and bene ts o the trans er should be con- the sending hospital into the discharge in ormation sent to the
sidered. While patients clearly bene t rom appropriate trans er to patient’s outpatient physicians.
specialty acilities in the case o stroke, trauma, and acute coronary Failure at any point reduces quality and delays intervention, leads
syndrome requiring percutaneous angioplasty, the bene ts o trans- to inappropriate allocation o resources, and increases the nancial
er in other cases are less clear. There ore, a care ul review should be burden on the patient, hospital, and health care system.
done by both the sending and receiving physicians to ensure that
the patient is appropriate or trans er and that the bene t o care
received at the receiving acility will outweigh the risks o trans er. In PRACTICE POINT
the case o the patient requesting the trans er, a thorough conver-
sation should occur to ensure that the patient (a) is not requesting
• •Intrahospital trans ers are at high risk or dropped patient
in ormation and delays in patient care. Risks may be minimized
the trans er merely because o poor communication at the sending
by (a) creating guidelines to ensure that patients are being
acility and (b) is not overestimating the impact o quality metrics
trans erred appropriately; (b) acilitating conversations
at the receiving acility. In the case o a patient who is perceived to
between the sending and receiving physicians, in addition
require specialized care, the sending hospitalist, the sending sub-
to any relevant subspecialists; (c) providing treatment
specialist (i available), the receiving subspecialist, and the receiving
recommendations to the sending physicians when appropriate;
hospitalist should con erence to determine medical necessity and
and (d) creating standardized checklists or documentation
plan o care, both to initiate care and ensure that appropriate care is
required upon trans er and upon discharge. Trans er centers
immediately implemented upon the patient’s arrival. Except in cases
and rapid trans er protocols can help streamline many o these
o urgency, overnight trans ers should be avoided as these have
processes.
been shown to have more negative outcomes.
82
■ ADMISSIONS FROM SKILLED NURSING FACILITIES Horwitz L, Meredith T, Schuur J, Shah NR, Kulkarni RG, Jeng GY. Drop-
ping the baton: a quality analysis o ailures during the transition
C
While patients rom skilled nursing acilities (SNFs) or subacute
rom emergency department to inpatient care. Ann Emerg Med.
H
rehabilitation acilities (SARs) usually present to the hospital through
A
the ED, they nevertheless represent a signi cant care transitions 2009;53(6):701-710.
P
challenge or the hospitalist. Not only are they usually medically Singer A, Thode H, Viccellino P, Pines J. The association between
T
complex, they also requently are unable to provide a coherent length o emergency department boarding and mortality. Acad
E
R
medical history or describe their medication regimen. In addition, Emerg Med. 2011;18:1324-1329.
they are less likely to be accompanied by a relative or caregiver than
1
Snow V, Beck D, Budnitz T, et al. Transitions o Care Consensus Policy
2
elderly patients who present rom their homes. In recent studies, up Statement American College o Physicians-Society o General
to 10% are trans erred to the ED without any documentation and Internal Medicine-Society o Hospital Medicine-American College
up to an additional 40% are missing in ormation. Most requently
C
o Emergency Physicians-Society o Academic Emergency Medi-
a
absent rom the documentation are baseline cognitive unction, cine. J Gen Intern Med. 2009;24(8):971-976.
r
e
current medications, and advance directive status. Because this
T
Admissions from the Ambulatory Health Care Center
r
population requently presents with non-speci c complaints such
a
n
as alls, dehydration, or con usion, these omissions resulted in more Carrier E, Yee T, Holzwart RA. Coordination between Emergency and
s
i
t
investigations, particularly head CTs. They are also at risk or medica- Primary Care Physicians. National Institute or Health Care Re orm
i
o
tion errors, unnecessary testing, and inappropriate/unwanted care.
n
Research Brie Number 3; February 2011.
s
While there have been ew studies looking at interventions, one
i
Eichner JE, Cooley WC. Coordinating the Medical Home with Hospi-
n
study showed that while providing the SNFs/SARs with trans er
t
talist Care. Hospital Pediatrics Online. 2012;2(2):105-108. Available
o
orms did not result in a signi cant use o the trans er orms, ollow
t
at http://hosppeds.aapublications.org. Accessed May 18, 2015.
h
up ound that the data transmitted still improved. There ore, hospi-
e
tal medicine groups should strongly consider partnering with the Van Blarcom JR, Srivastava R, Colling D, Maloney CG. The develop-
H
ment and implementation o a direct admission system at a
o
SNFs/SARs that requently send patients to their acility.
s
tertiary care hospital. Hosp Pediatr. 2014;4(2):69-77.
p
There is no consensus on in ormation that should be included on
i
t
a
trans er orms, but recommendations include reason or ED trans er, Transfers from Other Hospitals
l
:
past medical history, baseline cognitive unction, medication list and
H
Hernandez-Boussard T, Davies S, McDonald K, Wang N. Interhospital
e
mediation allergies, vital signs at time o complaint, advanced direc-
a
acility trans ers in the United States: a nationwide outcomes
l
tives, contact in ormation or the SNF/SAR’s health care provider,
t
h
and contact in ormation or the patient’s next o kin or medical study. J Patient Saf. 2014 [epub]:Nov 13
C
a
decision maker. Iwashyna T. The incomplete in rastructure or interhospital patient
r
e
trans er. Crit Care Med. 2012;40:2470-2478.
C
e
Admissions from Skilled Nursing Facilities
n
PRACTICE POINT
t
e
Gri ths D, Morphet J, Innes K, Craw ord K, Williams A. Communi-
• •Patients trans erred
r
rom SNFs/SARs are at high risk o
s
,
cation between residential aged care acilities and the emer-
medication errors and unnecessary testing and treatment.
E
m
gency department: a review o the literature. Int J Nurs Stud.
Providing partnering SNFs/SARs with standardized trans er
e
2014;51:1517-1523.
r
orms may help improve communication.
g
e
n
c
y
D
SUGGESTED READINGS
e
p
a
r
Admissions from the ED
t
m
e
Beach C, Cheung D, Apker J, et al. Improving interunit tran-
n
sitions o care between emergency physicians and hospital
t
,
O
medicine physicians: a conceptual approach. Acad Emerg Med.
u
2012;19:1188-1195.
t
s
i
d
e
H
o
s
p
i
t
a
l
T
r
a
n
s
e
r
s
83
CHAP TER
13 INTRODUCTION
In-hospital care has su ered as a result o the increased ragmenta-
tion in the delivery o care, speci cally secondary to new clinical
models such as the rise o hospitalist care in the United States, and
the move toward cutting hours or residency trainees in teaching
hospitals. This ragmentation has resulted in a greater need or care
coordination and a ocus on transitions, particularly or the hospital-
ized elderly population. For example, or a typical patient, a member
o the patient’s primary team is present in the hospital only 50% o
Care Transitions the time. Hospitalized patients are passed between doctors an aver-
age o 15 times during a single 5-day hospitalization.
within the Hospital: In addition to the ocus on end-o -shi t changes that are germane
to hospital medicine, there is increasing recognition that the ocus
The Hand-Off on hando behaviors needs to extend to cover the vulnerable

“service change,” which represents a more permanent change in
primary hospital care provider. While the current literature has
ocused on strategies or to e ectively “give” in ormation during the
hando , recent literature has also highlighted the need to examine
Vineet M. Arora, MD, MAPP the critical role o the receiver.
Jeanne M. Farnan, MD, MHPE The hando is
a uid, dynamic exchange that is subject to distraction,
interruptions, uctuates on aptitude o and con dence
in o -going and on-coming clinician and is contingent
on the on-coming clinician’s con dence in the quality,
completeness o the in ormation.
Cook et al (2000)
While the scope o Cook’s de nition re ers primarily to shi t change,
the term handof s has taken on a li e o its own, with the term being
used synonymously with a broader set o care transitions, such as
admission, discharge, and even communication between outpa-
tient physicians. In this chapter, we will ocus on the skills that are
essentially to those hando s which permeate the in-hospital setting
or hospitalists, with a special ocus on shi t and service change, as
well as introduce assessment strategies to ensure the sa ety and
ef cacy o these hando s.
TYPES OF INTRAHOSPITAL HANDOFFS
■ SHIFT CHANGE
Shi t change is the trans er o content and pro essional responsibility
rom one clinician to another at the end o the shi t. One important
distinction among shi t changes is whether the outgoing clinician is
returning to assume ongoing care o this patient or the hando is
just a temporary coverage or emergencies until the primary team
returns. In the case o the latter, the covering physician is o ten
accepting a hando only to manage overnight emergencies, but
planning and execution o care are largely on hold.
• • Signout: A type o shi t change that o ten pre erentially re ers to
a primary team who is assuming care o the patient and trans-
ers care temporarily to another clinician and that primary team
member will return to assume care o patient. Can also re er to
the written document used to trans er in ormation. It can also
re er to the time period when people are actively handing over
patients to another group o providers.
• • Cross-coverage: The care that a clinician provides when “covering”
a patient whose daily responsibility is assumed by another
clinician or team.
84
■ SERVICE CHANGE
TABLE 13-1 Questions to Risk Stratify Handoffs—If Yes,
C
A service change is a permanent trans er o content and pro es-
to Any, Inherently Higher Risk
H
sional responsibility at the end o one’s on-service time or rotation
A
to a new physician or team o providers who will assume ongoing (1) Is the patient physically moving?
P
care o the patients. This service hando is o ten more extensive (2) Is the hando permanent (more than just a ew hours or a
T
and includes description o the initial reason or the patient’s need night)?
E
R
or hospitalization, hospital course to date, current status, and antici- (3) Is the patient unstable?
pated plan o care, including discharge. The timing o the service
1
(4) Is this the irst time the receiver is hearing about a patient?
3
change is driven by the duration o the rotation and can happen
as requently as weekly to as long as monthly. Service changes may
not always include ace-to- ace opportunities or discussion and are
C
a
requently supplemented with a written account o the patients’
r
e
hospital course, o ten re erred to as an “o -service note.” decisions and judgments. Certain core elements o hando s include
T
the ollowing:
r
a
n
■ SERVICE TRANSFER
s
■ VERBAL COMMUNICATION
i
t
Service trans er is the change o service o a patient rom care o one
i
o
A hando typically, and ideally, has some element o verbal com-
n
group o clinicians to an entirely di erent group o clinicians, usually
s
rom a di erent specialty or ward, to receive a di erent service that munication, either ace-to- ace or over the phone. The goal o
w
verbal communication is o ten to build a shared mental model or
i
is unique to the receiver’s specialty or ward. This could include an
t
h
“escalation o care” due to worsening patient illness (trans er to the a patient, with a ocus on anticipatory guidance and tasks to be
i
n
intensive care unit) or trans er to a subspecialty service or a speci c done and the rationale which accompanies those tasks, as well as
t
h
management issue (trans er rom medicine team to surgical team their priority. Verbal communication also allows or other critical
e
actions such as questioning and reading back in ormation relayed
H
or procedure and postoperative care). As with the service change,
o
the service trans er is accompanied by a verbal exchange o patient and received. One key eature o verbal communication is that it
s
p
in ormation as well as a comprehensive trans er summary which should not “rehash” what is already available in printed or electronic
i
t
a
includes a written documentation o the patient’s reason or trans er records. It is important to strike the right balance between too much
l
:
and too little in ormation, particularly to ocus on what the receiver
T
and detailed in ormation about their hospital course.
h
really needs to know. While verbal ace-to- ace communication is
e
H
RISK STRATIFICATION OF HANDOFFS not always possible, it is certainly the ideal as it provides or these
a
opportunities. During shi t change in hospitals, this process is o ten
n
In considering the various risks associated with these hando s, a
d
called “signing out.” During an admission, this could take the orm o
-
white paper rom University Health System Consortium suggests
O
that the ollowing three questions can be used to triage risk to a report given over the phone between the emergency room physi-
patients during hando s: (1) Is the patient physically moving? (2) cian and the hospital-based physician.
Is the hando permanent (more than just a ew hours or a night)?
(3) Is the patient unstable? I the answer to any o these questions ■ WRITTEN COMMUNICATION
is a yes, then the risk is inherently higher. There ore, the highest risk There is usually some orm o a written communication (or “transi-
transitions may be a service trans er to the ICU or emergent sur- tion record”) that supplements the verbal hando with additional
gery—since the patient is unstable, moving, and it is a “permanent in ormation that could become important at a moment’s notice,
hando ,” meaning more than just a ew hours. Another example such as the patient’s primary care physician or code status. The
certainly includes the service change, in which the hando is per- written hando document generally either is a user-created docu-
manent and the clinician is likely not returning to assume care o ment that is computer-based, is auto-generated by the electronic
the patient. medical record (EMR), or is a hybrid o these. O ten, this written
In addition to these questions to strati y risk during hando s, communication is used as a peripheral guide or the conversation
another philosophy that has emerged is the concept o “com- during the verbal hando o in ormation. During shi t change in
mon ground”—or rather how much knowledge do the incoming many academic teaching hospitals, this written communication is
and outgoing clinicians already share about the patient? When known as the signout.
a receiver may not know a patient at all, the hando may be at
greater risk due to the high degree o uncertainty that can cloud ■ TRANSFER OF PROFESSIONAL RESPONSIBILITY
the initial evaluation o a patient. This is not only true during A hando is more than just the trans er o in ormation; it is also
service and shi t hando s, but other trans ers o care, such as a the trans er o pro essional responsibility. Acknowledgment o the
patient being admitted rom the emergency department to the accountability or a patient’s care is an important eature o success ul
hospitalist on the oor. Uncertainty is a de nite risk or patients hando s. Given the need or requent shi t handover in the current
and has been demonstrated to lead to patient harm or near misses hospital systems, this is a critically important step in the process. Prior
and inef cient work in both resident signouts and hospitalist work has indeed demonstrated that the being cared or by a cross-
service changes. In addition, uncertainty can also lead to rework, covering physician is associated with an increased risk o preventable
including procedures and testing, which may also negatively adverse events. The etiology o this risk is multi actorial. Not only do
impact the patient’s overall care. There ore, hando s are inherently covering physicians lack the primary knowledge regarding a patient
risky when the receiver does not have any a prior knowledge o the they are caring or, but they could also lack o pro essional ownership
patient (Table 13-1). o patients who they did not directly admit or care or a on a more
routine basis. While improving the process o hando s can increase
CORE COMPONENTS OF HANDOFFS the knowledge a covering physician, e orts to improve pro essional
Regardless o the type o in-hospital transition involved, hando s responsibility are equally important. In essence, a high-per orming
have the common goal o creating a shared mental model between team culture that unctions such that “every patient is our patient” is
the sender and the receiver in order to ensure that surrogate what is needed to support the necessary pro essional responsibility.
85
S e nde r S top pa tie nt S pe cific ve rba l excha nge Re ce ive r inte gra te s
orga nize s & ca re ta s ks to be twe e n s e nde r a nd new informa tion a nd
upda te s ha ndoff conduct re ce ive r (could be in a s s ume s ca re of
informa tion ha ndoff pe rs on or ove r phone )
P
pa tie nt(s )
A
R
T
Pre -hando ff Arrival Dialo g ue Po s t-hando ff
I
• La ck of time, • No s e t Se n d e r c o u ld • Forge t key ta s ks
poor time loca tion or • Provide dis orga nize d info or informa tion
T
ma na ge me nt, time
h
• Us e va gue or uncle a r • Not docume nt
e
fa tigue, or • Not a ble to la ngua ge a ctions ta ke n
S
p
work preve nts conta ct • Act on pla n
• Fa il to provide clinica l
e
upda ting
c
s e nde r or impre s s ion (wha t is without ta king
i
a
• La ck of re ce ive r
l
wrong), a nticipa tory new a rriving
t
y
clinica l • Compe ting guida nce (if/the n), pla n informa tion into
o
judgme nt to
f
obliga tions (to do), & ra tiona le (why) a ccount
H
cons truct (work or • Not inve s t in the
o
Re c e ive r c o u ld
s
prope r pe rs ona l) ca re of pa tie nt
p
ha ndoff • Not lis te n (dis tra ctions )
i
t
• Ha ndoff not (la ck of
a
• Mis unde rs ta nd
l
• Va gue a priority profe s s iona l
M
la ngua ge ove r ta s ks • Not cla rify (a s k re s pons ibility)
e
d
que s tions )
i
c
i
n
Figure 13 1 Phases o the hand-o .
e
a
n
d
S
y
s
t
e
CORE STEPS TO THE HANDOFF PROCESS in ormation trans erred is up-to-date, or more cultural issues such as
m
a lack o priority or the signout process itsel .
s
In addition to the core components o a hando , it is important to
o
consider the core steps to the process o hando s. In thinking about Failing to update the written communication can result in either
f
omissions (in ormation not present) or commissions (in ormation
C
hando s as a process, one can conceptualize our basic phases
a
provided is incorrect). For example, in one study o written signouts,
r
to the process. Modi ed rom a consensus paper or Emergency
e
Medicine hando s, these our phases would include the ollowing 80% contained at least one medication omission and 40% one com-
(Figure 13-1): mission. Over hal had the potential to cause signi cant harm to a
patient. Although omissions were more common, commissions,
1. Pre-handof : Sender organizes and updates written in ormation such as including medications that were not actively being used
or hando . This o ten critical step will be discussed urther in in the patient’s care, were more serious. Examples included antico-
hando ailures, as this is where many errors can occur. agulants, intravenous (IV) antibiotics, narcotics, and hypoglycemics
2. Arrival: Sender completes patient care tasks to conduct hand- (insulin, etc). While the advent o EMR, and the ability to auto ll
o (or participate in “signing out”). This step also includes the certain elds, including medications, has helped to minimize the
negotiation between sender and receiver or time and meeting medication omission and commission errors, prior data has estab-
place or the hando . lished that technological solutions alone are not robust enough to
3. Dialogue: A speci c verbal exchange that takes place between prevent hando related errors such as these.
sender(s) and receiver(s). This verbal exchange could either be For these reasons, written communication that is linked to the
ace-to- ace (o ten pre erred) or over the phone in cases when electronic health record is o ten pre erred although care must be
an in-person hando is not possible. taken to ensure the written section is updated daily as well. In addi-
4. Post-handof : Receiver integrates new in ormation and assumes tion to omissions and commissions in the written communication,
ongoing care o patient(s). the use o vague language such as “today,” “tomorrow,” or “yesterday”
can result in con usion as can the use o nonstandard abbreviations
DIFFERENTIAL DIAGNOSIS OF FAILED HANDOFFS that either are not understood or can be mistaken or something
Understanding the content and process o hando s is essential to else (eg, HL or hyperlipidemia, which is o ten perceived as Hodgkin
understanding how hando s may ail. Each step in the process o lymphoma). Written communication can be plagued by the TMI (too
hando s is prone to ailure as outlined here. much in ormation) phenomenon, in which extraneous and non-
essential in ormation are included in the written signout which do
■ PRE-HANDOFF FAILURES not provide in ormation use ul or those covering the patient during
Since the ocus o the pre-hando phase is to create and update a shi t hando and can serve to clutter the existing, use ul in orma-
the written communication or the hando , ailures in this phase tion. Finally, cutting and pasting in ormation into the written signout
lead to errors in the transition record. O ten the inability to carry document can also serve to perpetuate inaccurate in ormation.
out the pre-hando phase is due to lack o time, ine ective time
management, workload, or orgetting to do so. Systems issues can ■ ARRIVAL FAILURES
contribute to the inability to update the written in ormation, includ- Arrival ailures include not arranging a speci c location or time
ing not scheduling protected time near the end o a shi t to ensure to meet or a hando . Even with a telephone hando , i a time is
86
not speci ed, the sender or receiver may ail to make contact at over-estimation o e ectiveness occurred. One way to mitigate
the hando time. Moreover, experienced hospitalists may make the trap o the egocentric heuristic is to improve the “common
C
an assumption that they do not “need” a verbal hando , since ground” or the a priori knowledge that senders and receivers
H
their experience will be enough to guide them on what to do. It
A
have about the patient. For example, i the night hospitalist has
P
is equally possible that the high workloads night hospitalists ace covered the patient be ore, they have some “common ground”
T
may hinder verbal hando s, because the night hospitalist is cover- with the day hospitalist in the sense that they both have had
E
ing too many patients or a meaning ul verbal hando to occur. some knowledge and interaction with the patient rom which the
R
Likewise, other work demands (competing clinical work, unstable verbal hando can build. Improving common ground is a design
1
patient, etc) or personal issues (late to work or having to leave problem—a key question is can you enhance the structure o the
3
work early due to amily illness) can also compromise the arrival team or the schedule such that the receiver shares more a priori
phase o a hando . Systems issues can also contribute to arrival knowledge with the sender?
C
ailure, with shi ts that start and end at the same time, requiring In addition to ailures on the part o senders and receivers,
a
r
e
either one individual stay late or one arrive early, in order to proin ormation transmission can be hampered due to noisy, distract-
T
vide adequate time to e ectively transition patients. This is espe- ing settings that discourage conversation, the hierarchal nature
r
a
cially likely i the hando does not take explicit priority over other o medicine (which can discourage open discussion between
n
s
clinical tasks. For example, the sender could be ready to arrive to providers), language barriers, lack o ace-to- ace communication,
i
t
i
o
the hando but the receiver could be in the operating room. For and time pressures that lead to a hurried dialogue. Using this
n
s
service changes, this is especially problematic as residents and ac- ramework, service trans ers could be especially prone to commu-
w
ulty are o ten transitioning rom other rotations, such as vacations nication ailures due to dialogue ailures i the services are rom di -
i
t
or elective time, and so may not be physically present in the same erent disciplines (eg, emergency medicine and hospital medicine)
h
i
n
city, or have access to the EHR or even e-mail. Identi ying a time since the sender and receiver may have very di erent expectations
t
to turn over a service also requires that the individual handing o o the level o detail or type o content to be reviewed. In this set-
h
e
is aware o the identity o their replacement, and in some systems ting, the lack o a “shared mental model” or the hando is missing,
H
this can pose a challenge. and there ore it is likely that the hando may be unsatis actory to
o
s
When the trans er o content is separated in time and space rom one or both parties.
p
i
t
the trans er o pro essional responsibility, arrival ailures can o ten
a
l
:
arise. For example, during service trans ers, such as rom the ICU to
T
■ POST-HANDOFF FAILURES
h
the oor, it is possible that due to timing o bed availability a di er-
e
ent team may receive the patient than originally received the trans- Post-hando ailures can take many orms, and one o the most
H
common is the diminished pro essional responsibility toward
a
er hando . This is o ten exacerbated during times o bed shortages,
n
since days may elapse rom when a trans er is initiated to when a patients o receivers who are “just covering the patient.” One key
d
-
question is, “does the incoming clinician have the same invest-
O
patient actually receives a bed.
ment in the patient care as the outgoing clinician?” It is o ten very
clear when this is not the case. For example, night moonlighters
■ DIALOGUE FAILURES or nocturnists may adopt an attitude that the patients they are
Similar to arrival, the dialogue phase o hando s could result in caring or are “not their patients” and that their job is just to hold
ailure on the part o either the sender or the receiver. For example, down the ort until the day team arrives. In these instances, the
the sender could provide disorganized in ormation; use vague or de ault could be to do as little workup as possible or any acute
unclear language; or ail to provide enough clinical background to patient issues and de er to the primary team. While in many
enable uture decision making. On the other hand, receivers could cases this “temporizing approach” may not result in any harm to
ail to listen due to either inattention or external distractions. They a patient, there are clearly time when delays in clinical decisions
could also misunderstand the in ormation or ail to clari y any items could harm the patient. For example, a patient who is meeting
they misunderstood through the use o questions. Both senders and sepsis criteria may not receive antibiotics because the night physi-
receivers have a responsibility to ocus on the process, and are both cian could be de erring the choice to the primary hospitalist. Like-
critical actors in the trans er o in ormation. wise, in service changes, it is also possible that the physician who
Data rom routine studies o human communication suggest is leaving the service may no longer be invested in the ongoing
that senders o ten overestimate how well receivers will under- care o the patient or patients, especially those who have been
stand the in ormation that they are trans erring. Interestingly, this passed rom physician to physician in multiple service changes.
worsens the more amiliar two people are with one another. This Post-hando ailures in service trans ers could mani est in the
“egocentric heuristic” can lead to communication errors due to receiving service never “learning the patient” or expediting dis-
the use o vague language. For example, a husband may tell a charge or a patient without ully addressing their issues because
wi e, “Meet me there a ter work” but not clari y where “there” is o a relative lack o pro essional responsibility.
or whether he means a ter her workday ends or a ter his workday
ends. In his mind, he understood what he was trying to say, but
he did not e ectively communicate it. This same problem applies STRATEGIES FOR IMPROVEMENT
in hando s. A study o pediatric hando s in optimal conditions In the e ort to improve in-hospital hando s, The Joint Commission
(dedicated room, time, limited interruptions) demonstrates that made standardized hando communications the subject o a 2006
60% o the time, the most important piece o in ormation about a national patient sa ety goal requiring institutions to “implement a
patient was not communicated despite the sender believing it had standardized approach to hando communication, including an
been. In addition, the rationale or to-do actions was o ten not pro- opportunity to ask and respond to questions.” Critical elements
vided. A common example reported was that the covering intern in this standardized model should include an interactive, timely
was told to “check the CBC (complete blood count)” but not given process that contains up-to-date in ormation with minimal inter-
any reason or doing so or what to do with abnormal results. This ruptions. Evidence or these goals emerges rom the experience o
study occurred with optimal physical conditions and a dedicated other industries, trials o technological solutions, or communication
time or the hando process, and still communication ailures and practices in health care.
87
■ STANDARDIZED OR STRUCTURED TEMPLATES FOR that “anticipated events are clearly labeled” and “tasks to be done
WRITTEN HANDOFF INFORMATION are highlighted” or incoming hospitalists. Avoiding editorializing
The importance o the implementation o a standardized strategy and extraneous in ormation allows the hando to be a streamlined
ow o clinically use ul in ormation that is pertinent or the covering
P
is critical or both the verbal and the written component o the
A
intrahospital hando . The use o standardized language during (receiving) physician and allows the sender to create a picture or
R
the verbal hando helps to ensure transmission o consistent their shared mental model.
T
in ormation and allows or interactivity in the hando . One popular
I
model is the Situation Brie ng model (SBAR), which is a technique ■ USE OF READ-BACK
that originated in the U.S. Navy to ensure the relay o critical Read-back allows the physician receiving the hando to check the
in ormation. SBAR has been used success ully by allied health in ormation received rom the sender. The use o read-back is also
pro essionals, such as in nursing. Other mnemonics that have a Joint Commission requirement or receipt o critical lab tests. Use
T
h
been used include SIGNOUT?, ANTICipate, HANDOFF, and IPASS. A o read-back has been shown to reduce the number o laboratory
e
prior systematic review o hando mnemonics yielded 46 articles reporting errors during requested read-back o lab results. Although
S
p
detailing 24 hando mnemonics; ew were evaluated or validated per orming a read-back o the entire verbal hando could be
e
c
in research settings. cumbersome and undesirable, the use o ocused read-back can
i
a
l
Several institutions have success ully used structured templates, enhance memory or the high-priority items o a verbal hando ,
t
y
such as computer-aided and electronic health record (EHR)-aided namely tasks to do, and to clari y anticipatory guidance as already
o
f
signouts, to ensure the transmission o accurate and updated in or- highlighted. One misconception about read-back is that it does not
H
mation in the written component o the hando . In act, Petersen need to occur or every patient on the hando , but can occur as a
o
s
and colleagues demonstrated a trend toward reduction o prevent- part o a synthesis or a ‘chunk’ o patients, or or an entire hando .
p
i
t
able adverse events a ter the implementation o a computerized
a
l
signout system over 20 years ago! Lee and colleagues demonstrated ■ IMPROVING PROFESSIONAL RESPONSIBILITY
M
in a randomized, controlled trial that a standard signout guide that
e
DURING HANDOFFS
d
ensured the inclusion o critical content resulted in improved writ-
i
While it is hard to imagine telling people to be more “pro essional”
c
i
ten signout quality. However, using a standardized template or
n
during hando s, it is worth noting that some organizational cultures
e
communication does not mean it is updated and has the correct and systems do a better job o promoting this shared responsibility,
a
or most pertinent in ormation or the covering physician. Not only
n
or what has been described as “continuity-enhanced” hando s. For
d
is it critical to ensure that the in ormation included in the written example, systems in which the sender and receiver both meet the
S
y
signout is accurate, it is equally important to ensure adequate time patient rst hand since they unction on the same team are likely
s
t
to per orm this update. Technological solutions, when combined
e
more e ective at promoting common ground and pro essional
m
with systems changes, can yield the best results. responsibility than a system in which the receiver is a “hired gun”
s
o
that is “just covering.” While this type o system is not technically
f
■ FACE-TO-FACE VERBAL UPDATE WITH easible or essential or all patients, it may be bene cial or the sick-
C
a
INTERACTIVE QUESTIONING est patients, such as those that are in an ICU setting, who depend on
r
e
Studies o shi t Changes in other industries highlight that the use a high degree o primary knowledge regarding their hospital course.
o ace-to- ace (in person) verbal update with interactive question- Another model or improving pro essional responsibility during
ing is critical in conducting an e ective hando . Studies o health the post-hando period that alls short o having the sender and
care pro essionals demonstrate general agreement with this prin- receiver both “know the patient” is to ensure a structure o repeated
ciple. Moreover, ace-to- ace verbal update is o ten suggested as a interaction between the sender and receiver such that trust is built
recommendation or inpatient hando s. between both parties. This could be accomplished in multiple ways,
including a system that schedules hospitalists or blocks and avoids
having a new clinician each night or day. This could improve pro es-
PRACTICE POINT sional responsibility or a good hando and post-hando care since
both the sender and receiver will ace each other again in the uture
• •Major strategies to improve intrahospital hando s include
and have an opportunity to receive ollow-up or updates on any
standardized or structured templates, ace-to- ace verbal
clinical questions that were outstanding at the time o the initial
update with interactive questioning, an emphasis on
hando .
anticipatory guidance and tasks to be done, and use o
read-back.
EDUCATION, EVALUATION, AND SIMULATION
Given recent regulatory, accreditation and education requirements,
several educational curricula geared toward teaching trainees and
■ EMPHASIZE ANTICIPATORY GUIDANCE AND TASKS aculty e ective hando strategies have been piloted and evalu-
TO BE DONE ated. In act, educators have encouraged raming the approach to
Receivers o intrahospital hando s o ten state that they need only hando education as one would any other entrustable pro essional
the pertinent in ormation—what may happen and what to do activity, in a milestone-directed ashion tied to the measurement
about it. Un ortunately, the actual practice is o ten that the sender o competency. One large multisite study in pediatric teaching
provides too much in ormation or too little in ormation. As a result, hospitals tested the e ectiveness o the IPASS initiative, a hando
emphasis on these items can be especially help ul to hone receiver improvement “bundle” which includes a mnemonic to standard-
understanding o the patient. Indeed, one study shows that a ter ize oral and written hando s, hando communication training,
the receipt o an intrahospital hando , receivers are more likely patient sa ety culture training, and aculty development in hando
to remember “i /then” items or “to-do” items more than general observation and a sustainability campaign. The evaluation o this
knowledge items about a patient. Moreover, the Society o Hospital educational bundle was shown to decrease the rate o preventable
Medicine Hando s Task Force recommends that “insight on what to adverse events in the postintervention period, as well as signi cant
anticipate and what to do is the ocus o the verbal exchange” and reduction in error rates and a signi cant increase in the quality o
88
hando s. While this curriculum holds great promise, understanding continued provision o sa e care during these transitions, providers
what parts o the bundle worked most e ectively as well as why it should be aware o the types o transitions and the ways in which
C
did not work in all sites is important. Several other curricular bundles these transitions represent vulnerability or patients and their sa ety.
H
exist within the literature, based upon specialty ocus area and With this knowledge, employing strategies to ensure e ective
A
P
program need. In addition, longitudinal approach to hando educa- communication is critical to the delivery o sa e patient care during
T
tion, beginning in medical school, predicts improved per ormance transitions.
E
on standardized measures, and com ort with hando s as an early
R
trainee. However, the education needed is not just at the student SUGGESTED READINGS
1
or resident level. Faculty-based education is required in order to
3
ensure the ef cacy and quality o observed trainee hando s. Work Arora VM, Berhie S, Horwitz LI, Saatho M, Staisiunas P, Farnan JM.
done at the University o Chicago has demonstrated that the use Using standardized videos to validate a measure o hando quality:
C
o video-based scenarios in a aculty development program led to
a
the hando mini-clinical examination exercise. J Hosp Med. 2014;
r
e
valid reliable rating instrument, the mini Hando CEX. 9(7):441-446.
T
While several validated instruments exist or measuring in vivo
r
Arora VM, Manjarrez E, Dressler DD, et al. Hospitalist hando s: a
a
hando behaviors, the development o standardized, objectives
n
systematic review and task orce recommendations. J Hosp Med.
s
clinical encounters (OSCEs) to measure hando quality, are only
i
t
2009;4(7):433-440.
i
o
beginning to be recognized as a method o evaluation. Prior work
n
Chang VY, Arora VM, Lev-Ari S, D’Arcy M, Keysar B. Interns overesti-
s
has demonstrated that a simulation-based exercise or students, in
w
which a mock, or standardized hando was per ormed, improved mate the e ectiveness o their hand-o communication. Pediatrics.
i
t
con dence and sel -ef cacy when per ormed at the transitional 2010;125:491-496.
h
i
n
time period in the ourth year. Cheung DS, Kelly JJ, Beach C, et al. Improving hando s in the
t
Finally, although most educational and evaluation modules ocus
h
emergency department. Ann Emerg Med. 2010;55(2):171-180.
e
on the communication skills o the sender, we are just now realizing Greenstein EA, Arora VM, Staisiunas PG, Banerjee SS, Farnan JM.
H
the need or training on how to be an e ective receiver. As we have
o
Characterising physician listening behaviour during hospitalist
s
discussed, hando s are o ten plagued by interruptions, both clinical
p
hando s using the HEAR checklist. BMJ Qual Sa . 2013;22(3):203-209.
i
t
and otherwise, as well as reluctance to question colleagues, and
a
Hinami K, Farnan JM, Meltzer DO, et al. Understanding communication
l
:
competing demands on our attention. Work done within hospital
T
during hospitalist service changes: a mixed methods study. J Hosp
h
medicine has demonstrated that the longer the hando , and the
e
Med. 2009;4(9):535-540.
more patients signed over, the higher the likelihood o interrup-
H
Riesenberg LA, Leitzsch J, Little BW. Systematic review o hando
a
tion. In addition, this work has generated the HEAR checklist, an
n
observation-based instrument which allows or the evaluation o the mnemonics literature. Am J Med Qual. 2009;24(3):196-204.
d
-
O
receivers listening behaviors. Starmer AJ, Spector ND, Srivastava R, et al. Changes in medical
errors a ter implementation o a hando program. N Engl J Med.
CONCLUSION 2014;371:1803-1812.
Ensuring sa e and e ective hando s is critical to patient sa ety and UHC Best Practice Recommendation: Patient Hand O Commu-
the delivery o quality care. These hando s occur during times nication White Paper. University Health System Consortium;
o patient care transition in the hospital setting. To ensure the May 2006.
89
CHAP TER
14 INTRODUCTION
Hospital discharge is a critically important care transition. Due to
the complexity and potential or errors inherent in the discharge
process, this care transition continues to be an area o ocus or
many patient sa ety organizations, regulatory agencies, and quality
improvement initiatives.
The discharge transition represents a vulnerable time or patients
or several reasons. There is a shi t o responsibility rom the inpa-
Care Transitions at tient care team to the outpatient or postacute care providers, and
with that comes great risk or breakdown in communication. Medi-
cations and other treatment plans are o ten adjusted in the hospital,
Hospital Discharge and patients and caregivers are challenged with new sel -care tasks
and ollow-up responsibilities at hospital discharge. Adverse out-
comes are common in the postdischarge period, with studies show-
ing that about one-hal o patients experience a medical error and
Kelly Cunningham Sponsler, MD, SFHM approximately one in ve patients su er an adverse event. These
adverse events, which include adverse drug events and increased
health care utilization such as unscheduled hospital readmissions
and emergency department visits, are o ten judged to have been
preventable. Table 14-1 outlines some o the patient, clinician, and
system actors that contribute to unsuccess ul discharge transitions.
RISK STRATIFICATION
Due to the complexity o care transitions, all patients are potentially
at risk or an unsuccess ul hospital discharge. However, certain patient
populations may be at higher risk than others. Most studies o dis-
charge interventions have targeted geriatric patients or patients with
speci c disease processes, such as congestive heart ailure, as these
populations are known to have high rates o hospital readmission,
upward o 20% within 30 days. Other patient-speci c characteristics,
such as low health literacy, low socioeconomic status, and psychiatric
comorbidity are also associated with worse outcomes a ter discharge.
While it is important to have standardized care processes in place that
acilitate sa e transitions or all types o patients, it is advantageous to
have strategies or identi ying patients who may bene t rom more
intensive care transitions interventions.
There have been several tools and models developed to identi y
patients who are at highest risk or readmission. However, these
models are not able to ully and reliably predict hospital readmis-
sion, and many incorporate administrative data that can be burden-
some to collect. Thus, having a process to easily ag patients with
certain high-risk disease processes, psychosocial actors, and/or
requent health care utilization, is optimal.
A ew o the more commonly used scoring systems are listed in
Table 14-2. The LACE index incorporates hospital length o stay, acu-
ity o the admission, comorbidity measured with the Charlson score,
and emergency department visits in the previous 6 months to predict
the rate o readmission or death within 30 days. The 8P risk scale, part
o Project BOOST (Better Outcomes by Optimizing Sa e Transitions),
accounts or prior hospitalizations, problem medications, psychiatric
problems, principal diagnosis, polypharmacy, poor health literacy,
patient support, and unmet palliative needs. The HOSPITAL score
was more recently published and was designed to predict potentially
avoidable readmissions. Variables include discharge hemoglobin
<12 g/dL, oncology as the discharging service, discharge sodium level
<135 mEq/L, per ormance o any procedure during index hospitaliza-
tion, being admitted on a nonelective basis, number o admissions in
the last 12 months, and hospital length o stay ≥5 days.
90
PRACTICE POINT
TABLE 14-1 Factors Contributing to Unsuccessful Discharge
C
Transitions • There are many patient characteristics and comorbidities
H
that place patients at increased risk or postdischarge adverse
A
• Premature discharge
events, including readmission. Several tools are available to
P
• Inappropriate discharge setting
T
acilitate identi cation o high-risk patients who may bene t
E
• Unrecognized medical, unctional, social needs rom more comprehensive support during the discharge
R
• Speci ic clinical conditions transition.
1
Congestive heart ailure
4
■
■ Psychiatric comorbidities
Poor social support A STANDARDIZED APPROACH TO DISCHARGE PLANNING
C
•
a
Hospital-based providers typically care or a heterogeneous patient
r
• Low health literacy
e
population characterized by a wide range o medical and psychoso-
T
• Inadequate hando s
r
cial needs. As such, it is essential to have a standardized approach
a
■ Pending tests
n
to discharge planning that starts early during hospitalization and
s
i
■ Additional work-up
t
involves collaboration with a multidisciplinary team. Key elements
i
o
■ Incomplete or unreceived discharge summary
n
o discharge care coordination include identi cation o the appro-
s
• Delayed or unscheduled ollow-up priate discharge destination, proactive scheduling o ollow-up
a
t
Lack o advanced care planning appointments, care ul medication reconciliation, and engagement
H
•
o
o patients and caregivers.
• Failure to ensure comprehension
s
p
i
■ Disease-speci ic education
t
■ CHOOSING A DISCHARGE DESTINATION
a
l
■ Discharge instructions
D
Discharge planning should begin at admission and continue
i
Medication-related problems
s
•
throughout hospitalization in parallel to the medical evaluation and
c
h
■ Adverse drug events
treatment plan. Once hospitalized patients are medically ready or
a
r
g
■ Failure to obtain necessary medications discharge, they may be transitioned to a number o locations and
e
■ Therapeutic duplication levels o care. Discharge to a site that cannot meet a patient’s medi-
■ Poor adherence
cal, unctional and social needs can lead to adverse events, so it is
important to identi y these needs and match them to available ser-
• Preventable complications rom hospitalization
vices and support. The most common discharge locations include
home with or without caregivers, home with home health services,
inpatient rehabilitation acilities, skilled nursing acilities, long-term
While these existing tools are not per ect, they may help providers acute care hospitals, and extended care acilities. Patients may also
identi y many o the highest risk patients and could serve as a trigger be transitioned to hospice care that can be provided in the home,
or more robust discharge planning interventions. Medical judg- in nursing acilities, or in dedicated inpatient palliative care units.
ment, or “gut check,” remains a critical part o the risk assessment The Centers or Medicare and Medicaid Services categorize health
and should supplement any objective scores. care acilities into di erent groups based on the acuity and intensity
TABLE 14-2 Select Tools for Assessing Readmission Risk
Name Elements Comments

LACE index • Length o stay Predicts rate o unplanned readmission
• Acuity o admission or death within 30 d; scores range rom
0-19 with a air discriminative index
• Comorbidity (Charlson score) (C statistic 0.684)
• Emergency department visits in the previous 6 mo
The 8P screening tool • Problems with medications (polypharmacy, high-risk medications) From the Project BOOST toolkit; has not
• Psychological (depression) been rigorously validated
• Principal diagnosis (cancer, stroke, DM, COPD, heart ailure)
• Poor health literacy
• Patient support (absent or insu icient)
• Prior hospitalization (in last 6 mo)
• Palliative care (advanced illness, li e expectancy < 1 y)
HOSPITAL • Hemoglobin < 12 g/dL at discharge Predicts 30-d potentially avoidable
• Oncology as the discharging service readmissions with air discriminatory
power (C statistic 0.71)
• Sodium level < 135 mEq/L at discharge
• Procedure per ormed during hospitalization
• Index admission type (nonelective)
• Number o admissions in last 12 mo
• Length o stay ≥ 5 d
91
TABLE 14-3 Available Services and Requirements for Common Postacute Care Sites
Specific Certification or Nursing Physician Rehabilitation Diagnostic, Ancillary

P
Medicare Requirements Services Services Services Services
A
Skilled nursing “Quali ying event” o 3-night 2-4 h per patient Physician visit Physical, O -site laboratory
R
facilities (SNF) inpatient stay; skilled needs per day required every occupational, and radiology, limited
T
>1 h per day, 5 d per week; 30 d; o ten utilize speech therapy, ability to manage
I
initial physician visit required nonphysician approximately unstable patients
within 30 d o admission to providers or 1 h per day
acility medically necessary
visits
T
h
Inpatient 75% o patients all into 13 5-6 h per patient Face to ace visits Multimodal Lab and radiology
e
rehabilitation diagnosis categories; require per day by a rehabilitation services, at least available, some ability
S
p
facilities (IRF) multidisciplinary therapy; physician at least 3 h per day to handle unstable
e
>3 h o therapy per day, 3 times per week patients
c
i
a
5 d per week
l
t
y
Long term Average length o stay 5-6 h per patient Daily or near-daily Multimodal Lab and radiology
o
acute care ≥25 d; highly complex per day physician visits; services available, some ability
f
H
hospitals (LTAC) medical patients (ventilator consultant specialists to handle unstable
o
management, complex widely available patients
s
p
wound care)
i
t
a
Extended care Long term custodial care; <2 h per patient Physician visits Physical, O -site laboratory
l
M
facilities (ECF) reimbursement through per day every 30 d occupational, and radiology, limited
e
Medicaid speech, ability to manage
d
i
recreational unstable patients
c
i
n
therapy
e
Home health Medicare requires Face to Examples: wound Requires a physician Physical, N/A
a
n
Face Encounter orm and care, IVtherapy, (usually primary occupational,
d
physician certi ication o medication and care) to oversee plan speech therapy
S
y
homebound status disease education o care
s
t
e
m
s
o
o care that can be provided, and numerous ederal regulations exist team members to discuss disposition regularly, such as through daily
f
C
regarding the required services and patient eligibility or coverage o huddles or collaborative rounds, will help acilitate discharge planning.
a
r
these services. Table 14-3 provides a brie overview o the certi ca- Implementing a structured ormat to these discussions and clearly
e
tion criteria and types o services available. There are two Medicare identi ying roles and responsibilities will oster communication among
requirements o which inpatient physicians should be particularly team members and urther streamline the discharge transition.
aware. First is the need or a “quali ying event” o a preceding 3-night
inpatient hospitalization or Medicare patients who are re erred
to skilled nursing acilities. Second, or Medicare patients who are PRACTICE POINT
re erred or home health services, documentation o a ace-to- ace • Utilizing the expertise o a multidisciplinary team, such as
physician (or nonphysician provider) encounter must be provided, social workers, case managers, and physical and occupational
along with evidence that the patient is homebound and has skilled therapists, coupled with patient pre erences, goals o care and
needs that can be met on an intermittent, rather than continuous, skilled needs, will acilitate timely re erral to the appropriate
basis. See Section III: Rehabilitation and Skilled Nursing Care. level o postdischarge care.
Selecting the appropriate discharge disposition will o ten require
input rom several members o the care team, as well as consid-
eration o patient and amily pre erences. Involving physical and
occupational therapy is imperative or those patients who have ■ SCHEDULING FOLLOW-UP APPOINTMENTS
unctional de cits at baseline or who have experienced debil- Discharged patients typically have ongoing medical issues and
ity due to acute illness. Physical and occupational therapists can require continued examination, medication adjustment and
evaluate patients’ mobility and ability to per orm sel -care and will reassessment o their treatment plan to ensure that they continue
provide expert recommendations regarding the need or skilled to recover rom the acute illness that precipitated hospitalization.
services and equipment postdischarge. Thus, timely ollow-up with a primary care provider (or a specialist
Social workers and case managers are also important members o or acility-based physician, i appropriate) is essential in order to pro-
the care team and will assist with identi ying and coordinating the vide ongoing evaluation and management o medical conditions,
most appropriate discharge destination. They are skilled at assessing to identi y barriers to recovery, and to troubleshoot problems that
patients’ support systems and identi ying social barriers to a suc- arise a ter discharge. Studies have shown that lack o ollow-up is
cess ul discharge. They help clinicians and patients understand what associated with hospital readmission, though the evidence is mixed
services will be covered by the patient’s insurance. regarding the degree to which ollow-up impacts patient outcomes.
Other members o the care team may also provide insight into In a large analysis o Medicare claims data, 50% o the patients who
patients’ postdischarge care needs, including but not limited to were rehospitalized within 30 days a ter discharge to the commu-
nurses, pharmacists, respiratory therapists, geriatrics and palliative care nity had no associated bill or an outpatient visit during that time
consultants, and psychiatrists. Having shared understanding regarding rame, suggesting an opportunity to improve the timeliness and
patients’ estimated day o discharge and developing a mechanism or reliability o posthospital ollow-up.
92
Identi cation o a patient’s primary care physician should be medication reconciliation is associated with reductions in medica-
done at the time o hospital admission. Although the inpatient tion discrepancies and adverse drug events; there is less evidence
C
team should con rm that each patient has a ollow-up provider regarding the role o medication reconciliation in decreasing read-
H
and communicate with this clinician as medically appropriate, the
A
missions and emergency department visits. Existing literature sug-
P
actual task o identi ying and documenting the name and contact gests that involving clinical pharmacists in medication reconciliation
T
in ormation or the provider may be assigned to administrative per- and targeting high risk patients, or example, those taking more than
E
sonnel. Equally important is early identi cation o patients who do 10 medications or high alert medications such as anticoagulants
R
not have a primary care physician. Institutions and hospital-based and insulin, are associated with more avorable outcomes.
1
providers should have a process or assigning primary care providers Medication reconciliation is also an opportunity to evaluate poly-
4
to patients who are not already established with outpatient care. pharmacy, screen or high alert drugs and potentially inappropriate
For patients who are transitioning to acilities, a medical director medications, and identi y drug-drug or drug-disease interactions,
C
(and o ten other acility-based physicians and physician extenders) all o which can contribute to postdischarge adverse drug events.
a
r
e
will oversee general acute and chronic medical needs in the postdis- Clinicians should assess medication adherence, address barriers
T
charge period. The medical director will typically serve as a primary to adherence, such as inability to a ord medications or complex
r
a
care physician or patients residing in long-term care settings, but it dosing regimens, and communicate the updated medication list
n
s
is important to recognize that patients who are receiving short-term to the patient and ollow-up provider. The medication list should
i
t
i
o
postacute care in skilled nursing acilities or inpatient rehabilitation include explicit notation o which medications have been added,
n
s
will still need to have a primary care provider identi ed or later discontinued, or changed during hospitalization to minimize the
a
ollow-up needs. Outpatient ollow-up care with specialists may also potential or con usion and patient harm. See Chapter 9: Principles
t
H
be indicated. o Evidence-Based Prescribing and Chapter 73: Patient Sa ety and
o
It is recommended that ollow-up appointments be made by Quality Improvement.
s
p
the hospital-based team be ore the patient leaves the hospital, so
i
t
a
that there is clarity among care team members and the patient
l
■ PATIENT ENGAGEMENT AND EDUCATION
D
regarding the ollow-up plan. This is pre erred over delegating the
i
s
Patient and amily engagement is a undamental component o the
c
responsibility to patients to schedule their own appointments a ter
h
discharge. Case managers may be most knowledgeable about discharge planning process and may help bridge the discontinuity
a
r
inherent between inpatient and outpatient settings. Many actors
g
community resources that provide medical care or uninsured or
e
underinsured patients. contribute to observations that patients leave the hospital unpre-
Several actors impact timing o ollow-up appointments, includ- pared, including both provider barriers (time constraints, competing
ing severity o the patient’s acute illness, pre-existing comorbidities, priorities, ailure to utilize patient- riendly language or to con rm
the patient’s ability to manage medications and sel -care needs, comprehension o in ormation delivered) and patient barriers (low
social issues such as transportation and caregiver support, and phy- health literacy, cognitive impairment, reluctance to ask questions).
sician availability. For most patients, postdischarge ollow-up within Evidence suggests that engaged patients have higher levels
7 to 14 days is reasonable, provided they are given suf cient instruc- o satis action and improved health outcomes. Thus, health care
tions at discharge regarding “red ag” symptoms and who to con- organizations such as the Institute o Medicine, Agency or Health-
tact with questions and concerns. For patients who are at high risk care Research and Quality, National Quality Forum, and the Joint
or readmission and adverse drug events, ollow-up as early as 48 to Commission prioritize patient engagement and the principles o
72 hours ollowing discharge may be pre erable. Follow-up phone patient-centered care and shared decision making. Initiatives have
calls and home visits may be utilized, especially or high risk patients. ocused on encouraging patients to ask questions o their health
care providers, enabling patient access to medical in ormation,
supporting communication with care providers, and acilitating sel -
PRACTICE POINT management o illnesses. As patients and their caregivers may be
the only continuity across care settings, it is particularly important
• Patients need timely ollow-up with a primary care provider
to rein orce these behaviors during the vulnerable periods o care
ollowing hospital discharge, with high risk patients being seen
transitions.
earlier, ie, within 7 days o discharge. Appointments should be
There are several approaches that may help patients understand
made be ore the patient leaves the hospital to ensure access to
their disease processes, treatment plans and discharge instructions.
ollow-up care.
Patients recall and comprehend only about hal o the in ormation
provided in a medical encounter; thus, patient education should
occur in small sessions throughout hospitalization and main points
■ MEDICATION RECONCILIATION should be reiterated. Health in ormation technology applications
Medication errors are common during transitions o care, with hos- are being increasingly incorporated. Patients should be encour-
pital discharge being particularly hazardous. During hospitalization, aged to ask questions and should be provided with disease-speci c
preadmission medications are o ten changed or discontinued, and education materials that are easily understandable with regard to
new agents added. Fi ty percent o patients experience a clinically language, ont and reading level.
signi cant medication error, ie, and adverse drug event or a medi- Discharge instructions may be con using, especially when they
cation discrepancy with the potential to cause harm, in the 30 days contain excess content and are given just as the patient is leav-
ollowing hospital discharge. ing the hospital. Steps should be taken to create patient-centered
Medication reconciliation, the process by which a patient’s instructions that are clear, tailored to patient language and literacy,
medication list is obtained, compared, and clari ed across di erent and ocused on critical details o sel -management. Patients should
sites o care, is a strategy or decreasing medication errors during be given the ollowing in ormation:
transitions. Medication reconciliation is a care process supported • Reason or hospitalization, treatment received, names o clini-
by national and international organizations including the Joint cians involved in care i questions arise postdischarge.
Commission, the Institute or Healthcare Improvement and the • Pertinent test results as well as pending test results.
World Health Organization. Studies have consistently shown that • Diet and activity.
93
• Medications, including any changes in regimen and potential
side e ects. TABLE 14-4 Recommended Components of the Discharge
• Follow-up appointments. Summary
• Identi cation o the person to contact with questions or
P
• Primary and secondary diagnoses
A
concerns.
• Important test results
R
• List o concerning symptoms and how to respond.
T
• Pending results and responsible party
Patient education and discharge instructions must be reviewed
I
with the goal o ensuring comprehension o the material. Teach • Recommendations regarding additional work-up or
treatment plan
back is one method or assessing patient understanding, whereby
the provider relays the material then asks the patient to explain the • Patient’s condition at discharge (including cognitive and
concept or directions in his own words. This approach allows the unctional status and abnormal exam indings)
T
h
provider to identi y any misunderstandings and address them • Complete list o reconciled medications
e
be ore the patient leaves the hospital. Follow-up arrangements
S
•
p
e
• Identi ication and contact in ormation or the sending and
c
receiving providers
i
a
PRACTICE POINT
l
t
• Resuscitation status
y
• Discharge instructions should highlight in ormation the patient
o
• Documentation o patient education
f
needs to understand and manage medical conditions a ter
H
o
discharge and should be con rmed using teach back.
s
p
i
t
a
communication; mail may lead to delays in in ormation trans er.
l
M
With electronic medical records becoming more widely imple-
THE DISCHARGE HANDOFF
e
mented, the ability to leverage health in ormation technology has
d
A systematic review o studies investigating communication
i
the potential to improve the accuracy and ef ciency with which
c
i
between hospital-based and primary care physicians revealed that
n
discharge summaries are generated and communicated.
e
direct communication occurred in requently (3%-20% o hospital
a
n
discharges) and that the availability o the discharge summary was
d
low (12%-34% at the time o the rst postdischarge appointment). PRACTICE POINT
S
y
The timeliness, accuracy, completeness, and quality o the discharge
•
s
The discharge summary is the primary method o
t
hando can have signi cant impact on patient care, and as such, is
e
communication between the hospital care team and the
m
an important part o the discharge transition.
ollow-up provider. To ensure e ective communication,
s
o
the discharge summary content and ormat should be
f
■ DISCHARGE SUMMARY
C
standardized, and reliable processes should be developed to
a
ensure timely generation o the summary and prompt trans er
r
The hallmark o the discharge hando is the discharge summary,
e
which is a written transition record documenting a patient’s diag- to the receiving provider.
nostic ndings, hospital management, and postdischarge ollow-up
arrangements. This document is requently the only orm o commu-
nication that occurs between the inpatient team and the ollow-up ■ VERBAL COMMUNICATION
providers. Un ortunately, existing literature suggests that discharge
Verbal communication is considered an important component
summaries requently lack important in ormation, including diagnos-
o hospital-based hando s, however it is uncommonly utilized at
tic test results, hospital course, pending tests at discharge, discharge
hospital discharge. A “warm hando ,” that is, direct communication
medications, ollow-up plans, and patient counseling.
between the discharging and receiving provider with an opportu-
Several organizations, including the Joint Commission and the
nity or the receiving provider to ask questions, may be a valuable
Society o Hospital Medicine, as well as expert panels, such as rep-
supplement to the written discharge summary in certain situations.
resentatives convened through the Transitions o Care Consensus
Verbal communication should ocus on anticipatory guidance and
Con erence, o er expectations and recommendations or discharge
tasks to be done. For a discharged patient, this conversation may
summary communication. At minimum, the discharge summary
emphasize pending tests, recommended outpatient work-up, or
should include diagnoses; important test results; pending results;
proposed medication adjustments during the ollow-up visit.
recommendations regarding additional work-up; patient’s condi-
tion at discharge; the reconciled medication list; ollow-up arrange-
ments; and identi cation and contact in ormation or the sending ■ MANAGEMENT OF PENDING TESTS
and receiving providers. Table 14-4 lists the suggested elements o Patients are commonly discharged rom the hospital with test
the discharge summary, including those which some may consider results still pending, and results that return a ter discharge some-
to be optional. times require a change in care plan. This presents a potential patient
The ormat, timeliness, and availability o the discharge summary sa ety issue, as ailure to ollow-up these results may lead to delays
are just as integral as the content. The use o structured templates in diagnosis or appropriate therapy, excess testing, and patient
with subheadings is recommended to acilitate inclusion o in or- harm. Approximately one-third o patients have pending laboratory
mation that will be valuable to ollow-up providers, organized in a results, most requently microbiology such as blood cultures that
way that highlights the in ormation most pertinent to posthospital have not been nalized, and re erence laboratory tests. As previ-
care. Prompt completion o the discharge summary, pre erably on ously mentioned, discharge summaries o ten lack this in ormation
the day o discharge, is necessary to ensure that ollow-up clinicians and the critical guidance regarding ollow-up. Moreover, studies
and other care team members are aware o the hospital course have shown that physicians are o ten unaware o test results return-
and treatment plan. The summary should then be orwarded to ing a ter discharge. In one study o patients discharged rom the
the appropriate ollow-up providers via ax or secure electronic hospitalist services o two academic hospitals, 41% had pending
94
radiology or laboratory results at discharge, and nearly one in 10 o summary, and creation o a written a ter-hospital care plan. These
these results were judged to be actionable. In subsequent survey 10 in-hospital interventions are supplemented by a pharmacist post-
C
o the inpatient or primary care physicians or these patients with discharge telephone component to rein orce the discharge plan,
H
actionable results, physicians had been unaware o 61% o results.
A
review medications and address questions and concerns. In a ran-
P
Electronic systems may be able to automate inclusion o pending domized controlled trial o general medicine patients, the group
T
tests in the discharge summary, but explicit designation o the respon- who received the RED process had a lower rate o 30-day combined
E
sible party is needed, along with a way to acknowledge and address the emergency department visits and hospital readmissions. Because
R
results. A ew studies have investigated systems designed to improve the elements were bundled, the investigators were unable to clearly
1
noti cation o pending results, but more work is needed. determine which components contributed to the reduction in hospital
4
utilization, and to what degree. Many institutions have subsequently
modeled their discharge programs based on Project RED.
C
PRACTICE POINT
a
r
e
• Test results that are pending at hospital discharge present ■ CARE TRANSITIONS INTERVENTION
T
r
a patient sa ety concern, as a subset o these results require
a
The Care Transitions Intervention is a patient-centered program
n
action. Pending results should be identi ed by the discharging
s
which is designed to empower patients and their caregivers to take
i
provider and communicated to the receiving provider via
t
i
an active role in the discharge transition. This bundled intervention
o
discharge summary as an explicit part o the discharge hando .
n
utilizes a series o tools and a nurse transition coach and consists o
s
a
both in-hospital and postdischarge components. The our core ele-
t
ments, re erred to as pillars, include medication sel -management
H
o
BUNDLED INTERVENTIONS skills, creation o a patient-owned personal health record, scheduling
s
p
and completion o timely ollow-up, and knowledge o “red ag” indi-
i
Patient sa ety, a health care climate where hospitalized patients are
t
a
cations that suggest a worsening condition, with an accompanying
l
being discharged “quicker and sicker,” and increasing accountability
D
response plan. These pillars are realized through the coaching o an
o institutions and providers or high quality, cost-conscious care have
i
s
advanced practice nurse transition coach who meets with the patient
c
prompted interventions that may improve the sa ety o the discharge
h
in the hospital, during a ollow-up home visit, and through a series o
a
transition. Many studies o discharge interventions have ocused on
r
ollow-up phone calls. In a randomized controlled trial o community-
g
parts o the discharge process, such as medication reconciliation
e
dwelling, cognitively intact patients aged 65 years and older who were
or ollow-up appointments. However, there are a ew well-known
admitted to a hospital in Colorado, Coleman and colleagues ound
discharge programs that have been shown to acilitate smooth transi-
that intervention patients had lower readmission rates at 30 days and
tions and decrease readmissions. Several o the most well-known and
at 90 days compared to patients receiving usual care.
success ul have consisted o multi aceted interventions that include
both in-hospital and postdischarge elements. Facilitation by a dedi-
■ TRANSITIONAL CARE MODEL
cated transition provider, usually a trained nurse, is another compo-
nent common to many success ul interventions. The transitional care model is an advanced practice nurse-directed
discharge planning and home ollow-up intervention which has been
well-studied in a number o patient populations. Developed by Naylor
■ PROJECT BOOST and colleagues at the University o Pennsylvania, this comprehensive
The Better Outcomes by Optimizing Sa e Transitions program, also intervention relies heavily on the advance practice nurse to per orm
known as Project BOOST, was developed by a panel o nationally in-hospital assessments and collaborate with care team members to
recognized experts in care transitions, hospital medicine, and qual- design an individualized discharge plan or each patient. Addition-
ity and patient sa ety, with input rom key stakeholders including ally, the nurse supports clinical care, provides patient and caregiver
payers, regulatory agencies, pro essional societies and patient education, and coordinates home services and supplies. Following
advocates. BOOST encompasses a comprehensive discharge inter- discharge, the nurse conducts a series o home visits and is available
vention, supplemented by a step-by-step implementation guide by phone 7 days per week to address medication-related problems,
and toolkit. These materials are designed to provide a ramework symptom management, ollow-up appointments, and other indi-
or quality improvement (eg, orming a team, setting aims, de ning vidual patient needs. Randomized controlled studies o this model in
measurement strategies) while supporting evidence-based inter- a geriatric medical and surgical population as well as in geriatric heart
ventions (eg, scripts or ollow-up telephone calls, training videos or ailure patients demonstrated reductions in hospital readmissions,
teach-back, discharge checklists). A unique aspect o Project BOOST decreased health care costs and longer periods o time between dis-
is its mentored implementation program that coaches participat- charge and rst readmission compared to routine care.
ing sites through the planning and implementation process and
enables them to tailor best practices to their local settings. Among DISCHARGES TO POSTACUTE CARE FACILITIES
an 11-site cohort o both academic and nonacademic hospitals that The majority o discharge transition studies examine interven-
participated in the mentored implementation program, hospital tions between hospital and home, but many o the best practices
units that implemented BOOST tools were ound to have reductions can also be applied to patients who discharge to post-acute care
in average rates o 30-day hospitalization. acilities. In general, the discharge process should emphasize the
importance o medication reconciliation, timely completion o
■ PROJECT RED and communication o discharge summaries with key in orma-
Reengineered discharge, also known as RED, is a discharge intervention, management o pending studies, and patient education.
tion that was developed and implemented at Boston Medical Center. However, since this patient population typically is older with
With this intervention, nurse discharge advocates work closely with more unctional impairment at baseline, there may be a greater
patients to deliver a 10-component discharge package, consisting o emphasis on advanced care planning with the patient and amily,
individualized patient education, coordination o ollow-up appoint- unctional status, and anticipatory guidance to the receiving pro-
ments, review o pending tests, reconciliation o medications, coun- vider. Physician Orders or Scope o Treatment (POST) orms, which
seling on postdischarge problems, transmission o the discharge are physician-signed order orms that communicate treatment
95
pre erences or patients with serious illnesses who are trans er- Forster AJ, Mur HJ, Peterson JF, Gandhi TK, Bates DW. The incidence
ring across care settings, are o ten completed when patients are and severity o adverse events a ecting patient a ter discharge
discharged to postacute care. A ull set o trans er orders also must rom the hospital. Ann Intern Med. 2003;138:161-167.
accompany patients who are being trans erred to acilities. When
P
Halasyamani L, Kripalani S, Coleman E, et al. Transition o care
A
writing these orders, discharging providers should be mind ul or hospitalized elderly patients—development o a discharge
R
that daily physician visits are not expected at most postacute care checklist or hospitalists. J Hosp Med. 2006;1:354-360.
T
acilities; there ore, physicians should be very speci c when com-
Jack BW, Chetty VK, Anthony D, et al. A reengineered hospital dis-
I
municating patient needs or wound care, laboratory ollow-up,
charge program to decrease rehospitalization. Ann Intern Med.
and medication titration. See Section III: Rehabilitation and Skilled
2009;150:178-187.
Nursing Care.
Kripalani S, Jackson AT, Schnipper JL, Coleman EA. Promoting e ec-
T
tive transitions o care at hospital discharge: a review o key issues
h
CONCLUSION
e
or hospitalists. J Hosp Med. 2007;2:314-323.
S
Hospital discharge is a complex process that can be associated with
p
Kripalani S, LeFevre F, Phillips CO, Williams MV, Basaviah P, Baker DW.
e
adverse events such as medication errors, missed diagnoses and
c
De cits in communication and in ormation trans er between hos-
i
hospital readmissions. The discharge process should be made sa er
a
pital-based and primary care physicians. JAMA. 2007;297:831-841.
l
t
by implementing standardized approaches to risk strati cation, dis-
y
charge planning, the discharge hando , and patient education. Naylor MD, Brooten DA, Campbell RL, Maislin G, McCauley KM,
o
f
Schwartz JS. Transitional care o older adults hospitalized with
H
o
heart ailure: a randomized, controlled trial. J Am Geriatr Soc.
SUGGESTED READINGS
s
p
2004;52:675-684.
i
t
a
Coleman EA, Parry C, Chalmers S, Min S. The Care Transitions Inter- Roy CL, Poon EG, Karson AS, et al. Patient sa ety concerns arising
l
M
vention: results o a randomized controlled trial. Arch Intern Med. rom test results that return a ter hospital discharge. Ann Intern
e
2006;166:1822-1828. Med. 2005;143:121-128.
d
i
c
i
n
e
a
n
d
S
y
s
t
e
m
s
o
f
C
a
r
e
96
SECTION 4
Patient Safety and Quality Improvement
97
CHAP TER
15 INTRODUCTION
In the last 15 years, Quality Improvement and Patient Sa ety have
emerged as major ocus areas or health care systems around the
world. The landmark 1999 report, To Err is Human, de ned Patient
Sa ety as reedom rom accidental medical injury, which is o ten
the result o error. Errors are de ned as ailures o execution or plan-
ning. Unplanned events that arise rom medical care, whether due
to human or systems-based errors, are urther classi ed into near-
misses or adverse medical events; a near-miss (or “close call”) is an
Principles of Patient event that causes no harm but had the potential to do so, while an
adverse medical event causes patient harm.
Safety: Intentional Identi ying adverse medical events as a source o human su er-
ing, the World Health Organization in 2002 recognized that the need
Design and Culture to improve Patient Sa ety as a undamental principle o all health
systems. The concept o Patient Sa ety o ers a positive spin on
the more emotionally laden concept o medical error. Traditionally
regarded as the result o incompetence, poor preparation or lack o
motivation, medical error is now understood as a product o poorly
David J. Lucier, Jr., MD, MBA, MPH designed systems o care that contribute to harm. The modern
Alexander R. Carbo, MD, SFHM view o medical error is that Patient Sa ety can be produced only in
organizations that take a systems-based approach to the problem,
Saul N. Weingart, MD, PhD recognizing the inherent limits o human per ormance and the
need to engineer the care delivery process in a way that is based on
scienti c principles. Nowhere is this issue more pressing than in the
acute care hospital.
■ DEFINING THE PROBLEM

Patient Sa ety emerged as a public health problem ollowing the
November 1999 release o To Err Is Human by the Institute o Medi-
cine (IOM). This report described the epidemic o medical errors in
the United States, accounting or as many as 98,000 unnecessary
deaths per year. The IOM report described an approach to under-
standing this problem that relied on developments in human ac-
tors engineering and cognitive psychology. By ocusing on methods
to diagnose and improve systems o care, the report pointed to a
novel approach or addressing this epidemic.
The IOM report provoked a broad response. The President o the
United States directed the Federal health care agencies to review
and implement the recommendations outlined in the report. The
predecessor o the Agency or Healthcare Research and Quality
(AHRQ) issued $50 million in research grants. Accreditation agen-
cies such as The Joint Commission and the Accreditation Council
or Graduate Medical Education developed standards and goals
related to Quality Improvement and Patient Sa ety that are required
o hospitals as well as residency and ellowship programs. A group
o Fortune 500 companies organized themselves into a consortium
called the Leap rog Group in order to encourage large businesses to
purchase health care rom organizations that met high standards or
Patient Sa ety. Advocacy groups such as the Institute or Healthcare
Improvement and the National Patient Sa ety Foundation created
campaigns and collaborative partnerships to spread Patient Sa ety-
related improvements. Local, regional, state, and national govern-
ment organizations banded together to develop and cooperate
on initiatives to reduce medical errors. In short, the To Err Is Human
report helped to crystallize a movement in the United States and
abroad that brought a new intensity o purpose to enhancing
Patient Sa ety and reducing medical errors.
99
■ HOSPITALIZED PATIENTS o age are particularly vulnerable, perhaps due to their reduced
Much o the early work on Patient Sa ety ocused on hospitalized physiological reserve. An error a ecting a sick, elderly person may
patients. This occurred or several reasons. Inpatients were judged to be more likely to result in injury than in a younger person with
ewer comorbidities. The same is true or young children. The need
P
be particularly vulnerable by virtue o their acute illness, comorbidities,
A
and the intensity o the interventions delivered. Hospitalized patients to calculate weight-based medication doses con ers on children an
R
were more accessible to investigators or study. Improvements that increased risk due to medication errors. Other patients at high risk
T
a ected the system o care were more readily developed and deployed include those undergoing neuro, thoracic, or vascular surgery. These
I
in the hospital compared to settings such as ambulatory care, with are inherently risky procedures and o ten per ormed on individuals
ewer centralized resources to support measurement and improve- with multiple or serious underlying comorbidities. Patients admitted
ment initiatives. While Patient Sa ety in ambulatory and chronic care urgently are at higher risk than elective admissions. In addition, the
settings is an area o increasing importance, inpatient acute care ser- number o interventions a patient experiences increases the oppor-
T
h
vices remain the mainstay o sa ety study and intervention. tunities or a mishap. In the Adverse Drug Event Prevention study,
e
Given the central role o the acute care hospital in e orts to study Bates and colleagues reported that the highest rates o adverse
S
p
and improve Patient Sa ety, hospitalists are particularly well posi- drug events were among patients in the medical intensive care
e
c
tioned to serve as Patient Sa ety champions in their organizations. unit. This was due to the greater number o medications and doses
i
a
these patients received. Medication-related errors and adverse
l
Hospitalists are directly responsible or health care delivery, giving
t
y
them rsthand knowledge about how errors and injuries occur. drug events are an area o special interest to researchers and prac-
o
titioners, since these events account or the greatest proportion o
f
They understand how current systems may play a role in contribut-
H
ing to harm. Most importantly, they are likely to have an in ormed adverse events among admissions to the medical service. Studies
o
s
perspective about the types and methods o improvement that are that examined adverse drug events among hospitalized patients
p
i
identi ed a consistent list o medications that account or a dispro-
t
both easible and e ective. Frontline provider involvement is a criti-
a
l
cal component o success ul improvement projects, and hospitalists portionate share o serious incidents: anticoagulants, antibiotics,
M
are well equipped to participate in a meaning ul way. chemotherapy agents, narcotics and sedatives, and insulin.
e
d
i
c
■ EMERGING AREAS OF RISK
i
n
SCOPE OF THE PROBLEM
e
Adverse drug events have been a particularly ruit ul area o work
a
in Patient Sa ety, resulting in the dissemination o improvements
n
■ EPIDEMIOLOGY OF MEDICAL ERROR
d
in electronic order-entry systems, pharmacy sa e practices, and
S
Medical error was long regarded as a rare phenomenon. In the 1980s guidelines or use o high-alert medications. Researchers have been
y
s
and 1990s, however, sentinel cases brought widespread attention to tackling the problem o diagnostic error as well, driven in part by the
t
e
this problem. Among the most widely publicized was the case o
m
prominence o missed or delayed cancer diagnoses and miscom-
Libby Zion, a young woman who died at a New York Hospital in
s
munication with outpatient providers. Research has ocused on the
o
1984 a ter she was prescribed meperidine and a monoamine oxi- development o methods to understand lapses in critical processes
f
C
dase inhibitor—a atal combination. Ten years later, Betsy Lehman, o care, such as communicating and interpreting critical test results,
a
a young mother and Boston Globe reporter, died o an accidental
r
and ensuring timely completion o re errals. Other thought leaders
e
chemotherapy overdose due to an ambiguous medication order. have ocused attention on how doctors think. Can we train clinicians
A series o subsequent studies ound that errors were common, to avoid premature closure o diagnostic options by maintaining a
especially among patients admitted to the hospital through the broad di erential diagnosis? How can we help them to avoid com-
emergency department. The rst large, epidemiologic study o mon mistakes, such as con rmation bias or premature conclusions?
medical errors was reported in the New England Journal o Medicine Hospitalists are well suited to study and address the risks asso-
in 1991. The Harvard Medical Practice study examined over 30,000 ciated with hando s and transitions o care. Hospital medicine
medical records o patients hospitalized in New York State in 1984. practice is ri e with opportunities to transition patient care to other
Investigators learned that 3.7% o patients had an adverse event, hospitalists at the end o the shi t or the week, to coordinate care
de ned as an injury due to medical care. These patients had serious with subspecialists and with colleagues in nursing and pharmacy,
adverse events, including those that extended the patient’s hospi- and to interact with the re erring community practitioner. Research
talization, or resulted in death or disability. These medical injuries shows that hospital discharge is a particularly vulnerable time or
resulted rom surgical and medical care at similar rates, though patients, and a time when errors may occur or a variety o reasons.
the events that occurred on the medical service were more o ten Failure to reconcile mediations at discharge may lead to con usion
judged to be preventable. Indeed, about one in our events was on the part o patients, reducing outpatient medication adherence
ound to be the result o negligence: care that ell below community rates or increasing the risk o adverse drug events, in turn increasing
standards o medical care. readmissions. Hando s to community physicians may ail to occur
The Medical Practice Study was an a ront to the concept that i the hospital discharge summary is delayed, incomplete, or i there
medical injuries are rare events. Although critics challenged the is no system in place to transmit in ormation e ciently. Recom-
results, the ndings have proven robust. Replications o the Medical mended tests and procedures ollowing discharge are o ten missed.
Practice Study in Colorado and Utah (in the United States), Canada, Promising approaches that hospitalists can utilize to address these
United Kingdom, Australia, Spain, and France all show substantially problems include standardization o hando s through the use o
similar results. Five to 25% o hospitalized patients experience an templated signout orms, electronic communication with re erring
adverse event due to medical care during their hospitalizations, and providers, and hospitalist-sta ed postdischarge clinics designed to
many are preventable. Researchers extrapolated the Colorado and support high-risk patients closely a ter discharge but be ore their
Utah study results to calculate the 44,000 to 98,000 excess deaths primary care ollow-up appointment. Communication and coordi-
reported in the IOM’s To Err Is Human report. nation between hospitalists and outpatient providers is increasingly
identi ed as a risk area, and uture collaboration between providers
■ VULNERABLE PATIENTS in di erent care settings will be necessary to address it.
Although all hospitalized patients are at risk o medical errors, cer- As health care moves toward value-based payment and deliv-
tain groups seem to be at particularly increased risk. Both extremes ery systems, overtreatment and overutilization are increasingly
100
identi ed as areas contributing to Patient Sa ety risk. In the decade Everyone is prone to slips, lapses, and mistakes. We can reduce the
a ter To Err is Human, underuse has improved while overutilization requency o these errors through education and training. However,
C
has not. Overutilization o inappropriately broad spectrum antibiot- no one is in allible, and there ore, no one is immune rom error. In
H
ics can lead to bacterial resistance or opportunistic in ections, as in
A
act, certain conditions can increase the risk o harm. Workers are
P
the case o carbapenem-resistant Enterobateriaceae or Clostridium more likely to make slips and mistakes when they are tired or over-
T
dif cile; hospitalists have an opportunity to become champions o worked, bored, distracted, intoxicated, or ill.
E
antibiotic stewardship. Medical care can also cause indirect harm: Skills, rules and knowledge-based cognitive errors are o ten
R
excessive or unnecessary CT scans expose patients to radiation unavoidable, and there ore clinicians should not be held account-
1
and reveal “incidentalomas”— ndings o benign or uncertain sig- able or these innocent errors. In contrast, individual actors should
5
ni cance that would otherwise have gone undetected, but that be accountable or at-risk behaviors, which occur when an individual
prompt repeat imaging, laboratory studies, or biopsy. Hospitalists is unaware o the hazard associated with a behavior that deviates
P
must also be aware o incidentalomas that are missed when transi- rom the standard, and reckless behaviors, which occur when an
r
i
n
tioning to the outpatient setting, and ultimately can cause delays in individual knowingly acts dangerously. These behaviors are rare and
c
i
p
diagnosis o malignancy. More diagnoses, more testing, and more may merit disciplinary action.
l
e
treatments increase the potential or patient harm.
s
■ SYSTEMS BASED
o
The advent o Meaning ul Use legislation provided incentives to
adopt quali ying electronic health records (EHRs) rapidly, creating
P
Although all humans err, the impact o mistakes is more serious or
a
new Patient Sa ety vulnerabilities in many hospital and health sys-
t
individuals whose decisions and behaviors have a consequential
i
e
tems. While technological innovations like Computerized Physician
n
e ect on others. Military and commercial aviation, nuclear power,
t
Order Entry (CPOE) systems can improve e ciency and sa e care, and health care are examples o industries where error can be cata-
S
a
Health In ormation Technology (HIT) may have unintended conse- strophic. In these settings, researchers and organizational leaders
e
quences. A ter implementing CPOE, one institution ound that the
t
have begun to ocus on the systems in which individuals work in
y
:
adjusted mortality rate in their pediatric population increased rom order to design de enses that identi y, intercept, and prevent errors
I
n
2.8% to 6.3%. Newly described errors are being recognized, such be ore they result in harm.
t
e
as the right order being placed on the wrong patient through a
n
A system is de ned as a set o interdependent processes designed
t
CPOE system. Other examples o risk include ragmentation in EHRs,
i
to accomplish a common aim. Certain characteristics o systems can
o
n
lack o HIT interoperability, and poorly-designed user inter aces. allow or acilitate individuals’ per ormance o unsa e acts. These
a
l
Implementation o a new HIT system is a particularly vulnerable characteristics are o ten called latent conditions or latent actors.
D
time or patients, and must be monitored closely to mitigate patient
e
Examples o latent actors include poor training, duty schedules
s
harm. During new HIT implementation, hospitalists can have a large
i
that provide little time or sleep, lack o adequate supervision, lack
g
n
impact on patient sa ety by identi ying mal unctioning systems and o su cient supplies, and a culture that discourages cooperation
a
bringing these glitches to attention. and teamwork. Analysts o ten ocus on latent actors that represent
n
d
Creating e ective interventions relies on a solid understanding o design f aws or a particular process and that, in turn, allow unsa e
C
the nature o error in health care, the methods to assess risk in health acts to result in harm.
u
l
care organizations, and the tools that are used to develop Patient
t
Consider the case o a physician who ailed to ollow-up on a
u
Sa ety improvements. These topics are the ocus o the remainder
r
radiology report showing a new lung nodule. The unsa e act must
e
o this chapter. be understood in the context o the latent actors that contributed
to the error. Was the physician overworked, covering or vacation-
ing colleagues? Was there a consistent approach in place or the
THE NATURE OF ERROR IN HEALTH CARE practice or ollow-up o test results? Did a radiologist attempt to
contact the ordering clinician unsuccess ully? Multiple latent ac-
■ HUMAN BASED tors typically contribute to an accident—and ew are apparent until
Human error is a complex phenomenon, but one that has come into a ter an accident occurs. When a series o latent ailures align, harm
better ocus. Students o human error have argued that both human can result. This model o organizational ailure has been described
and systems actors contribute to error. By “human actors” we mean by British psychologist James Reason as the “Swiss cheese model.”
the environmental, work conditions, organizational, and individual Errors in medicine can be classi ed in other ways, into categories
characteristics that inf uence work per ormance. Experts conceive o such as diagnostic errors, medication errors, and communication
human per ormance in several categories: skills, rule-based actions, and transition errors. It is important to examine the underlying con-
and per ormance that rely on novel problem solving. When skills, tributions o human and systems actors to each o these categories.
knowledge, and rules break down or are misapplied, errors occur. These issues will be addressed in subsequent chapters.
These so-called active ailures can urther be subdivided into errors
o execution and errors o planning. EFFECTIVE PATIENT SAFETY PROGRAM DESIGN
Skills are stereotyped behaviors that require little conscious AND CULTURE
thought. When the appropriate skill is chosen but carried out No single model has emerged or an “ideal” Patient Sa ety program
incorrectly, the error is classi ed as a slip. For example, choosing in hospitals. There are, however, several key components o e ec-
an antibiotic to treat in ection but inadvertently setting the wrong tive programs. E ective programs must have methods to detect
rate on the in usion pump would be an example o a slip. Similar to errors, analyze events, and implement improvements. In addition,
a slip, a lapse is when the appropriate action is omitted or not car- e ective programs build a sa ety culture in the organization that os-
ried out, rather than being carried out incorrectly. When acing an ters an environment where reporting, analyses, and improvement
un amiliar scenario where one cannot readily apply an automatic initiatives can f ourish. Buy-in rom senior leadership and physician
or routine skill, knowledge-based and rules-based cognition takes engagement is a particularly important ingredient o the mix. E ec-
over. Mistakes can be knowledge- or rules-based, and result when tive Patient Sa ety programs are resourced adequately to ul ll their
an individual did what they intended, but it did not work as planned. mission. Many hospitals appoint and support Patient Sa ety o cers,
For example, recognizing the in ection but choosing to treat with a risk managers, and data analysts, data managers, and process
diuretic would be a mistake. improvement specialists. These pro essionals and their activities are
101
usually housed in a Department o Healthcare Quality, although the Improvement interventions vary rom changing one step in a pro-
speci c arrangements will vary rom organization to organization. vider’s workf ow to redesigning entire care delivery processes. The
scale and scope dictates the resources and method o implementa-
tion necessary to produce reliable, durable changes. Hospitalists
P
■ MEASUREMENT AND ANALYSIS
A
not only have opportunity to review adverse events via root cause
Hospitals use a variety o tools and techniques to measure Patient
R
analyses, they also can aid quality improvement personnel in under-
Sa ety. Most hospitals have voluntary systems or reporting near-
T
standing rontline provider workf ows.
misses and adverse medical events. This approach is in widespread
I
use, in part because o government and accreditation agency
requirements. Certain serious events must be reviewed internally ■ FOSTERING A CULTURE OF SAFETY—LEADERSHIP
and reported to the appropriate external oversight agency. Most AND FRONTLINE PROVIDER ENGAGEMENT
hospital pharmacy departments use a similar approach, reporting Other high-risk industries, like aviation and nuclear power, rec-
T
h
“interventions” that pharmacists per orm when they clari y or cor- ognized long ago that organizational leadership shapes culture.
e
S
rect a clinician’s order. Pharmacy interventions and sa ety incidents Obtaining buy-in rom hospital boards and executive leadership or
p
together represent an important source o data about errors and Patient Sa ety programs is essential to securing the engagement o
e
c
injuries, but these methods are subject to reporting bias. Busy clini- ront line providers.
i
a
l
t
cians o ten do not have time to complete these reports and may be The Institute or Healthcare Improvement (IHI), in their 5 Million
y
less likely to report their own errors than those per ormed by col- Lives campaign, recommended “Getting Boards on Board,”in an e ort
o
f
leagues upstream in the care process. In addition, clinicians tend to to “ ully engage the governance leadership in quality and sa ety.”
H
o
report very ew adverse events compared to nurses or other hospital Hospital boards can drive sa e care by using the ollowing approaches:
s
p
sta , but those that they do report tend to be more serious. In some • “Setting aims: Set a speci c aim to reduce harm. Make an explicit,
i
t
organizations, an increasing number o incident reports is inter-
a
public commitment to measurable quality improvements.
l
preted appropriately as a sign that sa ety is taken seriously by ront-
M
• Getting data and hearing stories: Select and review progress
e
line clinicians. Hospitalists should consider reporting adverse events, toward sa er care… at every board meeting, grounded in
d
particularly li e-threatening ones, as part o their duty to patient care.
i
c
transparency.
i
n
Given the limitations o incident reporting and pharmacy inter- • Establishing and monitoring system-level measures: Identi y a
e
ventions, hospital leaders and Patient Sa ety researchers have small group o organization-wide ‘roll-up’ measures o Patient
a
n
examined a variety o alternative approaches. Direct observation Sa ety that are continually updated and are made transparent
d
o clinicians at the point o care is a ruit ul strategy but requires a
S
to the entire organization and its customers.
y
tremendous amount o time and e ort to maintain. Chart review
s
• Changing the environment, policies, and culture: Commit to
t
e
methods are also well established but potentially resource intensive. establish and maintain an environment that is respect ul, air,
m
Since not all errors result in harm, recent measurement tools have and just.
s
been developed to ocus on harm events. The Institute or Health-
o
• Learning, starting with the board: Develop the board’s capabil-
f
care Improvement Global Trigger Tool allows or the identi cation o
C
ity and learn about how ‘best in the world’ boards work with
a
adverse events, based on clues seen in the medical record. executive and medical sta leaders to reduce harm.
r
e
Some organizations have access to sophisticated tools that can • Establishing executive accountability: Oversee the e ective
screen electronic medical records or possible errors or adverse execution o a plan to achieve aims to reduce harm, including
events. These tools examine medication records or events—such executive team accountability or clear quality improvement
as an order or antidote drugs like diphenhydramine, naloxone, and targets.”
epinephrine—that may signal the presence o an adverse event.
While many institutions strive to establish a nonpunitive environ-
AHRQ has developed a set o Patient Sa ety Indicators (PSIs) that
ment, as o 2014 up to 50% o employees still worry about negative
screen administrative records or adverse events based on diagnosis
repercussions rom reporting. Emphasizing a nonpunitive environ-
and procedure codes.
ment or adverse event and error analysis can help to establish a
Once a critical incident or set o incidents has been identi ed,
culture o sa ety, where all employees eel com ortable disclosing
health care organizations need to cull the lessons that can be learned
errors knowing that their disclosure will lead to improvement. The
rom these events. Traditionally, this has been the subject o the Mor-
importance o ront-line provider engagement is essential, and hos-
bidity and Mortality Con erence, though the ocus o M&Ms has o ten
pitalists play a critical role.
been on individual per ormance rather than systems actors that con-
Hospitalists understand the systems and processes o care within
tributed to errors. In contrast, root cause analysis is a systematic and
their local environment. They can identi y potential sa ety risks,
structured approach to identi y the latent conditions that contributed
escalate their concerns, advocate or improvements, and evaluate
to an error. Root cause analyses, per ormed properly, help organiza-
improvement designs critically. Hospitalists can weigh the easibility
tions to learn about the causes o errors and injuries and, in turn, to
o implementation plans and become local champions or process
develop initiatives that prevent these errors rom happening again.
or cultural changes.
■ IMPROVEMENT IMPLEMENTATION
QUALITY IMPROVEMENT
It is not enough to measure and analyze adverse events and
near misses; e ective Patient Sa ety programs integrate quality Although Patient Sa ety is a cornerstone o medical care, the notion
improvement techniques and personnel to implement interven- o “quality” in health care includes several other important com-
tions designed to prevent harm. Near-misses are particularly impor- ponents. The IOM has de ned six dimensions by which quality in
tant, as they signal weakness in the care delivery process that have health care can be evaluated:
not yet resulted in harm but might do so i le t unaddressed. • Sa e: avoiding injuries to patients rom the care that is intended
A ter an event is analyzed, the processes o care leading to that to help them.
event must be elucidated in order to design an e ective interven- • E ective: providing services based on scienti c knowledge to
tion. This step is critical. When the processes are mapped, exam- all who could bene t, and re raining rom providing services to
ined, and understood, appropriate interventions can be designed. those not likely to bene t.
102
• Patient-centered: providing care that is respect ul o and or anticoagulants and other high-alert medications. Rapid-cycle
responsive to individual patient pre erences, needs, and values, improvements have led to innovation in communication o critical
C
and ensuring that patient values guide all clinical decisions. test results and hando communication. As researchers understood
H
A
• Timely: reducing waits and sometimes harm ul delays or both the value o human actors principles and system-based design,
P
those who receive and those who give care. Patient Sa ety leaders have embraced and promulgated “best
T
• E cient: avoiding waste, including waste o equipment, sup- practice” interventions that rely on these concepts. Key principles
E
plies, ideas, and energy. include the concepts o standardization and reliability, appropri-
R
• Equitable: providing care that does not vary in quality because ate redundancy, use o communication, and teamwork tools. Best
1
o personal characteristics such as gender, ethnicity, geo- practice recommendations have been incorporated into recom-
5
graphic location, and socioeconomic status. mendations and standards put orth by The Joint Commission and
In this ormulation, sa ety is one o several components o high the National Quality Forum. Recognizing that there are limits to
P
human per ormance, researchers have investigated the use o orc-
r
quality care.
i
n
ing unctions as prompts. Attention has also been ocused on the
c
i
p
limits o human per ormance during times o atigue, with e orts to
■ IMPROVEMENT METHODS
l
e
reduce these e ects. Each o these concepts will be expanded upon
s
How does a health system improve its per ormance in one or
o
in subsequent chapters.
more o these domains? Many organizations rely on the Model or
P
a
Improvement, a well-described approach used to promote orga- CONCLUSION
t
i
e
nizational change. The basic steps in the Model or Improvement
n
Recognizing the need to mitigate patient harm rom preventable
t
include the ollowing:
adverse medical events, Patient Sa ety has become an organizational
S
a
• Setting aims priority o many health care institutions across the world. The epicen-
e
• Establishing measures
t
ter o work in the area continues in hospital care, given the multiple
y
:
• Selecting changes interventions delivered there and the vulnerability o the patient
I
n
• Small tests o change, as in Deming’s plan-do-study-act (PDSA)
t
population. New areas o Patient Sa ety risk are identi ed regularly
e
model
n
and become the ocus o scholarly research and improvement initia-
t
i
• Implementing changes tives. E ective Patient Sa ety programs are built on the understand-
o
n
• Spreading changes ing that medical errors are the result o human actors and systems
a
l
A ter setting aims, measures must be established to determine issues, obtain executive, board, and rontline provider engagement,
D
e
whether improvement results rom the changes that have been and establish a culture o sa ety by rein orcing a nonpunitive envi-
s
i
ronment. The leaders o these programs design systems that identi y
g
implemented. There are three undamental types o measures:
n
• Outcome measures: evaluate the end result o a given system
errors and learn rom them, build interventions to prevent errors and
a
n
or process. mitigate harm, and disseminate their work as best practices to the
d
• Process measures: evaluate the steps involved in a process.
greater medical community. Organizations that do these things well
C
create sa er, more e ective patient care.
u
• Balancing measures: evaluate whether changes in one area
l
t
u
result in (unintended) changes elsewhere.
r
e
Unlike measurement in research, measurement in quality SUGGESTED READINGS
improvement is used to bring new knowledge into daily practice.
Bates DW, Cullen DJ, Laird N, et al. Incidence o adverse drug events
It consists o small tests o change, with many sequential tests, and
and potential adverse drug events. Implications or prevention.
just enough data gathered in each round o testing to assess i
ADE Prevention Study Group. JAMA. 1995;274:29-34.
change results in improvement. The process is then continued with
the implementation and spread o change. Brennan TA, Leape LL, Laird NM, et al. Incidence o adverse events
Another improvement method that was developed originally or and negligence in hospitalized patients. Results o the Harvard
industry uses a detailed study o production processes to improve Medical Practice Study I. N Engl J Med. 1991;324:370-376.
e ciency by eliminating various orms o waste. The Toyota Produc- Conway J. 5 Million Lives Campaign: getting boards on board:
tion System is an example o this management philosophy, which engaging governing boards in quality and sa ety. Jt Comm J Qual
has become known generically as the Lean production system. Lean Patient Sa . 2008;34:214-220.
methodologies have made their way into health care, with a ocus Institute o Medicine, Committee on Healthcare in America. Cross-
on improving quality and reducing cost by streamlining the coming the Quality Chasm: A New Health System or the 21st Century.
plex care delivery process. Reducing complexity and waste reduces Washington, DC: National Academies Press; 2001.
resources and risk. A care process that takes 50 steps and 10 people
to accomplish is more expensive and potentially riskier than a Kohn LT, Corrigan J, Donaldson MS, eds. To Err Is Human: Building a
10-step process involving two people. While Lean is not engineered Sa er Health System. Report o the Committee on Quality o Health
to improve Patient Sa ety, it may do so indirectly by eliminating ine - Care in America. Washington, DC: National Academy Press; 2000.
ciencies that increase the potential or error and harm. Langley GL, Nolan KM, Nolan TW, et al. The Improvement Guide:
A Practical Approach to Enhancing Organizational Per ormance.
■ QUALITY IMPROVEMENT AND SAFETY RESEARCH San Francisco, CA: Jossey-Bass; 1996.
Quality Improvement and Patient Sa ety have intersected in recent Reason J. Human error: models and management. BMJ.
years, as quality improvement methods have been applied to 2000;320:768-770.
solve Patient Sa ety problems. For example, health leaders used Weingart SN, Wilson RM, Gibberd RW, et al. Epidemiology o medical
the PDSA model to develop medication sa ety improvements error. BMJ. 2000;320:774-777.
103
16
CHAP TER INTRODUCTION
“Patient-centered care” is a core principle in health care, and the
concept continues to move rom innovation to expectation and in
some cases, even regulation. But what exactly does it mean to pro-
vide patient-centered care, and how does one achieve this, at either
the individual or institutional level? This chapter explores the term
patient-centered, and current thinking about how to strengthen the
partnership between patients and providers in the delivery o care.
Throughout, emphasis is given to those innovations most relevant
Patient-Centered to the hospitalized patient.
Care DEFINING PATIENT-CENTERED CARE

While the term patient centered now appears commonly in both
medical literature and lay media, one may encounter a variety o
de nitions or this phrase. Perhaps the most “o cial” de nition is
Kenneth Sands, MD, MPH the one proposed by the Institute o Medicine (IOM) in the landmark
2001 document Crossing the Quality Chasm. It describes patient-
Lauge Sokol-Hessner, MD centered care as “…care that is respect ul o and responsive to
individual patient pre erences, needs, and values, and ensuring
that patient values guide all clinical decisions.” The IOM goes on
to describe patient-centered care as one o the six key “aims or
improvement” or quality o care. As such it is presented as an intrin-
sic value, undamental and irre utable, as opposed to a system prop-
erty with a known association with better outcomes. The concept
has been advanced in the orm o slogans such as “Nothing about
me without me,” “Every patient is the only patient,” “You’re a person
be ore you’re a patient,” and “Human First.” At its core, patient-
centered care can be described as treating each patient with the
respect and dignity that every human being inherently deserves.
Overlapping terms appear in the both the lay and medical lit-
erature, including patient partnering and amily-centered care. In this
chapter, the term “patient-centered” will be used to encompass the
general concept o making care delivery more responsive to the needs
and wishes o the individual patient and his or her amily. Family can
be de ned broadly as all the individuals whom the patient wants
involved in his or her care regardless o whether they are related
biologically, legally, or otherwise. From this de nition, it ollows that
i a patient has any amily, patient-centered care must include them.
The construct presumes that care delivery under current models
is not adequately patient-centered. The IOM “Chasm” report con-
ceptualizes the health care delivery system as in evolution, rom a
clinician-centric, poorly coordinated and non–evidence-based model
to a patient-centric, integrated system that consistently applies
scienti cally supported interventions. In “clinician-centric” models,
the patient plays a passive role in decision making regarding choice,
timing, and settings o care delivery; these are the exclusive domain
o the providers, and those same providers decide what in ormation
reaches the patient. Stories abound in the lay and medical literature
o patients eeling at the mercy o the medical system, unable to exert
control over their own care. However, it is hard to nd a quantitative
assessment o the current state o patient-centered care (or lack
thereo ) in the US health care system. Some insights can be gleaned
rom national results o the Hospital Consumer Assessment o Health
Care Providers and Systems inpatient survey distributed by the Cen-
ters or Medicare and Medicaid Services (CMS). Although scores have
been improving, 18% o patients still report “never or sometimes”
receiving communication about new medications and their side
e ects, and 9% answer similarly to questions about the responsive-
ness o hospital sta (composite scores, July 2012-June 2013 data).
104
The IOM describes the ully evolved stage o organizational and is given the option to participate. On rounds, the patient/
development as characterized by the patient and amily being part amily member is introduced to the members o the team, hears
C
o the health care team, with ull access to in ormation and the abil- the presentation o the clinical situation and is invited to partici-
H
ity to exercise as much control over care as desired. What speci c pate in developing the plan. Teaching, including discussion o the
A
P
actions can institutions and providers take to advance toward this condition and demonstration o physical ndings, occurs with the
T
model? These can be divided into three key properties: (1) ree f ow patient’s permission. For patients, such programs have been associ-
E
o in ormation, (2) partnering around individual patient needs, and ated with higher satis action, better clinical outcomes, and shorter
R
(3) involving patients in system design. lengths o stay. Health care workers, in turn, have reported higher
1
satis action with work and with the quality o some aspects o teach-
6
FREE FLOW OF INFORMATION ing in academic medical centers, but have also identi ed challenges
Patients are not truly partners in their own care i in ormation, either that have inhibited widespread adoption o such processes.
P
a
about themselves or the care they are receiving, is only selectively
t
i
■ COMMUNICATION AND RESOLUTION IN SETTING OF
e
available. The IOM suggests two “rules or redesign” to achieve
n
transparency: ADVERSE EVENTS
t
-
C
• Knowledge is shared and in ormation f ows reely. Flow o in ormation should not stop i care does not go as planned.
e
n
• Transparency is necessary. Patients who experience an adverse event are ethically entitled
t
e
to receive in ormation about that event, and typically respond
r
The rst rule recognizes that patients should have the ability
e
avorably to “I’m sorry” in those situations where apology is indeed
d
to receive complete and understandable in ormation about their
C
condition, in real time. The second “rule” establishes that patients appropriate. Un ortunately, open communication with patients
a
about medical error has been inhibited by a conventional wisdom,
r
e
are entitled to in ormation about the care itsel , including the
per ormance o the health care system and its providers, as well held or decades, that open disclosure and apology would increase
as the approach to care and its justi cation. While these concepts risk o litigation. There is a growing body o evidence that this is a
may seem sel -evident, the health care system has not tradition- alse premise, and that open communication ollowing error is not
ally been aligned with these rules. Prior to passage o The Health only ethically correct, but is likely to decrease malpractice costs
Insurance Portability and Accountability Act (HIPAA), exchange o since it provides an opportunity to bring issues to resolution without
medical documentation sometimes required that a patient obtain a involving the court system. A growing number o institutions are
subpoena. Today, care still remains ar rom transparent: published now implementing a systematized approach to harm events that
reports suggest the majority o hospitalized patients are unable to involves open communication, root cause analysis, and early resolu-
easily determine the name o the physician in charge o their care, tion i the analysis determines that the harm was preventable. Many
let alone the details o the care plan. Fortunately, recent innovations o these programs are now reporting a coincident improvement in
are improving in ormation exchange with patients and changing malpractice costs. Such programs are also more patient-centered in
longstanding traditions o care delivery. that time to resolution is quicker, and nonmonetary elements can
be part o the resolution. For example, patients su ering harm o ten
■ OPEN COMMUNICATION OF THE PLAN OF CARE look or an institutional commitment to decrease risk o a similar
Typically, the discussion, development, and implementation o harm event occurring in the uture.
a hospitalized patient’s plan o care occur without the patient’s Remaining patient-centered in the context o adverse events
involvement. Communication o the plan o care is a separate requires that an institution establish an unambiguous position on the
responsibility o the physician, occurring most o ten as an unstruc- topic, and communicate that position to the work orce. Mechanisms
tured verbal communication. There is thus no system that guaran- must be put in place to educate and support clinicians in the process
tees that the patient understands the plan, or has had the chance o disclosure and apology, which is o ten an ongoing event, requiring
to ask questions. Approaches that provide structure to these multiple communications as acts become available. For any given
exchanges, and thus more reliable sharing o in ormation, are clinician, personal involvement in disclosing an error to a patient will
appearing in the interest o both patient-centered care and patient be a rare event, so systems or “just in time”support and training must
sa ety. For example, the Veterans Health Administration introduced be available. Many institutions address this by creating a resource
“The Daily Plan,” a structured document reviewed with the patient group with speci c interest and training in best practice or com-
each day o his or her hospitalization. It contains in ormation such munication, empathy, and apology (where appropriate). Physicians,
as medications, scheduled procedures, and diet, reviewed with the nurses, social workers, patient sa ety pro essionals could all poten-
patient each day o his or her hospitalization. The expectation is this tially serve in such a role. In the setting o an adverse event, the expert
kind o intervention could improve provider-patient in ormation resource can support the clinicians involved and help determine the
exchange in both directions, provide an opportunity or patients to best timing, setting and participants or communicating the event
ask questions and share concerns, and identi y problems that might (see Chapter 20: Preventing and Managing Adverse Patient Events).
cause risks to sa ety or make the provider’s plan ine ective or in ea-
sible. Reported experience shows that the large majority o patients ■ ACCESS TO MEDICAL DOCUMENTATION
receiving such a plan perceive a better understanding o their hospi- Medical documentation has traditionally been the purview o the
talization, have a better ability to ask questions, and a higher level o clinicians and not the patient. The passage o HIPAA in 1996 estab-
com ort with their hospital stay. Similar positive ndings have been lished that patients must be permitted to review and amend their
seen by introducing structured patient involvement with hospital medical records, but access to the record is still largely based on
discharge planning. A key element o these new models is some an exception process, the record being provided when there is an
mechanism or “closed loop communication,” meaning there is active request by the patient, which in practice occurs rarely. On the
veri cation that the communication has been received, understood, other hand, surveys show that when given the option to view the
and any remaining questions have been answered. clinical record, the large majority o patients will accept. Many insti-
Several hospitals have actually embedded communication with tutions have responded with systems that allow patients to directly
patients into the work model by adapting bedside “rounds” to access elements o clinical documentation electronically. This is
include the patient and/or amily member. In a typical ormat, the o ten limited to objective content such as problem lists, medication
patient and/or amily member is oriented to the process o rounds lists, and test results.
105
Ready access to clinician documentation is now gaining momen- United Kingdom, or example, collects in ormation on symptoms
tum. A ormal, multi-institutional trial o “OpenNotes,” a system or and unctional status or all patients undergoing certain elective
electronically sharing outpatient clinical documentation, demon- surgeries.
strated that more than two-thirds o patients reported better under-
P
A
standing o their medical conditions, and elt more in control o their ■ ACCESS TO INFORMATION ON CLINICAL
R
care. At the same time, only 3% o physicians reported spending PERFORMANCE
T
more time answering patient questions outside o visits. The clinical There is a slow but undeniable trend toward sharing more in orma-
I
impact o more open access to documentation is now getting more tion about clinical per ormance with the lay community. Health
attention. For example, physicians in the OpenNotes trial report they care institutions have resisted this trend, on the basis that clinical
are more attentive to the accuracy o their documentation, and the per ormance data are too di cult to correctly interpret, cannot
approach to documentation o sensitive conditions such as cancer, be adequately risk adjusted, and/or will perversely impact clinician
T
h
mental health, and substance abuse. Results o the OpenNotes trial behavior. Thus, much o the initial e ort to make per ormance
e
were compelling enough that all three participating institutions data public has been involuntary, driven by regulators, creditors, or
S
p
have adopted the program as standard practice throughout their insurers, each o which may have its own unique requirements or
e
ambulatory operations. More research is needed in areas such as
c
public reporting o clinical per ormance. As a result, clinical in or-
i
a
the potential or open sharing o clinical documentation to better
l
mation available in the public domain can vary dramatically rom
t
y
identi y medical error, and/or to improve clinical outcomes. state to state. At a ederal level, the patient can nd a growing list
o
f
o measures o hospital per ormance disseminated by CMS via their
H
■ PATIENT-GENERATED HEALTH DATA website, http://www.medicare.gov/hospitalcompare/search.html.
o
s
Simultaneously, many institutions are now choosing to volun-
p
Patient-generated health data (PGHD) include biometric data (eg,
i
t
tarily share in ormation on clinical per ormance (Figure 16-1). This
a
home blood pressure readings), medical history, symptoms, and
l
trend is most readily apparent as a component o hospital websites,
M
administrative in ormation (eg, the name o a surrogate medical
but some hospitals are also choosing to share in ormation in the
e
decision maker). They are recorded by patients, as opposed to pro-
d
orm o mailings or posted material within the acility (illustration).
i
viders, and the growth o Internet-enabled devices has encouraged
c
i
Reasons to pursue this strategy likely vary by institution, but might
n
a movement to collect more PGHD. Proponents o PGHD hope that
e
it will engage patients more in their care and believe that it could include (1) the ability to provide context and explanation to in or-
a
mation already being shared elsewhere, (2) the ability to determine
n
shi t some o the burden o data collection rom providers during
d
time-limited encounters to patients outside o such encounters, and expand the port olio o in ormation being shared, (3) the belie
S
y
reeing patients and providers to spend their time together in more that sharing clinical per ormance in ormation is a good business
s
t
strategy, (4) the belie that sharing clinical per ormance in orma-
e
valuable ways, such as more deeply engaging in the shared deci-
m
sion-making process. However, there are many challenges. In ormation is consistent with institutional values. The result is that a cur-
s
rent survey o hospital websites will demonstrate a broad range
o
tion technology pro essionals must ensure that PGHD inter aces are
f
user riendly, acilitate high-quality data collection, and maintain o approaches to transparency: some provide a great breadth o
C
in ormation on per ormance, others almost none. Some hospitals
a
appropriate privacy. Simultaneously, it is important to ensure maxi-
r
e
mal interoperability between di erent electronic medical record provide metrics with a minimal amount o explanatory in ormation,
systems. Frontline clinicians must be involved in the design o while others appear to go to great lengths to make the in ormation
PGHD systems so that the new streams o data will t into their busy accessible and available to a lay audience.
workf ows. And all parties must develop ways o reviewing PGHD, Few data are available regarding the degree to which patients
especially prior to including them in patients’ medical records. use clinical per ormance data to make decisions regarding their own
care. The data that do exist suggest that public opinion strongly
avors the concept o public sharing o per ormance data, even
■ PATIENT-REPORTED OUTCOMES
though ew consumers currently direct their care on the basis o
Current standards or assessment o clinical outcomes are based objective, publicly reported metrics.
largely on objective measures and rarely include the patient’s
perspective. For example, outcome o a joint replacement surgery
might be evaluated on the basis o range o motion or occurrence
o a complication, without ormal capture o what the patient has
experienced in terms o pain relie or ability to pursue daily activi-
ties. For some time, evaluation o clinical outcomes in the research
domain has included such patient-reported outcomes (PROs). Now,
attention to PROs is growing as part o routine clinical practice,
spurred in part by provisions in the US Patient Protection and A ord-
able Care Act, which speci es a ocus on PROs and launched the
Patient-Centered Outcomes Research Institute. At the same time,
the increasing presence o electronic ormats or collecting and stor-
ing survey data is ueling the ability to collect PROs.
PROs might include in ormation about which outcomes matter
most to patients as well as actual measures o symptoms, unctional
status, or quality o li e. Collection o such outcomes is especially
relevant to conditions involving longitudinal care such as cancer,
heart disease, mental illness, or arthritis. There is growing interest
in developing standardized instruments or collecting PROs, and
ormal guidance rom the National Quality Forum on valid design Figure 16 1 Public display o per ormance on several clinical out-
o PRO-per ormance measures has been issued. Some o the best comes outside o an Intensive Care Unit. The ormat allows or the in or-
examples o broad implementation o PROs in clinical practice come mation to be continuously updated. (Courtesy o Beth Israel Deaconess
rom outside the United States. The National Health Service in the Medical Center, Boston, MA.)
106
PRACTICE POINT to lead to patient-centered care. For instance, imagine a patient
who just had a hip replacement and is asking to leave the hospital
C
Hospitalists can tangibly improve patient-centered care by:
as soon as possible. One approach would be to tell her she cannot
H
•
A
Implementing a structured approach to incorporating patients yet leave, and that the health care team will decide when she is
P
and amilies into the daily plan (eg, the use o “daily plan” ready to be discharged. A partnering approach would start by asking
T
templates and bedside rounds). about her concerns (Patient: “I am worried about getting a hospital-
E
• Fully disclosing adverse events to patients’ amilies in a timely acquired in ection”), align with those (Hospitalist: “We do not want
R
ashion. you to get an in ection or stay in the hospital any longer than is
1
• necessary”), help her understand the events that need to occur
6
Supporting and implementing hospital initiatives that
encourage open medical records, patient-generated health be ore it is sa e or her to leave the hospital (Hospitalist: “In order or
data entry, patient-reported outcomes, and widespread you to be able to unction when you leave the hospital, we eel it’s
P
important that you be able to…”), and nish by asking i the team
a
transparency o clinical outcomes data.
t
i
has addressed her concerns (Hospitalist: “How does that sound to
e
n
you? What questions do you have?”), thereby making her a more
t
-
C
active participant in her care.
e
In some cases, customizing care requires some preparation. For
n
PARTNERING AROUND INDIVIDUAL PATIENT NEEDS
t
e
Assuming the rst ideal has been met and there is complete example, encouraging patients to legally designate a surrogate
r
e
transparency and f ow o in ormation; how much control does the medical decision maker is a critical part o advance care planning;
d
without that in ormation, by the time the patient loses capacity to
C
patient have over the plan o care, or the ways in which that plan
a
make their own medical decisions, it may be too late to determine
r
is carried out? Patients o ten encounter systems that are unrespon-
e
sive to their individual needs. Truly patient-centered care requires what their pre erences are and who they would pre er to speak on
institutions to have the capability to ully elicit individual patient their behal . Other examples include reliably recording who the
pre erences, and the capacity to reliably customize care in response patient wants involved—and who they do not want involved—in
to those pre erences. their care, as well as how he or she pre ers to be addressed, and
which language is best or him or her. Customized patient-centered
care then means consistently using that in ormation when speak-
■ APPRECIATION OF INDIVIDUAL NEEDS
ing with the patient and amily, as well as arranging any necessary
At one level, the system can be made more responsive to the interpreter services in advance o an encounter. Electronic medical
patient simply by ensuring that the clinicians have a complete record systems must be designed to support busy clinicians by mak-
understanding o individual pre erences. Un ortunately, despite ing it easy to record such in ormation, and then ensuring it is easily
the central role o interpersonal communication in almost every retrieved at the point o care.
aspect o care delivery, evidence suggests that clinician-patient Being responsive to individual needs also requires ceding some
communication is o ten imper ect, and communication skills have control to the patient. For example, many hospitals are eliminating
only relatively recently been recognized as a core competency by restrictions on visiting hours in avor o open access or the patient’s
entities such as the Accreditation Council or Graduate Medical amily. Family presence during invasive procedures and resuscita-
Education and The National Board o Medical Examiners. Training tion events is endorsed by several pro essional societies. Programs
can in act lead to improved communication skills, allowing the clini- that allow patients or amily members to activate rapid response
cian to better identi y the unique needs and values o the patient, teams are now well described. Sometimes called “Condition H” or
while simultaneously improving patient understanding o his or her “Code Help,” the concept was initially advanced or the pediatric set-
clinical situation. ting, where serious adverse events have occurred in the presence o
Since needs and pre erences may be linked to individual ethnic a concerned parent who was unable to bring immediate assistance
and religious traditions, patient-centered care requires that provid- to the bedside. The same principal has now been extended to the
ers have an appreciation or this context. Cultural competence adult setting. A typical response team will include both a physician
re ers to the provider’s ability to bridge cultural di erences in the and nurse, and in some models may include critical care specialists,
provider-patient relationship, through understanding and respect social workers, chaplains, and/or patient representatives. It is not
o the patient’s belie s and awareness o one’s own biases. Cultural unusual or clinicians to initially express resistance to the concept,
competency is most o ten discussed as a strategy or decreasing the citing concerns that patients and amily members will overuse “code
persistent in erior clinical outcomes among minority populations help” or inappropriate, nonurgent issues. However, the experience
(see Chapter 37: Cultural Competency in Healthcare), but it is equally o institutions that have implemented response teams is that the
clear that cultural competency is an essential component o patient- option is used prudently, and allows earlier interception o potential
centered care. Like communication, cultural competency is teach- adverse events as well as other important issues such as inadequate
able; training programs in cultural competency are on the increase, pain control or communication problems. Success ul implementa-
and in act are mandated or physicians in many states. In addition, tion o a “code help” program requires patient and amily education
institutions with signi cant numbers o patients rom speci c ethnic about its purpose, a well-structured activation mechanism, and a
communities should look or ways to partner with those communi- prede ned set o individual responders. Protocols that de ne the
ties to develop both a shared appreciation o speci c needs and mechanism or recording the incident, and debrie ng with the
pre erences as well as appropriate institutional supports. patient and amily, should also be established.
Partnering around individual needs becomes particularly impor-
■ CAPACITY TO PREPARE FOR AND RESPOND TO tant when treatment options carry signi cant risks or consequences
INDIVIDUAL NEEDS to the patient. Conditions with multiple potential approaches to
Full appreciation o the unique needs o the individual is only management such as chronic back pain, depression, cancer, or
meaning ul i the system can customize care in response to those organ ailure, are all examples o conditions where patient pre er-
needs. The concept o partnering with patients is one way to think ence plays a major role in determining the best plan o care. Such
about the process o customizing care. Helping patients understand situations call or in ormed, shared decision making, a term or the
the reasons underlying their care is likely to make them engaged process o communicating with the patient about what matters
participants rather than passive recipients, and there ore more likely most to him or her in the context o the medical situation, with the
107
ultimate goal o making patient-centered medical decisions. Shared more broadly, the concept o patient-centered care extends beyond
decision making and partnering with patients replace two inappro- the approach to the individual patient, and includes as a tenet that
priate alternatives: telling patients what to do without their input, patients have a voice in the design o the care delivery system
and asking them to make choices rom a list o options without any itsel . In what might be described as a traditional model or hospital
P
A
guidance or advice. Education about shared decision making may administration, the patient is seen as a consumer o services, without
R
help clinicians and patients navigate this complex process and nd any ormal role as part o hospital operations. Consequently, despite
T
the right balance. the patient’s experiences with the institution, he or she has little or
I
The rst step is to ensure the patient understands his or her con- no ability to advocate or change, and the institution lacks the voice
dition, the prognosis, and the treatment options, to the degree they o the patient in the design o care processes. This is beginning to
desire. This can be accomplished using established communication change as hospitals move to involve patients in operational activi-
techniques such as “Ask-Tell-Ask”: (1) “ask” the patient how much ties, either by creating positions or patients or amily members on
T
h
they would like to know, how involved they would like to be in deci- existing hospital committees, and/or creating a separate “Patient/
e
sion making, and what they understand; (2) “tell” them about what Family Advisory Council” (PFAC) unction. Indeed the presence o a
S
p
they do not yet understand in patient-centered language; (3) “ask” PFAC is now mandated by regulation in some states. An institution
e
them how they eel about that in ormation, what questions they
c
may have a single PFAC or multiple PFACs based on a desire or
i
a
have and to reiterate their understanding to ensure everyone has a speci c patient involvement in discreet service lines.
l
t
y
shared understanding. A role or patient/ amily participation has been described or a
o
Be ore discussing treatment options, it is important or clinicians myriad o institutional processes, including strategic planning, acil-
f
H
to understand what matters most to the patient, in the context o ity redesign, research oversight, ethics, care coordination, education,
o
s
the medical situation. Although it is a seemingly vague construct, nance, credentialing, leadership search, in ormation technology,
p
i
understanding “what matters most” becomes increasingly impor- process improvement, patient sa ety, service excellence, and per-
t
a
l
tant as the stakes o the medical decisions rise. Asking about the sonnel practices. Table 16-1 demonstrates the range o involve-
M
patient’s hopes, goals, ears, and worries, and about any tradeo s ment possible or patient- amily advisors. Whether the plan is to
e
d
they are willing to make in order to reach their goals, brings clarity to involve patients on existing operational committees and/or to cre-
i
c
what matters most to a patient, and provides invaluable in ormation ate one or more PFACs, a success ul model or patient involvement
i
n
e
or the next step in shared decision making. in the design o care should begin with a vision and a plan that
a
Discussion o treatment options must include in ormation about addresses a number o key issues:
n
d
the anticipated bene ts and risks o each option, and should include • What are the goals or including a patient in this design
S
a recommendation rom the clinician based on knowledge o what
y
process?
s
matters most to the patient. Without such knowledge, the clinician
t
• What is the organizational model or patient involvement?
e
m
risks suggesting options that will not actually bene t the patient. • Who rom the organization will manage patient involvement?
s
For instance, recommending an aggressive chemotherapy regimen • What criteria will be used or selecting patient participants?
o
may be inappropriate or a patient who pre ers to maximize qual-
f
• What are the expectations o the patient participants?
C
ity o li e and minimize time spent in health care settings, even i What is the selection process?
a
•
r
that means they may die sooner than they would i they received
e
• How long is a term o service, and are there term limits?
that regimen. In ormed shared decision making can be supported • What is the orientation process or patients to a hospital
by decision aids, that is, structured materials such as a videotape administrative role?
or printed algorithm that help to illustrate treatment options and • What are the expectations or attendance?
associated risks and bene ts. • What criteria will be used to assess the per ormance o patient
participants, and/or the PFAC committee itsel ?
PRACTICE POINT An up- ront strategy to de ne these issues is likely to be rewarded
with a smooth unctioning program or patient involvement in
Hospitalists can better meet patients individualized needs by hospital operations.
taking the ollowing steps:
• Acquire and maintain communication skills and cultural FRONTIERS IN PATIENT-CENTERED CARE
competency. How will we know i we are becoming more patient-centered?
• Identi y who the patient does and does not want involved in Developing measures o patient-centered care is an important step.
his or her care, including his or her surrogate decision maker. One approach is to assess whether a hospital has the elements o
• Communicate in the patient’s native language. patient-centered systems described in this chapter. A sel -assessment
• Encourage and empower amily presence. tool created through a partnership o the Institute or Patient-
• Con rm that the patient- amily understand in ormation Centered and Family-Centered Care and the American Hospital
presented to them, via techniques such as “Ask-Tell-Ask,”and Association is available or this purpose (http://www.aha.org/aha/
decision aids to illustrate di cult concepts (such as a videotape content/2005/pd /assessment.pd ). But having the capability to be
or printed algorithms). patient-centered is not su cient; a system must also reliably deliver
patient-centered care or each patient.
What does it mean to be patient-centered on an individual level?
Certainly one can imagine situations where it is clear whether or not
INVOLVING PATIENTS IN SYSTEM DESIGN the care was patient-centered: or example, ailure to use an inter-
Systems o care delivery have largely evolved in response to the preter when it was requested by the patient would demonstrate a
needs o providers and the design o the payment system. The ailure to be patient-centered, whereas eliciting and honoring end-
patient thus encounters a care delivery model that is con using o -li e care pre erences would be a great example o patient-centered
to navigate, inconvenient, and severely ractured between care care. But in many situations, it may not be clear how to assess
delivery settings. Improved coordination o care or a patient can whether or not care has been patient-centered. Imagine a patient
help (see the section in this volume about “Transitions o Care”) but is admitted to the hospital or a viral illness. Despite its etiology, the
108
TABLE 16-1 Patient-Family Advisor Involvement at Beth Israel Deaconess Medical Center, Boston, MA
C
H
Committees with Patient-Family Advisors Projects with Patient-Family Advisor Involvement
A
Patient Care Assessment Committee o the Board o Directors (quality Communication apology and resolution program
P
and sa ety) improvements
T
E
Patient Care Committee o the Board o Directors Patient rights and responsibilities update
R
Ethics Advisory Committee In ormation, security and privacy policy
1
Medication Sa ety Subcommittee OpenNotes project
6
Drug Shortage Task Force PatientSite (patient portal) design
Critical Care Executive Committee Health care proxy material redesign
P
a
Critical Care Experience Task Force Improving the discharge experience through better
t
i
e
educational materials
n
t
Conversation Ready Initiative (End-o -Li e Planning) Communication be ore, during and a ter surgery
-
C
e
Moore Foundation Grant Committee Nurse competency training curriculum
n
t
Respect and Dignity Workgroup MOLST (Medical Orders or Li e-Sustaining Treatment)
e
r
educational material development
e
d
OpenNotes Workgroups Hospital room design
C
a
Service Excellence Steering committee Diversity planning group
r
e
patient requests an antibiotic or his illness, saying he believes it will SUGGESTED READINGS
help him improve. But instead o prescribing the antibiotic, the doc-
tor listens to the patient’s concerns, tries to ensure he eels heard, Berwick DM. What “patient-centered” should mean: con essions o
respect ully counsels him about the risks o antibiotics, prescribes an extremist. Health Af (Millwood). 2009;28(4):w555-w565.
other medications or the patient’s symptoms, and reassures him Delbanco T, Walker J, Bell SK, et al. Inviting patients to read their
that his symptoms will improve soon. In this situation, who is the doctors’ notes: a quasi-experimental study and a look ahead. Ann
judge o whether the care was “patient-centered”? Although most Intern Med. 2012;157:461-470.
readers would agree that the doctor provided excellent medical
Gerteis M, Edgman-Levitan S, Daley J, et al. Through the Patient’s
care, the patient may eel that his care was not “patient-centered”
Eyes. Picker/Commonwealth Program or Patient Centered Care.
because his request was not honored.
San Francisco, CA: Jossey-Bass; 1993.
Measures o patient-centered care should be de ned based on
both what matters most to patients and what is consistent with Johnson B, Abraham M, Conway J, et al. Partnering with Patients
acceptable medical practice. As a starting point, uture work could and Families to Design a Patient- and Family-Centered Health Care
begin to de ne and measure speci c elements o patient-centered System: Recommendations and Promising Practices. Bethesda, MD:
care, and through collaboration with patients and providers could Institute or Family-Centered Care and the Institute or Healthcare
identi y elements around which there is strong consensus. Applying Improvement; 2008.
these measures to populations could reveal the degree to which Lazare A. Apology in medical practice: an emerging clinical skill.
a system delivers patient-centered care. Such in ormation could JAMA. 2006;296:1401-1404.
reveal the opportunities or improvement and drive innovation by Peto R, Tenerowicz LM, Benjamin EM, et al. One system’s journey in
patients, providers, and institutional leaders. creating a disclosure and apology program. Jt Comm J Qual and
Much work remains to ully de ne best practice in patient- Patient Sa . 2009;35(10):487-492.
centered care, and the known steps to achieving optimally patient-
centered care are numerous and complex. Yet in spite o these Teutsch C. Patient-doctor communication. Med Clin N Am.
challenges, as providers, it is our responsibility to prioritize patient- 2003;87:1115-1145.
centered care. As a historical gure once said: “It is not incumbent
upon you to complete the work, but neither are you at liberty to
desist rom it.” [Avot 2:21, attributed to Rabbi Tar on.]
ACKNOWLEDGMENTS
Thank you to the ollowing Beth Israel Deaconess Medical Center
Patient/Family Advisors or their insight ul and ormative comments
on this chapter: Terri Payne Butler, Peter Tarsa, and Nicola Truppin.
109
17
Assessing and improving quality depends upon good measure-
ment. This chapter will discuss basic principles o measuring quality
in medical practice and how data can be used to identi y opportuni-
ties or improvement and support improvement e orts.
ASSESSING QUALITY THROUGH MEASUREMENT
■ PERFORMANCE ASSESSMENT FOR

Harnessing Data EXTERNAL COMPARISON
At all levels o the health care industry, rom national to local and
to Make Quality rom health systems to individual practitioners, per ormance assess-
ment is now routine. Care is assessed in multiple domains, such as
Improvement clinical outcomes, adherence to standards, the patient experience

o care, and the cost o care. Through external comparison, or
“benchmarking,” payers, the government, the public, and others
Decisions: attempt to judge the quality o health care providers and organiza-
tions. Per ormance measurement is more and more being tied to
Measurement and payment in the orm o value-based contracts, with nancial risks
or per ormance shi ting toward the providers o care, behooving us
Measures to understand how per ormance is assessed and improved. In order

or external comparisons to be meaning ul, several criteria need to
be met:
• • The measure should be accurately and reliably recorded across
sites. As an example, length o stay or an emergency depart-
Nathan Spell, MD
ment visit will have de ned starting and ending times that are
routinely captured, recorded, and reported. I the standard de -
nitions and practices are ollowed across all emergency depart-
ments, then length o stay comparisons should be meaning ul.
• • The measure should be valid and meaning ully re ect an
important outcome or process. Mortality is an example o an
outcome measure that has ace validity, but how it is de ned
will a ect its meaning ulness. I mortality is measured only
during the inpatient stay, it will appear better or hospitals that
are e ective at trans erring patients to other settings prior to
death. Measuring mortality over a longer term, such as 30 days,
6 months, or 1 year, may be more meaning ul.
Measures like mortality, patient unctional status, and unintended
hospital readmissions are examples o outcome measures. Outcome
measures tend to be the most salient to patients, the public, and to
health care providers. When using outcome measures or external
comparison, however, the validity o comparison o ten requires
adjusting or in uences on the outcome that are not determined by
the quality o care provided. Factors such as age, comorbid illness,
and social support o ten impact outcomes. Some o these actors
can be identi ed rom the available data sources and included in
risk adjustment algorithms. The degree to which an algorithm can
truly adjust or risk depends upon the quality o the data available or
the adjustment model and the power o the model itsel to re ect
these in uences.
Process measures re ect the steps in clinical care provided to
patients. Common examples o process measures used to assess
hospital per ormance are the CMS core measures. Generally, process
measures do not have to be adjusted or clinical or demographic
actors, as it is assumed that these actors have little in uence on our
ability to carry out the processes. To be valid or assessment, process
measures must also re ect meaning ul clinical care. Measurement
110
o medication reconciliation upon admission to the hospital is an analysis o aggregate data, and may there ore be a better source o
example o a per ormance measure o questionable validity. A pro- data or local improvement than the EHR itsel .
C
vider may achieve credit or per orming medication reconciliation
H
by use o an electronic tool embedded in the health record, but i an ■ COST DATA
A
P
accurate list o home medications has not been obtained, the act o Costs o care can be challenging to identi y, obtain, and understand.
T
medication reconciliation may not be clinically meaning ul. Detailed accounting o actual costs to provide individual compo-
E
R
nents o a given patient’s care is not available in many hospitals.
■ LOCAL QUALITY ASSESSMENT AND IMPROVEMENT Charges may be available, but charges do not have a direct relation-
1
7
External per ormance measures assess only a small portion o the ship to actual costs to provide care, and serve primarily or negotiat-
care provided by individual health pro essionals and organizations. ing contracts and generating bills with payers. I costs are o interest
What is measured locally should re ect the priorities o the organiza- in an improvement project, the hospital nance of ce can o ten
H
a
tion and the needs o patients. Accrediting bodies, such as The Joint provide guidance on how to generate close estimates, or how to
r
n
Commission, require that hospitals routinely assess and improve impute costs based on utilization o resources.
e
s
care or patients. Individuals and hospitals may be less able to per-
s
i
n
orm risk adjustment on internal data compared to data submitted ■ CLINICAL REGISTRIES
g
to external organizations, yet tracking clinical outcomes against past The hospital or clinical practices may participate in collecting data
D
per ormance can be a valid and use ul way to assess quality.
a
or registries o speci c diseases, conditions, or procedures. Pro-
t
a
essional organizations such as the American Heart Association,
t
o
SOURCES OF DATA FOR MEASUREMENT American College o Cardiology, and American College o Surgeons
M
AND IMPROVEMENT sponsor registries where clinical data rom patient care at one site
a
k
can be aggregated and compared to data rom other sites. O ten,
e
■ ADMINISTRATIVE OR “BILLING” DATA these data are more clinically detailed and rich than data extracted
Q
u
Administrative data are one o the oldest sources o data on hospital rom a data warehouse or an EHR since they usually re ect the
a
l
skilled interpretation o trained clinical abstractors who are able to
i
per ormance. To bill an external payer, hospitals must accurately
t
y
identi y the patient, principle and secondary diagnoses, procedures, read the charts and make judgments about the care, which is more
I
m
and any supply items charged. These data are generally obtained detailed than what is usually recorded in discrete data elds in the
p
rom the hospital registration system, nancial system, and medical EHR.
r
o
v
record. Names o physicians involved in care (attending physician,
e
USING DATA TO IDENTIFY OPPORTUNITIES
m
admitting physician, discharging physician, procedural physician,
FOR IMPROVEMENT
e
consulting physician) are generally recorded. Administrative data
n
Benchmarking against other organizations or established targets
t
will include the nature o the patient’s admission (elective, urgent,
D
or emergent), the source o admission (physician of ce, emergency can elucidate opportunities or per ormance improvement. A hospi-
e
tal or health system will have identi ed quality targets rom external
c
department, trans er rom other hospital or acility, etc) and the
i
s
benchmarking or through opportunities identi ed internally to
i
patient’s disposition at discharge (alive or dead, discharged to home
o
n
or other acility, etc). Limitations o administrative data are that it is improve care or value to patients. Where gaps exist between current
s
:
an incomplete re ection o the clinical in ormation, is dependent and desired per ormance, there are o ten opportunities or quality
M
improvement (QI) projects. Aligning QI work with these identi ed
e
upon the thoroughness and accuracy o provider documentation,
a
and is subject to interpretation by trained chart coders and abstrac- system priorities has many advantages. Access to data, project acili-
s
u
tors. Despite these limitations, much o the data or external per or- tation, change management assistance, and leadership support will
r
e
be easier to obtain.
m
mance measures originate in administrative data sets, as these data
Another use ul approach to identi ying opportunities is to
e
have been available or many years in standard ormats across the
n
develop reports rom the electronic data systems or data warehouse
t
health care industry.
a
to assess how closely the care provided matches pro essional stan-
n
d
■ CLINICAL DATA dards or best practice. For example, recommended management
M
o diabetes and hyperglycemia or most inpatients includes basal
e
The source o most clinical data or quality assessment and improve-
insulin with or without corrective dosing o short acting insulin.
a
ment is via an electronic health record (EHR). The EHR brings
s
u
A data report showing the rate at which patients receiving short
r
together patient demographics, orders, provider documentation,
e
acting insulin are also given basal insulin can be use ul to assess an
s
pharmacy data, laboratory results, radiology reports, and any clinical
opportunity to improve management.
in ormation documented by nurses and other health pro ession-
Unnecessary variation in clinical practice is now a recognized
als. O ten the EHR is inter aced with data rom separate electronic
problem in health care delivery. The Dartmouth Atlas project has
systems or patient registration, pharmacy, laboratory, and radiology.
demonstrated this phenomenon. The greater the ambiguity in care
Since EHRs are designed primarily to support clinical care and
practices, the greater the likelihood exists or variation between
patient interaction, they o ten do not contain all o the important
providers. Data reports displaying variation in costs, treatment, or
operational details about patient care. For instance, movement o
diagnostic approaches between providers or the same condition
patients within the hospital is not well re ected in most EHRs; the
may highlight opportunities or improvement.
hospital may have a separate electronic bed management system
that can be an important source o data or projects related to
patient ow. CHOOSING MEASURES FOR A QUALITY
A limitation o many EHRs is that the architecture supports the IMPROVEMENT PROJECT
care o individual patients but may not include an easy way to view
aggregate data across patient populations and time. The hospital or ■ USEFULNESS
health system may have a clinical data warehouse which holds data Though grounded in the scienti c method o hypothesis testing,
rom the EHR and other supporting electronic systems. Data ware- quality improvement techniques di er rom research methods
houses are structured to allow or more sophisticated reporting and by attempting to improve processes and outcomes in the local
111
environment, as distinct rom the intent to prove an underlying with a speci c date and time. With the operational de nitions in
generalizable truth. The measures chosen should be use ul enough hand, measures then need a speci c measurement plan. I reports
or the local improvement work and should not require the degree can be generated out o the electronic data systems, the measure-
o rigor necessary in research projects. To be use ul, measures ment plan would include setting up those reports and validating
P
A
should be easily obtained and checked requently enough to their accuracy by comparing to the chart documentation. A simpler
R
guide the improvement work. Per ormance should be shared with method in many cases is to measure a random sample. A measure-
T
the improvement team and others in real time, whereas research ment plan or sampling might include creating a work sheet or data
I
methods may call or blinding the involved clinicians to reduce bias gathering and outlining the sampling strategy, such as a conve-
in the results. nience sample o 10 consecutive charts, or two random discharges
each day o the week.
■ A BALANCED SET OF MEASURES
T
■ BUILD MEASUREMENT INTO THE WORKFLOW
h
For most quality improvement projects, there is a clear intended
e
measurable goal. Readmissions to the hospital within 30 days o For measures that cannot be easily pulled rom electronic systems,
S
p
discharge, wound in ection rates, and length o stay are good the measurement plan must identi y the collection method. Even
e
c
examples. Achieving progress on such measured goals, however, enthusiastic improvement team members do not relish extra work.
i
a
l
requires complex interventions occurring within the ecosystems o Assigning cases or retrospective chart review may meet more
t
y
hospitals and other clinical environments; as such, it is requently resistance than building measurement into the routine work o
o
f
use ul to measure other aspects o the complex systems when team members. Consider whether there are tools already in use
H
making changes, and create a “balanced set” o measures. A use- that can be modi ed to collect the additional data. In the readmis-
o
s
ul ramework is the Clinical Value Compass (CVC), which consists sion example, does the team already work rom a daily patient list
p
i
t
o our domains: clinical outcomes, unctional status or outcomes, on which they could record whether discharged patients have had
a
l
experience or satis action, and costs. Not every project will have appointments scheduled? Does an assistant already print discharge
M
potential measures in each o these domains, yet the CVC prompts paperwork rom which the ollow-up appointment status can be
e
d
consideration o perspectives beyond the main ocus, which is o ten recorded into a daily log?
i
c
i
on the clinical outcome o interest. Building on the example o
n
e
reducing readmissions to the hospital within 30 days o discharge, ■ TEST THE MEASUREMENT PLAN
a
one may choose to measure the combined costs o the principal
n
Because measurement is key or success and should not be taken or
d
and readmission stay. Patient satis action with the discharge pro- granted, testing the measurement plan on a small scale is a use ul
S
y
cess and overall care is relevant. I easible, a measure o patient step early in the project. The electronic report should be generated
s
t
unctional status at 30 days or longer may be a good complement
e
and validated early in the project (doing so allows you to trust uture
m
to the clinical outcome measure o the readmission rate. I a bal- electronic reports). For manual data collection, test the collection
s
anced set o measures has been chosen, it is also more likely that the
o
tool on a ew patients to veri y that all abstractors interpret the
f
team will detect unintended consequences o their improvement operational de nitions the way they were intended. The measure-
C
e orts. For example, i readmissions decrease, does the principal
a
ment plan should also ensure timeliness o the data. The pace at
r
e
admission length o stay increase? The outcome measure may also which the team can judge the results o improvement e orts will
be remote in time and not directly connected to day-to-day work. be limited by the requency o measurement. When easible, mea-
For example, 30-day readmissions are in uenced by many actors suring on a daily or weekly basis allows the team to make progress
and, by de nition, at least 30 days ollowing discharge must pass to quickly, compared to a monthly or quarterly measurement cycle.
judge the e ect o any intervention on the readmission rate. Add-
ing process measures that are relevant may be very use ul, such as ■ DISPLAYING THE DATA
ollow-up appointments scheduled prior to discharge, transmission
Optimally, data or improvement work are available requently and
o discharge in ormation to the primary physician, and con rmation
are used to drive change. Creating a run chart, which is a line graph
o patient learning through a “teach back” method.
o the data over time, visually demonstrates variability, trends, and,
when annotated with changes to the process, the success or ailure
A GOOD MEASUREMENT PLAN
o those changes. Regularly updating and sharing these results with
A ter project measures are selected, a measurement plan is neces- team members and other stakeholders helps maintain momentum
sary to make the measures available or improvement work. o the project.
■ MAKING MEASURES OPERATIONAL

SUGGESTED READINGS
Measures begin at the conceptual level. An example o a conceptual
measure is “the rate at which patients discharged rom the hospital Berwick DM, James B, Coye MJ. Connections between quality mea-
medicine service have a ollow-up appointment scheduled prior to surement and improvement. Med Care. 2003;41:I-30-I-38.
discharge.” To make a measure operational is to de ne it in such
Donabedian A. Evaluating the quality o medical care. Milbank Q.
a way that di erent people assessing the measure would come
1966;44:166-203.
to the same answer. To express this measure as an equation, the
denominator could be de ned as all patients discharged rom the Nelson EC, Splaine ME, Batalden PB, Plume SK. Building measure-
hospital medicine service, regardless o the recommendations or ment and data collection into medical practice. Ann Int Med.
postdischarge appointments or where the patients will be seen. 1998;128:460-466.
The numerator could be de ned as only those patients discharged Pronovost PJ, Nolan T, Zeger S, Miller M, Rubin H. How can
rom the service whose ollow-up appointments are within 30 days clinicians measure sa ety and quality in acute care? Lancet.
o discharge and that are present on the discharge instructions 2004;363:1061-1067.
112
CHAP TER
18 CASE 18-1
Midway through a wound debridement, the scrub nurse noted
that the sterilization indicators had not changed colors (indicating
an inadequate sterilization process). A subsequent root cause
analysis revealed that the sterile processing technician, at the
end of his shift, forgot to push the button to start the autoclave.
The next arriving technician did not notice that the sterilization
indicator on the cart had not changed color, so he took the cart
Standardization and with the unsterile trays, and placed them on the shelf for use.
Reliability INTRODUCTION
In 1999, the Institute o Medicine (IOM) highlighted two studies
rom the 1980s, which suggested that between 44,000 and 98,000
Richard S. Gitomer, MD, MBA, FACP patients die every year due to preventable medical errors. The
subsequent IOM report, Crossing the Quality Chasm, noted, “The
current systems cannot do the job. Changing systems o care will.”
The report went on urther to describe the six aims o sa ety, e ec-
tiveness, e ciency, patient-centeredness, timeliness, and equity.
With these aims, the IOM has de ned the ultimate vision or the US
health care system.
The limitations o the current health care system were urther high-
lighted by Elizabeth McGlynn’s study in 2003, which demonstrated
that patients only received 55% o the care warranted by medical evi-
dence. Furthermore, they ound that the likelihood that an individual
patient would receive all appropriate care was only 2.5%.
HUMAN FACTORS
■ THE INDIVIDUAL
A main contributor to the per ormance short all is the limitation o
human per ormance. Table 18-1 illustrates expected human error
rates in conditions o no undue time pressure or stress. Note that
“under very high stress when dangerous activities are occurring
rapidly,” the error rate can be as high as one in our (25%). There ore,
system designs that depend on per ect human per ormance are
destined to ail at a very high rate. Furthermore, systems designed to
unction in conditions o high stress with requent dangerous activi-
ties have a higher burden in order to ensure a avorable outcome.
As de ned by the Federal Aviation Administration “human
actors entails a multidisciplinary e ort to generate and compile
in ormation about human capabilities and limitations, and apply
that in ormation to equipment, systems, acilities, procedures, jobs,
environments, training, sta ng, and personnel management or
sa e, com ortable, and e ective human per ormance.” When con-
sidering human actors, it is help ul to consider the impact o the
human and the impact o the system separately. Reliable systems
must compensate or the limitations o human per ormance. In
addition, organizational characteristics can negatively or positively
impact human per ormance. When redesigning systems to improve
per ormance, it is help ul to understand the actors that may nega-
tively impact human per ormance, so that the design can account
or the expected vulnerability.
It is help ul to understand typical human limitations, which allow
or and encourage the creation o systems that may accommo-
date or such limitations. The rst is human memory; on average, a
typical human is only able to keep 7 ± 2 elements in their short-term
113
unction reliably. Human error can be addressed rom an organiza-
TABLE 18-1 Nominal Human Error Rates for tional perspective with a “person approach” or a “system approach.”
Selected Activities The person approach ocuses on the actions o the rontline sta
who commit errors. The errors, it is believed, are due to f awed
P
Activity (Assume no Undue Time Pressure
A
or Stresses) Rate mental processes that can be voluntarily corrected, with enough
R
motivation, attention, and vigilance. The institutional response is
Error o commission, eg, misreading a label .003
T
ocused on correcting the variation in human behavior. Frequently,
Error o omission without reminders .01
I
the responses ocus on engendering ear (disciplinary measures,
Error o omission when item is embedded in a .003 threat o litigation, retraining, naming, blaming, and shaming), so
procedure that the individual will ocus more intently on the task at hand and
Simple arithmetic errors with sel -checking .03 not make a similar error. O ten, new policies and procedures are
T
h
Monitor or inspector ails to recognize an error .1 written to ensure the correct behavior. In short, the person approach
e
Personnel on di erent shi ts ail to check the .1 implicitly assumes that bad things happen to bad people.
S
p
condition o hardware unless directed by a checklist In the system approach, however, human allibility is a unda-
e
mental premise; human errors are to be expected. The errors are
c
Error rate under very high stress when dangerous .25
i
a
activities are occurring rapidly seen as consequences o inadequate system design. It is believed
l
t
y
that most errors occur because system barriers and de enses that
o
are “upstream” to the event, lead to the undesired outcome. The
f
Adapted rom Park K. Human error. In: Salvendy G, ed. Handbook of Human
H
Factors and Ergonomics. New York, NY: John Wiley &Sons, Inc.; 1997:163. undamental premise or countermeasures in the system approach
o
s
ocuses on changing the conditions under which humans work,
p
i
rather than changing the human condition. When an error occurs,
t
a
l
the ocus is on how the de enses ailed.
M
memory. Systems that rely on human memory, there ore, are inher- The person approach is somewhat appealing on at least two lev-
e
ently imper ect. Reliable systems provide key in ormation at the time
d
els. It is emotionally satis ying to blame an individual or an adverse
i
c
when it is needed, rather than relying on memory. Humans also natu- event. In addition, divorcing the unsa e act rom the organization
i
n
rally cut corners or create “workarounds” when eeling rushed. Over
e
may be perceived as being in the best interest o the organizational
time, repeated short cuts or workarounds result in a narrowing sa ety
a
leaders. But, these bene ts come at a great cost. In 90% o aviation
n
margin. This is described as “normalization o deviance.” The natural
d
maintenance mishaps, the worker is ound blameless. In order to
S
tendency to cut corners and the lack o detectable consequences improve it is important to per orm detailed analysis o incidents, near-
y
s
alsely reassures the individual that they remain within an appropri- misses, and unsa e conditions. Within a “blame and shame” culture,
t
e
ate level o sa ety, or reliability, which rein orces uture unsa e behav-
m
the in ormation is not voluntary reported, out o ear o retribution.
iors. Diane Vaughan rst coined this term in describing the root cause
s
Conversely, the system approach recognizes human allibility and
o
o the 1986 shuttle Challenger explosion 72 seconds a ter li t-o . The system designs are success ul in spite o human error. In reliable
f
C
cause was a ailed “O-ring” on one o the uel cells. The Report o the organizations, admission o errors and near-misses is encouraged
a
“Presidential Commission on the Space Shuttle Challenger Accident”
r
and rewarded. Leadership in reliable organizations realizes that early
e
described how normalization o deviance led to the disaster. “NASA detection o latent conditions that promote human error is essential
and Thiokol accepted escalating risk apparently because they ‘got to creating reliable systems.
away with it last time.’” As Commissioner Feynman observed, the As such, most organizations now create barriers and de enses to
decision making was “a kind o Russian roulette… (The Shuttle) f ies prevent the errors when designing systems o care. However, each
(with O-ring erosion) and nothing happens. Then it is suggested, barrier and de ense is not per ect and has unique vulnerabilities.
there ore, that the risk is no longer so high or the next f ights. We Reason describes this as a slice o Swiss cheese. However, these
can lower our standards a little bit because we got away with it last vulnerabilities are dynamic. Sometimes the holes are larger; some-
time… You got away with it, but it should not be done over and times they are located in a di erent place. For example, i one o
over again like that.” Normalization o deviance occurs because o the barriers is a second nurse checking the dose o insulin, this step
the natural human tendency to slip into believing that in spite o the may be less reliable on a speci c day i the checking nurse was sleep
short cuts, adequate sa ety or reliability margins remain. In health deprived or distracted.
care, normalization o deviance is o ten a barrier when trying to In well-designed systems, when one barrier ails, a second is
implement and sustain basic and quality and sa ety interventions, able to catch the de ect and the outcome is not compromised. In
such as the “Universal Protocol”or central line insertion bundles. Over Reason’s model, the de ect may pass through one slice o Swiss
time, as strict adherence to the protocols wane, and “nothing bad is cheese, but is caught by the next slice. However, there are times
happening,” then urther slippage and workarounds ensue. when all the holes o multiple slices o Swiss cheese line up and the
Stress also signi cantly impacts human per ormance by causing outcome is compromised … the patient is harmed or the process is
tunnel vision and ltering. This causes a loss o pattern recognition not executed reliably.
that humans use to rapidly discern complex situations. Fatigue neg- Implicit in Reason’s system approach to human error is the
atively impacts human per ormance by impacting both short-term importance o culture. Reason notes that high-reliability organiza-
and long-term memory. The impact o atigue is similar to having tions have a reporting culture. It is essential or the sta to eel sa e
a blood alcohol level o 0.1%. Other signi cant actors that impact sur acing errors and near-misses. In high-reliability organizations,
health care worker per ormance include multitasking, interruptions, leadership accepts the accountability to create sa e environments
and environmental actors (such a poor lighting, noise, distractions, that acilitate success ul outcomes. The sta member is accountable
etc). All o these natural human limitations must be actored into any to make sa e choices by ollowing the processes that they helped
system design (or redesign) intended to improve quality or sa ety. to create. The sta is also accountable or sur acing existing and
potential opportunities or de ects. I all live up to these account-
■ THE ORGANIZATION abilities, the leadership response to error is supportive o the sta ,
James Reason described characteristics that impact an organiza- with a ocus on identi ying the source o the de ect and developing
tion’s capacity to support or impede an individual’s ability to a remedy to prevent its occurrence in the uture.
114
RELIABILITY In highly reliable organizations, decision-making authority seam-
lessly f ows to the person with the best in ormation to make the
C
As suggested in Crossing the Quality Chasm, new systems o care
decision. The deference to expertise might be exhibited on an aircra t
H
are required to achieve the level o reliability necessary to ensure
carrier where a seaman can stop the activities on the f ight deck
A
that all patients receive the care they deserve. The new systems o
P
because he sees a condition that might be unsa e or the landing
care will necessarily need to account or human actors on the orga-
T
planes. In the ICU, a nurse might “stop the line” i she does not see
nizational level, as well as at the individual process level.
E
all elements o the central line insertion bundle, regardless o who is
R
the central line insertion practitioner.
1
■ THE ORGANIZATION Other organizations may not have all the complexities o health
8
care, but the ve principles o mind ulness are directly applicable to
The health care industry has a level o complexity that matches,
health care. Organizations that exhibit these characteristics have a
i not exceeds other industries. Challenging the reliability o the
S
t
culture that promotes reliability.
a
system is that humans execute most key processes. The complex
n
d
systems, the dynamic environment, and the human involved with
a
process execution are major reasons or the reliability gap described ■ THE RELIABILITY GAP
r
d
i
at the beginning o this chapter. Principles gleaned rom other com-
z
Karl Weick paints a compelling picture o what reliability looks like at
a
plex organizations can help overcome the reliability gap.
t
the level o the organization. But, what does reliability look like at the
i
o
Karl Weick described key organizational characteristics evident
n
level o the process? Reliability is intentional, and there are principles
in highly reliable complex organizations. The organizations studied
a
that guide that work. The lack o reliability in health care is multi ac-
n
include, nuclear power plants, nuclear aircra t carriers, and commer-
d
torial. Understanding these causes helps shape the interventions.
cial airlines. He describes those organizations as having mind ulness
R
While not exhaustive, the ollowing three explanations highlight key
e
or “a rich awareness o discriminatory detail.” Individuals unctioning
l
barriers to process reliability.
i
a
in high-reliability organizations are aware o context, can discrimi-
b
1. While readily acknowledging human allibility in others, many
i
l
nate details, and how the current situation di ers rom expectations.
i
t
health care providers expect per ection o themselves. In addi-
y
The ve principles o mind ulness include:
tion, there is o ten a eeling that the only way to ensure reliabil-
• preoccupation with ailure, ity is to rely on no one else. From this high standard comes an
• reluctance to simpli y interpretations, over-reliance on vigilance and hard work. The reality o human
• sensitivity to operations, actors, however, prevents the individual rom per orming
• commitment to resilience, reliably.
• de erence to expertise. 2. Individual providers tend to look at their personal delivery o
Preoccupation with failure is a relentless ocus on potential ailure health care one patient at a time. Ideally, the provider cus-
modes and how they can be prevented. An example in health care tomizes a plan or the individual patient based on personal
might be a detailed examination o an order set looking or ambi- experience and the medical evidence. However, due to the
guities or potential error traps. A preoccupation with ailure helps limitations o human actors, and the paucity o high-level evi-
overcome the natural tendency to dri t into unsa e behaviors that dence or much o the care o the average practitioner, there is
result in normalization o deviance. high variation in how patients are treated rom one to the next.
Mind ul organizations resist the tendency to normalize unwanted While understandable, this lack o standardization comes at a
occurrences into expected events. This reluctance to simplify inter- cost … complexity. It is no longer possible or an individual to
pretations helps to maintain margins o sa ety and process reliability. deliver the ull spectrum o reliable care to their patient. Teams
Inappropriate simpli cation o the causes o adverse events, or a lack are essential or the delivery o reliable care. The more variation
o ocus on these events, over time, can result in major problems rom patient to patient in a care plan, the greater the level o
that might have be avoided i the initial interpretation had been complexity or the rest o the care team. The consequence o
more rigorous. For example, resisting attributing a medication error this unwarranted variation and increased complexity is lower
to a nurse ailing to use the “ ve rights” because she did not ocus reliability. This is not to say that standardization in a “cookie
(right patient, right medication, right dose, right route, right time). A cutter” approach is required to reduce complexity. Rather,
mind ul organization might ask i the process or executing the “ ve standardize what is “standardizable.” This relieves the care pro-
rights” is robust, and i there are adequate protections to ensure vider o the mundane and allows ocus on those parts o the
completion o the “ ve rights.” care plan that does require additional ocus or expertise. For
Sensitivity to operations ref ects a deep understanding o the pro- the rest o the care team, their work is less complex because
cesses at a rontline level. This ref ects knowing what really happens o the reduced unwarranted variation and they are in a better
in the messy world o reality, not what is policy, or what is supposed position to plan and anticipate.
to happen. In response to low pneumococcal vaccine rates, leaders 3. A third reason or the reliability gap is that many current pro-
in a mind ul organization do not assume that the nurses are not cesses ail to account or human actors. Process design based
doing their job, but examine the process and appreciate that the on human in allibility is inherently unreliable. The ollowing
cause is an overly complex screening tool. section describes an approach to process design that acilitates
Highly-reliable organizations realize that all systems can ail. But, accounting or human actors.
these organizations have a relentless ocus on not allowing that ail-
ure to compromise per ormance. A key element o the commitment
■ MEASUREMENT
to resilience is that the speci cs are not anticipated. A preoccupation
with ailure might result in the development o a medical rapid The ramework in Table 18-2 describes speci c interventions asso-
response team to attend to deteriorating patients be ore they arrest. ciated with predictable levels o reliability. Interventions that result
This team’s capabilities can be urther enhanced through simulation in 10–1 reliability rely on vigilance and hard work. Examples o these
exercises that expose the team to diverse scenarios. The simulations interventions include:
help reduce complexity by allowing team roles to be de ned ahead 1) Common equipment, standard order sheets, multiple choice
o time rather than in the chaos o the acute event. protocols, and written policies/procedures,
115
other f oors and directly address the barriers unique to the
TABLE 18-2 Levels of Reliability in Typical other f oors, rather than trying to address the barriers while also
Processes Infrastructure determining the correct process or delivering the instructions.
The segments should have the ollowing qualities:
P
Level of Reliability Typical Processes Infrastructure
A
<80% Chaos • Be based on a design theme that helps simpli y the improve-
R
ment activity (eg, knowledge, geography, willingness to
T
>2 de ects in 10
participate, patient characteristics, etc).
I
10-1 (80%) No articulated common process
• Contain a reasonable volume, so there are enough opportu-
2 de ects in 10 Reliance on training and reminders nities to do tests o change.
10-2 (95%) Intentionally designed • Have clear-cut de ned boundaries so there is no con usion
1-5 de ects in 100 Utilizes principles o human actors about the population o patients being addressed.
T
h
engineering 2. Standardization: Standardization yields several key bene ts.
e
S
10-3 (99.5%) Well-designed system with attention First, it improves reliability by reducing complexity, as described
p
to process, structure, and outcomes
e
1-5 de ects in 1000 above. Second, it helps provide an in rastructure where roles
c
i
and responsibilities are clearly de ned. A standard in rastruc-
a
l
t
ture also allows or simpli cation o training, and competency
y
o
testing. Standardizing key processes allows consistent imple-
2) Personal checklists,
f
mentation o evidence-based medicine. Lastly, it simpli es
H
3) Feedback o in ormation on compliance,
o
identi cation o de ects that can be analyzed or redesign, and
s
4) Suggestions o working harder next time,
p
acilitates per orming tests o change. The standardization step
i
5) Awareness and training.
t
a
should be re ned by serial tests o change. The reliability goal
l
Processes that result in 10–2 reliability use principles based on
M
or the standardization step is 80%. I the step is less than 80%
e
human actors and reliability science. These interventions tend to be reliable, the de ects will overwhelm the detection and mitiga-
d
more resource intensive. They include:
i
tion step, which tends to be more resource intensive.
c
i
n
1) Decision aids and reminders built into the system, 3. Detection and mitigation: The value o the detection and
e
2) Desired action the de ault (scienti c evidence), mitigation step is that it helps reduce the complexity o the
a
n
3) Redundant processes utilized, standardization step. Since the standardization step need only
d
4) Scheduling used in design development, capture 80% o the opportunities, it is not necessary to develop
S
y
5) Habits and patterns known and taken advantage o in the contingencies or lower requency occurrences. I the stan-
s
t
e
design, dardization step had to include those contingencies, it would
m
6) Standardization o process, based on clear speci cation and be too complex and there ore less reliable. The detection
s
o
articulation. and mitigation step includes two processes. The rst process
f
reliably identi es the de ects rom the standardization step
C
a
utilizing 10–2 level interventions. The second process mitigates
r
■ THE PROCESS
e
the de ect, using 10–1 level interventions. An example might
The reliability design strategy was described by a group at the Insti- be detecting ailures by running an electronic report daily or
tute or Healthcare Improvement (IHI). It consists o our steps: all patients who have not been screened or administration
1) Segmentation, o the pneumococcal vaccine. Those ailures would then be
2) Standardization, mitigated by a supervisor contacting the nurses caring or the
3) Detection and mitigation, patients who experienced the process ailure.
4) Redesign the process based on the de ects identi ed. 4. Process redesign based on de ects identi ed: In pursuit o
continuous improvement, the ourth step in the reliability
1. Segmentation: When embarking on an improvement project it
design strategy is to examine the de ects identi ed in the
is o ten help ul to divide the population into smaller groups in
previous step and eed the learning back into the design o
order to simpli y the tests o change. Process improvement is
the standardization and detection and mitigation steps. Once
taking general knowledge or procedures and making them rel-
the processes in the rst segment achieve an acceptable level
evant in the local context. Unlike research, it is not possible to
o reliability, the team should move on to the next easiest seg-
remove con ounding variables. However, with segmentation
ment. Now that the processes have been de ned and tested,
the con ounding variables can be controlled until the appropri-
work in the next segment can ocus on the barriers unique
ate time.
to the new segment. Attacking the segments rom easiest to
To illustrate the power o segmentation, consider implemen-
hardest allows the team to gain experience and urther re ne
tation o congestive heart ailure (CHF) discharge instructions.
the process, which simpli es the work in even the most di cult
The improvement team might consider segments based on
segment.
amiliarity o the medical and nursing sta with the care o
CHF patients. Clearly, there is a greater level o com ort on
the cardiology f oor than on the orthopedic f oor. One might CONCLUSION
consider a continuum o com ort as ollows: cardiology f oor, As highlighted by the IOM reports and McGlynn’s work, patients are
general medicine f oor, general surgery f oor, and orthopedic regularly experiencing preventable harm and are not receiving all o
f oor. Segmenting this way, and choosing to start the work on the care that is intended. Closing that reliability gap requires ocused
the cardiology f oor allows the team to ocus on the process o multidisciplinary team involvement to develop and oversee the
delivering the discharge instructions without having to man- process redesign. Understanding human actors and having a struc-
age the barriers o knowledge de cit and com ort with the tured model or increased reliability is essential. Reliable process
medications. Once the process or delivering the instructions design, regardless o the improvement methodology, is an e ective
is well de ned and operational, the team can move on to the approach to closing the reliability gap.
116
SUGGESTED READINGS ONLINE RESOURCES
C
McGlynn EA, Asch SM, Adams J, et al. The quality o health care deliv- Crossing the Quality Chasm. http://iom.nationalacademies.org/~/
H
A
ered to adults in the United States. N Engl J Med. 2003;348:2635. media/Files/Report%20Files/2001/Crossing-the-Quality-Chasm/
P
Nolan T, Resar R, Haraden C, Gri n FA. Improving the Reliability of Quality%20Chasm%202001%20%20report%20brie .ashx
T
Health Care. IHI Innovation Series White Paper. Boston, MA: Institute IHI Reliability Online Resources. http:/ /www.ihi.org/ IHI/
E
R
or Healthcare Improvement; 2004. To p ics/ Re liab ilit y/ Re liab ilit yGe n e ral/ Em e rg in g Co n t e n t /
SegmentPresentationandDesignTable.htm
1
Reason J. Human error: models and management. BMJ. 2000;320:768.
8
Vaughan D. The Challenger Launch Decision: Risky Technology, Culture, Just Culture Resources. https://www.justculture.org/tag/david-marx/
and Deviance at NASA. Chicago, IL: University o Chicago Press; Resar RK. Practical Applications o Reliability Theory. http://high-
S
1996. reliability.org/Practical_Reliability_Resar.pd
t
a
n
Weick KE, Sutcli e KM. Managing the Unexpected: Resilient Perfor- To Err is Human. http://www.nap.edu/read/9728/chapter/1
d
a
mance in an Age of Uncertainty. 2nd ed. San Francisco, CA: Jossey-
r
d
Bass; 2007.
i
z
a
t
i
o
n
a
n
d
R
e
l
i
a
b
i
l
i
t
y
117
19
Every organization needs a structure and a tool kit to support
employee and patient sa ety and continuous quality improvement.
Since what cannot be seen cannot be xed, robust identi cation o
adverse events and latent sa ety threats (risk to patients, amilies,
and employees) should be a priority o every hospital. Here, we
discuss structures and tools to identi y adverse events and risk.
Once identi ed, the hospital and sta must then prioritize e orts
and determine which techniques will be applied to reduce adverse
Tools to Identi y events and risk. Risk reduction e orts may ocus on error preven-
tion or on harm prevention, and both approaches are key to risk
Problems and reduction. Table 19-1 reviews de nitions or the key terms used
throughout the chapter.
Reduce Risks THE ROLE OF THE CULTURE OF SAFETY IN

IDENTIFYING RISK
In organizations with a robust culture o sa ety, it is evident to all
involved that sa ety is the top priority. People working in the area
Elizabeth H. Mack, MD, MS ocus on sa e practices and support one another in delivering sa e
care. Individuals in a mature culture o sa ety exhibit the principles
o a highly reliable organization, including “preoccupation with
ailure” as described by Weick and Sutcli e, such that there is a
constant obsession with mitigation o risk. Instead o ignoring small,
nagging concerns, workers share those concerns with others, and
team members rally to help resolve the concerns. When the culture
o sa ety is strong, people report concerns without ears o retali-
ation. Employees respond supportively to concerns about risks to
patient sa ety and do not seek to blame individuals when an error
occurs. When sa ety is a priority, leaders routinely inquire about
sa ety concerns and take the time to listen, seek to understand root
causes, demonstrate their commitment to sa ety through action,
and communicate back to sta on the organizational response to
adverse events and concerns. An organization’s culture o sa ety
can be measured by the Agency or Healthcare Research’s survey
on patient sa ety culture. The scores o this survey have been cor-
related with patient outcomes in multiple studies and can assist an
organization with monitoring their culture over time. Speci c ques-
tions may also be use ul to assess the e ect o culture on identi ying
problems (Table 19-2).
RISK IDENTIFICATION AS PART OF A SAFETY PROGRAM

Identi cation o patient sa ety risks requires an intentional multi ac-
eted approach. The National Quality Forum identi es and promotes
sa e practices in health care (see http://quality orum.org). Among
these best practices is the presence o leadership structures and
systems to ensure organizational awareness o sa ety ailures and
per ormance gaps that need attention. Government and com-
mercial health care purchasers, insurance providers, and hospital
accreditation organizations provide strong incentives or hospitals
to invest in in rastructure supporting the identi cation, analysis, and
correction o risk. Less obvious sources o harm, including deviations
rom generally accepted practice standards such as evidence-based
guidelines o care, are also important to detect. For example, iden-
ti ying variation within a subspecialty o blood product ordering,
compliance with tonsillectomy guidelines, or ordering o computed
tomography scans or suspected appendicitis is likely to lead to insti-
tutional cost savings and improved quality o care. Because in or-
mation about errors and risks comes rom many potential sources,
118
Knowing that the detected errors are but a sample o true prevent-
TABLE 19-1 Common Terms Used to Describe Errors able adverse events, hospital sa ety leaders must decide whether
C
and Harm there is value to applying additional resources to enhance detection.
H
Bringing together the multiple sources o in ormation allows or a
A
Term Explanation
P
broader view, and this e ort bene ts rom having a central sa ety
Adverse event Harm resulting rom medical care
T
committee or sa ety o cer to whom the in ormation unnels.
E
Preventable Harm due to medical care that could have
R
adverse events been avoided
METHODS OF IDENTIFYING ADVERSE EVENTS
1
Latent sa ety Factors that make error more likely but are not AND ERRORS
9
threat directly visible; also called blunt error
Some methods are reactive and retrospective, being generated
Near miss Error occurs but does not reach the patient
in response to speci c events that come to attention. In contrast,
T
o
Sentinel event Unexpected death or serious harm or risk systematic methods tend to identi y latent or hidden errors or risk
o
thereo
l
points, and are o ten both reactive and proactive. Both types o
s
t
Adverse Error resulting rom medication-related approaches are necessary in the overall strategy to identi y risks.
o
medication interventions
I
Certainly, there are some errors that will never come to the attention
d
event
e
o leadership, either because they are not recognized as errors or
n
Slip Failure o execution in which the technical
t
because employees do not eel com ortable reporting them.
i
y
action is observable
P
Lapse Failure o execution typically due to memory
r
■ REACTIVE METHODS
o
ailure
b
Event reporting systems
l
e
m
Voluntary event-reporting systems rely on employees to take time
s
to report a concern, trusting that action will be taken by leaders,
a
n
hospitals ace challenges to bringing this disparate in ormation and that retaliation will not occur. In a healthy culture o sa ety,
d
together, investigating and mitigating risk, and providing eedback employees report reely and openly, with ew barriers. Employees
R
e
to employees about actions taken in a meaning ul way. do not ail to report events that seem to represent temporary or
d
u
One major challenge is that only a small raction o preventable minor areas o potential harm, because they are aware that these
c
e
harm is detected by most methods, especially those based on vol- are important opportunities to learn. Indeed, a robust reporting sys-
R
untary human reporting. Computerized data mining methods tend tem will collect a signi cant number o system ailures or errors that
i
s
k
to greatly increase the number o potential errors identi ed, but the were success ully intercepted or that did not result in harm, known
s
speci city can be low. For example, a rule to identi y nephrotoxicity as near misses. Near misses are golden opportunities to identi y
rom medications may look or a rise in the serum creatinine during risk-prone conditions or processes and to intervene be ore harm
hospitalization. Most o the cases identi ed will not result rom an results. The ideal state would involve occurrence o near misses
adverse drug event, and additional resources will be required to only, and no incidents o preventable harm. An e ective reporting
urther investigate each case. system enhances the engagement o rontline sta in patient sa ety
by providing an identi ed channel or their observations. To be
e ective, employees must be aware that the reporting system exists,
reporting must be user- riendly, and it must be known that leaders
TABLE 19-2 The E ect o Culture on Identi ying Problems place value on the reports. Timely acknowledgment, expressions o
appreciation, and eedback regarding actions taken rein orce the
Questions to Assess the Culture o Sa ety
desired reporting behaviors.
Are people com ortable reporting errors they witness? Ease o reporting is key to maintaining a low reporting threshold.
Are people willing to report their own errors? Paper reports and verbal reports via telephone recording have
Do sta members reely discuss their concerns about patient advantages o speed, though the in ormation has to be transcribed
sa ety? and aggregated separately. Lengthy written or electronic reporting
Are supervisors receptive to these concerns? systems with a large number o required elds are not likely to be
Do people ear retribution or reporting or ear being blamed or used widely. Electronic reporting systems may prompt or more pre-
adverse events? cise and complete in ormation rom each report and may produce
structured reports rom which data are more easily analyzed. Elec-
Do sta members eel that leaders care about patient and sta
sa ety? tronic systems may also enable immediate noti cation o appropri-
ate personnel. For instance, an event reported as causing signi cant
Can people give examples o actions to improve sa ety that
patient harm may generate an automated communication to a
resulted rom reports o adverse events or concerns?
risk manager, sa ety o cer, or hospital leader, acilitating a timely
Questions to Assess Leaders’ Commitment to Sa ety response to the event. O note, most states provide legal protection
Are leaders visible where patient care is delivered, asking sta or event reporting such that the in ormation contained with the
about their concerns or sa ety? reports is not discoverable.
Do leaders ensure that sta members do not su er retribution How leaders and managers respond to aggregate data rom
or reporting? event-reporting systems will send strong signals to employees.
Is there an identi ied patient sa ety o icer who reports to system Because event reports are dependent on willingness to report and
leaders or to the board? are unlikely to ref ect true incident rates, leaders and managers
Is sa ety regularly on the agenda at meetings o hospital should exercise caution in in erring that a high number o events
departments, medical sta , and leaders? reported represents worse sa ety in one area versus another. In
Are adequate resources applied to identi ication and analysis o act, the number o reports may be more indicative o the culture
adverse events and risks? o sa ety than o sa ety itsel . However, it is human nature to con-
clude that higher numbers o reports indicate an area o unsa e
119
conditions. Managers may worry that reports ref ect poorly on their o a culture o sa ety. As described by Frankel et al, structuring the
per ormance and discourage use o the reporting system. Leader- content o rounds and recording the comments and concerns o
ship, particularly o the “C-suite” and the board o directors, should sta members may yield valuable insights. Past events deserving
pay attention to the use o data or learning rather than or judging o investigation and concerns about ongoing risks can be heard
P
A
in order or the reporting system to be e ective. directly by the leadership team. These reports stimulated by the visit
R
The options or reporter identity protection deserve intentional o leaders may not have been collected through other means. It is
T
thought when designing an event-reporting system. An open important to provide eedback to employees on items discussed in
I
system makes no attempt to protect the reporter’s identity, so sa ety rounds. I possible, it is use ul to develop a rotating schedule
colleagues and supervisors may identi y the reporter. This kind o to include all shi ts and all areas.
system may work in organizations where the culture o sa ety is
strong enough that there is no retribution or reporting and, in MORBIDITY AND MORTALITY CONFERENCES
T
h
act, reporting is rewarded, whether by peer appreciation or ormal
Morbidity and mortality (M&M) con erences are a time-honored
e
recognition. It is worth noting most health care organizations are
S
tradition in medicine. Discussion o patient deaths, complications,
p
not yet at this point in their sa ety culture journey. I reporters ace
or harms resulting rom care provides a learning opportunity or
e
criticism or retaliation in even a ew instances, willingness to report
c
the attendees. Too o ten, perhaps due to the con dential nature o
i
a
may be severely a ected.
l
t
many con erences, the learning stops at the door, and eedback to
y
A con dential reporting system allows identi cation o the
the larger system does not happen. Traditional M&M o ten involves
o
reporter only to responsible system administrators who will ollow
f
individual blame rather than a systems-based approach. Addition-
H
up on the event with the reporter. Con dential reporting may over-
o
ally, the discussion o ailures may be inhibited i the culture is not
s
come reluctance o some people to report and enhance detection
p
open, and the lack o a structured process o case review may lead to
i
o sensitive issues. For example, a person reporting inappropriate
t
wide variation in the conclusions drawn rom case review.
a
l
sexual comments or behavior may be reluctant to have his or her
Structuring case reviews, identi ying underlying causes o ailures,
M
identity known to the person whose behavior is being reported.
e
involving organizational quality and risk leaders, and assigning
d
Anonymous reporting serves to ully protect the reporter’s iden-
i
responsibility or making system changes can make the M&M con-
c
tity and thus may expand reporting o sensitive events and reporting
i
n
erence bene t the larger system.
e
in work areas where a climate o ear exists. I inadequate in orma-
a
tion was submitted to identi y the event, however, urther investi-
n
TRIGGER TOOLS
d
gation and learning are severely hampered. To the other extreme,
S
i the reporter gave ull in ormation in a detailed report and ew Adverse events and errors that sur ace through reporting systems
y
s
people are ully aware o the event, it may be impossible to maintain and complaints poorly represent the rate at which such problems
t
e
actually occur. Structured reviews o medical records can identi y
m
anonymity. Regardless o whether reporting system is open or con-
problems that have not been reported and provide a rate estimate.
s
dential, it is important in most health care organizations to allow
o
or the option to report anonymously. Where necessary to provide One such method has been developed by the Institute or Health
f
C
or anonymous reporting, e orts to improve the culture o sa ety Care Improvement. Triggers are explicit criteria or clues to the
a
presence o an adverse event, stimulating a deeper review o the
r
will ideally allow movement toward a more open reporting system.
e
record. For example, administration o naloxone as a trigger is likely
PATIENT COMPLAINTS AND CONCERNS to detect some causes o nonpreventable harm as well as prevent-
able harm. Errors in war arin management may be identi ed among
Patient complaints and concerns are another source o reports,
patients in whom vitamin Kis administered. Most commonly trigger
whether submitted by patients or amilies or by employees on
tools are used to detect adverse drug events, but they can also be
behal o them. The perspective o patients and amilies may be
used to detect other causes o potential preventable harm. Apply-
very valuable and complementary to the insights o sta members.
ing the tool to a random sample o hospital discharges can give an
These reports are more likely to ref ect the level o service, compas-
estimate o the rates o common sources o harm to patients. These
sion, communication, and partnering with patients and amilies.
rates can be tracked over time, and the in ormation obtained may
The reports are not likely to be anonymous and bring with them a
be used by the hospital to set improvement priorities.
duty to respond back to the patient or amily. As with reports rom
sta members, treating the in ormation as a learning opportunity
may shape how rontline workers respond to complaints. While the
■ MONITORING OF HIGH-RISK PROCESSES
volume and content o patient complaints and concerns may pre- Some processes o care are inherently risky, such as provision o
dict the likelihood o legal action, it may be di cult to draw actual sedation, per orming invasive procedures, and responding to car-
conclusions about patient sa ety rom rates o complaints. diopulmonary arrests. Particularly concerning are the low- requency,
high-risk events. Monitoring the processes o care and the outcomes
CLAIMS ANALYSIS may identi y deviations rom standard practice that may pose undue
risks and point to opportunities or improvement. Using the example
Analysis o medicolegal claims may be a tempting source o in or-
o cardiopulmonary resuscitation (CPR), the hospital may have a
mation, but it is less likely to generate use ul ideas about how to
mechanism to review that the right complement o personnel
improve patient sa ety, compared to other methods. Since the vast
responded to the emergency, that the team was alerted in a timely
majority o patients harmed through medical care do not bring
ashion, that a team leader was identi ed and roles assigned, that the
claims, this subset is idiosyncratic. Deep understanding o any given
patient was correctly assessed, that the correct resuscitation algo-
event degrades rapidly with time, and investigation o an event
rithms were ollowed, and that the medications and other therapies
should have occurred long be ore a claim is led.
were correctly administered. The results o CPR reviews and the out-
comes o the resuscitation attempts are then reported to the appro-
■ SYSTEMATIC METHODS
priate hospital representative or committee. Debrie ng immediately
Patient safety walk rounds a ter such events is likely to yield discovery o areas or improvement.
Scheduled rounding in patient care areas by leaders may be an In a hospital with a robust culture o sa ety, the insights rom
e ective method to accomplish several goals, including promotion monitoring are shared and lead to actions to address de ciencies.
120
Suppose, or example, resuscitation monitoring identi es that team members to share a mental model o the process steps.
patients with di cult airways are not managed as well late at night From the high-level diagram, the team can develop a detailed
C
as during the daytime hours when an anesthesiologist is available. understanding o the di erent steps and supporting processes.
H
This insight should be reviewed by hospital leaders who determine
A
When important gaps in the process cannot be lled in by
P
the actions needed to provide better airway management at night. team members, additional in ormation is sought.
T
One solution may be to train and certi y hospitalist physicians to 4. Identi y the ways that ailure may occur ( ailure modes).
E
manage di cult airways. Additionally, processes may be monitored Drawing upon the expertise o team members and upon data
R
using open auditing such as observing the surgical sa ety checklist where available, the team identi es vulnerabilities and underly-
1
preoperatively or using secret shoppers posing as amily members ing causes among the process steps. Failure modes may occur
9
to audit hand-hygiene practices. Health care processes must be under normal conditions or when the system is stressed.
monitored using multiple di erent methods to assure compliance. 5. Prioritize the ailure modes. The team must decide where to
T
o
ocus attention or improvement or process redesign among
o
■ IN SITU SIMULATION the ailure modes ound in the previous step. To prioritize, most
l
s
FMEA models apply a grading scale (eg, 1-10) to several aspects
t
Both Institute o Medicine reports, To err is human and Crossing
o
o each ailure mode and its causes:
I
the Quality Chasm, suggested that health care utilizes both crew
d
Severity: The e ect o each ailure is described and assigned
e
resource management and simulation to improve patient sa ety.
n
a grade. A low grade indicates a ailure that would have minor
t
In situ simulation involves practicing team-based care on actual
i
consequences or is easily recoverable; a high-grade ailure
y
patient care units and this orm o training provides a method to
P
identi y latent sa ety threats. e ect is catastrophic (would cause grave harm and cannot be
r
o
stopped once the ailure occurs).
b
l
Frequency: The probability o occurrence o the ailure mode
e
■ MINING ELECTRONIC DATA
m
or its causes is also graded. Low-grade requency suggests a
s
Health care acilities are rich in electronic data that can be mined or
very rare event and high grades suggest more requent events.
a
possible adverse events. Every hospital has administrative systems
n
Risk o escaping detection: Failures that are immediately
d
to support billing. Among the diagnosis codes are those that may
obvious allow or early detection and the opportunity to
R
indicate sa ety problems, such as codes or accidental puncture or
e
recover or to mitigate the e ects. Easily detected ailures
d
laceration o an organ and or oreign body le t in during a pro-
u
receive a low score or risk o escaping detection. Other ailures
c
cedure. Since October 2008, hospitals routinely indicate whether
e
may be detected through routine inspection some time a ter
conditions were present on admission. Reviewing diagnoses such as
R
occurrence. The highest score goes to ailures that are not
i
s
pressure ulcers and deep vein thromboses that were not present on
k
detected until the outcome ( ailure e ect) has happened.
s
admission may identi y opportunities or improvement.
Multiplying the grades o severity, requency, and risk o
Electronic laboratory and pharmacy systems can be used to
escaping detection or each ailure mode produces a risk prior-
identi y potential adverse events or urther review. Acute renal ail-
ity number. The ailure modes can then be ranked, with the
ure occurring a ter admission to the hospital may be identi ed by
highest risk priority numbers indicating where the team should
searching the laboratory data or patients in whom the serum cre-
ocus attention. Because this ranking process is imprecise, the
atinine has risen signi cantly. Practically, however, o ten the needed
results provide a guide rather than a prescription or the next
data are available within the electronic health record, but there are
steps.
not enough resources available to satis y the many demands or
6. Determine the action steps. I the FMEA leads the team to
data throughout the system.
the conclusion that the process is ar too unsa e to continue,
a complete redesign may be necessary. Or, the team may
■ FAILURE MODES AND EFFECTS ANALYSIS
conclude that bringing in a new piece o equipment is unjusti-
Failure modes and e ects analysis (FMEA) is a prospective technique ed given the associated risks. More commonly, the team will
to anticipate the ways in which a process may ail and to prioritize identi y ways to eliminate causes o ailures, to provide earlier
the e orts to prevent ailures. With roots in the military, FMEA is warning o ailures, or to mitigate the e ects o ailures in ser-
widely used in manu acturing and more recently is being applied vice o improving overall sa ety.
to health care delivery. While there are a number o models, includ-
ing health care ailure modes and e ects analysis used in the Veter-
ans Health Administration, FMEA generally ollows these steps. ■ RETROSPECTIVE INVESTIGATION OF EVENTS ROOT
1. Choose the target or analysis. Because conducting FMEA CAUSE ANALYSIS
requires expertise and the commitment o signi cant time and Root cause analysis (RCA) is complementary to FMEA. It is a retro-
resources, a hospital must select where to apply the technique. spective technique that provides a robust structure to review an
A requent consideration is to ocus where prior problems have adverse event or near miss. RCA is reactive; FMEA is proactive. But,
occurred. For example, an institution might use FMEA to iden- the ultimate goal o each technique is to identi y ways to prevent
ti y and prevent errors in the placement and use o nasogastric uture adverse events. See Table 19-3 or a comparison o these
tubes. FMEA can help a hospital sa ely implement a new tech- two techniques.
nology such as bar coding the steps o medication preparation, E ective RCA requires a detailed, intimate understanding o the
distribution, and administration. event being studied and necessitates the participation or interview
2. Assemble the team. To understand the process in ne detail, o people directly involved in the event. Because RCA is usually per-
it is critical to identi y and involve team members with exper- ormed in response to a recent adverse event, emotions o people
tise rom all disciplines that use, interact with, or maintain the involved may be ragile. A poorly per ormed RCA that permits blam-
process. A trained acilitator should be part o the team, as ing o individuals can seriously undermine the culture o sa ety. RCA
FMEA has a speci c structure. done well results in learning the underlying (root) causes o human
3. Describe the process to be analyzed in detail. This is best ailures, process ailures, or equipment ailures. Focus on correcting
done graphically, beginning with a high-level f ow diagram the root causes o ailure osters sa er care o patients and a sa er
that serves as a ramework or analysis. The f ow diagram allows work environment or employees.
121
or experienced in the technique. Setting the atmosphere by
TABLE 19-3 Comparison o FMEA and RCA providing an orientation to the process and laying out ground
Failure Modes and Effects rules or the conduct o the RCA can be critical to success.
Among the ground rules should be a prohibition against
P
Analysis (FMEA) Root Cause Analysis (RCA)
A
nger-pointing and personal attacks. While humans may have
Prospective technique to Retrospective technique to
R
predict the ways a process or analyze an incident or the erred, directing blame at an individual stif es learning. Rather,
T
equipment may ail and to underlying causes o ailure or every ailure point, ask “why” and “what conditions existed
I
plan prevention e orts and to identi y potential to permit this ailure?” Human error alone should not be an
solutions acceptable root cause. Another use ul ground rule is to avoid
Steps: Steps: speculation. Where gaps in understanding occur, the team
should seek additional in ormation. It is also important to
1. Choose target or analysis 1. Assemble the team
T
h
discuss up ront whether peer review statues will protect the
e
2. Assemble the team 2. Set the atmosphere in ormation discussed, and whether or not the in ormation will
S
3. Describe the process in 3. Describe the events in detail
p
be discoverable.
e
detail 4. Identi y root causes
c
3. Describe the events. Create a detailed f ow diagram or a
i
a
4. Identi y the ailure modes 5. Identi y solutions to prevent timeline o the sequence o events. The team seeks additional
l
t
y
5. Prioritize the ailure modes recurrences in ormation to ll in gaps in understanding.
o
4. Identi y root causes. Using structured questions as in the
f
6. Determine the action steps 6. Report indings and
H
recommendations to triggering and triage questions o the VA National Center or
o
s
leaders Patient Sa ety (see http://www4.va.gov/NCPS/rca.html) cre-
p
i
ates a more complete analysis by prompting consideration
t
a
l
o categories o causes, including environmental conditions,
M
equipment unction, policies and procedures, training, commu-
e
d
nication, and atigue. Again, digging deeper with each question
i
Conducting RCA requires a trained acilitator and an investment
c
and not accepting human error or procedural violation as a root
i
n
o time and resources. Hospitals will have to select where to ocus
e
cause are essential to identi ying preventable causes.
this tool. Serious adverse events, such as the unexpected death o
a
5. Identi y solutions to prevent recurrences. Using standard-
n
a patient related to an error, are obvious targets or RCA. However,
d
ization and reliability science will create more robust actions.
near misses that reveal a potentially serious process ailure should
S
Consider where similar vulnerabilities exist in the organization
y
also be considered or RCA. A patient who recognizes that his
s
and generalize the learning where possible.
t
e
chemotherapy has been mixed incorrectly because it is the wrong
m
6. Report f ndings and recommendations to leaders. This step
color may have prevented a serious medication error. RCA o this
s
will help secure leadership support or actions needed.
o
error may reveal weaknesses in the chemotherapy mixing process 7. Follow-up. It is important to complete the action items
f
C
that need to be xed. The Just Culture algorithm would encourage discussed in the RCA within a speci ed time period. Teams
a
us to handle investigation o events independent o the outcome
r
may consider stressing the system using in situ simulation to
e
that resulted. Failure in duty to ollow a procedural rule would result determine i the same vulnerabilities still exist or i the solutions
in the same treatment o the employee whether the patient died implemented have closed the gaps.
or caused no harm. Root cause analysis will generally ollow these
steps:
INCIDENT INVESTIGATIONS
1. Assemble the team. Some hospitals have an existing team
Most adverse events and near misses will not be investigated with
or conducting RCAs. This team rarely has rsthand knowledge
an RCA, simply because the hospital is not capable o responding to
o the event being reviewed and will probably not have an
all with this level o investigation. Still, the personnel per orming the
intimate understanding o the ocused work processes and
investigations should be seeking root causes o problems, detecting
environment o those involved in the event. The team will
trends, identi ying solutions, and generalizing to other areas. The
add members with such knowledge or will gain that knowl-
results o incident investigations should eed back into the sa ety
edge through extensive interviews. More commonly, a team
structure o the hospital through the sa ety committee or sa ety
is brought together speci cally or the RCA. In addition to the
o cer so that urther action may be taken. Providing eedback to
trained acilitator, the team should be interdisciplinary and
reporters is a key.
involve people who work closely with the processes being
evaluated. Some experts advise against including people
directly involved with the event as team members because o PEER REVIEW
their potential di culty with being objective and open in such Peer review processes are intended to adjudicate the competence
a orum. I those with direct involvement are invited to join the and pro essionalism o health care providers. As such, peer review is
RCA team, the acilitator must be sensitive to this conf ict and undamentally di erent rom the investigations already described.
create an atmosphere conducive to openness. I employees This approach ocuses not on problems in the health care system,
who were directly involved in the event do not participate but on individual per ormance. Peer review usually occurs as a phy-
in the RCA, they should be interviewed in depth in order to sician or other provider is newly hired or granted speci c hospital
understand the processes, culture, and root causes. Hospital privileges to con rm competence or in response to some concern
leaders, when possible, can strengthen team unction by raised about per ormance. Concerns may arise in response to a
participating and supporting improvement opportunities that particular incident or in response to data compiled over time. For
result rom the RCA. I a leader does not participate directly, instance, an unexpectedly high rate o procedural complications or
knowing that a leader will closely ocus on the team recom- resource utilization compared with peers may prompt closer review.
mendations can also lend weight to the RCA. Using a construct such as Just Culture may acilitate a air treatment
2. Set the atmosphere. In teams brought together or the RCA, when the temptation is to punish the employee in the ace o a bad
the leader or acilitator may be the only person ormally trained outcome.
122
Peer review may generate insights about an error-prone process Reacting quickly to every danger, though, may cause loss o ocus
or other system problems that can be ed into the hospital sa ety or on higher-priority risks. Ill-considered solutions can overburden
C
quality improvement structure. Judgments about the pro essional employees to the point o paralysis or can introduce additional
H
competence o individuals are managed through hospital medical
A
harms. As an example, computerized order entry systems permit
P
sta governance structures. checking or drug-drug interactions as a sa ety eature. System
T
As a strategy or reducing errors, peer review ollows a “person administrators can select the level o interaction (severe, moder-
E
approach” that is weak to the extent it assumes that human com- ate, minor) at which an alert interrupts the prescribing process. I ,
R
petence and behavior are responsible or most errors. Reliance on in response to an adverse drug event, the level is set to include all
1
human per ection f ies in the ace o what is known about human potential interactions, the number o interruptions may overwhelm
9
per ormance in a variety o tasks, and the personal approach to prescribers, creating alert atigue. As a result, cognitive errors may
error reduction undermines a culture o sa ety when individuals are increase and prescribers may ignore more serious drug alerts. The
T
o
blamed or errors and outcomes. The goal o peer review should net e ect may well be to increase risks.
o
be to ensure that health pro essionals are able to unction on a
l
s
level commensurate with the speci c privileges granted. A “system
t
o
approach” to error reduction assumes that most errors result rom CONCLUSION
I
d
f aws in the system in which people work. The ocus turns to creat- A robust sa ety program helps to manage the complex and danger-
e
n
ing a system that recognizes human allibility, discourages and does ous inpatient hospital environment. It sets strategies and tactics or
t
i
not acilitate workarounds, and prevents harm to patients. engendering a sa e culture, identi ying and investigating harms and
y
P
Peer review also encompasses the behavioral norms expected o risks, and prioritizing improvement e orts. The tools described in
r
o
health pro essionals. Behavior that is disruptive to sa e patient care this chapter support the execution o such a robust and compre-
b
l
or that threatens the sa ety o employees, such as verbal abuse o hensive sa ety program.
e
m
employees, lying about events in the care o patients, throwing pub-
s
lic temper tantrums directed at others, or making unwanted sexual
a
ACKNOWLEDGMENT
n
advances, must be dealt with by leaders. Failing to do so contributes
d
to ear among employees and distrust o leaders, undermining the The author would like to acknowledge Nathan Spell, MD, or his
R
e
culture o sa ety. contribution to the rst edition chapter.
d
u
c
e
RESPONDING TO IDENTIFIED ERRORS AND RISK POINTS SUGGESTED READINGS
R
i
s
■ PRIORITIZATION
k
Berenholtz SM, Hartsell TL, Pronovost PJ. Learning rom de ects to
s
Health care delivery is an inherently complex and dangerous eld. enhance morbidity and mortality con erences. Am J Med Qual.
Patients enter hospitals with conditions that may be either known 2009;24:192-195.
or unclear. The methods we apply to diagnosis and treat patients DeRosier J, Stalhandske E, Bagian JP, et al. Using health care
involve sharp objects, ionizing radiation, and toxins, delivered in a ailure mode and e ect analysis: the VA National Center or
team setting where we must plan and communicate clearly across Patient Sa ety’s prospective analysis system. Jt Comm J Qual Saf.
pro essional disciplines and across multiple hando s. The challenge 2002;27(5):248-267.
then lies in deciding which problems to tackle rst and which prob-
Frankel A, Graydon-Baker E, Neppl C, et al. Patient sa ety leadership
lems to set aside. When creating new processes or introducing new
walkrounds. Jt Comm J Qual Saf. 2003;29(1):16-26.
equipment, leaders must take the opportunity to apply FMEA or
other strategic sa eguards so the number o potentially chaotic and Gri n FA, Resar RK. IHI Global Trigger Tool for Measuring Adverse
unsa e processes does not continue to rise exponentially. Events. 2nd ed. IHI Innovation Series White Paper. Cambridge, MA:
Frontline employees bring their expertise about the risks they Institute or Healthcare Improvement; 2009.
encounter. They should collaborate with the sa ety o cer or com- Kohn LT, Corrigan JM, Donaldson MS, eds. To Err is Human: Building
mittee that has a broader view o the problem areas in the hospital. a Safer Health System. Washington, DC: National Academies Press;
The broader view comes rom ormal assessments o high-risk areas 2000.
or processes, as with the FMEA technique, and rom the accumu- Reason J. Human error: models and management. BMJ
lated experiences o risks and errors collected rom the reporting 2000;320:768-770.
system and other methods already discussed. When appropriate,
local problems can be addressed with local solutions by the people Root Cause Analysis page. VA National Center or Patient Sa ety web-
who do the work. The hospital sa ety o cer or committee will site. http://www.patientsa ety.va.gov/media/rca.asp. Accessed
decide where and how to apply the more resource-intensive e orts October 10, 2015.
to improve sa ety. Weick KE, Sutcli e KM. Managing the Unexpected: Assuring High Per-
formance in an Age of Complexity. San Francisco, CA: Jossey-Bass;
■ AVOIDING OVERREACTION AND UNINTENDED 2001.
CONSEQUENCES Wheeler DS, Geis G, Mack EH, LeMaster T, Patterson MD. High reli-
In the ace o an adverse event, it can be tempting to apply intense ability emergency response teams in the hospital: improving
ocus and impose a quick solution. Indeed, when a serious continu- quality and sa ety using in situ simulation training. BMJ Qual Saf.
ing risk o harm is ound, an immediate sa eguard is appropriate. 2013;22:507-514.
123
CHAP TER
20 INTRODUCTION
In November 1999, the Institute o Medicine (IOM) issued the report
To Err is Human, detailing a problem o preventable medical errors
that were killing as many as 98,000 inpatients per year. Subsequent
publications have estimated it may be as high as 400,000 per year.
Speci c types o medical errors highlighted in the IOM report
include error in the administration o treatment, ailure to order
and ollow-up on indicated diagnostic exams, and avoidable delays
in care and treatment. Many years later problems still exist: nearly
Preventing and 2 million patients a year develop in ections during their hospital-
izations, and 90,000 to 100,000 o those in ected die, while hand-
Managing Adverse hygiene rates range rom 30% to 70% at most acute care acilities.
The IOM report also estimated that medical errors cost the US
Patient Events: $17 billion to $29 billion a year, and called or sweeping changes to
the health care system to improve patient sa ety.
Improvements in patient sa ety have ocused on addressing the
Patient Sa ety and root causes o these preventable patient harm events, speci cally
events related to poor communication, lack o teamwork, ragmen-
the Hospitalist tation o care, and a lack o leadership rom the medical community.
In addition, patient sa ety experts have also implored physicians
and hospitals to approach patient harm events with transparent,
open, and honest communication between caregivers and patients
and amilies in order to learn rom mistakes and poorly designed
Timothy B. McDonald, MD, JD
systems.
This chapter reviews ways in which hospitalists may actively par-
ticipate in the prevention o patient harm and provide appropriate
management and assistance when patient harm does occur.
PREVENTING ADVERSE PATIENT EVENTS

Most patient sa ety experts would agree that the areas o highest
priority to proactively maximize patient sa ety t into three broad
domains: communication, teamwork, and leadership. Within each
o these domains lie critical concepts and issues about which the
highly reliable and sa e-practicing physician must remain mind ul.
■ COMMUNICATION
No chapter on the prevention o patient harm is complete without
a major ocus on the role communication—or lack thereo —plays
in serious patient sa ety events. The most common types o com-
munication o high priority in patient sa ety are listed in Table 20-1.
Handoffs
Year a ter year The Joint Commission (TJC) publishes data showing
65% to 70% o all sentinel events are rooted in communication break-
downs. It appears that since the implementation o the Accredita-
tion Council or Graduate Medical Education resident physician
work hour limitations the communication problems have increased,
especially in the area o handof s, when the responsibility o care is
passed rom one provider to the next. With this limitation o resident
physician work hours, the need and demand or hospitalists to “ ll
the gaps” in patient care has increased substantially. Associated with
that increase in demand, hospitalists in particular have recognized
the imperative o a standardized, user- riendly, and reliable method
o handing o care rom one provider to the next.
The content and process or handing o in the inpatient setting
has evolved as practitioners try to meet regulatory requirements
while maintaining simplicity, e ciency, and usability o the various
124
sa ety and legal perspective, it is also incumbent upon the hospital-
TABLE 20-1 Preventing Patient Sa ety Events ist to acilitate the correction o erroneous in ormation encountered
C
in the EHR.
H
Communication
A
Hando within and between services
P
MANAGEMENT OF CRITICAL TEST RESULTS
T
Document in the electronic health record Critical test results management cuts to the heart o the health care
E
Manage critical test results
R
business. US hospitals complete approximately 12 billion diagnostic
Teamwork tests every year. Most test results are within normal range and do
2
0
Multidisciplinary rounds not require ollow-up by the clinician. However, a small but impor-
In ection prevention tant number o test results, approximately1% to 5% o a hospital’s
test volume, are abnormal or critical. Hospitals and hospitalists have
P
Patient triage
r
a pro essional, legal, and ethical obligation to ensure that these
e
v
Rapid response teams results are communicated to the responsible physician and appro-
e
n
Leadership priate action is taken.
t
i
n
Hospital/Medical Center Committees Traditional systems to communicate and manage critical results
g
Sa ety culture are ull o potential points o ailure. In many hospitals, especially or
a
n
hospitalists, contact in ormation changes on a regular basis. Radiol-
d
ogy departments and the pathology lab may not have the correct
M
contact in ormation or the responsible physician. Faxes can be
a
n
equally problematic as the receiving machine might be o or out o
a
hando tools that are available. Various pneumonic tools, such as
g
paper. And once communicated, the right person might not receive
i
n
Situation, Background, Assessment, and Recommendation have
g
the in ormation. Un ortunately, radiologists and laboratory techni-
been devised to assist in the hando process but have come and
A
cians may spend hours or days trying to track down the appropriate
gone rom institutional policies and guidelines as providers struggle
d
physician or results communication. Not surprising, miscommuni-
v
with a reliable way to meet this important imperative. Hospitalists
e
cation o critical ndings have been identi ed as the causative actor
r
must play a role in designing and implementing a best practice
s
e
in 85% o radiology lawsuits. Appropriately, The Joint Commission
hando process appropriate or the context in which they work.
P
has deemed the management o critical test results as a national
a
Vendors o electronic health records (EHRs) have also entered the
t
i
patient sa ety priority and requires hospitals and health care pro es-
e
arena with EHR-based tools to acilitate the o ten onerous process
n
sionals to improve processes involved in such results. To improve
t
o handing o care o large numbers o patients. Regardless o the
E
the sa ety and quality o care their patients receive, hospitalists must
chosen method, all hospitalists must employ a reliable process to
v
e
play an integral role in the design and implementation o systems
transmit necessary patient in ormation rom physician to physician.
n
and processes to manage critical test results. At a minimum, in the
t
s
De ective or unreliable hando s substantially increase risk o patient
:
hospital setting they must actively participate in a process to ensure
P
harm and the associated liability.
a
the proper identi cation o responsible physicians and an e cient
t
i
e
means or involving those responsible physicians in the communi-
n
Documentation in the electronic health record EHR
t
cation and action based upon these results.
S
With the passage o the 2010 Health Care Re orm Act, it has become
a
increasingly clear that the use o electronic health records will
e
■ TEAMWORK
t
become much more ubiquitous in the coming years. While patient
y
a
sa ety bene ts o EHRs are well documented, only recently have
n
MULTIDISCIPLINARY ROUNDS
d
in ormatics experts been publishing the unintended, unsa e conse-
Data abounds on the value teamwork brings to the sa e and e ec-
t
quences related to their use.
h
tive delivery o health care. The days o a single physician e ectively
e
One o the most glaring examples o an unintended, unsa e con-
H
micromanaging a patient’s entire hospital stay are long gone.
sequence to EHR implementation is the abuse o “cut and paste” or
o
Research has shown that physicians can mitigate the negative
s
“copy and paste” unctionality, the process by which entire sections
p
e ects o the necessary ragmentation o health care delivery by
i
t
o nursing or physician documentation are copied and pasted rom
a
participating in multidisciplinary rounds during which physicians,
l
past to present notes. Numerous published reports demonstrate
i
s
nurses, pharmacists, and other allied health pro essionals discuss
t
cases in which erroneous in ormation has propagated, almost
“virally,” throughout a patient’s EHR through the use o copy and the daily plan or the patient and coordinate the transition o care to
paste. This process creates unsa e conditions or the patient such as the outpatient setting. Inclusion o the patient and amily in these
in the example o the erroneous propagation o a “ aux” allergy to rounds also provides bene t.
an important medication. Serious medical-legal consequences can Especially or complicated patients, multidisciplinary rounds
result or those who continue to misrepresent medical in ormation have been demonstrated to reduce length o stay, decrease the
through subsequent “copies” o the erroneous in ormation or or incidence o medication errors, prevent hospital readmissions, and
those who act upon this unreliable in ormation. The credibility o improve overall patient and amily satis action related to the hospi-
physicians comes into question when they are orced to de end mis- tal stay. As health care reimbursement models transition to a “pay
in ormation they have propagated throughout the medical record, or per ormance” or “pay or quality” metric, hospitals and hospital-
such as a temperature o 1101.5°F or a blood pressure o 1180/60. ists will nd multidisciplinary rounds undamental to the business
While the “copy and paste” unctionality provides use ul e - model o health care.
ciencies or documentation o long lists o medications or past
surgical procedures, hospitalists must be aware o the deleterious INFECTION PREVENTION
consequences o the inappropriate use o this unctionality and they All patient sa ety and quality organizations as well as health care
should serve as positive role models and mentors throughout the regulators have identi ed health-care-associated in ections (HAIs)
organization or others who document in the EHR. From a patient as a top priority. The human and nancial toll o HAIs accounts or
125
a large portion o preventable harm in the United States. There ore, From the skills perspective, a success ul hospitalist RRT team
the elimination o any signi cant proportion o HAIs is paramount leader must demonstrate competence in basic airway skills, use o
to control health care costs and preventing patient harm. The role emergency medications, and the ability to interpret electrocardio-
o the hospitalist as an essential, active member o the health care grams. The most important skill, however, rests in the ability o the
P
A
team is central to the e orts to reduce HAIs. hospitalist to unction as an e ective team leader with an ability
R
Three easily identi able areas o quality improvement related to communicate clearly, concisely, and calmly and demonstrate a
T
to prevention o HAIs include (1) hand hygiene, (2) preven- capacity to coordinate the activities o other care pro essionals they
I
tion o catheter-associated urinary tract in ections (CAUTIs), and might be meeting or the rst time.
(3) prevention o intravenous line-associated blood stream in ec- Hospitalists also add value by actively participating in hospital-
tions, especially those related to central lines (CLABSIs). With wide committees that review the outcomes o rapid response team
observed hand-hygiene rates in most hospitals hovering around an actions and other emergency cardiac care activities. Only then can
T
h
abysmal 40% to 50%, administrators struggle to nd solutions. Many they acilitate change in the processes the hospital puts in place
e
leaders have ound the solution to the hand-hygiene dilemma in the or recognizing and responding to the patient with unexpected
S
p
active engagement o hospitalists in their institution-wide e orts. changes in clinical condition.
e
Numerous success stories show that hospitalist engagement, by
c
i
a
actively promoting hand-hygiene within their team, has taken 40%
l
t
■ LEADERSHIP
y
to 50% compliance rates to a sustainable 90% or higher. From the
o
patient perspective, these increases in hand-hygiene rates translate Medical staff or hospital wide committees
f
H
into substantial reductions in methicillin-resistant staph in ections In recent years, The Joint Commission, the National Quality Forum,
o
s
and other HAIs. and the Centers or Medicare and Medicaid (CMS) have built stan-
p
i
The involvement o hospitalists on daily rounds in which they dards and endorsed sa e practices around medical sta and medical
t
a
l
are able to order the removal o nonessential Foley catheters sub- center leadership’s responsibility and accountability or the sa ety
M
stantially reduces the incidence o CAUTIs. The same holds true and quality within their organizations. This ocus provides the
e
d
or CLABSIs, in which the reduction in the days o use or invasive hospitalist with important opportunities to take leadership roles on
i
c
intravenous lines is also associated with a heath care-associated medical sta and hospital committees, working groups, and task
i
n
e
in ection. With all these e orts, hospitalists are ideally situated to orces that ocus on sa ety and quality. As physicians who concen-
a
a ect sa ety outcomes or their patients and others on their teams trate wholly on hospital-based care, no group is better positioned
n
d
and in the institutions where they practice. to inf uence outcomes than hospitalists. They should work as solu-
S
tion seekers or the best ways to standardize hando s, design the
y
s
■ PATIENT TRIAGE AND RAPID RESPONSE TEAMS (RRT) electronic health record, improve in ection control practices, create
t
e
m
As the rontline physicians accepting or coordinating inpatient accountabilities or the CMS Core Measures, and oversee team
s
hospital admissions, the hospitalist has an a rmative obligation to building and rapid response teams throughout the entire organiza-
o
tion. Failure to engage in these e orts represents signi cant lost
f
make certain newly admitted patients are placed on units and into
C
beds that are appropriate or the level o care they need. Nonethe- opportunities.
a
r
e
less, patients on appropriate wards or units will still deteriorate aster
than the care pro essionals anticipate. When that happens, in the
interest o patient sa ety, hospitals must have a process or rapidly Safety culture
summoning a team o pro essionals to assess the patient’s current As hospitalists are clearly some o the most visible physicians in any
state o deterioration and to assist in a “retriaging” process. organization, they bear a unique role and responsibility or promot-
Whether activated by other physicians, nurses, patients, or amily ing a robust “sa ety culture” within the organization and speci cally
members, hospitalists play an important role in the response to the in the units where they ocus their practice. The Agency or Health-
deteriorating patient in many institutions. E ective and valuable care Research and Quality identi es key eatures o a “sa ety culture”
physician members o a rapid response team, or any team assigned that includes the willingness o hospital sta to openly communi-
to respond to the clinically deteriorating patient, possess certain cate concerns about patient care on their units. Units where sta
necessary attitudes and skills. As a leader o the team, the hospitalist express com ort when questioning physicians or other authority
must approach each “call or help” with a high degree o “mind ul- gures when they disagree or have concerns are considered units
ness” in order to avoid premature closure based upon selective with a “sa er” culture. Other positive attributes o “sa e” units are
in ormation provided to them by the care pro essionals previously those in which mistakes are openly discussed and those discussions
caring or the patient. Regardless o whether the “call or help” or ocus on “systems” issues instead o blaming speci c individuals.
activation o the RRT seems appropriate a ter the initial response The di erence between the culture o one unit to that o another
and investigation, it remains critical that the physician leader o the in the same hospital o ten correlates with the di erence between
response team supports those who trigger the activation and helps the middle management communication styles o the associated
prevent ridicule or criticism o those who asked or help. units. Units with nursing managers and physician leaders who ocus
Hospitalists must remain open minded while gathering all poten- on open, honest, e ective, and nonpunitive communication related
tially important in ormation that might be use ul or arriving at solu- to adverse patient events score higher on sa ety culture surveys
tions or treatments to reverse the deteriorating trend in condition. than others. In addition, there are data to suggest that those units
While care ully listening to other team members and caregivers, with positive sa ety culture survey results have a lower incidence o
the leader o the team must be able to synthesize an approach that adverse patient events when compared to units with less positive
considers all the relevant actors, especially unexpected ndings, surveys. To that end, the hospitalist is uniquely situated to oster a
and avoid the temptation to disregard in ormation that might be culture o curiosity, inquiry, and appropriate challenging o authority
inconsistent with their preliminary diagnosis—the concept o pre- through role modeling and mentoring. By setting positive examples
mature closure. They must remain cognizant o their own con rma- or other physicians and sta , the hospitalist can lead in the e orts
tion biases and strive to keep them in check. Not all chest pain is a to ensure all o the units in which they work strive toward a sa e,
myocardial in arction and not all wheezing is asthma! patient-centered approach to medical care.
126
MANAGING PATIENT SAFETY EVENTS
TABLE 20-3 Immediate Response
C
■ THE PRINCIPLED APPROACH
H
Address current needs o patient and amily
A
Even when hospitalists do their best to proactively maximize the
Contact risk management or patient sa ety hotline
P
sa ety o their patients, unintentional harm still occurs. Importantly,
T
the integrity o the individual physician or institution rests on the Identi y care pro essionals or ongoing care
E
response to patient harm as much as it does to prevention. When Identi y key persons or patient/ amily communication
R
harm occurs there is a choice to deny, minimize, rationalize, and Preserve data, equipment, etc
2
0
blame others, including the patient, or to approach each harm event Document “just the acts” in medical record or risk management
with a commitment to an open inquiry and honest communication report
ollowing harm—the “principled approach.” The hospitalist is con-
P
r
stantly presented with this choice.
e
v
It is well recognized that there are a multitude o barriers to hon-
e
n
est communication ollowing harm that include ear o litigation,
t
hospitalists to report and encourage reporting harm events. These
i
n
humiliation, lost income, reputational damage, risk to privileges and
g
bene ts include (1) the activation o the internal patient sa ety and
license, and the uncertainty o outcome. Nonetheless, the “deny and
a
risk management processes including a crisis management plan,
n
de end” approach to patient harm and the delegation o managing
d
i indicated, (2) the preservation o data and in ormation, (3) the
harm to the legal community has arguably not prevented any o
M
opportunity to trigger immediate support or patient, amily, and
those eared outcomes and instead has damaged the reputation
a
care pro essional, (4) the initiation o a “quality committee” inves-
n
o the medical pro ession and prevented any learning ollowing
a
tigation and the “legal privilege” most states a ord such investiga-
g
patient harm events. The “principled approach” to patient harm
i
n
tions, and (5) the establishment o a communication link with the
is arguably the smarter approach because it e ectively addresses
g
harmed patient and his or her amily.
A
many o the reasons that patients sue (lack o communication, need
As with documentation in the medical record, when reporting a
d
or explanation, sense o dishonesty or “hiding something”) and pro-
v
patient sa ety event, hospitalists should take care to provide only the
e
vides a orum or learning and improving patient sa ety.
r
necessary actual in ormation to commence an investigation. They
s
e
The principled approach to patient harm relies heavily on the
should avoid documenting speculation, hasty conclusions with
P
hospitalist in at least six speci c areas: (1) the immediate response,
a
incomplete acts, or “ nger-pointing” in the report.
t
(2) reporting o harm, (3) communication, initially and in ollow-
i
e
n
up, (4) investigation, (5) identi cation o process and per ormance
t
Investigation
E
improvement opportunities, and (6) the necessary ollow-up. See
v
As advocates or quality medical care and patient sa ety, hospitalists
e
Table 20-2.
n
possess special skills or participating in the investigation o serious
t
s
Responding and reporting adverse outcomes (see Table 20-4). Patients and amilies want and
:
P
deserve the “ acts” a ter a harm event. An appropriate investigatory
a
Responding to and the reporting o patient sa ety events are the rst
t
i
process is needed to provide them with the necessary in ormation.
e
step in any principled process to patient harm (Table 20-3).
n
Reporting triggers the institutional response process, while the The hospitalist should commit to participating in any institutional
t
S
health care team responds to the immediate medical needs o the root cause analysis or other investigatory process ollowing a serious
a
harm event. Such investigations should try to avoid the traditional
e
patient. Most hospitals encourage care pro essionals, especially phy-
t
“shame and blame” approach to adverse events and instead ocus
y
sicians, to report any patient sa ety incident to its Sa ety and/or Risk
a
on system-based issues and identi cation o possible areas o
n
Management Department. Reports are o ten made by telephone,
d
handwritten, online, and in person. Hospitals are mandated by the improvement.
t
Nonetheless, prior to knowing all the acts, patients and amilies
h
Centers or Medicare and Medicaid and The Joint Commission to
e
provide or a reporting process or patient harm events. are still entitled to e ective communication in the early a termath
H
o a harm event.
o
It is incumbent upon the hospitalist to understand and appreci-
s
p
ate the reporting process used in any hospital where they work.
i
t
Communication
a
Importantly, the hospitalist needs to recognize the importance o
l
i
s
their role in taking care o the patient when harm occurs and ensur- Once harm occurs, honest and e ective communication helps
t
ing that neither the patient nor the amily is abandoned during this maintain trust between the patient and amily and care pro ession-
critical time. als. With their easy availability, o ten 24 hours a day, the hospitalist
is uniquely positioned to acilitate such communication. A ter all, or
Benefits of reporting the patient who has experienced an unexpected outcome, every
The bene ts o rapid institutional reporting o patient sa ety hour that goes by without e ective communication constitutes
events within the organization provide substantial incentive or all more harm.
TABLE 20-2 Principled Response to Patient Harm Events TABLE 20-4 Investigation
Respond immediately Per orm within context o authorized quality improvement
Report process
Investigate Involve interpro essional personnel as indicated
Communicate Utilize a validated root cause analysis (RCA) process or tool
Per ormance improvement Incorporate organizational quality and patient sa ety personnel
Follow-up Consider involving patients and amilies in RCA process
127
Honest and e ective communication a ter a harm ul adverse visit. It is critical or the hospitalist to understand the di erence
event is not just the right thing to do, but the smart thing to do as between “empathy” and “apology.” Empathy is the understanding
well. Patients and amilies sue, in large part, because they perceive or sharing o another person’s emotions and eelings whereas an
a lack o transparency, abandonment, or “cover-up.” A transpar- “apology” is admitting a mistake that caused harm. In medicine,
P
A
ent process with open lines o communication and disclosure o there is an ethical imperative to express empathy when patients su -
R
all pertinent in ormation can mitigate those patient and amily er harm but apologies should be reserved or situations when it is
T
perceptions. clear mistakes or errors have caused the harm. There ore, apologies
I
Disclosure o any actual error associated with the harm is a mul- should only be o ered when the acts are clear and agreed upon by
ti aceted process that requires care ul planning, preparation, and the stakeholders with knowledge o the event.
coordination by physicians and hospital administrators. Given its Example o expression o empathy: “I am sorry your pneumo-
complexity, physicians understandably ear that an inadequate or nia has progressed to the point where despite our best e orts we
T
h
poorly executed disclosure o medical errors will only serve to rus- now need to put a breathing tube in your windpipe to help you
e
trate rontline practitioners, ruin the reputation o the organization breathe.”
S
p
and individual practitioners involved in the incident, and encourage Example o apology: “I am sorry I did not check your abnor-
e
lawsuits.
c
mally low blood sugar result this morning. I I had seen the result, I
i
a
Success ul adverse event response programs that include “ ull dis- could have given you some extra sugar to prevent your seizure this
l
t
y
closure” rely heavily upon integration between the clinical depart- morning.”
o
ments and hospital risk management. This integration ensures that
f
H
the various stakeholders are “on board” or at least aware o the plan
o
PROCESS IMPROVEMENTS
s
or communication a ter adverse events. The stakeholder list must
p
The value in a principled, transparent approach to adverse patient
i
include the medical malpractice insurance carriers or the various
t
a
events lies in the ability to learn rom mistakes within a rigorous,
l
parties who might be a ected. In order to provide a consistent
M
approach to adverse events or providers, patients, and amilies, ref ective environment that promotes per ormance improvement
e
e orts designed to signi cantly improve the delivery o care. To
d
all o the steps involved in the response to harm should be preap-
i
c
proved by all appropriate stakeholders, be ore implementation. be e ective, hospitalists must play a role in these per ormance
i
n
improvement e orts that ollow suboptimal care. Patients and
e
The process o communication a ter harm occurs generally alls
amilies involved in adverse events caused by inappropriate care
a
in three phases: immediate, intermediate, and nal or ollow-up
n
are intensely interested in learning the ways in which similar events
d
phase (Table 20-5). The extent o hospitalist involvement in this
S
type o communication will depend upon the relationship o the are less likely to occur. Discussing these quality improvement mea-
y
s
hospitalist with the patient and his or her amily. Long-term close sures with them also helps to maintain the trust and bond between
t
e
patient and provider.
m
physician-patient relationships are conducive to multiple meetings
s
and discussions wherein the hospitalist will quickly discover that
o
“disclosure” is a process and not an event.
f
CONCLUSION
C
A liaison or the amily should be identi ed and appropriate
a
The evidence or abundant opportunities to improve patient care
r
empathy and assurance o nonabandonment is expressed at each
e
and prevent patient harm has become indisputable. Earlier esti-
mates on the annual number o preventable deaths in hospitals
ar underestimated the actual number o deaths that could be pre-
vented, especially rom health-care-associated in ections. Moving
TABLE 20-5 Communication orward it has become clear the mantra “no outcome, no income”
Immediate or “no pay or low-quality per ormance” is going to apply to a signi -
cant number o episodes o care. To that end, hospitalists are ideally
Express empathy
situated to positively inf uence issues in quality o care outcomes
Do not make promises you cannot keep through their active engagement on per ormance improvement
Disclose “known” acts e orts, their communication skills, and their ability to lead multidis-
Assure nonabandonment ciplinary e orts.
Identi y persons (liaison) or ollow-up In addition, hospitalists can take a lead in providing honest, open,
and e ective communication to patients and amilies a ter unex-
Intermediate
pected adverse event outcomes. This transparent approach can be
Ensure liaison presence
the catalyst, or the trans ormation o an organization’s culture or
Continue to express empathy e ectively responding to the needs o patients, providers, and the
Disclose acts discovered during investigation health care system. Adopting a policy and practice o transparency
Ask and answer clari ying questions related to harm events represents a major shi t in organizational
Apologize i consensus exists about error or substandard care ocus and will need the ull support o hospitalists to ully imple-
causing harm ment. This approach will require strong and persistent endorsement
Explain plan or prevention o uture harm by the kind o leadership that hospitalists can provide their organiza-
tions. The added value o transparency is ound in the opportunity
O er ongoing contact and communication
to rapidly learn rom, respond to, and modi y practices based on
Final or ollow-up phase harm investigation with these now transparent events.
Ensure liaison presence
Answer additional questions
SUGGESTED READINGS
Express empathy, apology, i indicated
Discuss per ormance improvement measures Boothman RC, Blackwell AC, Campbell DA Jr, Commiskey E,
O er ongoing contact and support Anderson S. Abetter approach to malpractice claims? The University
o Michigan Experience. J Health Life Sci Law. 2009;2(2):125-159.
128
Hickson GB, Federspeil CF, Pichert JW, Miller CS, Gauld-Jaeger J, McDonald T. Error disclosure within a principled approach to
Bost P. Patient complaints and malpractice risk. JAMA. adverse events. ASANewsletter. 2009;73(5):20-22.
C
2002;287(22):2951-2957.
H
McDonald TB, Helmchen LA, Smith KM, et al. Responding to
A
Institute o Medicine. To Err Is Human: Building a Safer Health System. patient sa ety incidents: the seven pillars. Qual Saf Health Care.
P
In: Kohn LT, Corrigan JM, Donaldson MS, eds. Washington, DC: 2010;19(6):e11.
T
National Academy Press; 2000.
E
National Quality Forum. Safe Practices for Better Healthcare–2010
R
Leape LL, Berwick DM. Five years a ter to err is human: what have we Update: AConsensus Report. Washington, DC: NQF; 2010.
2
learned? JAMA. 2005;293(19):2384-2389.
0
P
r
e
v
e
n
t
i
n
g
a
n
d
M
a
n
a
g
i
n
g
A
d
v
e
r
s
e
P
a
t
i
e
n
t
E
v
e
n
t
s
:
P
a
t
i
e
n
t
S
a
e
t
y
a
n
d
t
h
e
H
o
s
p
i
t
a
l
i
s
t
129
CHAP TER
21 INTRODUCTION
Achieving better health outcomes or patients and populations
requires a ocus on continuous quality improvement (QI). While
physicians pride themselves on being subject matter experts in
their ocused area o medical practice, such knowledge alone is
insu cient to produce undamental changes in the delivery o
health care. Physicians who practice in complex hospital and health
care systems must acquire another kind o knowledge in order to
develop and execute change.
Principles and W. Edwards Deming, an American statistician and pro essor who
is widely credited with improvement in manu acturing in the United
Models of Quality States and Japan, has described this knowledge as a “system o
pro ound knowledge” (Figure 21-1). This knowledge is composed
Improvement: o the ollowing items: appreciation or a system, understanding

variation, building knowledge, and the human side o change. These
concepts are just beginning to be taught to health care pro ession-
Plan-Do-Study-Act als and are essential or anyone who wishes to improve the health
care delivery system.
All hospitalists have witnessed changes that did not result in
undamental improvements within their hospital systems: the
Emmanuel S. King, MD, FHM computerized order set that was success ully implemented but
never revised based on prescribers’ eedback, the paper checklist
Jennifer S. Myers, MD, FHM or medication reconciliation that never gets lled out, or the new
rounding system that worked or the rst ew weeks but then ailed
to become a standard part o practice due to physician variation or
lack o commitment. These are all examples o f rst-order changes—
changes that ultimately returned the system to the normal level o
per ormance. In quality improvement work, individuals must strive
or second-order changes, which are changes that truly alter the
system and result in a higher level o system per ormance. Such
changes impact how work is done, produce visible, positive di -
erences in results relative to historical norms, and have a lasting
impact. Although the model or improvement described below may
seem simple, it is actually quite demanding when used properly;
and the process is essential to both learning and ultimately chang-
ing complex systems.
PLAN-DO-STUDY-ACT AS A TOOL FOR QUALITY

IMPROVEMENT
The Plan-Do-Study-Act (PDSA) model is a commonly used method
in quality improvement. Shewart and Deming described the model
many years ago when they studied quality in other industries. This
model rst appeared in health care when Berwick described how
the tools could be applied using an iterative approach to change.
Using a “test-and-learn approach” in which a hypothesis is tested,
retested, and re ned, the PDSA cycle allows or controlled change
experiments on a small scale be ore expansion to a larger system.
The our repetitive steps o PDSA—plan, do, study, and act—are
carried out until undamental improvement, which can be exponen-
tially larger than the original hypothesis, takes place (Figure 21-2).
PRACTICE POINT
• Use a “test-and-learn approach”to solve quality problems. The
PDSA—plan, do, study, and act— ramework is one popular
model to organize your approach to quality improvement
work.
130
to uncover underlying assumptions or biases be ore any testing,
and to enhance learning in the Study phase by providing a baseline
C
Appre cia tion point o comparison.
H
of a s ys te m
Teams new to QI requently will struggle with the question, “How
A
P
do we measure improvement?” De ning discrete process measures
T
is a good starting point when using PDSA. Process measures are
E
used to assess whether the cycle is being carried out as planned.
R
Building
Huma n s ide This is in contrast to outcome measures, which are used to track
2
knowle dge
success or ailure and ocus on the speci c outcome that the team
1
of cha nge
is trying to achieve.
P
■ DO
r
i
n
Unde rs ta nding
c
The Do phase in PDSA is a period o active implementation. It
i
va ria tion
p
involves eedback on the new process rom end users and rigorous
l
e
data collection. An overarching goal o this phase is to capture and
s
a
document not only compliance with the new process, but also devi-
n
d
ations, de ects, or barriers in the process. There are always aspects o
Figure 21 1 Deming’s System o Pro ound Knowledge. (Reproduced,
M
quality improvement projects that do not go as planned, and ex-
o
with permission, rom Langley GJ, et al. The Improvement Guide: A ibility and open-mindedness are critical to maximize learning rom
d
Practical Approach to Enhancing Organization Per ormance, 2nd ed.
e
improvement. The quality o the Do phase is intimately related to
l
s
San Francisco, CA: Jossey-Bass; 2009.) the quality o the Plan phase. A pit all or many novice QI teams is to
o
give in to the temptation to jump straight to implementing change
Q
■ PLAN
u
without spending a signi cant amount o time planning. A poorly
a
conceptualized improvement plan, an absence o a sound data col-
l
i
During the Plan phase, the team generates broad questions,
t
y
hypotheses, and a data collection plan. It is critically important lection model, or unclear accountabilities can have adverse ef ects
I
m
during this period to de ne expectations and assign tasks and on the implementation or “do” phase o a new initiative.
p
accountability to every team member. In the planning phase o the
r
o
■ STUDY
v
PDSA cycle, it is prudent to invest signi cant time and develop a
e
m
well- ramed question by reviewing related research and local proj- Analysis o available process and outcome metrics and a qualitative
e
ects and de ning meaning ul process and outcome measurements. appraisal o the process are the key activities in the Study phase.
n
t
Broad questions at the outset o a PDSA cycle can include “What are Time should be set aside to per orm a critical review o the data
:
P
we trying to accomplish?” and “What changes can we make that will collected and compare it to historical data (when available) and
l
a
baseline predictions. Close attention should be paid to possible
n
result in an improvement?” The ideal data collection tool answers
-
D
the question: “How will we know that a change is an improvement?” de ects in any element o the process, including the data collection
o
It is also help ul or the team to generate predictions o the answers plan. I such issues are uncovered, the team may need to revise the
-
S
to questions early on. This aids in raming the plan more completely, initial data collection tools and overall plan. Thought ul review o all
t
u
d
trials, even those that were clearly unsuccess ul based on metrics,
y
is a critical and valuable process or the team. In act, the “ ailures”
-
A
in a PDSA cycle can yield unanticipated and improved directions.
c
t
As the Study phase progresses, time should be spent considering
Plan
i a ollow-up PDSA cycle is planned and exactly what elements to
• Obje ctive
include in that cycle.
Ac t • Que s tions a nd
• Wha t cha nge s a re pre dictions (why? ) ■ ACT
to be ma de ? • P la n to ca rry out The nal component in a PDSA cycle is Act. The team should con-
• Ne xt cycle ? the cycle (who, wha t, vene or a eedback and action planning session. Frontline workers
whe re , whe n? ) in the system that is being changed should be included or honest
• P la n for da ta colle ction input. A team approach rather than a “top-down” approach acili-
tates an open review o successes and ailures. An action plan that
S tudy Do encompasses lessons learned in the rst three steps should then be
• Comple te the • Ca rry out the pla n put into motion. During this stage decisions are made about repeat-
a na lys is of the da ta ing certain test cycles a ter improvements are made or “spinning of ”
• Docume nt proble ms
new test cycles based on the original one.
• Compa re da ta to a nd une xpe cte d
pre dictions obs e rva tions
■ RAPID CYCLE, CONTINUOUS, AND SEQUENTIAL PDSA
• S umma rize wha t • Be gin a na lys is
In its most basic orm, the PDSA model described above can be
wa s le a rne d of the da ta
applied to change a single process. However, teams in health care
o ten con ront problems that require multiple changes, in parallel or
succession, in order or improvement to happen. Caution is advised
when initiating several PDSA cycles simultaneously, especially i
Figure 21 2 The Plan-Do-Study-Act Cycle. (Reproduced, with there are signi cantly dif erent data collection plans or i the team
permission, rom Langley GJ, et al. The Improvement Guide: A is inexperienced in QI methods. An alternative is a sequential PDSA
Practical Approach to Enhancing Organization Per ormance, 2nd ed. model in which one PDSA cycle eeds into the next. This approach,
San Francisco, CA: Jossey-Bass; 2009.) in which teams continually change and re ne their processes based
131
on data evaluation and eedback, is called “continuous quality rom the tool, to test the theory that this would improve compli-
improvement.” Experienced QI teams strive to utilize this approach. ance. The data collection plan was to track overall compliance with
Rapid cycle PDSA is a continuous QI process that lends itsel well the tool or a 2-week period. In order to isolate any improvement
to projects that are ocused on relatively small-scale changes. It is as a result o this one small change, no other changes were made
P
A
typically used by seasoned QI teams who are amiliar with the PDSA during this time.
R
model and who wish to implement rapid change.
T
■ DO
I
AN EXAMPLE OF PDSA IN ACTION The new version o the tool was implemented.
To illustrate the PDSA model or improvement, a real QI project is
presented here rom start to nish. A hospitalist group sought to ■ STUDY
implement a new discharge planning toolkit aimed at improving
T
Compliance rates signi cantly increased rom moderately high to
h
transitions in care through risk assessment at the time o hospital very high, and the risk actor screening data remained unchanged.
e
admission. A QI team was ormed with representatives rom health
S
Qualitative eedback rom rontline users was that the risk screening
p
care pro essionals involved in the discharge planning process. While
e
process was more streamlined and acceptable.
c
their ultimate goal was to reduce unplanned readmissions, their
i
a
rst team goal involved creating a new process to coordinate and
l
t
■ ACT
y
request risk-speci c interventions rom other teams (eg, nurse edu-
o
A brie but success ul cycle 2 ended with a plan to add an interven-
f
cators, pharmacists, a nurse or postdischarge ollow-up phone calls)
H
or patients deemed “high risk or hospital re-admission or transition tion checklist to the tool in the next phase.
o
s
in care problems” by a screening tool. In preparation or the project,
p
CYCLE 3
i
t
the QI team also per ormed a stakeholder analysis, which is a tool
a
l
that QI teams can use to identi y all o the individuals and groups
M
with a “stake” in the process being discussed.
■ PLAN
e
d
The goal o cycle 3 was to associate risk-speci c interventions
i
c
CYCLE 1 (education, ollow-up phone calls, and social work interventions)
i
n
e
with a patient’s individual risk actor pro le. To meet this goal, the
a
■ PLAN team implemented a new version o the tool that included the risk-
n
d
The initial PDSA cycle involved piloting a readmission risk screening speci c intervention requests and tracked request type and volume.
S
tool. A weekly meeting was convened that included representative A 2-week cycle was planned with continued weekly meetings dur-
y
s
ing this time.
t
users o the tool and assigned speci c responsibilities and tasks with
e
m
due dates to each team member. At baseline, the biggest barriers to
s
overcome were the perception that the new tool was extra work, ■ DO
o
f
introducing a paper-based tool in a largely electronic health care The team implemented a tool that allowed or interventions to be
C
environment, and lack o a tight in rastructure tying the requests requested at the time o risk actor screening.
a
r
e
to existing risk-speci c interventions. Based on these concerns, the
team reduced the number o interventions on the initial tool. A data ■ STUDY
collection plan was started and included both quantitative process A ter 2 weeks, the data showed stable high compliance with the
metrics (eg, compliance rates with the tool, requency o risk actors orm, stable risk actor data, but very low utilization o intervention
identi ed on the tool) and qualitative data rom the users o the tool. requests. At eedback meetings, rontline users stated that at the
time o admission, they were not ready to place a request or an
■ DO intervention. They elt that intervention requests should be dis-
The new tool was piloted or 2 weeks, during which time data was cussed in a multidisciplinary team on a ollow-up hospital day when
collected and eedback was solicited rom the rontline team. more in ormation was available.
■ STUDY ■ ACT
A ter 2 weeks, the data showed that overall compliance with the Discharge planners on the team suggested that the intervention
tool was moderately high, but that two risk actors, health literacy request process be integrated into daily discharge planning rounds,
and depression, had unexpectedly low percentages. On urther during which the entire patient care team (physician, nurse practi-
inquiry, members o the team admitted that when they per ormed tioner, registered nurse, discharge planners, patient service repre-
the risk screen, they paused on those two questions and requently sentative, and social worker) discussed each patient on the service.
le t them blank, concerned that it might take too much time during A nurse practitioner and patient service representative dra ted
the admission process and rustrate new users o the tool. paper orms that could be used to communicate requests or each
o the interventions to the appropriate personnel and to document
■ ACT completion o the task. The next phase would trial this new process.
The team decided to make another edit to the tool be ore the second
PDSA cycle. The health literacy and depression screening questions CYCLE 4
were removed based on eedback, with a plan to reintroduce them
when the tool was more embedded in the hospital admission work ow. ■ PLAN
Cycle 4 was ocused on implementing and studying the new dis-
CYCLE 2 charge rounds process to request risk-speci c interventions. The re-
quency o intervention requests in each category was added to the
■ PLAN existing process metrics. Since this was a more substantial change
The second PDSA cycle ocused on ollow-up data collection with than be ore and involved more than just one team o rontline users,
the health literacy and depression screening questions removed a 4-week cycle duration was chosen.
132
■ DO ALTERNATIVE MODELS OF QUALITY IMPROVEMENT
C
Clinicians continued to screen patients using the risk screening tool, In addition to the PDSA model described above, there are other
H
intervention request orms were kept on hand during discharge rameworks that have been used to design and execute qual-
A
rounds, and the patient service representative and discharge plan- ity improvement projects. Adopting one speci c ramework (as
P
ners prompted the teams to request interventions based on patient opposed to adopting several) allows an organization to learn a com-
T
risk actors. The requests were orwarded to the appropriate person- mon language and approach to improvement. Six Sigma and Lean
E
R
nel (registered nurse, pharmacist, nurse educator), who then docu- are two common rameworks that will be brie y described.
mented completion o the intervention on the orm.
2
Six Sigma was developed by Motorola in the mid-1980s and is
1
ocused on reducing variations in a process. Six Sigma is a popular
■ STUDY per ormance improvement methodology which uses a ve-phase
approach to problem solving, called DMAIC (De ne, Measure, Ana-
P
Compliance rates and risk actor data remained steady, but there
r
i
was a signi cant increase in intervention requests in all categories. lyze, Improve, and Control). This ramework guides users to de ne
n
c
However, documentation o completion o the intervention was their QI goals, measure the current process, analyze root causes o
i
p
the quality problem, improve the process on the basis o the previ-
l
low. It was determined that the documentation requirements were
e
s
un amiliar to the intervention teams, which was an oversight. ous steps, and nally control the process to ensure that variances
a
n
are corrected be ore they result in de ects and the new process
d
■ ACT becomes standard work.
M
Lean manu acturing (or just “Lean”) was adapted rom the Toyota
o
For the next cycle urther improvements in documentation o inter-
d
vention completion and a reintroduction o the health literacy and Production Systems and is ocused on continuously reducing waste
e
l
in operations and enhancing the value proposition to customers.
s
depression screening questions was planned.
o
The Lean approach is based on a ew key principles: de ning the
Q
■ LESSONS LEARNED problem rom the customer perspective, identi ying the activities
u
required to provide the customer with a product or service, produc-
a
This example illustrates several important points or success ul use o
l
ing the products or services only when needed by customers, and
i
t
y
the PDSA model. First, the engagement and involvement o the end pursuing per ection in the process.
I
m
users o new QI tools and processes is critical to the success o any
p
improvement project. These users are experts in the process who CONCLUSION
r
o
o ten know what should be tested next, and per ect champions
v
Plan-Do-Study-Act has remained a undamental tool or continu-
e
when changes are disseminated on a larger scale. While it may be
m
impossible to address or x every problem that they identi y, hear- ous quality improvement. Once com ortable applying this iterative
e
approach, hospitalists can af ect both small- and large-scale changes
n
ing their input, implementing changes based on their suggestions,
t
in their health care systems. Other QI rameworks that hospitalists
:
and giving praise or their involvement and patience is an important
P
should be aware o include Six Sigma and Lean methodologies.
l
a
skill or leaders o QI. Second, exibility and creative thinking, skills
n
Adopting one speci c ramework (as opposed to adopting several)
-
that are used requently in clinical care, are also essential in QI. In the
D
case study, several barriers were identi ed such as: concerns about allows hospitalists and organizations to learn a common language
o
-
and approach to improvement.
S
paper orms, perception that certain risk actors would halt the risk
t
u
screening process, and lack o in rastructure around the systematic
d
SUGGESTED READINGS
y
documentation o interventions. As these barriers became apparent,
-
A
the team remained exible and changed a part o the new process
c
Berwick D. Developing and testing changes in delivery o care. Ann
t
without compromising the integrity and team goals o the project.
Intern Med. 1998;128:651-656.
Langley GJ, Moen RD, Nolan KM, Nolan TW, Norman CL, Provost LP.
PRACTICE POINT The Improvement Guide: APractical Approach to Enhancing Organi-
Critical to the success o any quality improvement project: zational Per ormance. 2nd ed. San Francisco, CA: Jossey-Bass; 2009.
• Per orming a stakeholder analysis to help identi y all individuals Varkey P, Reller K, Resar R. Basics o quality improvement in health
and teams that have a “stake”in the quality problem that is care. Mayo Clin Proc. 2007;82(6):735-739.
being addressed.
M Engagement and involvement o the “end”users o new QI
tools and processes.

M Small tests o change that include data collection ollowed
by data analysis and decisions on how to proceed.
133
CHAP TER
22 INTRODUCTION
The practice o medicine is at heart an exercise o collecting, lter-
ing, summarizing, managing, analyzing, and acting upon in orma-
tion. This in ormation comes directly rom the patient’s narrative
history, but also rom amily and caretakers, and other providers. It
is also derived rom diagnostic interventions, including the physical
examination, laboratory tests, radiologic exams, and procedures.
Combined with re erence knowledge about physiology, pathology,
pharmacology, and other basic science disciplines, the physician
The Role of makes an expert assessment o the patient’s conditions and risks,
and then recommends an action plan. In ormation about this plan
Information must be communicated and coordinated with a larger team and

with the patient and their amily, executed, and then in ormation
Technology in about how the patient responds ed back, in order to make adjust-
ments over time. I this ow o in ormation is compromised or
hampered at any point in this cycle, then the potential or quality
Hospital Quality and and sa ety problems emerges. Given this intense in ormation-rich
environment that the clinician must navigate, especially in the inpa-
Safety tient setting, it is clear that the judicious application o in ormation

technology (IT) can greatly empower the hospitalist in providing
high quality and sa e patient care; and conversely, that injudicious
application o IT can promote errors and adverse outcomes.
In ormation technologies that impact patient sa ety and quality
Saverio M. Maviglia, MD, MSc
o care can be grouped into three major categories. First, there are
the interventions that impact care as it is delivered in real time—this
class is generally called decision support because it involves clini-
cians while they are making diagnostic and therapeutic decisions.
The second class o in ormation technologies, broadly known as
surveillance, monitors the immediate downstream care processes to
detect anomalies and unintended consequences so that e ective
corrective action may be taken quickly. The last general category
o IT or sa ety and quality is data mining, or retrospective analysis
o large repositories o data, such as patient registries, electronic
health records (EHRs), and administrative databases in order to
detect meaning ul patterns and signals that may help in orm ways
to improve one or more health care delivery processes. Data mining
overlaps with classical epidemiological health services outcomes
research.
DECISION SUPPORT
As de ned above, decision support is any type o in ormation
system that intends to direct, guide, or alter medical decision mak-
ing as it occurs in real time. This may occur via passive delivery o
knowledge, such as quick access to online digital re erences, drug
compendia, clinical calculators, or di erential diagnosis tools. In this
case, the user must voluntarily choose to activate the service. This
type o decision support is usually well received by busy clinicians,
because the clinician is motivated to get a question answered. How-
ever, passive decision support does not address latent in ormation
needs, or knowledge de cits unknown to the clinician.
Decision support may also occur via active knowledge delivery,
such as alerts to avoid unsa e or undesired behavior, or reminders
to promote desired behavior; the service is activated automatically.
Usually, the intended behavior is evidence based, such as avoid-
ing drug combinations that have been shown to result in adverse
e ects; but it can also be policy driven, such as to promote some
medications over others based on ormulary or insurance criteria.
As active decision support is o ten interruptive, clinician acceptance
134
o this in ormation is variable, depending upon the perceived use- system that oversees all digital transactions in a health in orma-
ulness o the in ormation provided and the manner in which it is tion system, such as new orders, new lab results, and new patient
C
displayed. encounter records; a repository o rules that de ne potential events
H
There are certain decision support systems which all somewhere
A
and the actions that should be taken; and a variety o e ector
P
between active and passive, which acilitate work ow. Examples systems to carryout the actions, such as texting or e-mail alerting.
T
include messaging systems such as signout applications and secure Collectively, these components orm what is commonly known as
E
e-mail or text paging, electronic medication reconciliation applica- an event engine. As an example, the monitoring system registers all
R
tions, and results management programs. Also in this category is new lab results, including an individual patient’s alling platelet level.
2
the organized presentation, or summarization, o patient data that A rule in the repository de nes a clinically signi cant rapid rate o
2
lowers cognitive burden. decline (eg, an absolute drop o 50,000 or a relative drop o 30% over
The ideal decision support intervention simultaneously acilitates 2 days, or o 75,000 or 50% over 3 days) in the right clinical context
T
h
the desired work ow(s) while impeding the undesired work ow(s). (the patient has a current active order or a heparin-containing med-
e
In other words, the most e ective decision support interventions ication), to generate a response (alert the responsible clinician the
R
o
make it “easy” to do the right thing, and “hard” to do the wrong thing. next time she logs into the system, or text page a backup clinician i
l
e
For example, compare two ways to implement decision support or this does not occur within 24 hours). It is clear rom this example that
o
optimal drug dosing. The rst, more common approach is to analyze the e ectiveness o this surveillance system is only as good as the
I
n
medication orders a ter they are entered; compare them to rules that breadth o events that can be monitored, the granularity with which
o
assess patient actors such as age, gender, and comorbidities such as rules can be authored to de ne clinically signi cant events, and the
r
m
renal dys unction; and then display a series o corrective alerts. The breadth o interventions available or actions.
a
second approach does as much o the patient-speci c calculations Surveillance systems can collect and analyze quality-related data,
t
i
o
as possible up ront, so that only the most reasonable medication as well as sa ety data. Such systems, sometimes called pro ling or
n
alternatives or a given indication are o ered in the rst place, with detailing systems, have been long utilized by pharmaceutical com-
T
e
de ault dose and requency precalculated to match the patient’s panies to direct and tailor marketing e orts, but they can also be
c
h
condition. Only the prescriber who chooses to override the de aults is used to track how o ten a hospitalist utilizes non ormulary medica-
n
o
interrupted to provide an override reason. O course, the more sophis- tions; or how well a provider is achieving quality o care metric goals,
l
o
ticated consultative approach to delivering decision support requires such as percentage o their patients who have a Foley catheter or
g
y
more data in computable orm about a patient, as well as more com- over 48 hours; or rates o resource utilization, such as magnetic reso-
i
n
plicated and nuanced rules, than the typical critical approach. nance imaging or headaches. This in ormation can be shared with
H
Evidence suggests that decision support can success ully in u- just the relevant provider, or with an entire practice, either deidenti-
o
s
ence provider behavior, improve process measures, increase quality ed or not. The most sophisticated pro ling systems present the
p
i
o care, and reduce errors and adverse events. However, the way data in a quality dashboard that the provider can query dynamically
t
a
l
decision support is implemented can dramatically a ect e ective- in real time and link the data to relevant actions, such as e-mailing
Q
ness. For example, when in uenza vaccination reminders were patients, agging them or callback appointments, or automatically
u
a
rst implemented at one inpatient site, the e ect was minimal; but re erring patients to a disease management program.
l
i
t
y
when the alert was changed to a complete prewritten order, and In addition to patient-speci c and provider-speci c event
a
the de ault was set to “accept” instead o to “decline,” the inpatient engines, surveillance systems have also been developed to work
n
d
vaccination rate increased rom 1% to 51%. at the population level. For example, there are monitoring systems
S
that track aggregated data about visits to regional emergency
a
e
rooms, including chie complaints, to detect early signals o disease
t
PRACTICE POINT
y
outbreaks such as rom in uenza or bioterrorism.
• Evidence suggests that decision support can success ully
in uence provider behavior, improve process measures,
DATA MINING
increase quality o care, and reduce errors and adverse events.
However, the way decision support is implemented can The nal class o IT that can be brought to bear on quality and sa ety
dramatically a ect e ectiveness; poorly implemented systems is the retrospective analysis o large datasets to look or trends and
can have unintended adverse consequences on health care patterns and their relationship or association with signi cant events,
providers or promote error. interventions, or behaviors. This traditionally has been called health
services research, and can ocus on either health care outcomes
(mortality, morbidity, readmissions, adverse event rates) or their
More problematic is the growing recognition that poorly imple- process-based proxies ( requency o deep vein thrombosis [DVT]
mented systems can have unintended adverse consequences on prophylaxis measures, rate o compliance with recommended
health care providers, to the point o inciting clinician revolt, such guidelines, or proportion o completed discharge summaries within
as occurred at Cedars-Sinai Medical Center in Los Angeles in 2003. 24 hours). Both are valid indicators o quality and sa ety e ective-
Worse yet, computerized provider order entry, coupled with deci- ness, though hard outcomes are o ten pre erred, but usually more
sion support, can potentially promote errors. There ore, the behavior dif cult to measure and in uence. For example, measuring the
o such systems must be continually scrutinized, and the in orma- e ect o an intervention on incidence o DVT would be ideal, but in
tion they provide must never be accepted blindly. practice would require signi cant manual data collection by chart
abstraction, and may be too in requent an event in the time window
allowed or study to make statistically sound conclusions. Instead,
SURVEILLANCE measuring how the intervention impacts the number o orders or
Surveillance IT is analogous to secondary prevention or care—it subcutaneous heparin, especially where such orders are placed via
is meant to detect complications o care early so that the conse- computer-based provider order entry (CPOE), is much easier and a
quences can be prevented or ameliorated. The most prevalent more common event.
example, which is still relatively rarely implemented in practice, is The newest direction this type o research has taken is the com-
adverse event detection. This requires an electronic monitoring bination o large datasets rom di erent disciplines, to look or new
135
and sometimes unanticipated or counterintuitive associations. THE HOSPITALIST’S ROLE
Because o the increasing likelihood o chance alone being respon-
There is great opportunity or clinicians such as hospitalists, who
sible or observing such relationships between data when multiple
are o ten experts in work ow and systems thinking (whether by
statistical tests are per ormed with the same data, this type o knowl-
P
ormal training or simply by experience), to help guide implementa-
A
edge discovery requires large collections o data, runs the risk o
tions o decision support interventions within their practice sites
R
uncovering statistically signi cant but clinically irrelevant patterns,
in order to increase the chance o success. Even a ter success ul
T
and should always be considered hypothesis generating rather than
implementation o in ormation technologies, there is ongoing need
I
con rming or re uting. A simplistic example is that i one were to
or clinicians to provide eedback about what works, what does not,
measure how requently lung cancer patients carry matches com-
and what could be done to improve the system. A higher level o
pared to patients without lung cancer, one might be tempted to
involvement o hospitalists is to provide subject matter expertise
conclude that carrying matches is a very dangerous activity. Another
to tweak rules and author new ones to make the decision support
T
h
example is that since test results are typically de ned to be in the
more speci c, relevant, and e ective. This never-ending work to
e
normal range when they lie within the 95th percentile o results
S
keep the content o rules in line with ever changing knowledge has
p
rom a healthy population, then a battery panel o 20 tests will have
emerged as a new eld o its own, called knowledge management.
e
at least one alse positive result almost two-third o the time.
c
i
a
l
t
y
MEANINGFUL USE PRACTICE POINT
o
f
In the United States, a recent potent driver o health IT e orts has •
H
Knowledge management requires keeping the content o
been Meaning ul Use (MU). Through the Health In ormation Tech-
o
rules in line with ever-changing knowledge, and creating new
s
nology or Economic and Clinical Health Act, up to $27 billion (or
p
ones to make decision support more speci c, relevant, and
i
t
over $60,000 per clinician) over 10 years is set aside to promote not
a
e ective. Even a ter success ul implementation o in ormation
l
just the adoption, but the “meaning ul use” o Electronic Health
M
technologies, there is ongoing need or clinicians to provide
Records in order to achieve improved patient outcomes. This is
e
eedback about what works, what does not, and what could be
d
being pursued incrementally via three stages enacted over time.
i
done to improve the system.
c
i
n
• Stage 1 (2011-2013): Capture health in ormation electronically
e
in a structured ormat; use that in ormation to track key clinical
a
n
conditions and coordinate care; implement clinical decision
d
support tools; use EHRs to engage patients and amilies and SUGGESTED READINGS
S
y
report clinical quality measures and public health in ormation.
s
Ash JS, Berg M, Coiera E. Some unintended consequences o
t
e
• Stage 2 (2014-2016): Encourage the use o health IT or continu-
m
in ormation technology in health care: the nature o patient
ous quality improvement at the point o care and the exchange
s
care in ormation system-related errors. J Am Med Inform Assoc.
o
o in ormation in the most structured ormat possible.
2004;11(2):104-112.
f
• Stage 3 (starting 2017): Promote improvements in quality, sa ety
C
a
and ef ciency leading to improved health outcomes by ocus- Bates DW, Cohen M, Leape LL, Overhage JM, Shabot MM, Sheridan T.
r
e
ing on decision support, patient access to sel -management Reducing the requency o errors in medicine using in ormation
tools, access to comprehensive patient data and improving technology. J Am Med Inform Assoc. 2001;8(4):299-308.
population health. Fieschi M, Du our JC, Staccini P, Gouvernet J, Bouhaddou O. Medi-
Compliance with the initial stages is incentivized with bonus pay- cal decision support systems: old dilemmas and new paradigms?
ments rom Medicare and Medicaid to both eligible providers and Methods Inf Med. 2003;42(3):190-198.
to hospitals. These payments decrease over time and eventually Jung E, Li Q, Mangalampalli A, et al. Report central: quality report-
disappear, to be replaced by payment penalties starting in 2015 or ing tool in an electronic health record. AMIA Annu Symp Proc.
noncompliant providers and hospitals. 2006:971.
There are growing data to support the claim that MU Stages 1 and Kaushal R, Shojania KG, Bates DW. E ects o computerized
2 have had a major impact on the use o EHRs and clinical decision physician order entry and clinical decision support systems
support (CDS) in the United States. For example, a longitudinal study on medication sa ety: a systematic review. Arch Intern Med.
o 493 non ederal hospital IT budgets rom 2009 to 2011 ound 2003;163(12):1409-1416.
increases in the percentage o hospital annual operating budgets
allocated to IT in the years leading up to these ederal incentives. Schedlbauer A, Prasad V, Mulvaney C, et al. What evidence supports
Adoption o EHRs by US hospitals has increased dramatically since the use o computerized alerts and prompts to improve clinicians’
the MU program was implemented. In 2011, 833 hospitals and prescribing behavior? J Am Med Inform Assoc. 2009;16(4):531-538.
57,652 pro essionals attested or MU. The ollowing year, an addi- Sittig DF, Wright A, Simonaitis L, et al. The state o the art in clinical
tional 1726 hospitals and 132,395 pro essionals attested. According knowledge management: an inventory o tools and techniques.
to CMS data presented in April 2013, over 77% o eligible hospitals Int J Med Inform. 2010;79(1):44-57.
and approximately 53% o eligible providers quali ed to earn the The Joint Commission. Sa ely implementing health in ormation and
bonus. The data also indicate major increases in CDS use compared converging technologies. Sentinel Event Alert, Issue 42; 2008.
to 2008 baseline.
Weiner JP, et al. “e-Iatrogenesis”: The most critical unintended
On the other hand, data suggesting that the MU program actu-
consequence o CPOE and other HIT. J Am Med Inform Assoc.
ally has improved health IT is mixed. For example, 60% o hospitals
2007;14(3):387-388.
responding to a survey reported that their EHRs did not contain all
eatures essential or high-quality care. And data that MU has posi-
tively impacted patients are indirect and inconclusive.
136
SECTION 5
Practice Management
137
CHAP TER
23 INTRODUCTION
According to the Society o Hospital Medicine, the number o hospi-

talists has increased rom approximately 5,000 hospitalists in 2005 to
more than 44,000 hospitalists in 2015. Despite this explosive growth
and the act that the majority o hospitals now have hospitalist
programs, not all o them have been success ul in establishing a
thriving organization that becomes part o the abric o the hospital.
The supply-demand imbalance or hospitalists continues. The etiol-
Building, Growing ogy or the imbalance is multi actorial. Contributing actors include
the small number o medical school graduates pursing hospital
medicine continues to be below the market needs and the relative
and Managing a ease o moving rom one hospitalist team to another. Couple these
actors with the increased level o physician stress secondary to
Hospitalist Practice understa ed programs and a continued push on scope o practice

with physicians who are younger than those in other specialties and
these actors together perpetuate the supply-demand imbalance
in the market today. The issues experienced by hospitalists are not
Robert A. Bessler, MD unique; other specialties including emergency medicine and critical
care have similar challenges with turnover, recruiting physicians and
temporary workers.
More emphasis is now placed on the patient experience o care
with the introduction o value-based purchasing. There has been a
ocus on educating patients about the role o hospitalists yet many
patients and their amilies continue to express con usion about
the role. It is still common or patients to misconstrue the term
“hospitalists” or “hospice.” Too o ten, hospitalists assume patients
understand their presence at the bedside. More e ort in explaining
the role o the hospitalist as the internal medicine physician or amily
medicine physician who is responsible or patient care while the
patient is in the hospital is essential. Once patients understand that
the hospitalist is the physician assuming responsibility or every-
thing rom admission to discharge, including making patient rounds
and ordering all needed tests and procedures it helps them under-
stand why the hospitalist is caring or them. An important compo-
nent o the dialog is that the patient understands that their primary
care provider (PCP) is in ormed o their progress and resumes care
or the patient postdischarge.
With the Centers or Medicare and Medicaid Services moving rom
a ee- or-service to a ee- or-value payor, the hospitalist takes on an
important role in coordination o care with a ocus on population
health. Today there is a deeper understanding o the importance o
managing population health to drive the health o the community
that a health system serves. Central to this movement is the need
or robust measurement systems that enable us to concentrate on
the outcomes o a population instead o individual silos within the
delivery system. Hospitalists are in unique position to deliver on
the Institute o Medicine’s “Triple Aim,” targeting better health or the
population, better quality and patient experience o care while low-
ering the cost o care. With more than 50% o all health care spend-
ing generated rom the acute care admission through the 90-day
postacute period, the hospitalists team is ideally suited to manage
care rom the emergency department (ED) to postacute care.
The highest per orming hospitalist groups can bring value to the
populations they serve through predictable outcomes. Hospitals
would bene t rom bringing hospitalists into the discussion about
population health and overall per ormance improvement in acute
and postacute care management. Many hospital Accountable Care
Organizations (ACOs) have not ocused on a postacute care strategy,
139
where much o the variability and costs occur in the 90-day period
ollowing discharge nor have they recognized the role hospital- TABLE 23-1 Building a Hospitalist Program: Key Factors
ists can play in tackling this issue. Improving per ormance across to Consider
the acute episode o care is best achieved with a comprehensive
P
Characteristics Examples
A
hospitalist in rastructure that incorporates physician development,
Recruiting • Is the location conducive to recruiting
R
leadership support and incorporation o evidence-based data to
hospitalists? Do they need to recruit a
T
measure per ormance and drive continuous quality improvement. leader?
I
High-per orming hospitalist teams that hardwire these elements
Compensation plan • What is the market rate?
into their practice will drive per ormance improvements and grow
their practice. Number o • What is the number o patients at
encounters/ 7 AM census?
This chapter explores the speci c components essential to build-
T
physician • What total number o patient encounters
h
ing, growing, and managing a thriving hospitalist practice with
e
staying power in light o the new ee- or-value environment. will physicians manage per day?
S
• What is the acuity o patients in the mix?
p
e
c
STRATEGIC PLANNING Schedule • Is a traditional block schedule easible?
i
a
• Do you o er additional vacation days?
l
It is important to have a strategic plan or the practice around
t
y
growth and the types o hospitals and programs best aligned with Management • What local support is required?
o
f
agreed upon goals and objectives. For example, strategic planning support • What regional support is required?
H
o
may require not aligning with all groups requesting support o Tools to support • How will hospitalists record charges?
s
p
the hospitalist team. I a group does not t your strategic pro le communications, • Is there a convenient method to
i
t
or geography, it may be best to decline the opportunity to man-
a
charge capture, communicate to PCPs?
l
age a program. Depending on the goals o the practice, certain scheduling, metrics
M
• How will you demonstrate
e
approaches may not promote patient satis action or continuity o
d
improvement in per ormance?
care goals. For example, when a hospital simply wants your team
i
c
• How will the group demonstrate quality?
i
to cover admissions during the “o hours” that residents are not
n
e
covering patients and then trans er patients back to residents or Clinical processes • What best practices does the group
a
development adopt?
n
surgeons during “peak hours”. These practices work against the
d
goals o improving the patient experience o care and can erode • How do the processes impact care?
S
y
coordination o care. Obstacles o geographic distance requiring a
s
t
day o travel o the core management team present an additional
e
m
burden that may make it best to pass up the opportunity to service
A ter a program is up and running, success ul practices may be
s
a hospital i key management team members cannot be present
o
aced with unprecedented growth. Hospital leaders will need to:
f
on a regular basis. Each hospitalist group should critically evaluate
C
• Set expectations and priorities or growth.
a
whether the growth in a new hospital makes sense based on the
r
De ne key stakeholders.
e
values and goals o the hospitalist practice, in addition to the hospi- •
tal seeking hospitalist services. • Plan or growth.
• Assess the evolving needs o the service, such as using advance
practitioner providers (NPs and PAs).
STRATEGIC PLANNING PROCESS
• Determine the skills in a hospitalist practice and the need or
Be ore starting a hospitalist practice, determine actors that predict additional provider training.
the success or ailure. Identi y the business and nancial motivators • Determine whether the requested skill set o providers by
required to build, expand, and manage a hospitalist service. These hospital administration coincides with the ability to recruit to
actors should incorporate the needs o the hospital and community the program.
the practice it serves. • Reassess the compensation model as the needs o the service
change. For example, hospitalists with the skills to provide
ICU procedures will cost more per shi t than general medical
PRACTICE POINT
hospitalists.
The hospitalist practice must start with a strategic planning From the building stage orward, there is a constant need or
process. outstanding management to ensure a hospitalist practice thrives
• What are the goals o the practice? by using the steps provided in the ollowing tables: (Tables 23-1,
• What are the needs o the hospital? 23-2, and 23-3)
• How easible is it to recruit to the location? • De ne the right leadership and structure.
• What outcomes and metrics are expected by the hospital? • Create an ownership mentality.
• Can the practice commit to the hospital’s per ormance • Setting up the right processes.
expectations? • Tracking and reporting actionable data.
• Provider education ocused on leadership excellence and
per ormance management.
In order to build a hospitalist practice, hospital leaders should: • Promoting outreach to the physician community and acilitating
• De ne the scope o services.
transitions o care.
• Articulate the vision, mission, values, and key value drivers
(KVDs) o the practice. BUILDING A HOSPITALIST PRACTICE
• Establish the employment model and compensation strategy Building a hospitalist practice starts with de ning the prospec-
to drive per ormance. tive hospital partner’s needs or a hospitalist program. In many
• Determine the size and cost o the program. community hospitals, a hospitalist program is created to care or
140
TABLE 23-2 Growing a Hospitalist Program: Core Values and TABLE 23-3 Managing a Hospitalist Program: Key Strategies
C
Goals for Effective Management
H
A
Characteristics Examples Characteristics Examples
P
Quality • Measure length o stay Recruiting • How does the group identi y new hires?
T
E
• Measure readmissions rate • Does the group use a recruiting agency?
R
• Measure CMS core measures Overhead • What percentage o revenue is allocated to
2
• Measure time o discharge support programs (overhead)?
3
• Measure case mix index • Do costs incorporate utilization o advance
practice pro essionals, nurses, support sta ,
Satis action • Measure patient satis action
B
and locum tenens?
u
• Measure nursing satis action
i
l
Training • What allocation o resources does the group
d
i
• Measure PCP satis action have or CME training?
n
g
• Measure specialist satis action • How are new group members trained?
,
G
• Measure administrative sta satis action
r
• How are leaders mentored?
o
w
E iciency • Determine how to improve admission and Growth • Does the group want to expand?
i
n
discharge e iciency • Is the group capable o taking on additional
g
Innovation • What tools can be developed to support the patients at the primary site?
a
n
team’s core values?
d
Service lines • Does the group ocus on acute care
Teamwork • Determine how the team interacts with
M
contracts with traditional hospitalists?
a
monthly and quarterly meetings. • Does the group provide intensivists services?
n
a
• How do you organize in teams?
• Are there other service lines to consider:
g
i
• What is the role o advance practice providers?
n
surgicalists, laborists, academic hospitalists,
g
Leadership • Is there a leadership development training post-acute care/transitional care?
a
path? Improvement • Where is the group’s ocus on quality?
H
o
• Is there a medical director or chie hospitalist strategies E iciency? Satis action?
s
p
on the site?
i
t
a
• Are there regional leaders or clinical and
l
i
s
business operations?
t
the unassigned patient population. But even the de nition o an
P
Financial • Does the group charge a ee or services?
r
unassigned patient is subject to much interpretation. For example,
a
c
• What are the overhead costs to manage the at many hospitals in the Puget Sound region o Washington State,
t
i
c
practice? an unassigned patient is any patient showing up in the emergency
e
• Is there a clear return on investment or the department and requiring admission who does not have a pri-
hospital to retain services o the group? mary care doctor that admits patients at the hospital. In contrast,
Integrity • What guidance does the team provide to the in Orlando, Florida, an unassigned patient is only de ned as a
physicians in the group? patient who has no primary care doctor. In Orlando, i a patient has
• How do we manage the impact o actions, a primary care provider but that doctor does not have admitting
values, methods, measures, principles, privileges, it is standard practice to call the primary care provider to
expectations, and outcomes o the team? identi y who will care or the patient in the hospital.
• What criteria are used to assess integrity o
candidates?
PRACTICE POINT
Research • Is the group involved in research?
• Is there support or data collection and The needs assessment, rom the perspective o the hospital might
analysis? include:
• What unding is available to the group to • PCP and/or surgical dissatis action
support research? • Admission and management o unassigned patients
PCP • How does the group measure PCP • Admission and management o over ow patients due to
satis action satis action? American College o Graduate Medical Education (ACGME)
• Does the group reach out to the PCPs? work hour restrictions
• How does the group track re errals rom PCPs? • High inpatient census and long average length o stay (ALOS)
Nursing • How does the group measure nursing • Low reported per ormance measures
satis action satis action? • External regulation (rapid response teams, code teams, etc)
• Does the group inter ace with nursing?
• How does the group track nursing impact on In addition to covering the unassigned patient population, many
outcomes?
hospitalist services cover those primary care providers who do not
Specialist • How does the group measure specialist want the responsibility o admitting their own patients. There are
satis action satis action? two main orms o coverage relationships: coverage arrangements
• Does the group reach out to the specialist? or 24 hours per day, 7 days per week; and coverage which is more
• How does the group track re errals rom like a house sta model in which the hospitalist admits the patients
specialists? but then turns the care back over to the PCP the next day. These lat-
ter models continue to decline in numbers because o dif culty with
141
P
2014 83% 7% 1% 10%
A
R
T
I
2012 88% 6%1% 5%
T
h
e
S
p
e
2010 90% 6%1% 4%
c
i
a
l
t
y
o
f
H
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
o
s
Inte rna l Me dicine Pe dia trics Me dicine /Pe dia trics Fa mily Me dicine
p
i
t
a
Figure 23 1 Specialty composition o survey respondents. (Source: 2014 State o Hospital Medicine Report. Reprinted with permission rom the
l
M
Society o Hospital Medicine.)
e
d
i
c
recruitment o high-quality providers motivated to build a meaning- general, leap rog compliance guidelines drive a dedicated intensiv-
i
n
e
ul career with a resident-type model. ist model, typically mandated in regional and tertiary hospitals.
a
Hospitalist programs may also be created to manage medical “Code coverage” also de nes the scope o the hospitalist practice.
n
d
specialty and surgical patients, usually a ter establishment o the Many hospitals provide a separate code team, made up o the emer-
S
initial hospitalist program. gency medicine physician or in-house intensivist plus respiratory
y
s
It is essential to determine which patients the hospitalist group therapy, nurses, technicians, and pharmacists. Increasingly, hospital-
t
e
m
will manage, the scope o services, and whether additional training ists are being asked to partake in responding to the code process
s
or some o the program members will be required. According to and arranging patient trans ers to the ICU. In general, emergency
o
the Medical Group Management Association and Society o Hospital physicians have more training and chances to keep their skills sharp
f
C
Medicine 2014 State o Hospital Medicine Report (n = 4867) (see around the procedures o a code, including intubation, starting
a
r
Figure 23-1). central lines, and transvenous pacing. Typically, while an emergency
e
Eighty-three percent o practicing hospitalists are trained in internal medicine physician may respond rst, a hospitalist with advanced
medicine, 10% in amily medicine, 7% in pediatrics, and 1% in med/ cardiac li e support training assumes leadership o the code.
peds. Data rom the American Medical Group Association (n = 3700) Whether the hospitalist scheduled or the night shi t is actually in
report hospitalist training to represent 89% internal medicine, 6% the hospital or at home on call or emergencies also de nes the scope
amily medicine, 5% pediatrics (did not ask about med/peds). When o practice. Hospital-employed and hospital-contracted models tend
looking at the combined MGMA (community hospitals) and AMGA to have in-house coverage while physicians who are part o a private
(academic hospitals) data, the distribution represents training span- ee- or-service group without a hospital contract tend to be available
ning 80% internal medicine, 8.5% amily medicine, 10% peds and 1% as an on-call physician available rom home. Variables that impact the
med/peds. The general trend represents and increased in hospitalists decision beyond economics include the volume o cross-coverage
with amily medicine training. patients, the number o admissions per night, coexisting resident cov-
In most community hospitals today, hospitalists manage ICU erage, and the response time o the physician, i on call rom home.
patients. While there are just over 10,000 intensivist physicians in the
United States, there is an increasing demand or critical care services ■ DEFINING THE TYPE OF EMPLOYMENT MODEL
to serve the aging population and extended li e expectancy. Although
There are several common employment models or hospitalist
the number o critical care physicians in training has been growing,
practices: employed by a private practice, by a hospital, by a mul-
it will be dif cult to meet the patient demand with the rapidly aging
tispecialty group, by a health plan/HMO, or a multisite or national
population. Research indicates the increased demand creates a short-
practice. Among the multisite or national practice subgroups
all o intensivists equal to 35% by 2020, requiring hospitalists to step in
there are staf ng solutions that specialize in emergency medicine,
to ll some o the demand. In general, the larger the hospital the less
anesthesia, and a host o other physician specialists. Some o these
ICU medicine a hospitalist per orms. Many hospitals have mandatory
multisite specialty practices will hire hospitalists who work as inde-
ICU consults a ter a set number o days or hours in the ICU or they
pendent contractors alongside the specialist. Among the national
provide speci c guidelines on managing ventilated patients. The most
hospitalist groups there is a wide spectrum o employment
popular model may be a hybrid arrangement in which access to a
arrangements ranging rom those o ering ownership and partner-
critical care physician occurs during the day and or emergencies but
ship to those that operate solely with independent contractors.
in-house at night. In such cases the hospitalist commonly does the
work around admissions and daily visits with a consult and a ollow-up
visit by the pulmonary critical care physicians. ■ DEFINING THE VISION, MISSION, VALUES, AND KEY
With the labor shortage being even more severe or critical care, VALUE DRIVERS OF YOUR PRACTICE
hybrid models, along with the advent o telemedicine, are likely to It is critically important to de ne the vision, mission and values o
take on even more ICU coverage responsibilities in the uture. In the practice rom its inception. The leaders and hospitalists should
142
take this task seriously. Schedule time to discuss and debate what is ■ MARKETING YOUR HOSPITALIST SERVICES
important to the group and leadership. The process o constructing
C
The best marketing generates word-o -mouth public relations based
your program’s mission and vision statement should not be taken
H
on how satis ed your patients are as well as the nursing and other
lightly. This process can take weeks to develop. Start by establishing
A
hospital sta . An e ective campaign requires all hospitalists on the
P
dedicated time and secure an environment that is conducive to team to be ully engaged with the practice’s vision, mission and values.
T
having uninterrupted, rank discussions. Enlist the input o all team In addition to the passive marketing that comes rom word-o -
E
members.
R
mouth marketing, it is important to develop a marketing plan. A
A mission statement explains the overall purpose o the hospital- typical marketing plan or a practice includes initiatives that drive
2
ist practice. The mission statement articulates what the organization
3
patient satis action to generating awareness in the community
does right now, in the most general sense. In this way, the mission through PCP outreach. Create a budget that supports the plan.
also sets parameters or what the organization, through omission,
B
does not do. Example o a mission statement: “The Hospitalist Group
u
PRACTICE POINT
i
l
o Hilltop builds healthy relationships between St. John’s Hospital
d
i
n
and primary care providers in the community through public educa- Your marketing plan should include segments that target the
g
tion and direct assistance services.” ollowing areas:
,
G
By comparison, the vision statement articulates the uture o the • Identi y your target markets: Decide which target markets
r
o
organization and the community that it serves. The vision state- you want to canvas. You can either target re errals in speci c
w
i
ment, when compared with the current reality o the organization geographic areas or by targeting outreach to specialists.
n
g
or the community, implies the work still needs to be accomplished. • Develop a public relations plan: Launch a new program with
a
In this way, it lends credibility and motivation to the mission
n
press releases, open house events, or broadcast the addition to new
d
statement. Example o a vision statement: “The Hospitalist Group physicians through yers or direct mail campaigns.
M
envisions a group practice that drives improvements in patient • Create a promotion/awareness plan: You can develop practice-
a
n
outcomes including evidence that re ects our value to hospitals in branded written articles on a variety o topics that convey answers
a
g
our community.” to patients’questions using topics such as What is a Hospitalist?
i
n
On a yearly basis the practice should de ne key value drivers or Improving Patient’s Health Literacy. Use these in a mailing to
g
that articulate the ocus o the organization and those areas that
a
your community or have the hospital place your articles in their
H
require organizational ocus in order or the business to grow. Key newsletter. Develop a social media campaign to highlight the
o
value drivers (KVDs) should be set by the leaders with input rom
s
culture o your practice to support recruiting and growth ef orts.
p
the entire team. KVDs must be easy to remember, measurable, and • Develop patient satis action tools: Create large, oversized
i
t
a
achievable. The behaviors that support the key values should also
l
business cards with photos o physicians, hospitalist brochures
i
s
be clearly de ned. In doing so, those in the practice will have a clear
t
with photos o engaged, riendly physicians; consider web-based
P
understanding o expectations even prior to joining the practice.
r
in ormation to share with patients.
a
These behaviors should be rein orced through the compensation
c
• Create recruiting advertisements or physicians: Provide
t
i
and promotion practices o the group to make the practice values
c
your recruiters with materials about the opportunity or special
e
meaning ul and alive on a daily basis. Typically teams evaluate prog-
in ormation about the location and hospital. Place them in
ress on KVDs monthly or quarterly.
hospitalist journals as print advertisements and classi ed ads.
• Conduct market research: Conduct market research in your
■ ESTABLISHING METRICS AND SETTING NEW GOALS local area to be sure you know what the local market is paying or
FOR PERFORMANCE AND OUTCOMES hospitalists and places they practice and who might be interested
in joining your practice in the area.
Standard outcome metrics including average length o stay, core
measures, case mix index, cost per case, and discharge ef ciency are • Pro le your team: Utilize a website and direct mail with photography
o your team or host an open house or educational event.
expected by hospital administration rom the hospitalist group. It is
essential to meet with the hospital and obtain agreement on which • Develop a social media strategy: Share the culture o your team to
initiatives the hospitalist team will ocus. Establish a data collection encourage prospective re errals or service and or recruiting.
and reporting mechanism and the requency o assessments. Prac-
tice metrics that are becoming increasingly important to hospitals
include the Healthcare Cost and Utilization Project (called “H-CUP”). ■ DESIGNING THE MODEL
HCUP is a set o health care databases, so tware tools, and products It is essential to determine the size o the practice needed. The vol-
developed through a Federal-State-Industry partnership and spon- ume o patients who will be seen on a daily, nightly, and monthly
sored by the Agency or Healthcare Research and Quality. Using the basis determines the size o the practice. Next, assess the number
HCUP databases collates data collection rom State organizations, o physicians required to meet the needs o the practice based on
hospital associations, private data organizations, and the Federal that estimated patient volume. The number o physicians depends
government creating a national data benchmark. on what is considered an acceptable workload o patients to man-
HCUP databases include the largest collection o longitudinal age per day, per night, and per month. To determine the number
hospital care data in the United States, with all-payer, encounter- o patients, de ne the average number o admissions per day. I
level in ormation going back to 1988. These databases enable the emergency department uses a tracking tool, review the data to
evaluation o cost and quality o health services, medical practice project the number o unassigned patients based on historical data.
patterns, access to health care, and outcomes o treatments at the In many hospitals, these data are not accessible prior to initiating
national, State, and local levels. a program. Historically, the ward clerks simply entered the admit-
In addition to the standard outcome measures and HCUP data, ting physician’s name in the hospital in ormation system without
it is use ul to track and report other practice related trends, includ- mention o the act that the patient did not have a primary care
ing PCP re erral volume and re erral patterns, patient satis action, physician. It is essential to have a way to track the types o patients
physician recruiting ef ciency, physician retention and 30-day same by re erral type (eg, by PCP, unassigned, or consultations) when the
diagnosis readmission rates (Figure 23-2). hospitalist program begins operation.
143
S tandard Outc o me Me tric s S ite Re po rt
1234—Ho s p ita l A
Qu a rte rs Ra n g e : 2014:Q2 - 2015:Q2
P
A
R
Adjus te d LOS % fo r Me dic are , Me dic al, Re g ular Patie nts
T
95%
I
2014:Q2 - 2015:Q2
Targ e t : 95% 2014:Q2 2014:Q3 2014:Q4 2015:Q1 2015:Q2 90%
ALOS Ratio (% HCUP) 72.6% 66.5% 73.3% 72.4% 72.2% 85%
Ac tual ALOS (days ) 2.90 2.77 2.94 3.17 3.08 80%
T
h
Expe c te d ALOS (days ) 3.99 4.17 4.01 4.38 4.27 75%
e
Dis c harg e s 107 102 93 136 118 70%
S
p
(Re g ular Inpatie nt)
65%
e
c
2014:Q2 2014:Q3 2014:Q4 2015:Q1 2015:Q2
i
a
l
t
Adjus te d CMI % fo r Me dic are , Me dic al, Re g ular Patie nts
y
o
f
2014:Q2 - 2015:Q2 106%
H
o
Targ e t : 102% 2014:Q2 2014:Q3 2014:Q4 2015:Q1 2015:Q2
104%
s
p
CMI Ratio (% HCUP) 98.8% 101.5% 100.8% 105.0% 105.8%
i
t
102%
a
Ac tual CMI 1.01 1.16 1.09 1.16 1.18
l
M
Expe c te d CMI 1.02 1.14 1.08 1.11 1.11
100%
e
d
Dis c harg e s 107 102 93 136 118
i
c
98%
i
n
2014:Q2 2014:Q3 2014:Q4 2015:Q1 2015:Q2
e
a
Obs e rvatio n Le ng th o f S tay (Ho urs )
n
d
S
30
y
2014:Q2 - 2015:Q2
s
t
2014:Q2 2014:Q3 2014:Q4 2015:Q1 2015:Q2 28
e
m
Ave rag e LOS (Ho urs ) 24.67 19.04 20.81 24.43 28.85 26
s
Obs e rvatio n Ho urs 2911 2171 2705 2981 3318
o
24
f
Dis c harg e s 118 114 130 122 115
C
22
a
(Obs e rvatio n)
r
e
20
18
2014:Q2 2014:Q3 2014:Q4 2015:Q1 2015:Q2
Pe rc e ntag e o f Patie nts with Obs e rvatio n S tatus

2014:Q2 - 2015:Q2 48%
2014:Q2 2014:Q3 2014:Q4 2015:Q1 2015:Q2 46%
Pe rc e nt 39.5% 41.3% 47.1% 39.7% 37.3% 44%
Obs e rvatio n Dis c harg e s 118 114 130 122 115 42%
Dis c harg e s 299 276 276 307 308 40%
(Inpatie nt and Obs e rvatio n) 38%
36%
2014:Q2 2014:Q3 2014:Q4 2015:Q1 2015:Q2
Figure 23 2 A dashboard o standard outcome metrics organized by volume o patients, quality, utilization, satis action trend, and market data
indicators including evaluating per ormance to HCUP data.
In addition to determining the volume o unassigned patients, census determined, calculate the number o the physicians per morn-
estimate the number o PCPs interested in turning over care. The ing required or the hospitalist program.
only risk o double counting is i no hospitalist program existed There is much debate over the most appropriate census or the
be ore a new program starting up. Typically, in that scenario, the physician who begins rounding at 7:00 a m. In general, based on a
primary care provider was also likely cover unassigned patients. typical mix o a ew ICU patients and the balance o the load being
A ter determining the number o admissions per year, divide the medical patients, a hospitalist can manage 15 patients sa ely and ef -
admissions by 365 days per year to obtain a rough estimate o the ciently. This number varies considerably due to the di erent agendas,
number o physicians required. Then take the average length o stay acuity o patients, concomitant responsibilities such as rapid response
or the patients and add 1 extra or the day o discharge. Take this teams, code teams, teaching, and goals o practices. To achieve the
number and multiply it by the number o admissions per day to deter- objectives o early discharge, multiple visits a day and a considerable
mine the 7:00 a m census. For example, i there are ve admissions per amount o committee involvement, hospitalists can maintain a cen-
day with an average 4-day length o stay, the 7:00 a m census would be sus in the range o 14 to 15 patients. I the goal is productivity, and
calculated as 5 × (4 ALOS + 1) = 25 patients at 7:00 a m. With the 7:00 a m in some cases the use o advanced practitioner providers (APPs), the
144
S tandard Outc o me Me tric s S ite Re po rt
C
1234—Ho s p ita l A
H
Qu a rte rs Ra n g e : 2014:Q2 - 2015:Q2
A
P
T
Re admis s io n Rate fo r Me dic are , Me dic al, Re g ular Patie nts
E
R
2014:Q2 - 2015:Q2 16%
2
Targ e t : 9.5% 2014:Q2 2014:Q3 2014:Q4 2015:Q1 2015:Q2 14%
3
Ac hieve d 15.9% 9.8% 9.7% 5.9% 5.9% 12%
Dis c harg e s 107 102 93 136 118 10%
B
u
8%
i
l
d
6%
i
n
4%
g
,
2014:Q2 2014:Q3 2014:Q4 2015:Q1 2015:Q2
G
r
o
Adjus te d Mo rtality Rate fo r Me dic are , Me dic al, Re g ular Patie nts (Ro lling Avg )
w
i
n
100%
g
2014:Q2 - 2015:Q2
a
Targ e t : 90% 2014:Q2 2014:Q3 2014:Q4 2015:Q1 2015:Q2 80%
n
d
Mo rtality Ratio (% HCUP) 54.3% 46.3% 35.6% 13.5% 25.2% 60%
M
Ac tual Mo rtality 1.21 1.20 0.93 0.46 0.89
a
40%
n
Expe c te d Mo rtality 2.23 2.59 2.62 3.39 3.54
a
20%
g
Dis c harg e s 330 418 430 438 449
i
n
(Re g ular Inpatie nt) 4%
g
2014:Q2 2014:Q3 2014:Q4 2015:Q1 2015:Q2
a
H
Quality Me as ure s
o
s
p
100.2%
i
2014:Q2 - 2015:Q2
t
a
100%
l
Targ e t : 100% 2014:Q2 2014:Q3 2014:Q4 2015:Q1 2015:Q2
i
s
99.8%
t
Ac hieve d ND 98.7% 100.0% 100.0% 100.0% 99.6%
P
r
99.4%
a
To tal Re s po ns e s ND 154 78 55 43
c
99.2%
t
i
99%
c
e
98.8%
98.6%
2014:Q2 2014:Q3 2014:Q4 2015:Q1 2015:Q2
HCAHPS Phys ic ian Co mmunic atio n TOP Box %

2014:Q2 - 2015:Q2 84%
Targ e t : 80% 2014:Q2 2014:Q3 2014:Q4 2015:Q1 2015:Q2 82%
Ac hieve d 82.0% 73.1% 78.6% 79.2% 83.3% 80%
To tal Re s po ns e s 167 78 28 48 48 78%
76%
74%
72%
2014:Q2 2014:Q3 2014:Q4 2015:Q1 2015:Q2
Figure 23 2 (Continued)
volume per hospitalist may be as high as 20 patients per day. Some physicians is o or 7 days. There are also hybrid arrangements in
practices de ne the census as the number o encounters per day, which the physician works about the same total number o hours
which include new admissions as well as discharges. per month but with shorter periods o time on duty. In such a
In a pure productivity-driven private practice model, the night model, a 7:00 a m census o 25 to 30 patients would likely have six
shi ts are o ten covered rom home (eg, only coming back to the ull-time physicians. In contrast, a private group model may take
hospital or emergencies). This typically also means that the day-shi t every ourth night o call rom home, which could be managed with
doctors might share night call, even a ter working all day. In many our ull-time physicians on the team. The marketplace supply and
practices today, the night shi t is covered by a separate physician, a demand or physicians and goals o various clients (eg, a hospital,
nocturnist, due to the volume o admissions at night and the vol- HMO, or payor) o ten dictates the type o model required.
ume o cross-cover work needed.
In general, the billing revenue o a nocturnist hospitalist is lower
than a day-shi t hospitalist. ■ DETERMINING THE COST AND DIRECT COST OF THE
A highly prevalent hospital-employed and national group prac- HOSPITALIST PROGRAM
tice model includes a schedule in which the hospitalist physician Calculating the cost o a hospitalist program includes direct labor
is on duty or 12 hours, 7 days a week and the ollowing week the costs: salaries o the providers, bene ts cost, malpractice coverage,
145
and billing costs. The volume o patient visits, the payer mix,
and the distribution o CPT codes reported determine the direct
• ACGME.
patient care revenues o the practice. The medical director who • Public per ormance reporting, obtaining ≥ 90% core measure
typically has responsibility or driving hospital outcomes deter- scores.
P
A
mines any additional revenues. According to a survey conducted What is your work environment saying about the practice?
R
by the Society o Hospital Medicine in 2014, 89.3% o hospitalist
•
T
programs required a subsidy or ee to help with the payer mix Patient sa ety, quality, satis action.
•
I
o the unassigned patients, night call coverage in-house, and or Ef ciency o care.
those organizations that ocus on driving per ormance through • Career satis action that integrates core values.
service o erings. The ranges o ees hospitals pay range rom • Service excellence and patient sa ety.
$0 per year to $250,000 per physician annually. Fees are typically • Continuous quality improvement and innovation.
T
•
h
based on scope o work and payer mix. Pro essional growth, leadership, and scholarship.
e
S
■ SETTING THE COMPENSATION MODEL What are the expectations o hospital management?
p
e
•
c
In conjunction with determining the cost o the program, a com- Caring or unassigned/uncompensated patients.
i
a
pensation model must be established. In the past decade, two • Reducing ALOS or top 10 DRGs by hospitalist discharge volume.
l
t
y
signi cant challenges drove hospitalist compensation: an imbal- • 24/7 service demands.
o
•
f
ance o supply and demand, coupled with the rapid rise o salaries Reducing practice variation o hospitalists.
H
that began escalating in 2001. This phenomenon has created a •
o
Hospitalist training on palliative care, end-o -li e, and other
s
signi cant compression in salaries. O ten the least experienced phy-
p
medical specialties.
i
t
sician’s compensation is closely aligned with the most experienced •
a
Development o a comanagement consulting service or a
l
physicians in the practice. This compensation compression creates
M
preoperative testing center.
a dichotomy in the reward system on physician skill and experience
e
• Improvement o patient ED to oor times.
d
levels creating challenging team dynamics. There are two primary
•
i
c
Care o admitted patients in the ED.
i
models: a productivity model or a salary model. Many salary models
n
• Managing the chest pain unit or rapid admission team.
e
also include a component o compensation ocused on productivity
•
a
and quality metrics as well as outcomes. Improvement o chart documentation or core measures (such
n
d
Recruiting a team o physicians and hiring a leader is a critical core as smoking cessation counseling).
S
competency or every hospitalist practice as discussed in Chapter 25. • Improvement o billing or services provided.
y
s
• Leadership o rapid response teams or ill inpatients.
t
Acquiring e ective recruiting techniques is an area o investment
e
•
m
that should not be minimized or overlooked in the development o Development o a transitional care program to address
s
a strong hospitalist practice. continuity o care in postacute acilities or providing care in the
o
f
patient’s home.
C
GROWING A HOSPITALIST PRACTICE
a
Does the practice have these evaluations and measurements in place?
r
e
■ SETTING PRIORITIES FOR GROWTH
• Report card or hospitalists.
Once the practice launches, priorities must be established or the • Primary care physician survey.
growth o the hospitalist program. I the unassigned patients are • Multiyear strategic planning, quarterly reports.
already covered in the practice, the next step could be a myriad
o other opportunities, including contracting with PCP practices.
• Hospitalist career satis action survey.
It is essential to understand the scope o growth and prepare • Hospitalist annual retreat with management to establish goals.
in advance o the patients’ arrival. Many practices have ailed or • Develop a 3-year plan or a hospitalist service that mirrors the
imploded by taking on more growth than they could handle. hospital’s multiyear plan.
I there is a desire to handle 15 more patients per day with a • Create a meaning ul, motivating, and achievable blueprint or
7 days on/7 days o model, it might be as simple as guring out clinical enterprise.
the need to hire two more physicians. However, i the program is • Proactively support mission o patient care, quality
already quite busy and adding three to our new admissions per improvement, and patient-centered care.
day is in the growth plan, adding an admitting shi t may be called
or as well.
■ DEFINING KEY STAKEHOLDERS
The key stakeholders in the practice need to be clearly de ined.
PRACTICE POINT Certainly, the doctors and advanced practice team members
Use these common areas o practice management and determine in the practice are key stakeholders, but in many practices the
whether you are prepared to grow. hospital administration is also a key stakeholder. Identi ying pri-
Re ect about your hospitalist practice: orities is much like a game o chess. For example, i you choose
to help solve another primary care group’s needs be ore helping
• What are your priorities? the orthopedic group with comanagement needs there may be
• What are your goals and core values? repercussions. You should expect that the hospital administrator
• What e ort can you invest to grow? want to weigh in on how this decision impacts the hospital and
its development plans.
What are the expectations o external interests?
• Per ormance measures. ■ INCORPORATING ADVANCED PRACTITIONER
• Satis action o outside primary care physician groups. PROVIDERS
• The Joint Commission requirements. Another key decision or program growth is how to incorporate
advanced practice providers in the practice. While this topic
146
is covered in the literature, there are plenty o mixed opinions MANAGING THE HOSPITALIST PRACTICE
on the use o advanced practice providers in the inpatient set-
C
ting. We have ound two main areas o optimal bene it in our ■ SELECTING THE RIGHT LEADERSHIP AND STRUCTURE
H
practices.
A
There is a shortage o high-quality physician leaders in the
P
The rst bene t or incorporating nurse practitioners (NPs) and United States. To properly manage the practice, it is critical to
T
physician assistants (PAs) is in very small programs o our ull-time appoint the most capable physician leader and establish an
E
physicians with a daily census that can have dramatic swings around e ective practice structure. The hospitalist leaders’ roles are
R
the average. The cost o an NP or PA provider is about one-hal the complex; they not only serve the hospitalists’ team but also play
2
labor cost o a physician, and this can be a cost-e ective way to
3
signi icant roles within the hospital. In these roles, hospitalist
leverage the existing physician coverage. directors are the most connected to how things work on a daily
There is also a bene t rom the use o advanced practice provid- basis. Strong hospitalist physician leaders must lead by example.
B
ers (APPs) in very large programs, particularly in the management o
u
They must have e ective organizational skills, be great com-
i
l
surgical patients or their comorbid conditions. Many practices have
d
municators, and seek win-win situations or the hospitalist team,
i
n
incorporated APPs due to the physician shortage and a ailure to medical sta , and hospital. Hospitalist leaders also need to be
g
recruit and retain high-quality physicians. One unique challenge is
,
aware o the pro essional goals o their members and delegate
G
that many APP’s value comes rom their experience. An APP practic- some responsibilities so that each member can also lourish and
r
o
ing in the acute care setting or 10 years is much more likely to be
w
ind a pro essional niche within the organization. Hospitalist
i
able to unction as a hospitalist than a new graduate APP. For hos-
n
directors may become isolated in their role, so it is important to
g
pitalist physicians, there is clearly value and competency in new a ensure that they have advocates or mentors who can promote
a
physician starting to work directly upon completing their training. It
n
their agendas as well as provide counseling related to hospital
d
is crucial to understand both the state and hospital-speci c by-laws politics. See Chapter 6: Leadership.
M
associated with the use o NPs and PAs. Without such understand- Many hospitalist programs include a version o shi t work. This
a
n
ing, the proposed program plan could be rejected by the hospital. type o schedule combined with the Generation Yculture in medical
a
g
For example, i the rules state that the NPs’ work must be signed o school today, centered on work hours and patient volume restrictions,
i
n
and reviewed by a hospitalist it does not create the same work orce have led to a unique challenge in hospital medicine. Many physi-
g
multiplier as a site where on the right patients, the NP can operate cians seek direct employment models. They place a very high level
a
H
relatively independently. o value or time o . This can make it challenging to engage them
o
in what matters to make a practice success ul. Ensuring the right t
s
p
■ TYPES OF PHYSICIANS IN THE PRACTICE begins with the initial interview when per ormance expectations are
i
t
a
clearly articulated.
l
Another area o importance in growing a hospitalist practice
i
s
t
involves the types o physicians utilized. It is becoming more com-
P
■ CREATING AN OWNERSHIP MENTALITY
r
mon to have amily medicine-trained hospitalists practicing along-
a
c
side internal medicine hospitalists in the same practice. Much has Like any small business, an ownership mentality is essential to
t
i
c
been debated on this topic and today nearly 10% o hospitalists the success o the hospitalist practice. It is ideal to introduce the
e
nationwide are amily medicine trained. Factors that go into the importance o the ownership mentality expected during the hiring
determination to hire them include their com ort level with ICU process. Those applicants who give solid examples o times in their
patients and their experience managing the higher-level acuity career where they got involved, highlight scenarios when they did
patients. Another challenge is their ability to navigate the local poli- things because they thought no one else could do it better, and
tics associated with an internal medicine outpatient practice re er- are passionate about those experiences is telling o their potential.
ring its inpatient practice to a amily medicine physician. We have These are typically indicators that the physician is the type o hos-
ound that the experience o the provider trumps all board certi ca- pitalist who can make the practice excel. De ning the behaviors
tion. There are plenty o internal medicine physicians unquali ed to that support the values o the practice and then evaluating and
practice as hospitalists as well. rewarding those behaviors goes a long way to rein orcing what is
important. For example, i participating in hospitalist committees is
■ THE PROS AND CONS OF CAPS ON SERVICES important and it can be rewarded as part o how the productivity
dollars are allotted. Leading a hospital committee or playing a lead-
During the hospitalist practice growth phase, the group must be
ership role within the medical sta could be rewarded to an even
able to handle all o the new patients it agreed to accept or have a
greater extent.
Plan B. Plan B might include a oodgate that closes in the orm o
a cap. This has been achieved at some hospitals to maintain sa e
and e ective volumes. Two types o caps exist including those ■ SETTING UP THE RIGHT PROCESSES
requiring a backup system. The backup system can be the exist- Part o managing the practice is ensuring that the right processes
ing hospitalists at a very high labor cost to a hospital or the new are in place. Processes should be established or physician sched-
group o primary care physicians who have asked or coverage; this uling and daily case management meetings. Hospitalist processes
group may need to agree to provide occasional coverage at the should be highly sophisticated to drive improvements in utiliza-
hospital. The latter group o physicians tends to be a short-term tion, documentation, discharge planning, and prospective quality
patch; they can quickly lose their skills and credentialing in the metric monitoring. All o these processes require a tremendous
inpatient setting. Ideally, i the hospitalist group has agreed to amount o time, energy, and in many cases, technology and in ra-
accept a new group o patients, they need to have the capacity structure to drive clinical and nancial per ormance or the hospi-
24 hours per day, 7 days per week. A “sick call” rotation to cover tal and hospitalist practice. The scope o processes is beyond the
anticipated maternity and paternity leaves as well as unexpected scope o this chapter, but this is a core competency that should
absences may have the bene t o allowing hospitalists to ocus not be overlooked in the management o an e ective hospitalist
on career development, especially quality improvement initiatives program. Hospitalist leaders can promote simple solutions that
when they are not seeing patients, and not overwhelming them make it easy or clinicians to communicate at transition points
with service obligations. such as setting up dedicated phone lines in primary care practices
147
so that the hospitalists do not waste valuable time trying to reach than chart review but is o ten based on the discharging physi-
PCPs. Delegating postdischarge phone calls to hospitalist nurses, cian, which may or may not re ect an individual’s per ormance.
APPs or case managers helps create capacity or the hospitalist Supplemented with primary care satis action data and chart review
team. o key quality indicators (eg, trans ers to intensive care units and/or
P
A
readmissions) the hospitalist service can initiate rapid cycle improve-
R
■ THE VALUE OF DATA TRACKING AND REPORTING ments and educational initiatives. The impact o initiatives may be
T
tracked over time.
The well-known saying “you can’t manage what you don’t measure”
I
is quite true in hospital medicine. It is essential to de ne what is ■ PHYSICIAN OUTREACH TO THE COMMUNITY OF
relevant to the practice and measure outcomes that matter most. PHYSICIANS
Many hospitalist teams require sophisticated technology solutions
and partnerships with the hospital to obtain data. Benchmarking A success ul practice ideally includes the community o physicians
T
h
per ormance and then creating an action plan to improve upon raving about the hospitalist group they partner with or their inpatient
e
coverage. Many practices have imploded by not building relation-
S
areas is an e ective approach. Evaluating per ormance on a monthly
p
and quarterly basis is a best practice. ships with the community o primary care providers and specialists.
e
c
Metrics that matter on the revenue side o the equation include: An outreach plan and daily communication on shared patients are
i
a
essential to building the bridges necessary to the medical sta . Noth-
l
t
Volume
y
• ing is more important to this communication than a phone call at
o
• Work RVUs total discharge linking the patient back to the community physician and, as
f
H
• Work RVU/CPT code a bonus, having the opportunity or the hospitalist to let the commu-
o
• Payer mix
s
nity doctor know what a great job he or she did or the patient during
p
• Collections per encounter (accrued)
i
the hospitalization. Faxing, electronic messaging through an EMR, or
t
a
• Charges
l
e-mailing in an HIPAA-compliant way all o the necessary in ormation
M
• Collections per month on discharge is also an essential element to this process.
e
• Readmissions
d
i
c
On the labor cost side o the equation, there are a myriad o labor ■ TRANSITIONS OF CARE BACK TO THE COMMUNITY
i
n
e
cost metrics. However, cost per shi t is a commonly used metric With the advent o the ocus on transitions in care in the bundled
a
that helps to manage the per ormance o the practice, inclusive o payment care improvement program rom CMS, it is now essential
n
d
bene ts or the providers. to link the inpatient hospitalization to the continuum o care in the
S
Outside o nancial per ormance o the practice, there exist three postacute environment—including home care or those patients
y
s
main areas o per ormance monitoring: quality, utilization, and
t
at highest risk or readmissions. Many success ul practices are
e
m
satis action. already on the ore ront o this by placing partner physicians in
s
For quality, most practices manage core measure per ormance these acilities on the same hospitalist plat orm as their acute care
o
and readmission rates. In 2016, most hospitals ace 30-day readmis-
f
partners. It is essential to have clear and deliberate hando plans
C
sions penalties. According to Modern Healthcare’s analysis o the and processes in place, especially i the patient’s next site o care
a
r
CMS data, 38 out o 3,400 hospitals will be subject to the maximum
e
is not home. The goal is to get patients home and stay home or
3% reduction in reimbursement. This requires added ocus and pro- as long as possible.
cesses to address the issues, as the etiology surrounding readmis-
sions is so multi actorial. Mortality and complication rates have an ■ INTEGRATION TO THE POSTACUTE PHYSICIAN
added importance in the years ahead as well. Finally, or those patients discharged rom the hospital back to their
Utilization is a challenge or the average practice without a home, it is essential that the hospitalist ensures a smooth transition.
deep investment in in rastructure. Discharge order time by physi- The window o time rom when patients are discharged rom acute
cian and utilization o the ollow-up CPT codes to discharge codes care until they have a ollow-up visit with their PCP is an especially
are other signi cant measures o throughput or the average risky time rame or adverse outcomes and readmissions. Instituting
hospitalist. The more ollow-up visits a doctor has within a group a patient callback program or having a team that has a plan to check
compared to the same number o discharges o his or her peers on patients postdischarge, having a way to track this data and teams
could indicate the physician holds on to patients longer. Clearly, to that communicate along the care pathway is a necessary and essen-
be statistically valid, a large enough sample size will be needed to tial part o care or a high-quality hospitalist program.
compare to peers. Other important metrics include cost per case
by major DRG group. CONCLUSION
The third pillar o a high-quality hospitalist program is to measure
The creation o new hospitalist practices is likely to continue in the
satis action. Patient, PCP, nurse, and hospital administration satis ac-
coming decade with the changes orthcoming in our health care
tion are key areas to track. Surveys and call centers are e ective tools
delivery system. Building, growing, and managing a success ul and
to monitor and analyze per ormance and test the e ectiveness o
thriving hospitalist practice is possible by ocusing on the essential
improvement initiatives.
elements outlined in this chapter.
By identi ying the parameters or measurement, any hospitalist
service can develop a hospitalist scorecard to clari y the vision and
strategy o the service by gaining consensus regarding what will SUGGESTED READINGS
be measured and reported. The scorecard requires setting targets,
Dye CF. Leadership in Healthcare: Essential Values and Skills. Chicago,
aligning strategic initiatives, allocating resources, and establishing
IL: Health Administration Press; 2010.
milestones. Once these goals are set, a process o communication
and education o the members o the service must take place about Kau man KP. Best Practice Financial Management: Six Key Concepts or
the goals and linking rewards to per ormance measures. There are Healthcare Leaders. Chicago, IL: Health Administration Press; 2006.
pit alls relating compensation to quality o care i the candidate Lee BD, Herring JW. Growing Leaders in Healthcare: Lessons rom the
measure is not attainable or i the per ormance measure link to Corporate World. Health Administration Press. Chicago, IL: Health
awed data. Computer-generated data is much easier to obtain Administration Press; 2009.
148
Levoy B. 222 Secrets o Hiring, Managing, and Retaining Great Employ- ONLINE RESOURCES
ees in Healthcare Practices. Sudbury, MA: Jones and Bartlett
C
Publishers; 2006. Gummadi S, Geyer C, Rossi C. Employment Opportunities or
H
A
Maccoby M. Leaders We Need: And What Makes Us Follow. Boston, MA: Hospitalists Trained in Family Medicine [abstract]. J Hosp Med.
P
Harvard Business School Publishing; 2007. 2014;9(suppl 2). http:/ / www.shmabstracts.com/ abstract/
T
e m p lo ym e n t -o p p o rt u n it ie s- o r-h o sp it alist s-t rain e d -in -
E
Murphy S. Building and Rewarding Your Team: A How-To Guide or
amily-medicine/. Accessed June 1, 2015.
R
Medical Practices. Englewood, CO: MGMA; 2008.
ICU staf ng shortages linked to aging population. ATS Bulletin. http://
2
Nemeth C. Improving Healthcare Team Communication. Burlington,
3
ats-365.ascendeventmedia.com/highlight.aspx?id=1135&p=45.
VT: Ashgate Publishing; 2008.
Accessed December 1, 2011.
Simone KG. Hospitalist Recruitment and Retention: Building a Hospital
B
US Department o Health and Human Services, Health Resources
u
Medicine Program. Hoboken, NJ: Wiley-Blackwell; 2009.
and Services Administration Report to Congress. The Critical
i
l
d
Zahaluk DW. The Ultimate Practice Building Book: How To Regain Care Work orce: A Study o the Supply and Demand or Critical Care
i
n
Control O Your Practice, Achieve A Competitive Advantage In Your
g
Physicians. http:/ / bhpr.hrsa.gov/ healthwork orce/ reports/
,
Local Market, And Reconnect With The Joy O Medicine In The New
G
studycriticalcarephys.pd . Accessed May, 2006.
r
Healthcare Economy. Bloomington, IN: Tra ord Publishing; 2007.
o
w
i
n
g
a
n
d
M
a
n
a
g
i
n
g
a
H
o
s
p
i
t
a
l
i
s
t
P
r
a
c
t
i
c
e
149
CHAP TER
24 INTRODUCTION
According to the Society o Hospital Medicine (SHM), more than
44,000 clinicians have chosen hospital medicine or their career,
approximately 40% o hospitalists are employed directly by hospi-
tals, and the practice o hospital medicine is the astest growing
medical specialty, growing at a rate o 5% to 10% annually.
Because o the strong demand or hospitalists, recruiting compe-
tition is erce. Consequently, hospitalist practices must view recruit-
ing and retention as core competencies. This chapter provides best
Best Practices practices and key principles or recruiting and retaining quali ed
hospitalists.
in Physician In addition to updated in ormation provided by Kirk Mathews

and Dr. John Nelson, this chapter includes in ormation rom the
Recruitment and 10 Key Principles and Characteristics o an E ective Hospital Medi-

cine Group developed by the SHM Key Characteristics Workgroup.
The 10 key principles are e ective leadership, engaged hospitalists,
Retention adequate resources, e ective planning and management in ra-
structure, alignment with hospital and/or health system, support o
care coordination across the care setting, taking a leadership role in
clinical issues, scope o clinical activities, a patient/ amily-centered,
Steven B. Deitelzweig, MD, MMM team-based model, and recruitment and retention o quali ed clini-
cians. While all o these principles are important aspects o a suc-
R. Kirk Mathews, MBA cess ul group, this chapter ocuses primarily on the 10 characteristics
John Nelson, MD, MHM that support Principle 10—recruitment and retention: sourcing and
recruiting candidates who are committed to a career in hospital
Society o Hospital Medicine Key
medicine and who are board certi ed or board eligible, developing
Characteristics Workgroup 1 a good orientation program, providing resources or pro essional
growth, paying competitive compensation, ensuring that employ-
ment agreements are valid and air, measuring job satis action, and
monitoring clinical competency and pro essionalism.
SOURCING AND RECRUITING CANDIDATES

The sourcing and recruitment process consists o seven steps:
1. Preparing a job description
2. De ning the pro le o a quality candidate
3. Finding candidates
4. Managing the application process
5. Interviewing candidates
6. Making selection decisions
7. Extending job o ers
■ 1. PREPARING A JOB DESCRIPTION

A good job description can help attract the right candidates, create
appropriate expectations, assist in evaluating employee per or-
mance, and more. A well-written job description will contain the
ollowing elements:
• Position title
• Department
• Reports to
• Overall responsibility
• Key areas o responsibility
• Terms o employment
• Quali cations and credentials
1
Members o the Society o Hospital Medicine Key Characteristics Workgroup are
Patrick Cawley, MD, Steven Deitelzweig, MD, Leslie Flores, MHA, Joseph Miller, MS,
John Nelson, MD, Scott Rissmiller, MD, Laurence Wellikson, MD, and Winthrop
Whitcomb, MD.
150
When preparing a job description, it can also be use ul to re ect the board o directors. Alternatively, a hospitalist could
upon the group’s culture and include candidate characteristics that participate in other regional or national activities in which
C
are compatible with that culture in the quali cations section. For there is interaction with hospitalist peers on issues/topics
H
example, i the group unctions in a very team-oriented manner, an
A
relevant to the specialty.
P
otherwise “per ect” candidate who strongly pre ers to work autono- Become a Fellow in Hospital Medicine. SHM established
T
mously and avoid team meetings would not be a good t. the Fellowship in Hospital Medicine program to recognize
E
A good job description is not just a laundry list o responsibilities. hospitalists who have committed to the specialty. The three
R
It should re ect the culture, mission, and priorities o the group. levels o recognition are Fellows, Senior Fellows, and Masters.
2
It should be reviewed and updated or each hire so it re ects the Criteria include 5 years as a practicing hospitalist, no disci-
4
group’s current situation and goals. plinary action that resulted in the suspension/revocation
Finally, the job description must avoid any re erence to race, color, o credentials or license, and endorsement by two active
B
religion, age, gender, national origin, nationality, or physical or men- SHM members. Additional criteria are the demonstration o
e
s
t
tal disability as such re erences are illegal. personal dedication in any or all o the ollowing: teamwork,
P
quality improvement, and leadership.
r
a
■ 2. DEFINING THE PROFILE OF A QUALITY CANDIDATE
c
Earn the Focused Practice in Hospital Medicine MOC Credits.
t
i
c
Every good candidate sourcing strategy begins with the establish- The American Board o Internal Medicine and the American
e
Board o Family Medicine established the Focused Practice
s
ment o candidate parameters, including training requirements
i
n
(MD, DO, internal medicine, amily practice, etc), board certi cation, in Hospital Medicine (FPHM) program as part o their MOC
P
and a commitment to a career in hospital medicine. processes. The FPHM program “assesses, recognizes and sets
h
y
standards or the speci c knowledge, skills and attitudes o
s
• Training Requirements: Practice leaders must have a good
i
c
understanding o their customers—the re erring physi- general internists who ocus their practice in the care o hos-
i
a
pitalized patients.”The FPHM MOC program does not result
n
cians—when establishing candidate training requirements.
in a subspecialty designation; it recognizes internal medicine
R
For example, the decision to hire amily practice physicians
e
and amily medicine physicians who practice as hospitalists.
c
as hospitalists should only be made with the approval o the
r
u
re erring medical sta .
i
t
■ 3. FINDING CANDIDATES
m
• Board Certi cation: All ull-time and regular part-time hos-
e
pitalists should be board certi ed or board eligible in an One o the most important recruiting principles is to ensure that the
n
t
applicable medical specialty or subspecialty. Certi cation by group’s hospitalists are actively engaged in sourcing and recruiting
a
n
a medical specialty board has become an accepted structural new group members. Involving the sta hospitalists with the sourc-
d
measure o physician quality and competence. In the United ing and recruitment o new physicians can be valuable or several
R
e
States, 24 medical specialty boards certi y physicians in vari- reasons:
t
e
ous specialties and subspecialties. To become board certi ed, • They might be able to identi y good sources o candidates.
n
t
a physician must receive supervised, in-practice training or
i
• The common identity and engagement o the group’s hospi-
o
n
several years a ter medical school. In addition, specialty boards talists are rein orced as they seek to identi y and recruit physi-
require passage o a written exam. The exams are intended to cians who would be a good t with their practice.
assess medical knowledge and clinical judgment. • Physician candidates are well in ormed about the position
Medical board certi ications are time limited, typically or because they meet their potential colleagues in person and
10 years. Doctors whose certi icates are time limited must can get an honest perspective on what it would be like to work
success ully complete recerti ication requirements under in the group.
a program called Maintenance o Certi ication (MOC). The • Both parties, the hospitalists in the group and the physician
policy o the American Board o Medical Specialties states candidates, can make in ormed decisions.
that “maintenance o competence should be demonstrated A candidate sourcing plan is similar to an investment port o-
throughout the physician’s career by evidence o li elong learning lio—diversity is good. In addition to recommendations rom sta
and ongoing improvement o practice.” Each specialty board is hospitalists, the ollowing recruitment strategies and tools can help
implementing this policy in its own way, but all are committed practices identi y candidates.
to a program that requires that the physician to do the ollowing:
• Existing Medical Staf : Asking medical sta leadership to
Maintain a license in good standing with state licensing suggest sta physicians who might be good candidates could
boards yield good results. Leaders are likely to know which physi-
Periodically do surveys o patients and o peers cians are the best clinicians and who might be thinking about
Periodically show evidence o knowledge and judgment changing their practice situation.
Show evidence o a commitment to li elong learning and • Medical Staf Re errals: Newly trained physicians will be
involvement in a periodic sel -assessment process aware o physicians behind them in the next graduating class,
Periodically show evidence o sel -evaluation o per or- and recently relocated experienced physicians will know or-
mance in practice mer colleagues who might be good candidates. Some employ-
• Commitment to Hospital Medicine: To maintain the stability ers o er a nancial incentive or candidates hired as a result o
o the practice, a signi cant proportion o the ull-time hospi- a medical sta re erral.
talists in the group should demonstrate a commitment to a • Residency Program Relationships: Establishing relationships
career in hospital medicine. A hospitalist can demonstrate a with the program directors o residency programs within
commitment to a career in the specialty in several ways: 200 miles can yield outstanding low-cost results. O ering to
Participate in specialty pro essional activities. The SHM is provide a noon lecture or some other resource to the program
the pro essional society or hospitalists. A hospitalist can director can be an excellent way to begin to develop the rela-
join SHM and take advantage o the educational programs, tionship and to gain exposure to the residents. This exposure—
resources, and meetings and/or seek a leadership role at even exposure to rst- and second-year residents—is valuable
SHM through participation in committees, task orces, and/or or uture recruiting.
151
• Other Local/Regional Hospitals: Recruiting physicians rom visit. Per orming this kind o in-depth assessment be ore extend-
other local hospitals can be e ective because these physicians ing an interview invitation will minimize the possibility o learning
have already decided to live in the area. I the hospital is a something about the candidate during the onsite interview that will
riendly competitor or a patient re erral source, political and/or eliminate him/her rom consideration.
P
A
business caution is advised. This approach to recruiting is oreign to most employers but is
R
• Online Hospitalist Job Boards: Because candidates do most best illustrated by considering how major universities recruit their
T
o their job investigations online, online job boards are an star athletes. College coaches never invite athletes to visit their
I
important recruiting tool. The SHM Career Center (www. campus or of cial recruiting visits be ore they thoroughly assess
hospitalmedicine.org/careercenter) is an excellent candidate them. College coaches watch hours o game lms o the prospec-
sourcing tool. Other popular job boards are hospitalistjobs tive recruit. They interview the candidate by phone. They interview
.com, practicelink.com, jobs.todayshospitalist.com, nejmcareer- the recruit’s high school coach, guidance counselor, and sometimes
T
h
center.org, and jama.careers.adicio.com. even the girl riend or boy riend! Only a ter they are convinced the
e
• Journal Advertising: Journal ads should be eye-catching and athlete is someone they want to recruit do they invite him/her to
S
p
brie and should prompt the candidate to take action—either campus or the of cial visit. During the visit, the recruit will brie y
e
place a phone call or visit a website to learn more. Many medi-
c
visit ace to ace with the head coach, and the rest o the visit is
i
a
cal journals also o er online job boards. organized to convince the athlete that his/her athletic and academic
l
t
y
• Direct Mail: Because direct mail campaigns usually generate goals can be achieved at that university.
o
a very low return, an inexpensive and color ul postcard is the To bring this same level o thoroughness to the hiring process or
f
H
most cost-e ective tool, and purchasing a very targeted mail- a hospitalist, here are the aspects o each candidate that must be
o
s
ing list is necessary. assessed prior to an onsite visit:
p
i
• Career Fairs: Career airs are designed to give second- and
t
• Training and Clinical Skills: Identi y and question any gaps
a
l
third-year residents exposure to various employers and to
in the candidate’s training. Most candidates will be able to
M
give employers the opportunity to begin recruiting relation-
e
provide a thorough explanation or gaps in their training and/
d
ships with candidates. Career airs are most e ective i they are
or work history without dif culty, but employers should veri y
i
c
within relatively close geographic proximity o the practice.
i
n
the candidate’s explanation whenever possible. Also veri y that
e
the candidate has the skills (including any procedural skills)
a
n
required or the practice.
d
PRACTICE POINT • Prior Work Experience: Discuss all prior work experiences
S
y
• One o the most important recruiting principles is to ensure with the candidate. Asking the candidate about the reasons
s
t
e
that the group’s hospitalists are actively engaged in sourcing or terminating each employment relationship can uncover a
m
and recruiting new group members. pattern o discontent or employer dissatis action.
s
• Communication Skills: Good verbal communication skills are
o
f
critical or a hospitalist. Many employers make the mistake o
C
a
• Recruitment Firms: Some groups use recruitment rms to talking too much when assessing a candidate. Set the stage
r
e
extend their reach on dif cult-toll searches. Contingency- with a question and then allow the candidate to speak as long
based companies only charge a ee i they are success ul. as s/he wishes.
Although using a contingency-based company is a risk- ree • Work Ethic and Workplace Attitude: The ollowing questions
approach, contingency recruiters may shop marketable candi- can help determine a candidate’s work ethic and workplace
dates to a variety o potential employers with no real obligation attitude:
to any o those employers. Retained recruiting rms typically
charge a monthly ee in return or their e orts to ll a position. How many patients do you eel you can com ortably see in
However, retained rms rarely o er a guarantee o success, so one day?
the employer takes the risk o making a signi cant investment Describe a situation where you elt overworked.
without success ully hiring a candidate. Both types o rms can How hard do you like to work?
and do produce results, but be ore signing a contract, employ- When do you typically begin your workday?
ers thinking o using a recruitment rm should rst establish a Describe a time when you were asked to work overtime
relationship with the recruiter who would be assigned to their without compensation.
search and make sure they have an acceptable level o com ort Describe any volunteer work that you have per ormed
with the recruiter’s approach. including outside the eld o medicine.
What personal grati cation do you get rom your work?
■ 4. MANAGING THE APPLICATION PROCESS • Re erences: The process o checking re erences has become
less valuable because employers and program directors ear
legal recourse i they provide a negative report. However,
PRACTICE POINT because checking re erences is a required element o any
recruitment, here are tips to enhance the process:
• The most important aspect o managing the application
process is to make every e ort to thoroughly assess key Ask or speci c re erences. Employers should always ask to
aspects o each candidate prior to extending an invitation speak with the candidate’s supervisor or residency program
or an onsite visit. director. Other potentially valuable re erences are hospital
administrators, re erring physicians, unit nurses, and ED phy-
sicians who have worked with the candidate.
Because onsite interviews are time consuming and expensive, Speak directly with the re erence whenever possible. Much
properly managing the application process is imperative. The most can be gained rom listening care ully when the re erence is
important aspect o managing the process is to make every e ort to questioned. Pregnant pauses and voice in ections can speak
thoroughly assess key aspects o each candidate prior to an onsite volumes about what that person believes about the candidate.
152
Describe the practice. Employers should describe their • Control the Agenda. The interview team should be com-
practice and ask how the candidate might per orm in posed o the very best the practice has to o er, and they
C
such a practice. Attempt to engage the re erence in a should be well prepared or their role. The employer should
H
meaning ul dialogue about the candidate’s strengths and hand-select this team, bypassing chronic complainers.
A
P
weaknesses. • Invite the Candidate’s Spouse. Practices recruit an entire
T
amily, not just the physician. O ten the spouse represents
E
more than 50% o the decision to accept or decline an o er, so
R
■ 5. INTERVIEWING CANDIDATES
every e ort should be made to set a date or the visit when the
2
The purpose o the ace-to- ace interview is to con rm an evalu- spouse will be able to attend.
4
ation that has already been conducted. This approach allows the • Do Not Invite the Candidate’s Children. Encourage the candi-
employer to ocus on what should be the real goal o the onsite date to leave children at home i possible. Parents are likely to be
B
visit—convincing the candidate that s/he wants the job! distracted i the children attend. I the candidate insists in bringing
e
s
When a physician candidate has passed the preinterview assess-
t
the children, arrange or child care during portions o the visit.
P
ment and the likelihood is high that s/he will receive an o er during • Plan or One Onsite Visit. I the preinterview assessment and
r
a
the onsite visit, the candidate should then be invited to an onsite
c
the onsite visit are well structured and thorough, the employer
t
visit that includes the ollowing elements:
i
c
should be prepared to make a decision a ter the visit. Some
e
employers think that i they like the candidate on the rst inter-
s
• Ice-Breaker Event: O ten a meal, co ee, or an open-house
i
n
type o event is best or an ice breaker. The event should be view, they can bring him/her back or a second visit. In the current
P
casual and in ormal with plenty o opportunity or discussion recruiting environment, they may not get a chance at a second
h
y
and social interaction. trip. It is important the interview be structured to ensure both
s
i
c
• Business Discussion: This is the time to review the nancial the candidate and the group leaders obtain all the in ormation
i
a
elements o the position, including a review o the employment required to make a decision about moving orward—or not.
n
Remember that Time Kills All Deals. Because recruiting com-
R
agreement. The employer should make sure the candidate •
e
petition is strong, employers must act with a sense o urgency,
c
has the opportunity to ask any questions about the nancial
r
u
arrangements including income, bene ts, and vacation. responding to candidate questions immediately, providing
i
t
a sample agreement during the interview, and sending an
m
• Community Introduction: The community tour should
e
include introductions to key people in the community who executable agreement immediately ollowing the visit. Every
n
candidate should receive a ollow-up phone call within a day
t
might play a role in the candidate’s li e a ter relocation. For
a
example, i the candidate has children o school age, a drive by or two o the visit.
n
d
the school would be nice, but making an appointment or the • Expect the Unexpected. Employers should be exible enough
R
candidate to meet the school principal is better. to allow or unexpected and unavoidable changes to the inter-
e
t
view agenda.
e
• Time with Future Colleagues: Every onsite visit should pro-
n
t
vide time or the candidate to communicate with other physi-
i
o
cians in the practice about how the practice unctions. The ■ 6-7. MAKING SELECTION DECISIONS AND EXTENDING
n
candidate should be encouraged to ask any questions about JOB OFFERS
practice operations. As already stated, it is critically important or employers to respond
• Hospital Tour: A hospital tour should include time with both to candidates’ questions and concerns in a timely manner ollowing
physician and nonphysician hospital leaders. For example, the onsite visit. Negotiations will be simpli ed i the practice has
even a brie hello and handshake with the hospital CEO can valid and comprehensive employment/independent contractor
convey a valuable message to the candidate that the hospital- agreements. A ormal contract is important to both the hospitalist
ist program is important to the hospital. and the employer. A contractual relationship requires the parties to
• Wrapup Discussion: The wrapup discussion is the time to ask think clearly about their expectations and obligations. The contract-
or the candidate’s initial eedback on the visit, correct any alse ing process should allow the parties to articulate what they want
impressions, and ask the candidate i s/he wants the job. out o the arrangement and to discuss important practical issues.
Furthermore, even the best o relationships may change. The parties
may change their minds about the type o contract terms to which
PRACTICE POINT they wish to be bound. A ormal contract ensures that even during
• Many employers make the mistake o talking too much when periods o disharmony, the parties will be required to abide by the
assessing a candidate. Set the stage with a question and then agreed-upon contract terms.
allow the candidate to speak as long as s/he wishes. In 2011, the American Medical Association published the Anno-
tated Model Physician-Hospital Employment Agreement that suggests
physician-hospital agreements address the ollowing topics:
Here are tips to keep in mind when planning a candidate visit. 1. Preliminary considerations and basic agreements
• Be Honest. No practice is per ect and no community is uto- 2. Term
pia, so employers should not try to hide de ciencies rom 3. Duties o the physician
the candidate. However, take care not to overemphasize 4. Employer’s obligations
any potential concerns. This can happen inadvertently when 5. Physician compensation
every person who speaks with the candidate mentions a 6. Reimbursement o expenses
negative aspect o the community. They are doing so with 7. Employer-paid bene ts and time o
the candidate’s best interest in mind, but repetition might 8. Loyalty and con dentiality covenants
give more weight to the concern than is warranted. One way 9. Termination
to avoid this problem is to assign one person to mention 10. Disability or death
the concern and make sure the entire interview team under- 11. Remedies
stands who has received the assignment. 12. Miscellaneous
153
When a candidate wants to negotiate one or more elements o areas to ensure that the plan does not inadvertently violate any
the contract, the employer should implement “big picture nego- legal requirements.
tiations.” Both parties are best served when each considers speci c • Can Be Modi ed over Time: Practices o ten evolve to the
contract requests in light o all requested contract concessions. The point that the compensation plan is not a good match with
P
A
employer should in orm the candidate o this policy and instruct the current situation and may even inhibit the ability o the
R
him/her to thoroughly review the employment agreement and practice to make adjustments in scheduling and other areas.
T
present all concerns/requests at one time. This approach will pro- For example, compensating hospitalists at a set dollar amount
I
tect the employer rom making a signi cant concession without per shi t worked may be reasonable today, but in the uture the
knowing i another signi cant request will ollow. I the candidate schedule may need to change so shi ts are longer or shorter
presents a long list o contract requests, the employer should ask and no longer match the single shi t rate in the compensation
which, i any, o the requests are deal breakers and address these rst plan. Because it is typically dif cult and time consuming to
T
h
to avoid the rustration o working through a long list o relatively make signi cant changes to a compensation plan, a practice
e
minor concerns only to see the deal all apart at the end. These might decide to simply keep the compensation plan as it is
S
p
negotiations should always be conducted in an unemotional, busi- and orgo making the needed scheduling changes. In this case,
e
nesslike manner. Such conduct will have the added bene t o giving the compensation plan is an impediment to e ective practice
c
i
a
the candidate con dence that the practice leaders handle dif cult operations. In many practices, a compensation plan based on
l
t
y
issues with pro essionalism. hours o work or productivity will be easier to adjust in response
o
to changes in scheduling and patient volume.
f
H
COMPENSATION • Rewards Good Work: Ideally, a compensation plan should
o
s
To attract and retain quali ed hospitalists, compensation must be encourage and reward the per ormance and behaviors that the
p
i
practice desires. A portion o total compensation may be tied
t
market competitive. According to The State of Hospital Medicine in
a
l
2014 report prepared by the SHM, median compensation or adult to the doctor’s citizenship in the practice or to per ormance
M
hospitalists increased 8.2%— rom $233,855 to $252,996—since the on quality measures or other domains. The amount o money
e
d
2012 report. at stake and the thresholds that trigger payment must be
i
c
planned care ully to ensure that they in uence behavior and
i
When a practice hires a new hospitalist or per orms an annual
n
e
review, the level o compensation can set the tone or long-term are not viewed as too easy or too dif cult to reach.
a
satis action or disappointment. I the practice pays below market
n
d
compensation, it runs the risk o losing good talent and/or creating
S
PRACTICE POINT
y
resentment. I the practice pays above market compensation, it can
s
t
create an entitlement mentality among some hospitalists, generate •
e
When a practice hires a new hospitalist or per orms an annual
m
unnecessary expenses or the practice, and create potential compli- review, the level o compensation may set the tone or long-
s
ance issues. term satis action or disappointment. I the practice pays below
o
f
An e ective practice has a compensation philosophy that re ects market compensation, it runs the risk o losing good talent
C
the values o the practice, conducts suf cient research to ensure
a
and/or creating resentment. I the practice pays above market
r
e
that the compensation program is market competitive, has a com- compensation, it may create an entitlement mentality among
pensation communication strategy, and explains individual com- some hospitalists, generate unnecessary expenses or the
pensation decisions credibly. practice, and create potential compliance issues.
ATTRIBUTES OF A WELL-DESIGNED COMPENSATION PLAN

Groups should compute the value o total compensation to the METHODS OF HOSPITALIST COMPENSATION
hospitalists: a summary o salary, bene ts, and other costs required A reasonable way to think about the various methods or ormulas
to retain, train, support, and reward a hospitalist. Although the used to compensate hospitalists is to think o them as consisting o
amounts and methods o hospitalist compensation have evolved, up to three distinct elements. These three elements can be com-
the characteristics o a desirable compensation plan have not bined in varying proportions to create a compensation method.
changed. These characteristics are as ollows: • A Fixed Component, O ten Called a Base Salary: A xed
• Is Easy to Understand: Any compensation plan should be salary has the advantages o being known to all parties in
simple enough that the hospitalist can explain it rom memory. advance, eliminating uncertainty in budgeting and providing
Complicated ormulas used as the basis or paying a quality certainty or the hospitalist. However, a xed salary model
bonus or end-o -year pro t distribution are problematic i the alone is not a good choice or most practices because it does
hospitalists do not ully understand them. Disputes and resent- not reward or encourage good per ormance. Hospitalist com-
ment may arise when a doctor has misunderstood the ormula pensation surveys show a xed salary was the most common
and anticipated a larger bonus than the one that was paid. method in use in 1997, but since then its popularity has steadily
• Is Easy to De end in Public: Despite e orts to keep the details declined in avor o a combination o xed and variable ele-
o a compensation plan private, the in ormation can become ments based on production, per ormance, or both.
public through members o the practice or during a mal- • A Productivity Component Based on Patient Care Work-
practice suit. Avoid incentives that could be embarrassing or load or Production: Here is a sample compensation plan or-
damaging. For example, a nancial reward or reducing length mula based entirely on production: (collected professional fees +
o stay could be interpreted as an incentive to send patients other support [if any]) – overhead. The “other support” in the or-
home be ore they are ready. mula is most commonly money paid by the hospital in which
• Complies with All Laws and Regulations: Because the regu- the hospitalists see patients. Production compensation has the
lations governing physician compensation and the nancial advantages o allowing each doctor to work more or less than
relationships between doctors and hospitals or other entities others in the group, and it encourages everyone in the group
are complex and always changing, compensation plans should to be attentive to optimal staf ng and scheduling and to busi-
be always reviewed by personnel knowledgeable in these ness practices such as proper current procedural terminology
154
(CPT) coding. A signi cant weakness in compensating hospital- ORIENTATION FOR NEW STAFF
ists solely on production is that it incentivizes volume and not
C
An e ective orientation program will make a positive rst impres-
quality. Potential consequences are the occasional hospitalist
H
sion on new clinicians, acilitate more rapid assimilation, contribute
who takes on an unreasonable and unsa e workload and pro-
A
to job satis action and retention, and re ect eedback rom hospital-
P
viding nancial rewards or increased lengths o stay.
ists who have participated in the program.
T
• A Per ormance-Based Component: Compensation elements
An orientation program provides an opportunity to welcome the
E
related to per ormance may be based on quality targets such
R
new clinician; introduce him/her to colleagues and key support sta ;
as patient and/or re erring physician satis action, proper CPT
2
convey the group’s vision, mission, and values; clari y job expecta-
coding, committee participation, and meeting attendance. An
4
tions and responsibilities; amiliarize the new clinician with the work
e ective plan requires that a meaning ul dollar amount o sal-
environment; and mentor recent graduates o training programs.
ary be at risk, and care ul choices o metrics and thresholds on
Potential topics to address in a hospitalist orientation program
B
which payment is based must be made. Good metrics include
e
include the ollowing:
s
t
those that hospitals are already required to measure and report,
P
such as per ormance on quality measures and patient satis ac- • History and philosophy o the practice
r
a
• Policies and procedures
c
tion. The number o metrics should be small enough that the
t
i
• Hospitalist job description
c
dollar amount o payment available or each remains signi cant.
e
Consequently, no more than three or our metrics should be • Compensation program, including incentives
s
i
• Medical record standards
n
part o per ormance compensation. Changing the metrics in
P
the plan every year or so is reasonable. For thresholds that trig- • The hospital’s EMR system
h
• Key members o the medical and hospital sta
y
ger payment, i achieving the threshold is too easy, there will
s
i
• Coding/documentation
c
be little or no e ect on behavior. I it is too dif cult, the doc-
i
a
tors may ignore it. Setting a single per ormance threshold that Good practices are to pair the new clinician with an established
n
triggers payment o all available dollars o ten causes problems.
R
hospitalist or a period o shadowing and to conduct an orientation
e
Instead, a predetermined scale— rom no payment to the maxi- tour that includes the ollowing:
c
r
u
mum amount available—is a better, more equitable approach.
• Hospital departments (emergency department, pharmacy,
i
t
For example, i patient satis action or hospitalist patients is
m
nursing, case management, etc)
e
currently 70% and the goal is 75%, the portion o dollars paid
• Hospital administration
n
could be determined by subtracting the baseline (70% in this
t
• Key re erring physicians (PCPs)
a
example) rom actual per ormance (assume it is 73.8%). In this
n
Other best-practice elements o a hospitalist orientation are the
d
example, patient satis action improved 76% ([3.8 ÷ 5] × 100),
R
so 76% o the available dollars would be paid as a measure o ollowing:
e
t
percentage o improvement toward the goal.
e
• A welcome meeting
n
• An announcement in the local newspaper
t
i
o
SOURCES OF DATA ON HOSPITALIST COMPENSATION • An update o the practice website and brochure
n
AND WORKLOAD • Printed business cards
The Medical Group Management Association conducts the largest • Assignment o a mentor or 60 to 90 days
survey o hospitalist compensation, productivity, and pro essional
ee collections each year; results are published each summer based
EDUCATIONAL OPPORTUNITIES
on data collected at the end o the prior calendar year. The SHM
State of Hospital Medicine Survey is published in the summer on even Pro essional education and development are a ormal requirement
years and includes detailed data on hospitalist schedules, incentive o the pro ession o medicine, as speci ed by the American Medi-
compensation metrics, scope o clinical services, advanced practice cal Association (AMA), the various specialty boards, state licensing
clinician (NP and PA) staf ng, and other topics. agencies, and the Accreditation Council or Continuing Medical
Education (ACCME). To maintain their licenses and/or specialty
accreditations, physicians must obtain a suf cient number o con-
PRACTICE POINT tinuing medical education (CME) credits during a speci ed period.
• An e ective practice has a compensation philosophy that ACCME and the AMA de ne CME as “educational activities that
re ects the values o the practice, conducts suf cient serve to maintain, develop, or increase the knowledge, skills, and
research to ensure that the compensation program is market pro essional per ormance and relationships that a physician uses to
competitive, has a compensation communication strategy, and provide services or patients, the public, or the pro ession.”
explains individual compensation decisions credibly. The AMA PRA (Physician’s Recognition Award) Category 1 CreditTM
system has become the CME standard or licensing boards and spe-
cialty organizations nationwide and is recognized by all US jurisdic-
tions. For an activity to be designated or AMAPRACategory 1 Credit,
RETENTION it must be certi ed by an accredited CME provider.
Physician turnover is expensive and disruptive, underscoring the An e ective practice provides its hospitalists with resources or
critical importance o good recruiting. Every practice should work pro essional growth and enhancement, including access to CME.
hard during the recruitment phase to ensure the candidate is a good A wide variety o educational activities quali y or CME certi cation,
t or the practice. However, the work does not end once the new including lectures and meetings and activities based on recorded or
hire is brought on board. Retention is an ongoing process with ele- published content.
ments directed to brand-new hires and to seasoned veterans alike. The practice should have a policy or encouraging and nancially
Important components o a retention strategy are an orientation supporting the continuing education and pro essional develop-
process or new sta , resources or pro essional growth and devel- ment o its hospitalists, including CME credits, and should have
opment, job satis action monitoring, and a documented method or a procedure or con rming that all hospitalists in the group have
monitoring clinical competency and pro essionalism. participated in the CME program.
155
HOSPITALIST JOB SATISFACTION E ective groups will systematically—at least yearly—monitor job
satis action and well-being through meetings with individual hospi-
One o the most important aspects o retention is to measure,
talists and/or anonymous surveys. Practice leaders should also rou-
monitor, and oster hospitalists’ job satis action, well-being, and
tinely communicate examples o initiatives undertaken to address
P
pro essional development.
A
hospitalist job stress, dissatis action, and/or burnout.
Job stress and dissatis action among physicians can lead to burn-
R
Here are additional strategies or enhancing job satis action:
out and a range o undesirable outcomes, including unplanned
T
turnover, absenteeism, judgment/action errors, and con icts/alien- • Create a work/li e/compensation balance that makes sense
I
ation rom pro essional colleagues. Furthermore, the potential or and do everything possible to keep it in balance. I any
more tangible adverse outcomes such as accidents, litigation, and one element o this three-sided scale is signi cantly out o
increased worker compensation cases may exist. Research has balance with the other two or an extended period o time,
documented that work stress and dissatis action also can lead to turnover is inevitable.
T
h
physical illness. Finally, job stress and dissatis action may lead to a • Provide a physician eedback orum o some kind. Frustration
e
will build when physicians eel that their issues are not being
S
poor balance between work and personal li e and the reliance on
p
maladaptive coping strategies such as drug and alcohol abuse and heard. They may not always get the answers they wanted,
e
c
dependence. There ore, it is important that the group leadership but air-minded physicians will understand i they are at least
i
a
l
assume responsibility or addressing the job satis action o the hos- allowed such a orum.
t
y
pitalists in the practice. • Use team-building exercises or events, such as social events
o
f
The SHM publication AChallenge for a New Specialty: AWhite Paper and pro essional retreats, to keep individuals rom eeling
H
isolated within the group.
o
on Hospitalist Career Satisfaction outlines the ollowing job stresses
s
p
and dissatis actions that hospitalists ace:
i
t
a
CLINICAL COMPETENCY AND PROFESSIONALISM
l
M
Patients and the public in general expect their health care pro-
e
viders to be clinically competent and to act pro essionally. The
d
Nature of Nature of the Work External
i
c
group leadership is accountable or the care delivered by the
i
the Work Environment Influences
n
physicians and other providers in the practice. Clinicians can
e
• High acuity/ • Volume o work • Impact o ACGME
a
make one or more signi icant errors, exhibit poor judgment,
n
complexity o • Time pressure work rules on behave poorly with patients or other clinicians, or demonstrate a
d
illness/lack o patient care/
S
• Night and pattern o poor or unsa e care. A documented, structured process
y
predictability teaching
s
weekend or identi ying and addressing such issues will not only prevent
t
• Li e and death • Legal and
e
coverage the well- unctioning members o the practice rom becoming
m
implications o responsibilities regulatory
resent ul o a practice that allows a colleague to get away with
s
clinical decisions concerns
o
• High census providing poor care but can also, in many cases, remediate the
f
• Provider • Financial
C
conditions troublesome clinician.
a
interdependency pressures rom
r
• Intermittent Clinician competency and pro essionalism should be addressed
e
and payers
communication demand (beeper at two points in time: when the clinician applies or privileges and
always going o ) Career/ is initially credentialed and when a de ciency is identi ed related
• Limited patient organizational
• Workplace to current clinical competence, practice behavior, or the ability to
in ormation issues
con licts and per orm a procedure. Practice leadership should monitor clinicians
• Administrative interruptions • Reimbursement
to identi y de ciencies in care and implement appropriate remedies
and based on o ice
• Workplace or health care providers ound to be de cient such as remedial
documentation model
requirements discrimination training, proctoring, reassignment, and, i remedial actions ail,
• No established
• Lack o termination rom the practice.
• Medical legal risk track or
understanding promotion
• Potential hostility o the role o
rom patient’s the hospitalist • Little control PRACTICE POINT
amily by hospital over key issues
Personal issues administrators (workload, • One o the most important aspects o retention is to measure,
schedule, case monitor, and oster hospitalists’job satis action, well-being, and
• Pro essional • Hospitalists pro essional development.
types)
advancement working on a
temporary basis • Con lict between
• Financial
while waiting service mission
pressures and other equally
to pursue other CONCLUSION
• Pressures rom important
career plans An understa ed hospitalist program or one with a number o
spouse/ amily responsibilities
• Medical sta dissatis ed clinicians will quickly become unstable, resulting in
• Unrealistic job • Limited
con licts signi cant clinical and nancial consequences. It is imperative that
expectations pro essional
• Ergonomics hospitalist program leaders become pro cient in recruiting and
• Inability to say recognition
(poorly designed and unding retaining the physicians they need to provide the outstanding care
“no” work space and/ or scholarly their patients deserve.
or equipment) activities
• Limited • Leadership
workspace structure within
the hospital (not
“at the table”)
156
25
Multidisciplinary care re ers to the active collaboration between
various members in the health care system to deliver optimal
care or every hospitalized patient. Success ul teamwork is a core
competency that can be taught and incorporated into patient care
processes. The Association o American Medical Colleges (AAMC),
the Accreditation Council or Graduate Medical Education (ACGME),
and the Society o Hospital Medicine (SHM) require speci c team-
work-related competencies or medical students, residents, and
Teamwork in hospitalists. Hospitalists can improve multidisciplinary care o hospi-
talized patients by demonstrating group dynamic skills, conducting
Leadership and e ective multidisciplinary team rounds, evaluating per ormance,

providing eedback, teaching about error and how teamwork and
Practice-Based communication can reduce error, and by leading quality improve-

ment initiatives.
The US health care system is a highly organized and complex
Management system. Over the last three decades o the public sa ety move-
ment there have been landmark studies and published reports
about individual and systemic ailures that have not only cost
lives but wasted billions o US dollars while delivering unsa e care.
Scot T. Smith, MD Although di erent solutions may be debated, it is clear that the
US health care system will need to be redesigned to deliver the
Scott F. Enderby, DO, SFHM highest quality o care possible while being ever mind ul o ways
Robert A. Bessler, MD to improve e ciencies and reducing the cost o care delivery.
Sweeping change requires e ective teamwork on every level:
hospital networks, hospital, hospitalist service, skilled nursing, and
home care—a more comprehensive approach to multidisciplinary
patient care.
In general, most physicians have little ormal training relating
to complex hospital systems or human error and lack insight into
their own limitations during conditions o stress, lack o sleep, or
conf icting demands. Strong hierarchy, power di erentials, lack o
clarity requiring speci c tasks and roles, and lack o coordination
are common teamwork and communication ailures in health care.
Lessons learned rom the aviation industry can be applied to the
delivery o hospital care, and hospitalists can take steps to reduce
the likelihood o (1) individual error resulting rom physiologi-
cal and psychological limitations o human beings and (2) team
errors resulting rom ailure to act or deviation rom established
standards. Although it is not possible to eliminate individual error,
systems can be designed that reduce the likelihood o error and
make hospitals a sa er environment or patients. Working in teams
and serving as the hub o communication network in the hospital,
hospitalists are ideally poised to change the culture o “how we do
things around here” by serving as clinical role models and as lead-
ers o patient sa ety on the multidisciplinary care team and truly
coordinating care throughout the acute episode o care. Without
e ective teamwork and medical leadership, however, these com-
plex systems have been shown to be less e ective in producing
quality outcomes.
THE MULTIDISCIPLINARY HEALTH CARE TEAM

The hospitalist team is a unit o pro essionals that directly provides
care and so most directly impacts the patient experience and the
quality o care. Composition o the team varies, but a team typically
may include hospitalists, consulting physician(s), advanced care
practitioners, nurses, case managers, social workers, and pharma-
cists, amongst others.
157
Individuals have particular tasks based on their particular PRACTICE POINT
specialties, but the hospitalist team depends on each other or
situational awareness and goal success strategies. Situational Provide care that satisf es the patient
awareness is a common, accurate understanding o the patient’s • Improve the patient’s perception o your team. Patients o ten
P
eel that their care team is not coordinated or that various
A
condition, needs, clinical trajectory, and eelings based on the
members are not talking to one other.
R
multiple perspectives o team members. The team only obtains
• Speci cally, address your coordination with nurses, other physicians
T
situational awareness when these perspectives are communicated
I
within the team. Without the perspectives o team members, no on your team and specialists, pharmacists, therapists, etc.
individual—including the hospitalist—truly has situational aware- • Let patients know that you are aware o and approve o what
ness. High-quality, coordinated care is compromised when deci- others on the team are doing.
sions are made, orders written and even discussions with patients
T
h
and their amilies occur without the perspective o other members
e
o the care delivery team. It is vital to quality, patient-centered care Limiting the risks associated with prolonged unnecessary hospitaliza-
S
tion should be a stated goal o the hospitalist team. In support o that
p
that a single message is delivered rom everyone on the team.
e
The hospitalist is the hub o communications and transitions in goal, each team member—as they round on their patients and decide
c
i
whether to discharge now or not—must determine whether the bene ts
a
care and leads the health care team to ensure the patient is at the
l
t
o continued hospitalization outweigh the inherent risks o continued
y
center o decisions.
o
hospitalization. Each team member should be encouraged to articulate
f
H
the rationale or their decision to the rest o the care team.
o
PRACTICE POINT
s
p
PRACTICE POINT
i
t
Limit your blind spots and those o your team
a
l
Unless you must, do not deliver care without situational awareness. Discharge the patient sa ely and as soon as possible
M
Create workf ow scenarios that allow multiple team members’
e
Hospitalization exposes patients to a host o physical and psycho-
d
perspectives be ore making decisions, writing orders, and logical risks including:
i
c
• Blood stream in ections
i
interacting with the patient.
n
e
At a minimum, round with the patient’s nurse be ore you see • Respiratory in ections
a
n
the patient. • Urinary in ections
d
Communicate your perspective to the team. The team relies on • Adverse drug events
S
y
the hospitalist or clinical perspective. At a minimum ask:
• Pressure ulcers
s
t
● What problems are being addressed and is each problem
e
• Falls
m
getting better or worse?
• Functional decline
s
• What is being done or the patient (tests, evaluation by
o
• Anxiety
f
consultants, interventions, medication changes, etc) and why?
C
a
• What does the patient need to be sa er?
r
e
• What does the patient need to eel better? A hospitalist typically leads the inpatient care team and is increas-
• What does the patient need or sa e, timely discharge? ing more involved in postdischarge care coordination, particularly
• What is the next site o care and how is this being coordinated with Accountable Care Organizations (ACOs) and CMS’ Bundled Pay-
postdischarge? ment Care Improvement (BPCI) initiative. As the leader, the hospitalist
• What are the risks o readmission? How can we work together is responsible or goal clarity, role clarity, communication, and team
as a team to reduce avoidable readmissions to the hospital? cohesiveness. E ective teamwork requires the willingness o the team
members to work toward a shared goal. Goal clarity requires explicitly
stating what de nes success or the team and a quality outcome or
the patient. And, now with ACOs and BPCI the hospitalist team has
Goal success is an optimal patient experience. The team relies on an important role to help reduce readmissions by coordinating transi-
each member to provide best practice care by limiting unneces- tions in care and ensuring that the risks that could bring the patient
sary variation in practice, enhancing patient satis action with the back to the hospital are identi ed and mitigated.
hospitalization experience, e ective communications throughout
• What medical conditions are or are not being treated as an
the team, and discharging the patient as sa ely and as soon as pos-
inpatient?
sible. Reducing per ormance variability o the team ultimately helps
• What is the goal o treatment?
to address the patient’s problems and ensures better outcomes.
• What is the endpoint o hospitalization?
By using evidence-based protocols, order sets, checklists and other
• What is the reason or each test, intervention, change?
institutional-speci c processes, the team can help standardize
• What care coordination is required postdischarge to limit
per ormance.
avoidable readmissions?
PRACTICE POINT PRACTICE POINT

Limit unnecessary variation in practice Be explicit about goals
Reliable systems make improved quality more likely: Write the goal in the patient chart. For example:
● I there isn’t clear evidence supporting a particular choice, the • “Chest pain: The patient has multivessel CAD and demand
team should consistently use institutional therapeutic choices ischemia. He declines intervention other than medication
(eg, antibiotic choices, VTE prophylaxis, etc). changes. I am titrating nitrates and beta-blockers. My goal or
• Set consistent times or rounds. discharge: pain ree at rest and while walking slowly in room,
• Use institutional order sets. tolerating medication without orthostatic symptoms.”
158
Role clarity requires explicitly identi ying who will do what on TEAM COMMUNICATION WITH PRIMARY CARE
the team. PROVIDERS, POSTACUTE PROVIDERS, PATIENTS,
C
AND FAMILIES IS A BIDIRECTIONAL PROCESS
H
• Who on the care team is responsible or which aspects o
A
care? Hospitalists generally de ne themselves as specialists o the medical
P
• Who is discussing which issues with the patient? care o hospitalized patients. Mistakenly, however, hospitalists may
T
• Which consultant is managing which problem? believe that they do not need either the input o the outpatient
E
practitioner(s) or to partner with them. This is another silo mentality
R
• Who is writing orders or what?
that neither osters high-quality care nor patient satis action. In this
2
E ective communication requires sharing essential in orma-
new era o accountable care, the hospitalist has an opportunity to
5
tion across the hospitalist team and in this new environment o
set the patient up or success by having a strategy or postacute
pay or quality; hospitalist teams are also leading change in the
care ollowing discharge. From a patient perspective, care does not
T
postcharge care strategies and communications. The team leader
e
begin with admission and end with discharge, and many patients
a
is responsible or demonstrating techniques that encourage
m
wish that their practitioners had an ongoing role in their care dur-
speci c teamwork, modeling behaviors that ensure that roles are
w
ing hospitalization and today many do have ways to extend care
clearly de ned, sharing in ormation in a timely and accurate ash-
o
ollowing discharge. Many hospitalist teams are now partnering
r
ion, discussing and mutually agreeing to plans o care. E ective
k
with a postacute care providers who manage transitions in care that
i
teamwork can reduce the number o medical errors and avoid-
n
occurs in skilled nursing or rehab centers and even a team o home
L
able readmissions through improved communication and better
e
health providers who check-in on high-risk patients a ter discharge.
a
coordination o care.
d
Ensuring that primary care and postacute providers are involved at
e
• What method will the team use to communicate? Examples
r
various points in the continuum o care is a major quality issue that
s
h
may include reading each other’s chart notes or discussing is being addressed increasingly by hospitalist teams.
i
p
ace-to- ace or by phone. Based on experience, we know that the quality o communica-
a
• How o ten will members communicate? Will this occur once a
n
tion rom the hospitalist to the community physician is paramount.
d
day or throughout the day? Will there be consistent expecta- The inpatient health care team must understand and synthesize
P
tions or communication such as during daily rounds, prior to
r
outpatient goals into the initial hospital care plan and then proac-
a
c
the end o the day or only on an as-needed basis? tively communicate with the outpatient team as the hospitalization
t
i
c
• How can tone and language be used e ectively to decrease proceeds. The hospitalist-leader should ocus the inpatient team on
e
-
barriers and misunderstandings? processes in the transition o patients to other settings, ensure accu-
B
a
racy, thoroughness and timeliness o documentation, and optimize
s
e
d
communication be ore discharge actually takes place to increase
M
PRACTICE POINT the likelihood that discharge orders are carried out as intended.
a
The rst step in understanding outpatient goals is identi ying the
n
Structure your communication
a
leader o outpatient care and to ensure that those at high risk or
g
Make team communication more reliable by:
e
readmission have a speci c plan to minimize this risk. This simple
• Using checklists
m
act can be challenging. Some patients have established primary care
e
• Setting predictable rounding times
n
providers; some have multiple specialists; many have no outpatient
t
• Using agreed-upon care protocols care providers. The hospitalist team must develop a discharge plan
• Using structured communication such as SBAR that supports outpatient goals to the greatest extent possible; then
communicate that plan to the patient’s outpatient and postacute
Situation: the speci c problem:
care teams.
“Mrs. Johnson has a headache and is hypotensive.”
Background: the speci c history that may relate to the current
situation: PRACTICE POINT
“She ell last night and did not have a CT o her head; she does Outpatient and postacute care goals
take anticoagulants.” On admission ask, actively listen, and communicate key in ormation:
Assessment: the analysis o the problem: • Talk to the patient. Who is the primary provider? What are the
“I believe she has a bleed in her head.” patient’s own goals o care? I they require postdischarge care,
what is the most appropriate next site o care or the patient, i
Recommendation/Request: the team member makes a it is not home?
recommendation and request o another: • Talk directly to the outpatient or postacute team and con rm
“Dr. Smith, please see her immediately.” the goals.
• Talk to the patient’s support team (eg, amily, riends, etc)
identi ed and authorized by the patient. What are their goals?
Cohesiveness requires recognition o a shared purpose, • Are the patient’s long-term goals identi ed and documented in
de ined roles, and task interdependence. As the acilitator o an advanced directive or POLST orm? The hospitalization o ers
optimal team unction, the leader limits disruption and rag- an opportunity to approach the subject o li e choices and
mentation o the team that can occur with any dynamic and should be part o the dialog or all patients, and especially or
complex process. This requires active listening and requent those who may bene t rom palliative or hospice care.
communication with all members o the team, sharing decision-
making responsibilities, and proactively providing opportunities
or everyone to contribute according to their abilities, including Transitions in care are inherently risky events and e ective
patients and amilies. discharge planning begins on the day o admission. Inadequate
• How do we help each other succeed? preparation can compromise care, contribute to medication errors,
• How do we eliminate what pulls us apart? and create a sense o discontinuity or the patient and those who
159
provide their longitudinal care. Hospitalists must ocus on improv- On a daily basis, make sure multidisciplinary team rounds include
ing not only the substance o transitions, but also the experience a review o :
o the transition or patients, their amilies, primary care physicians • Provisional diagnosis
and postacute providers. From day one o hospitalization, com-
P
• Planned diagnostic work
A
municate with amily, advanced practitioners, nurses, primary care
• Management plan
R
providers, and others who will provide longitudinal care o the
• Anticipated date and time o discharge and to what setting
T
patient during and a ter hospitalization. The hospitalist should
• What to do i something goes wrong
I
determine, with the receiving community provider, how com-
munication should take place, (telephone, e-mail, page, or ax); Answer any questions and con rm understanding and consensus.
how o ten; by whom; what communication should occur; and Continuously update the primary nurse and other members o
determine the level o involvement o the outpatient practitioner the team when:
T
h
in discharge decisions rather than simply relying only on discharge
• There is a change in plan
e
summaries to trans er in ormation.
• There are results o a diagnostic workup
S
p
Importantly, the hospitalist-led team must also communicate to
• There are new diagnoses
e
patients and their amilies in clear next steps with regard to their
c
i
• There is a new complication
a
continued diagnosis, treatment, timing o anticipated discharge, and
l
t
• There is a postdischarge plan to a site other than home
y
care beyond the hospital. To do this e ectively, the team should have
o
a shared understanding about all o the issues impacting the patient, Listen to concerns and address them.
f
H
diagnostic ndings, and management plans. The patient’s primary
o
On the day o discharge summarize prior conversations:
s
care nurse should not be overlooked as a key communicator o in or-
p
• Medication changes and the reasons or changes
i
mation to patients and amilies. The team leader should include the
t
a
• New medications
l
patient’s nurse in rounds and update that nurse regularly. Likewise,
M
patients and amilies are important members o the care team and • Diagnostic studies, the results and pending results
e
• Consultations per ormed during hospitalization and speci c
d
should be in ormed and actually have an opportunity to ask questions
i
c
and give consent to treatment. Patients should not receive conf icting ongoing recommendations postdischarge
i
n
• Who to contact i there is an unexpected problem
e
in ormation rom doctors, nurses, consultants, and other members o
a
the team. The in ormation provided should be structured in straight- Provide written materials to complement verbal instructions
n
d
orward simple language in accordance with the patient’s literacy, that the patient should bring with him or her to the primary
S
utilizing interpreters when English is not the primary language.
y
care provider’s o ce or ensure the postacute care acility has the
s
Patient satis action surveys provide in ormation that can be used
t
appropriate in ormation to transition the patient’s care.
e
m
to improve team per ormance. Using eedback rom Hospital Con-
s
sumer Assessment o Healthcare Providers and Systems (HCAHPS)
o
survey results, physicians can better understand how patients per-
f
C
ceive the quality o communication coming rom their hospital phy-
a
THE ROLE OF THE DIRECTOR OF A HOSPITALIST SERVICE
r
sician. Whether it be how well the patient elt their doctor listened
e
to them, how well the doctor explained things in a manner that they Typically a hospitalist team is led by a chie hospitalist or medical
could comprehend, or whether they elt their doctor treated them director who provides daily clinical management o the team. In this
with courtesy and respect, we all bene t rom understanding how role, the chie hospitalist is also the liaison to the hospital adminis-
we are perceived by our patients. This type o direct eedback can tration. The chie hospitalist has responsibility or the per ormance
be used in educating hospitalists on the importance o providing o the hospitalist service, provides administrative support or the
patient-centered care. It also enables leaders to mentor team mem- service, develops schedules or hospitalists that ref ect manageable
bers and to set expectations or the desired per ormance, educating workloads, and typically serves on hospital committees including
the service on how to e ectively provide patient-centered care, and pharmacy and therapy, critical care, sa ety, utilization review, The
mentoring hospitalists who have lower scores. Joint Commission, HCAHPS Survey, and compliance review boards
to name a ew.
The comprehensive hospitalist service is a team o physicians,
PRACTICE POINT advanced care practitioners, and business managers. Depending
on the institution, the hospitalist team may be employed by the
Consistent, understandable patient education
hospital or part o an independent group o hospitalists, a mem-
On admission identi y the patient-authorized amily contact and ber o a large multispecialty group practice, or a liated with a
make sure to update him or her daily as well as the patient, i larger regional or nationally-based hospitalist organization. A small
appropriate, o : percentage o hospitalists are locum tenens physicians who ll an
• Provisional diagnosis important temporary role when teams require temporary sta ng.
• Areas o uncertainty Hospitalist leaders have the opportunity to set clinically appropri-
• Planned diagnostic workup, consultation ate and standardized care. As a result, teams o hospitalists have the
• Management plan ability to improve the quality o care o a larger group o hospitalized
• Anticipated date and time o discharge and to what setting patients by delivering measurable and consistent quality care. To be
• What to do i something goes wrong and patient does not e ective, hospitalist leaders should:
respond as anticipated • Set Clea r Expectations. E ective leaders provide teams with
• How to avoid readmissions—things to do and what to look or clear per ormance expectations. When the team understands
in the immediate postdischarge period that may be signs that how the leader measures excellence, they know what is
the patient needs medical attention expected o them.
Answer any questions and con rm that the patient’s amily In addition to their clinical responsibilities, all hospitalists
contact understands what you have said. should understand the hospital’s key initiatives and the areas
in which the hospitalist team is going to be accountable to
160
key hospital administrators including the chie executive o - necessary and ultimately be more e ective directors o the
cer (CEO), chie nancial o cer (CFO), chie medical o cer patients’ care. Providers need adequate time to communicate
C
(CMO), and chie nursing o cer (CNO). Using metrics provides with the various stakeholders in a thorough ashion, ocus on
H
an objective method or communicating consistently on team
A
the accuracy o documentation and deliver high-quality, e -
P
per ormance measured against team goals. Hospitalist leaders cient care. When hospitalists have a well-managed workload,
T
should ocus their hospitalist teams on delivering measurable it results in better work-li e balance and a more sustainable and
E
quality improvements by reducing practice variability. rewarding career.
R
• Delega te Responsibility. The hospitalist leader delegates • Commit to Develop Team Members. Hospitalist leaders pro-
2
responsibilities to team members and establishes open lines
5
vide pro essional development opportunities or their team
o communication. When teams have a clear understanding members. E ective teams bene t rom growth and advance-
o expectations, there is no room or ambiguity. Establish- ment and rom group incentives that drive per ormance
T
e
ing open and honest communication encourages teamwork excellence. Most physicians lack ormal leadership training.
a
and collaboration. Un ortunately, many hospitalist programs
m
Providing didactic skill development opportunities, access to
w
are carried on the back o the medical director. This model pro essional coaches, and an ongoing support network are
o
is destined to ail as it usually is the result o a lack o overall keys to success ul pro essional growth or all team members,
r
k
physician engagement within the group. Even with limited including the leader. Providing 360° reviews o the chie hos-
i
n
clinical responsibilities, the medical director must have superb pitalist are also essential to help the leader understand their
L
e
delegation skills to reduce the potential or burnout. blind spots and opportunities or improvement as well. These
a
d
• Empower Team Members. Excellent leaders empower hospi- evaluations and important to conduct annually as the team
e
talist team members to get involved. When teams eel sup-
r
composition can change and new members have an opportu-
s
h
ported to make decisions they become more e ective as a nities or eedback. It can also be an opportunity or the leader
i
p
group. This also helps to develop uture leaders by providing to assess improvement year over year.
a
n
them the opportunity to learn decision-making skills within the • Embrace Diversity o Teams. Today’s hospitalist teams are
d
ramework o the team. diverse and require leaders who recognize and embrace di -
P
r
• Deal with Con icts Swi tly. E ective leaders must be capable o erent points o view. With a variety o cultural backgrounds,
a
c
dealing with conf icts immediately and removing roadblocks that skill levels, training and team roles, it is important that the
t
i
c
can impair the e ectiveness o the team. Train team members in leader promote acceptance and openness when dealing with
e
-
how to engage in di cult conversations using an approach or
B
di erent situations. Team leaders can also help diverse teams
a
methodology that is common to the team, such as the program
s
with communication skills, especially where English may not
e
outlined in the book “Crucial Conversations,” by Joseph Grenny,
d
be the provider’s rst language. There are a number o support
M
Kerry Patterson, and Ron McMillan. Having a common ramework programs available to give providers the coaching and training
a
and language helps the team deal with conf icts. needed to be as e ective as possible in their communications
n
a
• Ensure Resources Are Availa ble. Leaders ensure their teams and pronunciation when talking with each other and their
g
e
have the necessary resources to do their work e ectively. In patients.
m
Hospital Medicine the leader has to compete with the other • Mea sure a nd Recognize Per orma nce. Measuring per or-
e
n
hospital priorities and resources. The success ul leader can mance and providing objective eedback to team members
t
navigate competing priorities without the emotion that o ten drives continuous improvement. E ective leaders provide
overtakes individual members o the team in the desire to help training and coaching to team members. They also nd ways
change a process. to recognize and reward per ormance excellence and improve-
• Recognize the Impact o Workloa d on Qua lity. When the hos- ment or both the team and individuals. Most physicians eel
pitalist workload is based on manageable encounters, hospital- they have “arrived” a ter medical school and residency. Most
ists are more apt to deliver very consistent, high-quality care. are not used to a per ormance evaluation as part o being
There is no national body o evidence that supports the ideal on a physician team. Tools, such as an annual evaluation can
workload. The right workload depends on a myriad o actors be a great opportunity or the leader o the team to provide
including the ollowing: eedback on speci c areas that will make the individual and
● Does the physician already know the patients on his or her the team more e ective. The hospitalist should be presented
rounding list? How many new patients are there? the metrics that matter to the patient and to the institution
● What kind o support sta is available to help the physicians? consistently.
● What administrative duties does the physician have during • Recruit the Right People to Your Hospita list Service. Recruit-
the work day? Is there specialty backup coverage? ing the right team is another important actor in developing an
● Does the physician have to do the procedures? e ective hospitalist team. Taking time to evaluate candidates
● How e ciently does the hospital run? or both clinical and technical competency, as well as chemis-
For the average hospitalist starting their day at 7:00 a m, a try with the team, is critical or ensuring the success o e ective
workload o 14-15 patients and admitting or consulting on a teams. Evaluate the candidate’s communication style and skills
ew more patients during the remainder o the day is ideal. The o the rest o the team. A well orchestrated interview agenda
hospitalist’s appropriate workload must also take into account allows all team members to have time to interact with the
the acuity o the patients and the wRVUs represented in the candidate. It is important to have a cross-section o data points
care o patients. In the case o a workload that does not include to assess compatibility o the candidate with the existing team.
new patients, it is o ten easy or an individual provider to man- Finally, be sure to reach out to individuals who worked with
age a higher volume o patients. In a smaller hospital with no the candidate previously to get a comprehensive picture o the
ICU coverage or a lack o specialty support, that same volume candidate you are considering.
typically takes much more e ort and time. A well-managed • Solicit Feedback rom Members o the Tea m to Improve
workload optimizes the provider’s ability to coordinate care Operations. A strong sense o team has been associated with
with the entire team, make additional visits with patients when higher retention.
161
Physicians have many allegiances, including their outside the health care team’s communication with patients, instruction
interests and amilies, their hospital, their medical group, and about medications, quality o nursing services, adequacy o plan-
their team. I a team is unctioning well, members want to stay ning or discharge, and pain management. Responses are reported
and make the team stronger. in six composite domains, largely ocused on the e ectiveness o
P
A
According to the 2015 salary survey by Today’s Hospitalist, communication. The HCAHPS score is meant to ref ect the patient’s
R
the mean age o hospitalists is now 44 years o age [http://www. perception o the quality o the care they received in the hands
T
todayshospitalist.com/survey/15_salary_survey/e03.php], an advance- o the doctors, nurses, and by the hospital. The ollowing three
I
ment o nearly a year older. There continues to be a signi cant need broad goals shaped the HCAHPS instrument to produce data about
to develop leaders and not wait or leaders to evolve. A respected patients’ perspectives o care that allow:
and e ective hospitalist leader is a prerequisite to achieving a highly 1. objective and meaning ul comparisons o hospitals on topics
unctioning hospitalist team. The eld o Hospital Medicine has that are important to consumers,
T
h
recognized this need and the Society o Hospital Medicine (SHM) 2. the creation o new incentives or hospitals to improve quality
e
recently established the Fellow in Hospital Medicine with three o care, and
S
p
levels (Fellow, Senior Fellow, and Master Fellow) recognition. While 3. accountability in health care by increasing the transparency
e
c
speci c traits are identi ed in the FHM, SFHM and MHM charters, o the quality o hospital care provided in return or the public
i
a
l
they are typically skills acquired by hospitalists who have bene ted investment.
t
y
rom mentors or coaches and/or completed additional training.
o
The average scores ref ect the entire hospital experience, but the
f
SHM also provides additional certi cation in leadership in Hospital hospitalist team can signi cantly inf uence HCAHPS survey results
H
Medicine (leadership undamentals and advanced leadership).
o
and improve patient satis action scores. The 2008 New England
s
The e ective hospitalist leader must inherently be an e ective
p
Journal of Medicine article, “Patients’ Perception o Hospital Care in
i
t
manager. Managing a team includes allocating resources skill ully,
a
the United States,” by Ashish K. Jha, et al. concludes that the cur-
l
meeting deadlines and obligations, and serving multiple stakeholders.
M
rent level o communication and care leaves plenty o room or
It also requires e ective communication, conf ict management, and
e
improvement. In this study, 63% o hospitals received a rating o
d
de t delegation ability. Masters-degree-level work, such as a Masters
i
9 or 10 rom patients, and 89% scored their experience at 7 or better.
c
i
in Medical Management (MMM), Masters in Business Administration
n
The quality o hospital care continues to be highly variable, signaling
e
(MBA) or Masters in Healthcare Administration (MHA) can help hone an opportunity or the hospitalist team to take the leadership role in
a
these skills. Alternatively, SHM advanced leadership courses o er strat-
n
driving quality improvements and ratings that reach 90% or more
d
egies and tools or personal leadership excellence and or developing consistently.
S
y
a winning team and strengthening your organization. While the objectives o the CMS’HCAHPS instrument were care ully
s
t
e
considered and the tool skill ully designed, the results pose challenges
m
ALIGNING HOSPITALISTS WITH HOSPITAL GOALS in interpretation. Many consider the HCAHPS data to be f awed due
s
o
to the multivariate nature o the patient’s course o care. Nevertheless,
The success ul hospitalist service has teams assigned to drive
f
tracking and communicating data on clinical per ormance, however
C
per ormance that are aligned with hospitals goals and report on
a
f awed, is a starting point and has previously prompted improve-
r
the results to hospital administration regularly. The most common
e
ments in the quality o clinical care in hospitals.
hospital initiatives are in the areas o quality, operations, satis action,
When the hospitalist actively participates in the patient’s transi-
and nancial per ormance. The hospitalist teams that survive and
tion o care rom hospital to home or to a skilled nursing acility,
thrive in the next decade must master the role o managing e ec-
data shows that satis action scores skyrocket. In addition, typically
tive hospitalist teams that drive real value and measurable results or
the amily, case manager, nursing sta , and primary care physicians’
their hospital partner.
satis action scores also increase. There also should be expectations
To improve outcomes, it is necessary or hospitalists to have
set or hospitalist relations and assessments with critical partners
clear and unctional processes that incorporate utilization o best
including emergency physicians, primary care providers, and spe-
practices. They should be encouraged to identi y ways to reduce
cialists. These relationships must be monitored on a regular basis to
variables in patient care, and importantly, they must be good stew-
ensure cooperative integration is achieved.
ards o their time and resources and ocus on what matters most.
With bundling o payments and more scrutiny on readmissions,
E ective prioritization and time management are critical or driving
hospitals may prioritize resources to reduce readmission rates. Hos-
improvements in all outcomes, whether clinical, satis action, or
pitalist leaders can develop communication standards to reduce
nancial per ormance.
practice variation, identi y patients at increased risk or readmission,
Each hospitalist service should determine benchmark per or-
and work with other hospital leaders to redesign the systems in place
mance expectations or their hospitalist team around admissions,
that do not promote sa e transitions. For example, the University
discharge planning, processes or signing o care between shi ts,
o Pennsylvania attributed a drop in readmissions rom a high o
and creating and managing care pathways. While there are a
15% to 5% in the short period rom the all o 2008 to February 2009
plethora o data sets that can be measured to drive per ormance,
to the implementation o tools rom the SHM Project BOOST, includ-
it is essential or hospitalist service teams to identi y metrics that
ing the “7P” checklist. The checklist tool simpli es the major modi -
can be used to drive per ormance improvement and consistently
able risk actors to consider or readmission to the hospital. There
measure them. Whether measuring continuous quality improve-
are seven risk actors tied to suggested interventions or problem
ments, satis action, e ciency or readmissions data, hospitalist
medications, principal diagnosis, depression, polypharmacy, poor
teams cannot improve unless per ormance is measured. Once
health literacy, patient support, and prior hospitalization.
the per ormance is measured, it can be managed. Per ormance
measurements also can be used to provide reward systems to
rein orce the behaviors. PROJECT BOOST
The Hospital Consumer Assessment o Healthcare Providers and Project BOOST (Better Outcomes or Older adults through Sa e Tran-
Systems (HCAHPS) survey developed by the Centers or Medicare sitions) is one e ort to improve the care o older patients as they
and Medicaid Services (CMS) asks questions directed largely at the transition rom inpatient care to an outpatient acility or home. The
patient’s experiences in the hospital. The survey probes e cacy o SHM, working with Blue Cross and Blue Shield o Michigan and the
162
University o Michigan, is launching a multisite implementation o whom the amily and patient see more than specialists or other
the program seeking to: hospital sta . It is incumbent on the hospitalist to ensure that
C
the hospital’s success is largely dependent on the hospitalist
H
• Avoid unplanned or preventable hospital readmissions and
A
emergency department visits within 30 days o hospital and his or her interaction with others.
P
discharge. • Employing a Service-Focused Mindset. Hospitalists are one o
T
• Improve acility patient satis action scores. the most visible groups in the hospital. The team that is service-
E
oriented and prioritizes quality patient care and communica-
R
• Improve patient satis action associated with discharge.
• Improve communication between inpatient and outpatient tion high on the list will be a sought-a ter change agent in the
2
hospital. The hospitalist team that can demonstrate e ective
5
providers.
• Improve patient and amily education about disease manage- communications and satis action as well as a commitment
ment and risk issues. to patient education is an invaluable resource to the hospital.
T
e
• Identi y patients at high risk and mitigate that risk with multi-
a
m
disciplinary risk mitigation tools and strategies; including
w
CONCLUSION
o
● Discharge coordination/communication with follow-on By virtue o their presence, hospitalist teams have changed the system
r
k
providers
o health care delivery in the United States. Hospital Medicine is now a
i
n
● Patient and caretaker disease and disease management educa- major cost center in the US health care system. Although the variety o
L
tion including a “teach-back”strategy to ensure comprehension
e
impending solutions to remedy our nation’s health care ills range rom
a
● Medication reconciliation including a review of interactions
d
insurance re orm to health care IT solutions, one common important
e
between discharge medications and previously prescribed med-
r
resource has emerged, namely, the active leadership o hospitalists
s
h
ications interactions
engaged to design and implement sweeping improvements in the
i
p
● Essential team members include nurses, case managers, patient quality, satis action, and e ciency o care delivered or hospitalized
a
educators, hospitalists
n
patients. Today, hospitalist teams play a more signi cant role in rec-
d
To be e ective, Project BOOST requires multidisciplinary team- ognizing the needs o patients and their amilies and have developed
P
r
work, coordinating seamless transitions o care by utilizing the
a
ways to demonstrate accountability. Hospitalists impact the majority o
c
combined expertise o team members. This has signi cant eco-
t
clinical decisions made on behal o hospitalized patients and there ore
i
c
nomic and quality implications. In a study by Cauwels, Jensen and directly determine how medical resources are utilized and drive health
e
-
Winterton a hospitalist group highlights progress made in reduc-
B
outcomes and costs on medical care collectively.
a
ing readmissions rates through Project BOOST. In this report, case
s
e
managers implemented BOOST, working with hospitalist teams.
d
M
They achieved statistically signi cant decreases in readmissions in SUGGESTED READINGS
a
30 days or all patients regardless o treating specialty or reason
n
a
or admission. Bohmer R. Designing Care: Aligning the Nature and Management of
g
e
Hospitalist teams must be integrated into these processes. When Health Care. Boston, MA: Harvard Business Review Press; 2009.
m
the team is evaluated as part o the process, individual outliers are
e
Cauwels JM, Jensen BJ, Winterton TL. Giving readmissions numbers
n
identi ed and receive the necessary training and mentoring to a boost. SD Med. 2013;66(12):505-507, 509.
t
improve individual per ormance. Equally as important, the hospital-
ist can help improve institutional per ormance. High-per orming Jha A, Orav J, Zheng J, et al. Patients’ Perception o Hospital Care in
teams are indispensable to high-per ormance hospitals. Examples the United States. N Engl J Med. 2008;359:1921-1931.
include the ollowing: Khatri N, Baveja A, Boren SA, Mammo A. Medical Errors and Quality o
• A High-Functioning Tea m Becomes the Li eblood o the Care: From Control to Commitment. California Management Rev.
Hospital. Hospitalists who improve the quality o the hospi- 2006;48(3):115-141.
tal, improve the quality o patient care. Physicians who work Lee TH. Turning Doctors into Leaders. Harvard Business Rev.
in such hospitals are involved in everything rom the dietary 2010:50-58.
needs o patients, to the workf ow, to access to CT results Lee TH, Mongan JJ. Chaos and Organization in Health Care.
24 hours per day. Many hospital CEOs cannot imagine li e with- Cambridge, MA: MIT Press; 2009.
out a hospitalist team helping to carry their hospital orward.
Patterson K, Grenny J, McMillan R, Switzler A. Crucial Conversations.
• An Ef ective Team Builds the Bra nd o the Hospita l. An
2nd ed. New York, NY: McGraw-Hill; 2012.
engaged and committed team is essential in advancing the
hospital’s brand. By reaching out to local providers, improving Sehgal NL, Green A, Vidyarthi AR, Blegen M, Wachter R. Patient white-
re errals, and making surgeons want to practice at their hospi- boards as a communication tool in the hospital setting: a Survey
tal, the hospitalist contributes signi cantly to the overall brand o practices and recommendations. J Hosp Med. 2010;5(4):234-239.
value o the hospital in the community. It is o ten the hospitalist What Makes a Leader? Harvard Business Rev. 2004:43-52.
163
CHAP TER
26 INTRODUCTION
Hospitalists ace the potential or con ict every day. They work in
highly complex organizations and in order to be success ul they
must interact e ectively with a wide variety o individuals in what
is o ten a challenging, emotionally charged environment. Hospi-
talists must learn to navigate not only the ormal organizational
bureaucracy o rules, systems, and processes, but also the in ormal
political hierarchy that in uences power and decision making.
O ten, they must do so with little or no ormal training in con ict
Negotiation and management at an early stage in their medical careers. In addition,
they may encounter con icts between what others would like them
Con lict Resolution to accomplish and their own workload demands and pro essional
expectations.
Hospital Medicine is also a young, evolving specialty that has
enjoyed unprecedented growth by serving the needs o multiple
Leslie A. Flores, MHA competing stakeholders. Although the specialty is maturing, it is still
populated by a high proportion o recent residency graduates and
early-career clinicians who may not have a complete understanding
o the specialty or even have career advancement on their radar
screen. The potential exists or the service obligations—both clinical
and in the area o institutional per ormance improvement—o
hospitalists to overwhelm opportunities or pro essional develop-
ment, and this may promote career dissatis action, turnover, and
symptoms o burnout. Leaders o hospitalist services may nd them-
selves isolated as they advocate or the pro essional development
and job satis action o group members while meeting the service
expectations o their employers or supervisors. The pro essional
medical society or hospitalists, the Society o Hospital Medicine,
is rapidly developing exible support resources or hospitalists
relating to business and clinical practice, engagement and career
satis action, core competencies, and role expectations. Until these
standards become widely disseminated and health care services
become better designed and hence less prone to error, hospitalists
will continue to work in a hospital environment where they will
increasingly be expected to per orm as change agents at a time
when change may not be welcomed by their hospitalist colleagues
or others at their institutions.
For the purposes o this chapter, it will be important to distinguish
between disagreements and con icts. Disagreements happen regu-
larly in human interactions, and occur whenever two or more indi-
viduals have di ering opinions about something. A disagreement
need not devolve into a con ict, and many do not. Con icts arise
when a party perceives that another party has negatively a ected or
will negatively a ect agendas that the rst party cares about. Con-
icts are de ned as processes that occur when tensions develop,
that is, the emotions associated with a disagreement become so
elevated that they impede the ability o the parties to interact with
each other e ectively.
Almost all con ict is a result o unmet expectations. For hospital-
ists, this commonly arises when there is a lack o understanding or a
di erence in expectations about their role. Hospitalists may assume
that primary care physicians have explained to patients that some-
one else will be seeing them in the hospital. Patients and amilies,
however, may not understand why their primary care physician
is not present in the hospital and directing their care. Emergency
Medicine physicians may expect the hospitalist to respond promptly
to take a complicated social admission o their hands whereas
hospitalists may eel that it is the role o the emergency room
164
physicians to discharge patients who do not require admission. work requirements. Men and women may have di erent expecta-
Emergency Medicine physicians and sta may expect or patients tions o their work, and o ten have di erent ways o responding to
C
be triaged to hospital oors (to reduce emergency department stress, emotion, and con ict. In the United States, men o ten tend to
H
length o stay) be ore critical in ormation is available, or may expect
A
use a competing or orcing style when aced with con ict, whereas
P
hospitalists to care or patients in the emergency department when women o ten tend to use compromising, accommodating, and
T
no beds are available. Meanwhile, oor nurses may expect hospi- avoiding.
E
talists to be immediately available to address nonurgent requests. Akey aspect o cultural di erences is the degree to which a person
R
There may be di erences o opinion among specialists and general- tends to identi y most strongly with the group o which he or she
2
ists regarding diagnosis, workup, and treatment or the role o the is a part (a “collectivist culture”) as opposed to identi ying with the
6
hospitalists in comanagement o specialty patients. All physicians sel (an “individualistic culture”). Individualistic cultures, which are
expect to be treated pro essionally, to have some autonomy over the dominant cultures ound in North America and Western Europe,
N
clinical decision making, and to have a reasonable work-li e balance. value autonomy, creativity, and personal initiative. Much o the rest
e
g
Hospital administrators and employers, however, may demand that o the world is composed o collectivist cultures, which instead value
o
t
hospitalists to per orm nonphysician tasks or solve problems or con ormity and harmony. A meta-analysis o studies on culture and
i
a
t
other physician groups without taking into account the perspec- con ict resolution styles ound that people in individualistic cultures
i
o
tives o the hospitalists or sta ng needs or time-consuming tasks. tend to choose orcing as a con ict style more o ten and people
n
a
When such expectations go unmet, people get rustrated or angry. who come rom collectivistic cultures tend to choose withdrawing,
n
They o ten respond in ways that then heighten rustration or anger compromising, or problem-solving styles instead.
d
C
on the part o others. Emotions on both sides become elevated, and
o
the stage is set or a con ict.
n
THE POTENTIAL BENEFITS OF CONFLICT
l
i
c
Con icts are inevitable in human interactions. The increasingly
t
R
PRACTICE POINT complex and collaborative nature o the work that hospitalists do
e
s
as team-based care models have emerged increases the risk that
•
o
Almost all con ict is a result o unmet expectations. For
l
interpersonal con icts may arise. The increasing cost pressures and
u
hospitalists this commonly arises when there is a lack o
t
competition or scarce resources that exist in an era o national
i
o
understanding or a di erence in expectations about their role.
n
health care re orm have increased stress and opportunities or con-
ict or all health care pro essionals. These con icts can be destruc-
tive i not e ectively managed. But a healthy approach to con ict
The most common reasons that expectations go unmet include: management acknowledges that not all con ict is entirely negative.
• Lack o Cla rity About What is Expected, or About How the There are potential bene ts that may be derived rom con icts
Expectation Will Be Met. It is easy to assume that because under certain circumstances. DeChurch and Marks (2001) reported
one’s expectations are clearly understood by onesel , they that the ways in which groups handle con ict help to determine
are clear to others as well. Hospitalists may assume a patient whether or not bene ts were realized, noting that “the relationship
understands the proposed treatment plan, but the patient between task con ict and group per ormance was positive when
or amily member may ail to understand the implications or con ict was actively managed and negative when it was passively
likely discharge plans. Even when expectations are care ully managed.” This suggests that Hospital Medicine physicians will be
explained, the other party may hear or interpret them di er- well served to develop e ective con ict management skills that
ently than the speaker intends. The other party may also react can help them increase the likelihood that the con icts they will
more to the emotional aspect o the discussion or who is doing inevitably ace may yield positive results. In order to do so, it will be
the talking rather than to the content. important or hospitalists to think strategically about how one may
• Lack o Agreement About What is Expected Or How to extract the maximum bene t rom con icts that do occur. Some o
Achieve It. The high degree o complexity in error prone the potential bene ts o appropriately managed con ict include:
health care systems, stress and pressure, and the need or rapid • Catalyst or Change. Con icts can orce needed change
change are important sources o potential con ict. Sometimes by sur acing problems that otherwise might not be recog-
each party’s expectations are clearly understood by the other nized, and by elevating latent issues to a level that demands
party, but they simply disagree with each other about the attention. This can be especially valuable in tradition-bound,
desired outcome, the method o achieving it, or both. This change-resistant organizations.
can occur i the parties have competing needs or interests. • Improved Outcomes. Similarly, con icts can ultimately yield
For example, although resident work hour restrictions are improved outcomes when they acilitate learning in the search
clearly delineated in the academic setting, stress and pressure or better solutions and bring to the ore ront use ul in orma-
develop or hospitalists when the increased service obligations tion and emotions that lie below the sur ace.
resulting rom such restrictions con ict with their expectation • Balance. Healthy con ict helps to ensure that balance is main-
or pro essional advancement. All parties may agree on the tained among competing needs and perspectives.
importance o improving patient ow rom the Emergency • Increased Accountability. Because con icts involve strong
Department to the inpatient oor, but may disagree about emotions, healthy con ict resolution usually involves care ul
the speci c methods to be employed by Emergency Medicine articulation o what the parties have agreed to do to resolve
physicians, hospitalists, and others to achieve this goal. Chang- it, and a signi cant degree o accountability to ensure that the
ing hospital processes to promote improved quality or greater agreements are ollowed through.
e ciency o ten demand changes to hospitalist work ow that • Improved Relationships. When people skill ully manage a
are stress ul or the hospitalists. con ict in healthy, respect ul ways, it can actually serve to
In addition, age, gender, and cultural di erences may play a role strengthen their relationship going orward. They end up
in the development and management o con ict. A generational understanding each other better, and building greater trust
gap may result in di erences in work expectations, a paternalistic because they have demonstrated that they can overcome
view o who is actually in charge, or resistance to changing to new di erences.
165
the conversation: “What is this con ict about? What steps will
TABLE 26-1 Five Key Principles o E ective resolve it? What points do I need to be sure to make? What will
Con lict Management I say to get my points across? What will the other person say?”
P
1. Commit to con ronting People skilled in con ict management realize that the conditions
A
2. Attend to the conditions matter just as much as—in act, maybe more than—the content
R
does. What types o conditions matter? The physical conditions mat-
T
3. Identi y one’s personal contribution
ter a great deal. Is the conversation taking place in a private place
I
4. Consider what is underlying others’ behavior instead o in public? Are the people involved in the conversation sit-
5. Clari y ting or standing so they can engage each other at eye level, or is one
person sitting with the other standing over him? Is there a desk or
other impediment between the participants? Is the room too large
T
h
or too small, too hot or too cold to be com ortable?
e
KEY PRINCIPLES IN CONFLICT MANAGEMENT Psychological conditions matter even more. The hospitalist who
S
p
This chapter o ers ive key principles that represent a good start wishes to be skilled at con ict management must learn to pay atten-
e
c
or those who wish to build better con lict management skills tion to what the other person or people involved in the con ict
i
a
l
(Tab le 26-1). However, more detailed treatments o all o these are experiencing emotionally. Are they eeling attacked or are they
t
y
principles and others are contained in the re erences at the end eeling sa e? Do they eel that the hospitalist respects them and has
o
f
o this chapter. their best interests at heart, or do they eel that their interests will
H
be ignored or belittled? Do they sense that the hospitalist is going
o
1. Commit to Confronting. Most people tend to shy away rom
s
to push her agenda or opinion and ignore theirs, or do they believe
p
con ict. It is tempting to believe that the problem will go away
i
t
the hospitalist is willing to listen and take their point o view into
a
by itsel i le t alone; that others will so ten their positions,
l
consideration? Do they eel that the hospitalist’s opinion matters, or
M
orget about the issue, or change their minds, i given enough
that dialogue should occur at a “higher level” with senior physician
e
time. But when pressed, most people will acknowledge this
d
leaders to the exclusion o hospitalists?
i
is simply a convenient excuse or avoiding a con rontation
c
i
Be ore the actual content—what the con ict is about and how
n
that they ear could become uncom ortable or out-and-out
e
unpleasant. Another important reason that people avoid it should be resolved—can be e ectively addressed, the skilled
a
n
con ict is their ear that openly con ronting the situation will con ict manager must take steps to set up conditions that allow all
d
make things worse, rather than better. They may worry about parties to eel com ortable, sa e, and heard. The necessary steps to
S
y
handling the con rontation badly and unintentionally causing creating these positive conditions involve ensuring mutual respect
s
t
among the parties, and identi ying or creating a mutual purpose.
e
the situation to deteriorate, or they may ear that the con ict is
m
intractable and that no matter how care ully and skill ully the In other words, do others believe the hospitalist sees them as indi-
s
situation is handled, the outcome will be negative. viduals worthy o the respect and consideration due to every human
o
f
being, and do they believe that the hospitalist is mind ul o their
C
In act, con icts cannot be resolved i they are not con ronted.
interests as well as his own in seeking an acceptable resolution?
a
r
They may be glossed over or pushed into the background, but not
e
truly resolved. And such con icts are likely to sur ace again, o ten in
unanticipated and damaging ways. Thus, a willingness to acknowl- PRACTICE POINT
edge the existence o a con ict and to step up and con ront it is a
precondition to e ectively managing the con ict.
• Be ore the actual content—what the con ict is about and
how it should be resolved—can be e ectively addressed, the
skilled con ict manager must take steps to set up conditions
PRACTICE POINT that allow all parties to eel com ortable, sa e, and heard. The
necessary steps to creating these positive conditions involve
• A willingness to acknowledge the existence o a con ict and ensuring mutual respect among the parties and identi ying or
to step up and con ront it is a precondition to e ectively creating a mutual purpose.
managing the con ict. This requires an open and honest
discussion o the issue, usually ace to ace, with the goal o
understanding the root causes (the unmet expectations) that
led to the con ict and addressing them. 3. Identi y One’s Personal Contribution. Con icts occur when
emotions get in the way o resolving disagreements. This is
true not only o others with whom a hospitalist may come
In this context, the term “con rontation” is not intended to mean in con ict, but o the hospitalist himsel . Another important
an angry, emotional exchange o verbal attacks. Instead, “con ron- competency or skilled con ict managers is the ability to step
tation” re ers here to an open and honest discussion o the issue, back rom their own emotions and assess their personal con-
usually ace to ace, with the goal o understanding the root causes tribution to the situation; in other words, what impact are their
(the unmet expectations) that led to the con ict and addressing own biases, assumptions, emotions, and actions having on the
them. The remaining principles in this secion are intended to assist con ict itsel , and on their approach to managing it? Do they
the con ronter, once the decision to con ront has been made, to truly intend to seek mutually acceptable solutions or do they
care ully plan the con rontation (when time permits), and to handle just want to win?
it success ully. For example, the person seeking to manage a con ict must pay
2. Attend to the Conditions. Patterson et al (2002) note that attention not only to what others are experiencing emotionally but
there are two components to every success ul crucial conver- also to what he is experiencing emotionally himsel . He needs to
sation: the actual content o the conversation, and the condi- ask, “Am I eeling sa e or am I under attack? Do I believe the others
tions under which the conversation occurs. Most people, when involved in this con ict will listen to me and take my interests into
planning to con ront or actually engage in a con rontation consideration, or not?” However, simply identi ying one’s own emo-
(a “crucial conversation”), think primarily about the content o tional state is not adequate. E ective con ict managers should also
166
have the sel -awareness to understand how their emotions will tend but others may be le t con used or unaware o how strongly the per-
to in uence their behavior in the con rontation. These tendencies son eels about the issue because o his silence tendencies. On the
C
are described as a person’s “style under stress.” other hand, answering “true” to some or all o questions 7 through
H
The Style Under Stress Inventory3 in Table 26-2 is based on the 12 means the person tends to go to violence when eeling unsa e
A
P
concept o conversational sa ety, and will assist individuals in assess- in a conversation. These people will o ten try to orce their opinion
T
ing their own personal style under stress. In completing the ques- on others by controlling the conversation and either prevent others
E
tions, one should answer “T” or true or “F” or alse, based on one’s rom speaking or belittle their contributions when they do.
R
most common tendencies when in con ict situations in the work Both silence and violence can be extremely damaging, when the
2
setting. People eel sa e in a crucial conversation i they believe that goal o the conversation is to con ront disagreements and work
6
they will be listened to respect ully and i they do not eel attacked toward mutually acceptable solutions. When people understand
or ignored. They eel that the other parties have their interests at their own silence or violence tendencies, they can begin to pay
N
heart, or at least that others’ interests and their own are not dia- attention to how they are responding during con ict situations.
e
g
metrically opposed without room or nding common ground. The They can look or evidence that they are not eeling sa e and then
o
t
inventory is designed to help people understand how they tend to step back to assess the impact their silence or violence is having
i
a
t
behave when they do not eel sa e in a crucial conversation. on the conversation and adjust their interactions accordingly. As
i
o
Individuals responding “true” or several o the rst six questions awareness o these tendencies grows over time, people can begin
n
a
are said to be going to silence when under the stress o a challenging to anticipate situations in which sa ety may be at risk and to pro-
n
con ict situation. This means they will tend to try to downplay or actively develop plans to manage their own tendencies to go to
d
C
sugarcoat an issue, or even avoid it outright by changing the subject silence or violence.
o
or disengaging when they do not eel sa e. In such cases, they may When thinking about one’s personal contribution to a con ict
n
l
believe that they have raised an issue and articulated their concerns, situation, one should also be cognizant o individual assumptions
i
c
t
and biases about others involved in the con ict, and especially one’s
R
belie s about others’ intentions. For example, it is usually help ul to
e
s
o
consider the problem o intent versus impact. When analyzing a con-
l
u
TABLE 26-2 Style Under Stress Test ict, one should consider asking, “Is it the impact (ie, the outcome)
t
i
o
o the other person’s behavior that is bothering me so much, or is it
n
1. Rather than tell people exactly what I think, T F what I believe about the person’s intentions?”
sometimes I rely on jokes, sarcasm, or snide This distinction is important because humans tend to overem-
remarks to let them know I’m rustrated.
phasize dispositional actors such as personality type or motives,
2. When I have got something tough to bring up, T F and to discount situational actors such as external stressors, when
sometimes I o er weak or insincere compliments interpreting the behavior o others; this phenomenon is known by
to so ten the blow.
psychologists as the undamental attribution error or correspondence
3. Sometimes when people bring up a touchy or T F bias. Because o this bias, the emotions a person experiences about
awkward issue I try to change the subject. a disagreement, and thus the level o con ict that ensues, may be
4. When it comes to dealing with awkward or T F heightened as a result o presumed negative intentions on the
stress ul subjects, sometimes I hold back rather part o others (“that surgeon is just lazy”) and discounting the cir-
than give my ull and candid opinion.
cumstantial actors that may be in uencing others’ behavior (“that
5. At times I avoid situations that might bring me T F surgeon is under real pressure to produce good outcomes, and
into contact with people I’m having problems does not have the training or experience to manage these complex
with.
medication regimens”).
6. I have put o returning phone calls or e-mails T F The undamental attribution error may be exacerbated by a
because I simply did not want to deal with the related tendency known as the actor-observer bias in which one
person who sent them.
tends to attribute others’ behavior to their dispositions but to attri-
7. In order to get my point across, I sometimes T F bute one’s own behavior to the circumstances (“that amily member
exaggerate my side o the argument. lost her temper because she’s a demanding jerk, but I only lost my
8. I I seem to be losing control o a conversation, T F temper because she pushed me over the edge”). Sel -awareness is
I might cut people o or change the subject in critical or e ective con ict management, especially awareness o
order to bring it back to where I think it should
one’s own assumptions and biases.
be.
9. When others make points that seem stupid T F
to me, I sometimes let them know it without PRACTICE POINT
holding back at all.
10. When I’m stunned by a comment, sometimes I T F
• Sel -awareness is critical or e ective con ict management,
say things that others might take as orce ul or especially awareness o one’s own assumptions and biases.
attacking, comments such as “give me a break!”
or “that’s ridiculous!”
4. Consider What is Underlying Others’ Behavior. One o the
11. Sometimes when things get heated I move rom T F
keys to e ective con ict management is the ability to analyze
arguing against others’ points to saying things
that might hurt them personally. why others respond the way they do in con ict situations
(taking into account both dispositional actors and situational
12. I I get into a heated discussion, I’ve been known T F
actors), and to modi y one’s interactions accordingly. The
to be tough on the other person. In act, they
might eel a bit insulted or hurt. concept o conversational sa ety applies to the other parties
involved in a con ict situation, as well as to onesel . Skilled con-
Excerpted with permission rom Patterson K, Grenny J, McMillan R, et al. ict managers become adept at not only reading and adjusting
Crucial Conversations: Tools or Talking When Stakes are High. New York, NY: their own behaviors, but also at looking or signs that others
McGraw-Hill; 2002. are not eeling sa e. Hospitalists may become more accepting
167
o the anger expressed by patients’ amilies, the critical com- In addition to the overwhelming in uence o emotion on how
ments rom other medical sta members, or the sugar-coated others respond to con ict situations, Hei etz and Linsky (2002) have
change o subject by the hospital executive when they realize argued that there are power ul and universal human needs that
that these behaviors o ten result rom others’ ear(s) that they in uence behavior, sometimes in dys unctional or disruptive ways:
P
A
will be treated with disrespect, attacked, or ignored. I they can
R
then work to address those underlying ears (part o paying Every human being needs some degree o power and
T
attention to conditions) be ore launching into the content o control, a rmation and importance, as well as intimacy
I
the conversation, they will be more success ul. and delight…. We all have hungers, which are expres-
sions o our normal human needs. But sometimes those
hungers disrupt our capacity to act wisely or purpose ully.
PRACTICE POINT Perhaps one o our needs is too great and renders us
T
•
h
One o the keys to e ective con ict management is the ability vulnerable. Perhaps the setting in which we operate exag-
e
to analyze why others respond the way they do in con ict gerates our normal level o need, ampli ying our desires
S
p
situations (taking into account both dispositional actors and and overwhelming our usual sel -controls. Or, our hungers
e
c
situational actors), and to modi y interactions accordingly. might be unchecked simply because our human needs
i
a
are not being met in our personal lives.
l
When someone acts in ways that contribute to a heightened
t
y
level o con ict, it is worth considering whether that person has
o
f
underlying human needs that are going unmet and that are When someone acts in ways that contribute to a heightened level
H
contributing to his or her challenging behavior.
o
o con ict, it is worth considering whether that person has underly-
s
p
ing human needs that are going unmet, and that are contributing
i
t
to his or her challenging behavior.
a
l
M
Another way o thinking about this issue is to anticipate that the 5. Clarify. When con ronting another person about a con ict
e
more signi cant the con ict, the greater the chance that people situation, e ective communication skills are essential. It is
d
i
will respond to it emotionally rather than logically. While it is not
c
important to clari y what one is attempting to convey; it is also
i
n
a clinically accurate model, it may be use ul to think o peoples’ important to clari y the other person’s point o view. Important
e
brains as having a logical core, surrounded by a layer o emotion communication skills include
a
n
(Figure 26-1). Every interaction a person has, no matter how logi-
d
• Setting the stage. Keeping in mind the principles o mutual
cal it is, passes through this emotional lter on its way in or out.
S
respect and mutual purpose, it may be valuable to start out
y
For most people and under normal circumstances, the layer
s
by communicating one’s own positive intentions to the
t
e
o emotion surrounding the logical core is relatively thin and the
m
other person(s) in a way that builds toward these goals.
in ormation rom the interaction passes through it in both direc-
s
• Managing expectations. Hospitalists should clearly communi-
tions, in ormed by the emotion but not substantially altered by it.
o
cate what their own expectations were in the situation that
f
In a con ict situation, however, the emotional layer surrounding the
C
gave rise to the con ict, and seek to understand what the
a
logical core in ates like a balloon. In this situation, the expanded
r
other person’s expectations were. This will create a ounda-
e
emotional layer takes over and prevents logical conversation and
tion or urther dialogue about the di erences between
data rom passing through. The person is responding rom her or
what each party expected and what actually occurred.
his emotion, rather than rom logic. A hospitalist may be attempting
• Active listening. Active listening skills involve not just hearing
to have a very logical conversation with a amily member, assum-
what the other person says, but also
ing that she is addressing the amily member’s logical core. But the
h actively engaging the other person with eye contact and
hospitalist’s logical words cannot get through the amily member’s
body language;
in ated emotional layer. The hospitalist is talking logic, and the am-
h working to enable the other person to eel com ortable
ily members are responding rom emotion; so no wonder they are
sharing potentially di cult in ormation;
unable to relate to each other. In such situations, it is necessary to let
h listening “between the lines” or what is not being said, as
some air out o the balloon—to give the emotional layer a chance
well as what is being said;
to de ate—be ore it will be possible to re-engage the logical core
h acknowledging the reality and legitimacy o the other
in problem solving or con ict resolution. Sometimes this requires
person’s emotions;
stepping away rom the conversation or a while and coming back
h paraphrasing and re raming to ensure understanding o
to it later.
the other person’s perspective;
h asking questions and probing to understand root causes;
h staying ocused on the other person, rather than one’s
own planned response.
• Joint problem solving. Engaging all parties to the con ict in joint
Emotiona l laye r problem solving will help to clari y what needs to happen to
Emotiona l laye r
resolve the con ict, and what the alternatives are or moving
orward out o con ict. It will also help build mutual support o
Logica l Logica l and commitment to the agreed-upon approach.
core core • Articulating next steps. Establishing a clear path o next steps
and assigning responsibilities are vital components o a clear
and e ective communication process. It is worth talking both
about the expected outcome, and about the method or pro-
cess by which the outcome will be achieved: it is not uncom-
mon or new con icts to arise inadvertently when two parties
Norma l c irc ums ta nc e s Conflic t
believe they understand what will happen, only to clash over
Figure 26 1 Logic and emotion diagram. how it will be accomplished.
168
PRACTICE POINT • Empathize. As with Covey’s Talking Stick example, patients
and amilies need to eel understood. It is not necessary to
C
• When con ronting another person in a con ict situation,
agree with them, but it is important to acknowledge their
H
e ective communication skills are essential. It is important to
eelings and to attempt to understand the issue rom their
A
ocus on ensuring clarity, both in what one is attempting to
P
perspective.
convey, and in understanding the other person’s point o view.
T
• Apologize. Byham points out that even i one does not wish
E
Important communication skills include setting the stage,
to admit ault, it is important to apologize or the situation,
R
managing expectations, active listening, joint problem solving,
and or the act that the patient’s expectations were not met.
2
and articulating next steps.
• Take responsibility or action. Once the emotional balloon has
6
been de ated, it is o ten possible to re-engage the patient
or amily member on a logical basis. A good way to make
N
this transition is to take some concrete action, either to
e
STRATEGIES FOR EFFECTIVE CONFLICT MANAGEMENT:
g
resolve the problem on the spot or to demonstrate a desire
o
CONFLICT RESOLUTION AND NEGOTIATION
t
to improve the situation.
i
a
1. The Talking Stick. Stephen Covey (1989) highlighted the
t
3. Principles o Ef ective Negotiation. Hospitalists requently
i
o
importance o empathetic communication in describing the nd themselves in potential con ict situations in which nego-
n
principle, “Seek rst to understand, then to be understood.”
a
tiation is an e ective strategy or addressing the issue. These
n
Covey (2004) urther described the use among Native Ameri- may include ormal negotiations such as the development
d
C
can cultures o the Talking Stick as a tool to help people o pro essional service agreements, employment contracts,
o
resolve di erences by creating greater mutual understanding or incentive compensation metrics, or they may be less or-
n
and respect.5 The Talking Stick is passed rom one person to
l
mal interactions such as working with specialists to de ne
i
c
t
another, and only the person who is holding the Talking Stick admitting responsibilities or co-management services. Strong
R
is allowed to present her or his perspective. This ensures that negotiation skills are also valuable or hospitalists working on
e
s
only one person talks at a time, and increases the ability o oth-
o
medical sta committees or quality improvement projects
l
u
ers to listen because they are not permitted to argue or make when the diverse interests o many parties must be reconciled.
t
i
their own points until the person holding the Talking Stick has
o
In traditional negotiations, each party stakes out a ormal posi-
n
nished. tion and then proceeds to bargain rom that position, using various
The most power ul aspect o the Talking Stick, however, is that tactics to “win” points that bring the nal compromise outcome
the person holding it does not relinquish it until she is satis ed that closer to this position. By contrast, the Principled Negotiation model
she has been ully understood by the others. It is the responsibility o developed by Fisher and Ury (1981) ocuses on understanding all
the listeners to listen care ully and with empathy, and to ensure that the parties’ underlying interests and on identi ying objective, air
the speaker eels understood—not necessarily agreed with—just options that can satis y everyone. The our tenets o Principled
understood. Once the speaker is satis ed that others understand Negotiation are as ollows:
him, she passes the Talking Stick on and assumes the responsibil- • Separate the people rom the problem. This principle addresses
ity to listen and make the next speaker eel understood. Covey the role o emotions and relationships in in uencing one’s
describes the value o the Talking Stick as ollows: perceptions about the negotiation. The authors suggest that
This way, all o the parties involved take responsibility negotiators seek to identi y when relationships (either as
or one hundred percent o the communication, both riends or adversaries) may be getting in the way o seeking the
speaking and listening. Once each o the parties eels best outcome, and that negotiators address these emotional
understood, an amazing thing usually happens. Negative aspects directly and openly with the goal o moving beyond
energy dissipates, contention evaporates, mutual respect them into objective and collaborative problem solving.
• Focus on interests, not positions. It is crucial to look beyond the
grows, and people become creative. New ideas emerge.
Third alternatives appear. ormal stance a person has taken and attempt to understand his
underlying interests, the “root causes” o his position. By under-
One does not need to use a physical Talking Stick to gain these standing all parties’ basic interests (both one’s own and the other
bene ts. It is possible to establish a ramework or interacting in person’s), one increases the chances o identi ying new perspec-
which the parties agree that they will alternate the responsibilities tives or solutions that will meet both parties’ interests.
o talking and listening until both eel ully understood. This pro- • Invent options or mutual gain. The authors argue that once
cess can be very e ective in acilitating the resolution o con icts emotional and relationship issues have been separated rom
between hospitalists and other specialists regarding scope and the substantive problem, and all parties’ underlying interests
service issues. Some parties may be able to do this independently, are understood, the role o the parties is to invent better
while others may bene t rom acilitation by a third party mediator. options. The steps in this process are: separating the identi ca-
2. Unhappy Patients and Families: Take the HEAT. Some o tion o options rom the act o judging them, looking or many
the most challenging con icts that hospitalists must manage options rather than a single answer, ocusing on options that
are those that involve the unmet expectations o patients and result in mutual gains, and then coming up with ways to make
amilies. Keeping in mind the role o emotion in con ict, Byham the decisions easy.
• Insist on using objective criteria. Finally, Fisher and Ury acknowl-
(1993) recommends the ollowing approach or those who are
responsible or addressing the needs o unhappy patients and edge that despite one’s best e orts, negotiators will sometimes
amilies, as summarized by the acronym “Take the HEAT”: ace situations in which interests are truly in intractable con ict
• Hear them out. Active listening without interrupting, dis-
and mutually acceptable options may not be available. In these
agreeing, or de ending is the crucial rst step. Angry patients cases, e ective negotiators will insist that decisions be made
and amily members need to be able to express their emo- using objective, usually externally validated, criteria.
tions in order to let some o the air out o the emotional In addition to the tenets o Principled Negotiation outlined above,
balloon. it is important to recognize that when the issues are complex, even
169
the best and most care ully documented negotiation will probably Jones EE, Nisbett RE. The Actor and the Observer: Divergent Perceptions
ail to anticipate every nuance that may arise going orward. For o the Causes o Behavior. New York, NY: General Learning Press;
example, when hospitalists negotiate and memorialize a “service 1971.
agreement” with a group o specialists to de ne who will admit
P
Patterson K, Grenny J, McMillan R, et al. Crucial Conversations: Tools
A
which types o patients, invariably a patient will present who does or Talking when Stakes are High. New York, NY: McGraw Hill; 2002.
R
not t neatly into any o the categories speci ed in the service
T
Ross L. The intuitive psychologist and his shortcomings: distortions
agreement. I the potential or this to occur is not acknowledged
in the attribution process. In: Berkowitz L, ed. Advances in Experi-
I
and planned or up ront, additional con icts may arise despite the
mental Social Psychology. vol. 10. Orlando, FL: Academic Press;
parties’ care ul e orts. The most valuable asset in such situations is a
1977:173-240.
strong underlying relationship o mutual trust and respect that will
enable the parties to resolve these issues on a case-by-case basis. Thomas KW, Thomas GF, Shaubhut N. Con ict styles o men
T
and women at six organization levels. Int J Con ict Manage.
h
The bottom line is that even the best negotiation skills and most
e
clearly dra ted documents cannot substitute or strong relationships. 2008;19:148-166.
S
p
Triandis HC. Individualism and Collectivism. Boulder, CO: Westview
e
CONCLUSION
c
Press; 1995.
i
a
l
Con ict is inevitable in human interactions, and the potential
t
y
or serious con ict will grow as the complexity o interactions REFERENCES
o
f
increases. The extremely challenging milieus in which hospitalists
H
practice are ri e with misunderstandings, disagreements, and unmet 1. Patterson K, Grenny J, et al. Crucial Conversations: Tools or Talking
o
s
expectations, placing hospitalists at risk or con ict on a daily basis. when Stakes are High. New York, NY: McGraw Hill; 2002:45-51 and
p
i
t
There ore, the ability to understand and e ectively manage con ict 68-74.
a
l
should be a core competency or all hospitalists. The rst step in
M
2. Patterson K, Grenny J, et al. Crucial Conversations: Tools or Talking
building e ective con ict management skills is to understand the
e
when Stakes are High. New York, NY: McGraw Hill; 2002:32-34 and
d
causes and potential bene ts o con ict. Next, hospitalists should
i
51-62.
c
i
learn and apply key principles o con ict management; and nally,
n
3. The authors provide a ree expanded version o the Style Under
e
hospitalists need to develop competence and con dence in imple-
a
Stress Inventory online at https://www.vitalsmarts.com/styleun-
menting use ul strategies or managing di erent types o con ict.
n
derstress.aspx, along with additional guidance on interpreting the
d
S
results.
y
SUGGESTED READINGS
s
t
4. Hei etz RA, Linsky M. Leadership on the Line: Staying Alive through
e
m
Covey SR. The 7 Habits o Highly Ef ective People. New York, NY: Simon the Dangers o Leading. Boston, MA: Harvard Business School Pub-
s
and Schuster; 1989:235-260. lishing; 2002:164.
o
f
5. Covey SR. The 8th Habit. New York, NY: Simon and Schuster;
C
DeChurch LA, Marks MA. Maximizing the bene ts o task con ict: the
a
role o con ict management. Int J Con ict Manag. 2001;12:4-22. 2004:197-201.
r
e
Gilbert DT, Malone PS. The correspondence bias. Psychol Bull. 6. Byham WC. Zapp! Empowerment in Health Care. New York, NY:
1995;117:21-38. Random House; 1993:145-146.
Holt JL, DeVore CJ. Culture, gender, organizational role, and 7. Fisher R, Ury W. Getting to Yes: Negotiating Agreement Without
styles o con ict resolution: a meta-analysis. Int J Intercult Relat. Giving In. Boston, MA: Houghton Mif in; 1981:16-98.
2005;29:165-196.
170
SECTION 6
Billing, Coding, and Clinical
Documentation
171
CHAP TER
27 INTRODUCTION
The medical record o an individual patient serves numerous unc-
tions. Ideally, the record should provide a comprehensive historical
vehicle promoting excellence in care delivery to a patient, tran-
scending communication barriers, and acilitating care coordination
among multiple disparate providers and acilities (such as hospitals).
However, the medical record also serves as the basis or a variety
o nancial, legal, and administrative unctions including the docu-
mentation or both pro essional and acility ee reimbursement,
Professional quality and sa ety assessments (including pay or per ormance), mal-
practice litigation and disability determinations, and community-
Coding and based care and public health initiatives.

Currently, the medical record o an individual patient is rag-
Billing Guidelines mented, with various pieces shared only sometimes among numer-
ous providers. Hospitalists typically provide episodic, acility-based
care. Fortunately, the proli eration, adoption, and increasing interop-
for Clinical erability o electronic medical records (EMRs), and their evolution
into personalized health records, still holds promise or consolida-
Documentation tion and availability o all relevant clinical in ormation to each pro-
vider participating in the care o a single patient.
■ SOME GENERAL PRINCIPLES

Carol Pohlig, RN, BSN, CPC This chapter ocuses upon the documentation requirements incum-
bent upon hospitalists or pro essional ee billing o their clinical
Scott Manaker, MD, PhD services. Some general principles o clinical documentation warrant
discussion despite this ocus, and apply to both paper and EMRs.
The documentation o pro essional services should always com-
prise the essential components o a patient’s chie complaint,
history, physical examination (PE), and medical decision making
(MDM). The concerns o both patient and provider should be clearly
recorded, including expectations (realistic or not) and satis action
(and dissatis action). All diagnostic test orders and results should
reside in the chart, as well as documentation o various speci c ser-
vices (eg, physical, occupational, speech, or rehabilitation therapy;
home health services, durable medical equipment needs, and social
work evaluations).
PRACTICE POINT
• The clinical documentation o pro essional services should
always comprise the essential components o a patient’s chie
complaint, history, physical examination, and medical decision
making. The concerns o both the patient and provider should
be clearly recorded, including expectations (realistic or not) and
satis action (dissatis action).
Corrections
At some point in time, every medical record requires a correction. In
a paper document, draw a single line through the inaccurate por-
tion and write a correction nearby, dating and signing the revision.
The original entry thereby remains legible or uture re erence. For
example, consider misidenti cation o a right swollen knee joint,
when actually it is on the le t: when the physician recognizes the
mistaken documentation, the right side should have a single line
drawn through it (ie, overstrike text appearing as right) and a note
written nearby indicating that the le t side is the accurate side. Sign
and date these changes on the day o correction. Methods and
173
appearances o corrections and amendments in EMRs continue and procedures. Even physicians receiving capitated payments or
to evolve, but all incorporate password-protected signatures with participating in various advanced payment models still typically
electronic date- and time-stamped entries. report CPT codes or their services.
Hospitalists predominantly report E/M codes (CPT 99201-99499),
P
A
Late entry documentation which or Medicare exceed $32 billion annually and account or
R
Like corrections, addenda or late entries can be made at any time more than 40% o the Medicare physician ee schedule allowed
T
and labeled as such, with a currently dated and timed signature. charges. Other bedside procedures and diagnostic testing, some-
I
Also, explain why the entry was made late and not contemporane- times per ormed by hospitalists, are also ound in CPT. In selecting
ously. For example, ollowing a hospital visit, a hospitalist responds the proper E/M code, the site and nature o service determine the
to a rapid response call and neglects to document the visit in the visit category; and the key components o history, physical examina-
hospital medical record. When recognized the next day, document tion, and MDM determine the speci c level o CPT code within a
T
h
the hospital visit as a late entry and speci y that the care was deliv- visit category.
e
ered the previous day. Many payers are now adopting Medicare’s recognition and regu-
S
p
lation o nurse practitioners and physician assistants as advanced
e
c
■ AN OVERVIEW OF HOSPITAL INPATIENT practice providers, independently able to provide, document, and
i
a
bill or E/M services. The number o advanced practice providers
l
PAYMENT SYSTEMS
t
y
per orming hospitalist services is rapidly increasing. There ore, when
o
As the most prominent payer in the United States, Medicare pay-
we re er to providers throughout this chapter, we include both phy-
f
ment systems are o paramount importance to understand. Fortu-
H
sicians and these quali ed advanced practice providers unctioning
o
nately, Medicare policies are established by the Centers or Medicare
s
in an independent billing role.
p
and Medicaid Services (CMS) and available online through the CMS
i
t
a
On-Line Manual System, so that everyone can in principle under-
l
THE KEY COMPONENTS OF PHYSICIAN E/M
M
stand how claims are processed and payments are made. Medicare DOCUMENTATION
e
policies are largely consistent across the United States, and many
d
Selection o an E/M level ocuses upon the content o the three key
i
private payers ollow CMS’s lead. However, remember that regional
c
i
components: history, PE, and MDM. Time is considered a ourth key
n
Medicare Administrative Contractors develop local coverage deter-
e
minations rom CMS payment policy, and can create idiosyncratic component, but only a ects the E/M level when counseling and/
a
or coordination o care dominate more than 50% o the physician’s
n
interpretations o the documentation guidelines or evaluation
d
and management (E/M) services. One such example would be the total visit time (see below). When counseling and/or coordination o
S
care involves less than 50% o the physician’s total visit time, time is
y
nuanced medical allowances or a detailed exam (see Physical Exam
s
t
only considered a contributory actor and does not determine the
e
section) implemented by Novitas Solutions. Similarly, various third
m
party payers have the right, as allowable under their contract with E/M level.
s
Two sets o documentation guidelines have been elaborated
o
physician groups, to create unique documentation mandates and
f
payment policies. Facilities also generate requirements, such as a by Medicare and largely adopted by other payers. The earlier 1995
C
guidelines are the most widespread, and generally applicable to
a
required history and physical be ore a procedure, which do not
r
e
comprise medically necessary, billable physician services. hospitalists along with most medical and surgical specialists. The
Medicare pays hospitals or inpatient services using an inpatient later 1997 guidelines elaborate specialty-speci c physical exami-
prospective payment system (IPPS), which relies primarily on the nations, as well as clearly articulate detailed physical examination
diagnosis in order to group the services delivered to an inpatient into requirements lacking in the 1995 guidelines. An added variation also
a Medicare severity-adjusted diagnosis related group (MS-DRG). Many exists between the two guidelines in aspects o history. The 1995
diagnostic categories have two or three severity levels, di erenti- guidelines will be described in detail throughout this section, and
ated by the presence or absence o a speci ed set o complications the 1997 guidelines are highlighted below in a separate section or
and comorbidities (CCs) or major complications and comorbidities completeness.
(MCCs). Hospitals and other acilities requently request hospital-
ists to clari y, expand, or speci y their clinical documentation to ■ HISTORY
ensure the assignment o a hospitalization to the proper MS-DRG. The elements o history include the chie complaint (CC), history o
This single MS-DRG payment covers all acility services during the present illness (HPI), review o systems (ROS), and the past, amily,
inpatient stay. Many payers have adopted a similar mechanism o and social histories (PFSHs). A chart note may not segregate these
providing a single, xed payment or an entire hospitalization, o ten elements into unique subtitled areas, but rather the in ormation
re erred to as a case rate. may be interspersed amid the written, typed, or even dictated
Some non-Medicare payers still reimburse hospitals and acilities narrative.
with a xed payment or each day, commonly described as a per
diem rate. The daily payment potentially varies depending upon the Chief complaint CC
types o services provided to the patient (eg, intensive care versus Typically, the reason or the visit is o ten quoted rom the patient’s
skilled nursing care). Other payment models continue to emerge as own words as a sign or symptom, such as, “my belly hurts.” Always
a result o health care re orm, including the evolution o account- document a CC in the progress note, even absent an acute com-
able care organizations and the proli eration o patient-centered plaint, such as, “pneumonia ollow-up.” Avoid statements lacking a
medical homes. speci c clinical re erence (eg, “post-op visit Day #3”).
■ AN OVERVIEW OF PHYSICIAN PAYMENT SYSTEMS History of present illness HPI

Physician services are typically reported using the American Medi- The HPI conveys in ormation about the CC, rom either the ori-
cal Association (AMA) Current Procedural Terminology (CPT), ourth gin (at an initial encounter) or the interval between sequential
edition, which lists descriptive terms and identi ying codes to report patient encounters. This in ormation is arbitrarily allocated into
medical services and procedures. CPT provides a uni orm language eight elements: location, quality, severity, duration, timing, context,
to accurately describe all medical, surgical, and diagnostic services modi ying actors, and associated signs/symptoms. The HPI is then
174
quanti ed as brie (one to three elements) or extended ( our or Determination of history level
more elements). For example, consider this extended HPI: “Patient
C
The number o historical elements present in the chart note deter-
complains o increased (severity) pedal (location) edema that began
H
mines the level o history (Table 27-1). I all o the requirements
2 days ago (duration). Less able to walk. No chest pain (associated
A
are not met or a given level o history, select the level associated
P
signs/symptoms).” with the def cient element. For example, a comprehensive history
T
requires documentation o the CC, ≥4 HPI elements, ≥10 ROS, and
E
Review of systems ROS
R
a complete PFSH. I the ROS only includes documentation or nine
The ROS re ers to signs or symptoms experienced in conjunction systems, a comprehensive history cannot be selected; report a ser-
2
7
with the CC. Fourteen systems are recognized: constitutional, eyes, vice that requires only a detailed history: CC, ≥4 HPI elements, 2-9
ears/nose/mouth/throat, cardiovascular, respiratory, gastrointesti- ROS, and a pertinent PFSH.
nal, genitourinary, musculoskeletal, integumentary (which includes
P
r
the breast), neurologic, psychiatric, endocrine, hematologic/lym-
o
Other circumstances
e
phatic, and allergic/immunologic. Medical necessity, as deemed
s
A PFSH obtained during an earlier encounter does not need to be
s
by the treating provider in light o the patient’s current or previous
i
o
rerecorded i the provider documents review and updating o the
conditions, determines the number o systems required or review.
n
previous in ormation. Update the history by describing any new
a
A ROS may be problem pertinent, extended, or complete. A prob-
l
in ormation or noting the absence o change, along with the date
C
lem-pertinent ROS documents one system directly related to the CC.
o
and location o the earlier PFSH; this earlier PFSH must be contained
d
An extended ROS requires documentation o two to nine systems,
i
in a printable area o the medical record. CPT requires only an
n
that is, the system that is directly related to the CC, along with one
g
interval history or subsequent hospital or subsequent nursing acil-
or more additional systems. A complete ROS documents 10 or more
a
ity visits, and it is usually unnecessary to record in ormation about
n
individual systems. When obtaining a complete ROS, to decrease
d
the PFSH, which is unlikely to change in these settings.
the amount o time spent listing each system individually, both the
B
Most auditors disallow a single statement as both an HPI element
i
l
1995 and 1997 (see below) E/M documentation guidelines allow the
l
i
and ROS element. The ROS and/or PFSH may be recorded by ancil-
n
physician to comment on the positive and pertinent negative sys-
g
lary sta , or on a orm completed by the patient. The provider must
G
tems, with an additional comment that the “remainder is negative.”
document review and con rmation o this in ormation recorded by
u
However, insurers may not accept alternative phrases, and Medicare
i
d
others, either by a re erence to the history orm in the progress note
(and some non-Medicare) contractors seek individual documenta-
e
or by initialing and dating the orm, making any necessary annota-
l
i
tion o each system when less than 14 systems are reviewed.
n
tions, additions, or corrections.
e
s
I unable to obtain history rom the patient, the record should
Past, family, and social histories PFSHs
o
describe the patient’s condition or the circumstance that precludes
r
The past history includes documentation o previous illnesses, hospi-
C
obtaining a history, and what attempts the provider made to obtain
l
i
talizations, surgeries, medications, allergies, and immunizations. The
n
the in ormation. For example, “… patient sedated and paralyzed,
i
c
amily history provides in ormation regarding potential hereditary unable to obtain additional history. Family currently unavailable to
a
l
illnesses. The social history may list details o the patient’s substance contact; in ormation obtained rom the sta and available medical
D
use (tobacco/alcohol/illicit drugs), sexual history, employment sta-
o
records.” However, reviewers expect providers to incorporate his-
c
tus, level o education, marital status, or living arrangements.
u
toric in ormation to the extent possible, rom all reasonably available
m
A pertinent PFSH includes a comment in any one o the three sources (eg, old records, emergency medical services documents,
e
histories (ie, past, amily, or social). Full credit or a complete PFSH
n
other provider documentation, or conversations).
t
requires a comment in each history (ie, past, amily, and social)
a
Finally, although the physician may collect all o the in ormation
t
i
when reporting initial hospital, observation, or nursing acility care,
o
required or a complete ROS, the most common underdocumenta-
n
consultations, and new o ce, home, and domiciliary visits. In con- tion error is ailure to document at least 10 systems. The second
tradistinction, emergency department (ED) services or established most common mistake is a missing amily history or social history.
patient visits in the home, domiciliary, o ce, or other outpatient
area require one comment in two o the three histories or credit as
a complete PFSH. ■ PHYSICAL EXAMINATION (PE)
Providers may review and comment that the “ amily history is non- Individual PE elements will be assigned to body areas (head and
contributory” and still receive credit or the amily history rom some ace, neck, chest, abdomen, genitalia/groin/buttocks, back/spine,
insurers. Certain Medicare contractors, such as Palmetto GBA, prohibit and each extremity) or organ systems (constitutional, eyes, ears/
this terminology and require speci c documentation regardless o nose/mouth/throat, cardiovascular, respiratory, gastrointestinal,
clinical relevance (eg, “ amily history negative or liver disease”). Also genitourinary, musculoskeletal, integumentary, neurologic, psychi-
note that with subsequent services, both or hospital care and nursing atric, and hematologic/lymphatic/immunologic). Providers may
acility visits, only an “interval history” is required and redocumenta- document speci c ndings (eg, “abdomen so t”) or make a general-
tion o the PFSH is unnecessary rom a billing perspective. ized comment (eg, “HEENT normal”). Abnormal ndings must be
TABLE 27-1 Levels of History
History Level HPI ROS PFSH

Problem- ocused Brie (≤3) None None
Expanded problem- ocused Brie (≤3) Problem pertinent (1) None
Detailed Extended (≥4) Extended (2-9) Pertinent (1)
Comprehensive Extended (≥4) Complete (≥10) Complete (2 or 3)*
All three or new patient encounters; two o three or subsequent or ED encounters.

*
175
speci cally documented, such as “S3”; however, a comment indicat-
ing “abnormal” without elaboration is insu cient. TABLE 27-2 Valuation of Diagnostic and Treatment Options
The PE documented in the medical record is categorized as
Number of Diagnoses/Treatment
problem- ocused, expanded problem- ocused, detailed, or compre-
P
Options Points per Problem
A
hensive. One comment in one area constitutes a problem- ocused
Sel -limited/minor problem (stable, 1 (max = 2 problems)
R
exam. The distinction between the expanded problem- ocused
improved, or worsening)
T
and detailed examination under the 1995 Guidelines is the great-
I
est ambiguity in physical examination documentation. Both the Established problem (stable or 1
improving)
expanded problem- ocused and detailed exams require documenta-
tion o two to seven systems. However, “detailed” is de ned as an Established problem (worsening) 2
extended examination o the a ected body area or organ system, New problem, without additional 3 (max = 1 problem)
T
workup
h
in addition to other symptomatic or related organ systems. The
e
number o required comments regarding the a ected body area New problem, with additional workup 4
S
planned
p
or organ system to consider the examination detailed has never
e
been de ned by either CPT or Medicare. Attempting to decrease
c
i
a
ambiguity and variability among auditors, Novitas Solutions scores
l
t
y
a detailed exam using the “4 × 4” rule: our elements examined in
o
our body areas or our organ systems (totaling 16 documentation is categorized as a new problem to the hospitalist newly treating the
f
patient during an admission or ketoacidosis.
H
elements). In contrast, other contractors suggest using the 1997
o
Established patients may also have new problems. For example,
s
guidelines (discussed later) or detailed exam requirements.
p
an asthmatic with a resolving f are may experience heartburn. This
i
The comprehensive examination is a general multisystem exami-
t
a
additional new complaint o heartburn may be considered new i
l
nation or a complete examination o a single organ system.
M
Medicare requires the minimum documentation or the general commented upon in the progress note and no prior care plan by the
e
hospitalist team or gastroesophageal ref ux exists.
d
multisystem examination to include one comment in each o eight
i
Physicians receive credit only or issues considered in the care
c
systems; o course, additional comments in each system and more
i
n
plan. Diagnoses merely listed in the assessment and plan without
e
than eight systems may be described, as clinically indicated. For
elaboration o the care, or simply ascribing the care to others
a
example, a comprehensive examination may be documented as
n
(eg, “diabetes—per endocrinologist”) are considered part o the
d
ollows: “P = 76, BP = 120/80, RR = 12 (constitutional); HEENT normal
patient’s problem list in the PFSH and do not add to the complexity
S
(eyes and ENMT); neck supple (musculoskeletal); regular rate and
y
o MDM. Additionally, new hospitalizations warrant new care plan
s
rhythm (cardiovascular); lungs clear (respiratory); normal bowel
t
e
development, and physicians can receive new problem credit even
m
sounds (gastrointestinal); no rashes (integumentary), normal gait
i the patient has been previously hospitalized by the same group.
s
(neurological).” The requirements or a comprehensive single organ
o
system still remain unde ned or use with these 1995 guidelines. This is a nuance o inpatient and observation care only.
f
C
a
Data considered
r
e
■ MEDICAL DECISION MAKING (MDM)
The second category o determining the MDM complexity is the
The complexity o MDM drives selection o a level o service. MDM is amount and/or complexity o data reviewed or ordered by the
categorized as straight orward, low, moderate, or high. Three catego- provider during the patient encounter. Both the type and source o
ries must be considered to determine the level o MDM complexity: in ormation considered are valued (Table 27-3).
the number o diagnoses, the amount and complexity o data, and Ordering and/or reviewing o pathology/laboratory, radiology,
the risk to the patient. and medicine data each provide separate but equal credit. Irrespec-
tive o the test volume in each category, only one point is allocated
Number of diagnoses considered
This rst category identi es the number o diagnoses and/or man-
agement options considered in the encounter, based upon the
documentation. Up to our points are assigned to each problem, TABLE 27-3 Valuation of Data Considered
with more points assigned or new problems than or established
problems, and a new problem requiring additional workup (ie, diag- Amount and/or Complexity of Data
nostic testing) given the maximum our points. Established prob- Ordered/Reviewed Points
lems identi ed as worsening receive a higher value than stable or Review and/or order o clinical test(s) 1
improving problems. A sel -limited or minor problem (eg, sunburn) Review and/or order o test(s) in the 1
receives minimal credit as these issues typically do not warrant a pathology/laboratory section o CPT
de ned plan o care (Table 27-2). Review and/or order o test(s) in the 1
New problems require initiation o a care plan, while established radiology section o CPT
problems may require modi cation or continuation o a care Review and/or order o test(s) in the 1
plan. An established problem has been previously considered by medicine section o CPT
the physician or provider group (to allow or cross coverage and Decision to obtain old records and/or 1
hando s between same specialty providers in the same group) in obtain history rom someone (nonhealth
the strictest o interpretations o the guidelines. Note that credit is care provider) other than the patient
also given or a problem considered, although not primarily under Review and summarize old records, obtain 2
treatment by the physician. For example, in a patient receiving additional history, or discuss the case with
steroids or an inf ammatory disease, the hospitalist receives credit another health care provider
or noting the potential adverse consequence upon serum lipids, Independent visualization o actual image, 2
even i a cardiologist is primarily treating the dyslipidemia. Notice, a tracing, or specimen
chronic condition such as diabetes, cared or by an endocrinologist,
176
per category (ie, pathology/laboratory, radiology, or medicine) or considered) is eliminated, and the lower o the two remaining
the encounter. For example, the provider ordering a dozen serologic categories (number o diagnoses) determines the moderate MDM
C
collagen vascular studies in the morning may also review the three complexity or the note. While most contractors utilize standardized
H
results received in the a ternoon; nonetheless, only one point is concepts when assigning points, beware o contractors and other
A
P
granted or this care. A single, separate point may be assigned each auditing programs who may impose di erent standards.
T
to pathology/laboratory, radiology, and medicine data, respectively,
E
are cumulative in nature, and the chart note should re er to all the ■ 1997 GUIDELINES
R
data reviewed or ordered to capture all o the provider work. In Medicare issued a second set o revised documentation guidelines
2
other words, i the chart note comments upon a radiology result
7
or E/M services in 1997. MDM and the level categories remained
(one point) and an echocardiogram order (one point), two points unchanged rom the prior 1995 documentation guidelines, as
may be awarded or the amount o data in that encounter. Inde- detailed above. While ambiguity plagues many aspects o the 1995
P
pendently visualizing images, tracings, or specimens is considered
r
guidelines, excessive proscription limits the 1997 guidelines. The
o
separately, and additional to reviewing the ormal interpretation,
e
1997 guidelines made a single minor revision to the history, while
s
as long as the chart note clearly documents this occurrence (ie, “…
s
the physical examination content received extensive modi cation.
i
o
my review o the EKG tracings showed…”). Without such speci c The requirements or each level o physical examination were heav-
n
a
re erence distinguishing personal review o the images and o the ily revised, and specialty-speci c single organ system examinations
l
C
ormal interpretation, an auditor only provides minimal credit or de ned. In response to widespread complaints rom the physician
o
merely reviewing the report. A re erence to data review without
d
community, CMS allowed physicians to document using either the
i
n
a comment about the result (eg, “CXR reviewed…”) will yield no 1995 or 1997 guidelines, according to their individual pre erence,
g
credit; there ore, be sure to include a comment on the ndings (eg, and directed auditors to review the physician documentation based
a
n
“CXR reveals cardiomegaly…”). on the set o guidelines most avorable to the physician or the E/M
d
Providers also receive credit or the additional e ort o obtaining code reported or the encounter. Most physicians, including hospi-
B
i
historical in ormation rom sources other than the patient, such as
l
talists, nd the 1995 guidelines most applicable.
l
i
n
records rom previous encounters or hospitalizations, and conversa-
g
tions with other health care pro essionals or amily members. The History
G
u
chart note should speci cally mention the source, along with the The 1997 guidelines do not limit the provider to identi ying indi-
i
d
in ormation reviewed (eg, “…spouse con rms loud snoring”).
e
vidual actors associated with the CC (eg, duration, timing, and
l
i
n
context). Rather, a provider may document the status o one or
e
Risk to the patient more chronic conditions; this option is most use ul or subsequent
s
hospital visits. A brie HPI documents one or two conditions, while
o
The third MDM category assesses the patient’s risk o complications,
r
morbidity, or mortality, with respect to the presenting problem, an extended HPI documents a minimum o three conditions. Some
C
l
reviewers allow this option only i applying the 1997 guidelines to
i
diagnostic procedures ordered, or management options chosen.
n
i
the entire note. However, beginning with dates o service Septem-
c
Four levels o patient risk exist (minimal, low, moderate, and high),
a
ber 10, 2013, Medicare gives credit or the status o chronic condi-
l
with examples o each risk type included in Medicare’s “Table o
D
Risk” (Table 27-4). The limited number o examples serves as an tions rom the 1997 guideline and the physical examination rom
o
c
illustrative re erence or common clinical scenarios, but not as a the 1995 guideline mixed within the same note.
u
m
comprehensive list.
Physical examination
e
When determining the level o risk, consider comorbidities as well
n
t
as the plan o care in assessing a patient’s risk, thereby potentially The 1997 guidelines allow a provider to select either a general mul-
a
t
tisystem examination or any one o the single organ system exami-
i
increasing the complexity o MDM. Similarly, diagnostic studies or
o
n
alternatively, procedures under consideration or excluded based nations. Hospitalists typically utilize the general examination, which
upon excessive risk, impact the complexity o MDM (eg, “…will speci es examination elements to per orm and document. Negative
de er MRI, as potential morbidity o transport to magnet exceeds or normal comments remain acceptable or the 1997 guidelines,
risks o empiric treatment”). Although many bulleted items on the along with the mandate to speci y comments on any abnormal
table may pertain to a particular chart note, the single bulleted item ndings. Documentation in the medical record o 1 to 5 speci ed
in any risk category associated with the highest risk determines the (re erred to as bulleted) physical examination elements comprises a
patient’s risk level. For example, a note documenting the monitoring problem- ocused examination; 6 to 11 bulleted elements de nes the
o liver unction tests or potential hepatotoxicity, or blood counts expanded problem- ocused examination; and a detailed examination
or possible cytopenias, comprises high risk (drug therapy requiring requires 12 or more bulleted items.
requent monitoring or toxicity). However, remember risk does not For the 1997 comprehensive general multisystem examination
equal complexity: risk is but one category (among three) o MDM (Tables 27-6 and 27-7), the provider must per orm all the elements
and not the sole contributor to complexity. speci ed in at least nine organ systems or body areas but only needs
to document a minimum o two elements rom each o those nine
Assigning MDM complexity systems or areas. For single organ system exams, all bullets must be
per ormed; however all bullets in all shaded boxes and at least one
Based upon the chart note, points are assigned or diagnoses man-
bullet in each unshaded box must be documented.
aged and data considered, and patient risk is assessed. The nal
result o MDM complexity hinges on the two highest valued catego-
ries. In other words, two o the three categories must meet or exceed ■ DETERMINING LEVEL OF SERVICE
the requirements assigned to a specif c level o complexity to select that For both the 1995 and 1997 guidelines, assign a speci c level to
level, as illustrated in Table 27-5. each o the three key components. Rate history and examina-
To illustrate the assignment o MDM complexity, consider the tion each as either problem- ocused, expanded problem- ocused,
ollowing example. The chart note considers three stable estab- detailed, or comprehensive. Rate the complexity o MDM as either
lished diagnoses (three points), several blood tests (one point), straight orward, low, moderate, or high. CPT correlates speci c levels
and high patient risk. The lowest o the three categories (data o the key components with certain levels o most E/M services.
177
TABLE 27-4 Table of Risk
Level of Risk Presenting Problem(s) Diagnostic Procedure(s) Ordered Management Options Selected
P
Minimal One sel -limited or minor problem (eg, cold,
• Laboratory tests requiring
• Rest
•
A
insect bite, tinea corporis) venipuncture Gargles
•
R
Chest x-rays Elastic bandages
T
• •
ECG/EEG
• Super icial dressings
•
I
Urinalysis
•
Ultrasound (such as
•
echocardiography)
KOH prep
•
T
h
Low Two or more sel -limited or minor problems
• Physiologic tests not under
• Over-the-counter drugs
•
e
One stable chronic illness (eg, well-controlled
• stress (eg, pulmonary unction Minor surgery with no
•
S
p
hypertension, noninsulin dependent diabetes, tests) identi ied risk actors
e
cataract, BPH) Noncardiovascular imaging
• Physical therapy
•
c
i
a
Acute uncomplicated illness or injury,
• studies with contrast (such as Occupational therapy
•
l
t
(eg, cystitis, allergic rhinitis, simple sprain) chest or abdominal CT) IV luids without additives
y
•
o
Super icial needle biopsies
•
f
Clinical laboratory tests
•
H
o
requiring arterial puncture
s
p
Skin biopsies
•
i
t
a
Moderate One or more chronic illnesses with mild
• Physiologic tests under stress
• Minor surgery with identi ied
•
l
exacerbation, progression, or side e ects o (such as cardiac stress test, risk actors
M
treatment etal contraction stress test) Elective major surgery (open,
e
•
d
Two or more stable chronic illnesses
• Diagnostic endoscopies with
• percutaneous, or endoscopic)
i
c
Undiagnosed new problem with uncertain no identi ied risk actors with no identi ied risk actors
i
•
n
e
prognosis (eg, lump in breast) Deep needle or incisional
• Prescription drug
•
a
Acute illness with systemic symptoms
• biopsy management
n
d
(eg, pyelonephritis, pneumonitis, colitis) Cardiovascular imaging
• Therapeutic nuclear medicine
•
S
Acute complicated injury (eg, head injury
• studies with contrast and IV luids with additives
•
y
with brie loss o consciousness) no identi ied risk actors Closed treatment o racture
s
•
t
e
(eg, arteriogram, cardiac or dislocation without
m
catheterization) manipulation
s
Obtain luid rom body cavity
•
o
f
(such as lumbar puncture,
C
thoracentesis, paracentesis)
a
r
e
High One or more chronic illnesses with severe
• Cardiovascular imaging
• Elective major surgery (open,
•
exacerbation, progression, or side e ects o studies with contrast with percutaneous, or endoscopic)
treatment identi ied risk actors with identi ied risk actors
Acute or chronic illnesses or injuries that
• Cardiac electrophysiological
• Emergency major surgery
•
pose a threat to li e or bodily unction tests (open, percutaneous, or
(such as multiple trauma, acute MI, pulmonary Diagnostic endoscopies with
• endoscopic)
embolus, severe respiratory distress, progressive identi ied risk actors Parenteral controlled
•
severe rheumatoid arthritis, psychiatric illness Discography
• substances
with potential threat to sel or others, peritonitis, Drug therapy requiring
•
acute renal ailure) intensive monitoring or
An abrupt change in neurologic status
• toxicity
(eg, seizure, TIA, weakness, sensory loss) Decision not to resuscitate or
•
to de-escalate care because o
poor prognosis
Initial patient encounters (initial hospital care, CPT 99221-99223;

TABLE 27-5 Levels of Medical Decision Making initial and subsequent observation care, CPT 99218-99220 and
99234-6; consultations, CPT99241-99245 and 99251-99255; and new
Diagnosis or o ce visits CPT 99201-5) require consideration o all three key com-
Treatment ponents. Consider only two o the key components or subsequent
Complexity Option Points Data Points Risk Level
hospital (CPT 99231-99233), observation visits (CPT 99224-99226), or
Problem- ≤1 (minimal) ≤1 (minimal) Minimal established outpatient (CPT 99211-5). The lowest component o the
ocused
two or three key components required determines the visit level. For
Low 2 (limited) 2 (limited) Low example, a level three initial hospital service (CPT 99223) includes a
Moderate 3 (multiple) 3 (multiple) Moderate comprehensive history, comprehensive exam, and high-complexity
High 4 (extensive) 4 (extensive) High decision making (Table 27-8). I the documentation merely sup-
ports a detailed level o history, yet meets the requirements or a
178
TABLE 27-6 1997 General Multisystem Physical Examination
C
H
System/Body Area Elements of Examination
A
Constitutional Measurement o any three o the ollowing seven vital signs: (1) sitting or standing blood pressure, (2) supine blood
•
P
pressure, (3) pulse rate and regularity, (4) respiration, (5) temperature, (6) height, (7) weight (may be measured and
T
recorded by ancillary sta )
E
R
General appearance o patient (eg, development, nutrition, body habitus, de ormities, attention to grooming)
•
2
Eyes Inspection o conjunctivae and lids
•
7
Examination o pupils and irises (eg, reaction to light and accommodation, size, and symmetry)
•
Ophthalmoscopic examination o optic discs (eg, size, C/D ratio, appearance) and posterior segments (vessel
•
P
r
changes, exudates, hemorrhages)
o
e
Ears, nose, External inspection o ears and nose (overall appearance, scars, lesions, masses)
•
s
s
mouth, and Otoscopic examination o external auditory canals and tympanic membranes
•
i
o
throat
n
Assessment o hearing (eg, whispered voice, inger rub, tuning ork)
•
a
l
Inspection o nasal mucosa, septum, and turbinates
•
C
o
Inspection o lips, teeth, and gums
•
d
i
n
Examination o oropharynx: oral mucosa, salivary glands, hard and so t palates, tongue, tonsils, and posterior
•
g
pharynx
a
n
Neck Examination o neck (masses, overall appearance, symmetry, tracheal position, crepitus)
•
d
B
Examination o thyroid (enlargement, tenderness, mass)
•
i
l
l
Respiratory Assessment o respiratory e ort (eg, intercostal retractions, use o accessory muscles, diaphragmatic movement)
i
•
n
g
Percussion o chest (dullness, latness, hyperresonance)
•
G
Palpation o chest (tactile remitus)
u
•
i
d
Auscultation o lungs (breath sounds, adventitious sounds, rubs)
•
e
l
i
Cardiovascular Palpation o heart (location, size, thrills)
n
•
e
s
Auscultation o heart with notation o abnormal sounds and murmurs
•
o
Examination o :
r
C
carotid arteries (pulse amplitude, bruits)
•
l
i
n
abdominal aorta (size, bruits)
•
i
c
a
• emoral arteries (pulse amplitude, bruits)
l
D
p edal pulses (pulse amplitude)
•
o
c
e xtremities or edema and/or varicosities
•
u
m
Chest (breasts) Inspection o breasts (symmetry, nipple discharge)
•
e
n
Palpation o breasts and axillae (masses or lumps, tenderness)
•
t
a
Gastrointestinal Examination o abdomen with notation o presence o masses or tenderness
•
t
i
o
(abdomen) Examination o liver and spleen
n
•
Examination or presence or absence o hernia
•
Examination o anus, perineum, and rectum, including sphincter tone, presence o hemorrhoids, rectal masses
•
Obtain stool sample or occult blood test when indicated
•
Genitourinary Male:
Examination o the scrotal contents (hydrocele, spermatocele, tenderness o cord, testicular mass)
•
Examination o the penis
•
Digital rectal examination o prostate gland (size, symmetry, nodularity, tenderness)
•
Female:
Pelvic examination (with or without specimen collection or smears and cultures), including
Examination o external genitalia (general appearance, hair distribution, lesions) and vagina (eg, general
•
appearance, estrogen e ect, discharge, lesions, pelvic support, cystocele, rectocele)
Examination o urethra (masses, tenderness, scarring)
•
Uterus (size, contour, position, mobility, tenderness, consistency, descent or support)
•
Adnexa/parametria (masses, tenderness, organomegaly, nodularity)
•
Examination o bladder (eg, ullness, masses, tenderness)
•
Cervix (general appearance, lesions, discharge)
•
Uterus (size, contour, position, mobility, tenderness, consistency, descent or support)
•
Adnexa/parametria (masses, tenderness, organomegaly, nodularity)
•
(Continued )
179
TABLE 27-6 1997 General Multisystem Physical Examination (Continued)
System/Body Area Elements of Examination

P
Lymphatic Palpation o lymph nodes in two or more areas:
A
R
Neck
•
T
Axillae
•
I
Groin
•
Other
•
Musculoskeletal Examination o gait and station
•
T
Inspection and/or palpation o digits and nails (clubbing, cyanosis, in lammatory conditions, petechiae, ischemia,
•
h
e
in ections, nodes)
S
Examination o joints, bones, and muscles o one or more o the ollowing six areas: (1) head and neck; (2) spine, ribs,
p
e
and pelvis; (3) right upper extremity; (4) le t upper extremity; (5) right lower extremity; and (6) le t lower extremity.
c
i
The examination o a given area includes:
a
l
t
y
Inspection and/or palpation with notation o presence o any misalignment, asymmetry, crepitation, de ects,
•
o
tenderness, masses, or e usions
f
H
Assessment o range o motion with notation o any pain, crepitation or contracture
•
o
s
Assessment o stability with notation o any dislocation (luxation), subluxation, or laxity
•
p
i
t
Assessment o muscle strength and tone ( laccid, cog wheel, spastic) with notation o any atrophy or abnormal
•
a
l
movements
M
Skin Inspection o skin and subcutaneous tissue (rashes, lesions, ulcers)
e
•
d
i
Palpation o skin and subcutaneous tissue (induration, subcutaneous nodules, tightening)
•
c
i
n
Neurologic Test cranial nerves with notation o any de icits
•
e
a
Examination o deep tendon re lexes with notation o pathological re lexes (eg, Babinski)
•
n
d
Examination o sensation (by touch, pin, vibration, proprioception)
•
S
Psychiatric Description o patient’s judgment and insight
y
•
s
t
Brie assessment o mental status including:
e
m
Orientation to time, place, and person
•
s
o
Recent and remote memory
•
f
C
Mood and a ect (depression, anxiety, agitation)
•
a
r
e
TABLE 27-7 Levels of 1997 Physical Examination
Level of Exam Performance and Documentation

Problem- ocused 1 to 5 elements identi ied by a bullet.
Expanded problem- ocused At least 6 elements identi ied by a bullet.
Detailed At least 2 elements identi ied by a bullet rom each o 6 areas/systems OR at least 12 elements identi ied by
a bullet in 2 or more areas/systems.
Comprehensive Per orm all elements identi ied by a bullet in at least 9 organ systems or body areas and document at least
2 elements identi ied by a bullet rom each o 9 areas/systems.
TABLE 27-8 Levels of Initial Hospital Care
Initial Hospital Care History Examination MDM Time

99221 Detailed or Detailed or Straight orward or low 30 min
comprehensive comprehensive
99222 Comprehensive Comprehensive Moderate 50 min
99223 Comprehensive Comprehensive High 70 min
180
TABLE 27-9 Levels of Subsequent Hospital Care
C
H
Subsequent Hospital Care History Examination MDM Time
A
99231 Problem- ocused Problem- ocused Straight orward or low 15 min
P
T
99232 Expanded Expanded Moderate 25 min
E
99233 Detailed Detailed High 35 min
R
2
7
comprehensive examination and high-complexity decision mak- assessments. Such tools are o ten custom designed in the context
P
ing, report only a level one initial hospital service (CPT 99221). In o chart reviews and institutional initiatives, or implementing cod-
r
o
contrast, i a subsequent hospital visit note contains a complete ing and billing audits associated with practice compliance plans.
e
examination and high-complexity MDM, then report CPT 99233 and CMS has not developed or endorsed any ormal audit tool or E/M
s
s
i
history need not even be considered (Table 27-9). When selecting services, although virtually all Medicare and non-Medicare contrac-
o
n
visit levels or services that only consider two key components, MDM tors (eg, Novitas and First Coast Service Options, Inc) make their E/M
a
l
should be one o those two key components. While not stated in score sheets available to assist providers in adhering to documenta-
C
o
the documentation guidelines, medical necessity underlies every tion guidelines.
d
physician service and is most appropriately demonstrated through
i
n
g
MDM. Reporting subsequent services with MDM as a key compo-
SUGGESTED READINGS
a
nent thereby precludes the allegation o an unwarranted high-level
n
d
subsequent encounter based upon merely a comprehensive, but
American Medical Association. Current Procedural Terminology (CPT)
B
medically unnecessary, history and examination or management o
i
l
2016. Chicago, IL: American Medical Association; 2015.
l
i
the presenting problem (eg, common cold).
n
g
Centers or Medicare & Medicaid Services. Evaluation and Man-
G
■ TIME COUNSELING/COORDINATING CARE agement Services guide; November 2014. Available at http://
u
i
www.cms.gov/Outreach-and-Education/Medicare-Learning-Net-
d
CPT assigns most E/M codes a typical time to render a service, but
e
work-MLN/MLNProducts/downloads//eval_mgmt_serv_guide-
l
importantly, the service duration need not last that length. For inpa-
i
n
ICN006764.pd .
e
tient services, time accrues as unit or f oor time in addition to ace-
s
to- ace time. When more than 50% o the total service time involves Centers or Medicare and Medicaid Services. Medicare Claims Process-
o
counseling and/or coordination o care, the provider may select a ing Manual. Bethesda, MA: Department o Health and Human
r
C
code ref ecting the total time spent with the patient, rather than Services; 2014. Available online at www.cms.hhs.gov/manuals/
l
i
n
the three key components. Time and the corresponding counseling downloads/clm104c12.pd .
i
c
a
details must be documented in the medical record when selecting Kuhn T, Basch P, Barr M, et al. Clinical documentation in the 21st
l
D
the E/M code on the basis o time (eg, “25 o 35 minutes spent urg- century: executive summary o a policy position paper rom the
o
ing the patient to undergo a diagnostic biopsy”). O course, record American College o Physicians. Ann Intern Med. 2015;162:301-303.
c
u
patient responses to counseling and all relevant history, examina-
m
Levinson DR. Coding Trends o Medicare Evaluation and Management
tion, and MDM as necessary or good patient care. A subsequent
e
Services. O ce o the Inspector General, Department o Health
n
hospital service involving 35 minutes spent by the provider permits
t
and Human Services; 2012. Available at http://oig.hhs.gov/oei/
a
selection o 99233 (Table 27-9).
t
reports/oei-04-10-00180.pd .
i
o
n
Levinson DR. Improper Payments or Evaluation and Management
TEMPLATED NOTES Services Cost Medicare Billions in 2010. O ice o the Inspector
Templated chart notes, whether ormatted as preprinted paper General, Department o Health and Human Services; 2014. Avail-
progress notes with check boxes or electronically constructed as able at http://oig.hhs.gov/oei/reports/oei-04-10-00181.pd .
combinations o macros and click-boxes, are common, appropriate Manaker S, Merlino D, Pohlig CA. Coding or Chest Medicine 2016:
documentation tools. Such tools enhance legibility and acilitate Pulmonary, Critical Care and Sleep. Northbrook, IL: American College
e cient documentation in accord with the E/M guidelines. Un or- o Chest Physicians; 2015.
tunately electronic notes o ten become the source or increasingly Novitas Solutions, Inc. Novitas Solutions: Specialty Exam
prevalent “cut-and-paste” errors, highlighting the need to balance Scoresheets. Available online at http://www.novitas-solutions.
e ciency with accuracy in providing sa e, e cient care to complex com/ webcenter/ aces/ oracle/ webcenter/ page/ scopedMD/
patients. Make each note speci c to the patient on that encounter sa d 60252a _5537_4c5d _9350_ca 405e 36e 159/ Pa g e 133.
date. Modi y in ormation and language brought orward rom any jsp x?conte n tId =00024402&_a rLoop =475344439251000#!
previous encounters so the current documentation demonstrates %40%40%3F_a rLoop %3D475344439251000%26contentId
the distinct clinical service o today. Do not include excessive data %3D00024402%26_ad .ctrl-state%3D141uivweue_335.
or repetitious in ormation that is not relevant to the current service;
such content is o ten misconstrued as being included merely to Siegler EL, Adelman R. Copy and paste: a remediable hazard o elec-
increase the billing level. tronic health records. Am J Med. 2009;122:495.
Siegler EL. The evolving medical record. Ann Intern Med.
■ AUDIT TOOLS 2010;153:671-677.
Audit tools are use ul adjuncts to ensuring adherence to E/M billing Siegler JE, Patel N, Dine CJ. Prioritizing paperwork over patient care:
guidelines, patient sa ety and quality initiatives, and other process why can’t we do both? J Grad Medical Educ. 2015;7:16-18.
181
CHAP TER
28 CONSULTS
■ INTRODUCTION
Medicare stopped paying or the consult codes [99241-99245,
99251-99255] January 1, 2010. The consult codes paid more than
their corresponding inpatient ollow-up [99231-99233] and outpa-
tient visit [99201-99215] alternatives, but came with very speci c,
but strict and con using, situational and documentation criteria to
allow or their reimbursement. Although Medicare stopped paying
Consultation, or the consult codes, they did not stop paying or consult services.
The concept o needing the expertise o another provider is still
essential to the practice o medicine. What has changed is how
Comanagement, those services need to be reported or reimbursement to Medicare.
Some private payors still reimburse or the consult codes, using the
Time-Based, and old Medicare situational and documentation guidelines, but these
numbers are decreasing. Until 100% o insurers stop paying or
Palliative Care Billing

the consult codes [99241-99255], two di erent payment systems
or consults exist. Thus, the same work provided to two identical
patients may be reimbursed di erently depending on the patient’s
insurer.
Yvette M. Cua, MD
■ MEDICARE PATIENTS AND THOSE INSURERS
FOLLOWING MEDICARE CONSULT RULES
The set o codes submitted or a consult to a Medicare patient is
based on several variables: inpatient versus outpatient status, dispo-
sition a ter being seen in an emergency room (eventual admission
vs observation status vs discharge home), and whether a patient is
new or established to the physician specialty and billing group.
(See Figure 28-1.)
■ INPATIENTS
Initial consults or patients in acute care hospitals and skilled nurs-
ing acilities (SNF) are billed with the same initial inpatient visit
codes used or admissions [99221-99223 (initial inpatient visit),
99304-99306 (initial SNF visit)]. All subsequent consult visits are
billed using the subsequent inpatient or SNF visit codes respectively
[99231-99233, 99307-99310]. The nal day seeing the patient as a
consultant will still be billed using the subsequent inpatient and
SNF codes, even i the consultant is contributing to or ully providing
the patient’s discharge management. Only the admitting physician/
billing group can bill or discharge management services [99238-
99239]. Medicare will only pay or medically necessary visits and not
just daily “routine” visits because the patient is still in the inpatient
hospital or SNF.
I a new problem or question sur aces, or the consultant is recon-
sulted during the same inpatient admission, with or without a time
lapse since the last consultant visit, a subsequent inpatient/SNF
visit code [99231-99233, 99307-99310] would be submitted or this
service.
I a consultant is asked to see a patient whom they provided
a consult on during a previous admission, even with the same
problem(s) as last admission, the consultant would still bill an initial
inpatient/SNF visit code [99221-99223, 99304-99306] or this work.
In the event that the documentation or the initial consult does
not meet the minimum documentation criteria or the lowest level
o initial inpatient service [99221], Medicare will allow this work to
be reported with subsequent inpatient visit codes [99231-99232].
The documentation criteria or these services parallels those o the
182
they may occupy a hospital bed in the same room as an inpatient.
Alg o rithm fo r s e le c ting the pro pe r CPT c o de s e t fo r
Initial consults or these patients are billed using outpatient visit
C
me diCARE c o ns ults in a ho s pital s e tting
codes [99201-99215]. Initial observation visit codes [99218-99220,
H
What is the patie nt’s s tatus ?
A
99234-99236] should not be submitted or this work. Within the out-
P
patient set o codes are new patient [99201-99205] and established
T
ER patie nt Inpatie nt Obs e rvatio n Outpatie nt patient [99211-99215] codes. A consultant now needs to determine
E
Loca tion d oe s n’t Amb ula tory s urg e ry
i a patient is “new” versus “established” to select the correct code
R
ma tte r, tha t is , outp a tie nt c linic
floor, ED, set to use. A patient is considered “new” to an individual physician
2
MICU, CCU, OR
i they have not received any E/M services rom any member o that
8
Has pt. be e n s e e n ≤3 y physician’s billing group and speci c (sub)specialty within the past
Was the pt. by anyo ne fro m yo ur 3 years, regardless o the location o the service. Note that billing or
C
dis c harg e d ? g ro up + s ubs pe c iality ? the pro essional component o a service, or writing an order on a
o
n
patient does not a ect the “new” or “established” status o a patient.
s
Ye s No Ye s No
u
Below are some common examples to illustrate nuances with
l
t
a
this de nition.
t
i
ER Initial Es tablis h New
o
1. An internal medicine consultant evaluates a Medicare obser-
n
vis it Inpt vis it o utpt o utpt
,
99281-5 99221-3 99211-5 99201-5
vation patient on the neurology service. The consultant has
C
o
never seen the patient be ore, but the patient was seen in the
m
emergency room by an internal medicine partner in their bill-
a
What is the patie nt’s s tatus ?
n
ing group, moonlighting 2 years ago. This consult would be
a
g
reported with an established outpatient code [99212-99215].
e
Inpatie nt Obs e rvatio n
2. A hospitalist evaluates a patient in observation status at the
m
Loca tion d oe s n’t
e
ma tte r, tha t is , request o the ENT service. This patient has a history o monthly
n
floor, ED,
admissions to the hospital medicine service or severe chronic
t
,
MICU, CCU OR
T
Has pt. be e n s e e n ≤3 y obstructive lung disease (COPD) although not recently. The ini-
i
m
by anyo ne fro m yo ur tial medicine consult would be billed using the new outpatient
e
g ro up + s ubs pe c iality
-
code set [99201-99205].
B
a
3. A patient, seen routinely by an endocrinologist in a multispe-
s
Ye s No
e
cialty practice or years, is placed in observation status on the
d
,
general surgery service with a consult request to an internal
a
Initial Es tablis h New
n
Inpt vis it o utpt o utpt medicine hospitalist in the same multispecialty group as the
d
99221-3 99211-5 99201-5 endocrinologist. The patient has never been seen by the inter-
P
a
nal medicine group in the past. This consult would be reported
l
l
i
Figure 28 1 Algorithm or selecting the proper CPT code set or medi- with a new outpatient code [99201-99205].
a
t
CARE consults in a hospital setting. (Copyright 2010, Yvette M. Cua,
i
4. A cardiologist read an electrocardiogram (ECG) on a patient last
v
e
MD. Used with permission.) week and billed or the ormal ECG report, but has never seen
C
a
the patient be ore, and no one in their group has ever provided
r
e
an E/M service to this patient. Subsequently, the cardiologist
B
evaluates this patient who is being observed on the internal
i
l
lowest two levels o consult service or the old code set [99251-
l
i
medicine service. The cardiologist would bill their initial consult
n
99252], which explains the rationale or this practice. Keep in mind
g
using the new outpatient codes [99201-99205].
that the ability to submit an initial consult service with a subsequent
inpatient visit code does not hold true or admissions. For an initial I a ollow-up consultant visit is necessary, even on discharge
inpatient visit (ie, admission) reported by the primary attending day, all o these patients are now established patients, and this
o record, i the documentation or this service does not meet the work would be billed using the established outpatient visit codes
minimum criteria or the lowest level o service [99221], subsequent [99212-99215].
inpatient visit codes cannot be submitted or this work. This work
would be reported with E/M code 99499 which means that the
documentation or the service does not meet criteria or recognized ■ EMERGENCY DEPARTMENT PATIENTS
E/M services. This would all under manual review and be le t up to For consults on patients seen in the emergency department, the
the Medicare contractor to determine what i any payment would patient’s ultimate disposition will determine which set o codes to use
be given. Check with local Medicare intermediaries or deviation to bill or the consult.
rom this rule. At least one carrier, Noridian, as o the time o publica- Admission to an inpatient service. Only one CPT code will be
tion o this chapter, will allow use o subsequent inpatient visit codes submitted or an evaluation by a consultant in the emergency
[99231-99232] or this admission scenario. department, ollowed by admission to the inpatient service, an
Since more than one physician will now be submitting an initial initial inpatient visit [99221-99223]. I that patient is admitted
inpatient/SNF code, the primary attending o record or the patient to a di erent subspecialty’s inpatient service, an initial inpatient
needs to attach modi er –AI (A “eye,” not A “one” or A “el”) to visit code [99221-99223] is still reported or this work.
their initial inpatient/SNF service. Consultants should not use this Placement into observation status. I the above patient is eventu-
modi er. ally placed into observation status on the consultant’s service,
only one CPT code will be submitted or both the work o the
■ OUTPATIENTS consult and the placement into observation, an initial obser-
Occasionally, hospitalists are consulted on patients in an outpatient vation visit code [99218-99220, 99234-99236]. I the patient
unit o the hospital, such as the ambulatory surgery center. Also, is placed into observation status on a di erent subspecialty’s
observation status patients are considered outpatients even though service, this work will now be reported with an outpatient
183
visit code [99201-99215] based on whether they are “new” or new Medicare rules, only one initial inpatient visit [99221-99223]
“established” to the consultant’s billing group and subspecialty. can be reported and reimbursed per patient per admission, thus
Discharged home. Finally, i the emergency room patient is the reimbursement or this work is substantially lower than i it were
eventually discharged home, an emergency room visit code reimbursed through the consult codes [99251-99255].
P
A
[99281-99285] will be submitted or the work o the consult. The nal criterion o REPORTing recommendations speaks to
R
good medical documentation. Although a consultant may opt to
T
call the requesting physician with urgent recommendations or
I
■ PAYORS WHO STILL RECOGNIZE THE OLD good quality patient care, these recommendations need to exist in
CONSULT CODES writing, accessible by the requesting physician. In the inpatient set-
For individual state Medicaid programs and or private insurers, ting, the consultant note entered into a shared medical record will
check with plan representatives to see i they ollow the new Medi- satis y this criterion.
T
h
care rules, or i they still pay or the consult codes [99241-99255]. For
e
those that still recognize consult codes, check with the individual ■ INPATIENTS
S
p
payors as well as your compliance o ce to see i they ollow the Initial consult visits to inpatients in both acute care settings as well
e
old Medicare consult situational and documentation guidelines or i
c
as SNFs are billed using inpatient consult codes [99251-99255]. All
i
a
they have their own internal set o requirements.
l
subsequent visits are billed using inpatient/SNF ollow-up codes
t
y
In order to bill or a consult based on the old Medicare guidelines, respectively [99231-99233, 99307-99310].
o
several criteria have to be met. The Mnemonic “The 3 R’s” re er to
f
H
these criteria. The consult has to be REQUESTED by another pro- ■ OUTPATIENTS
o
s
vider who wants the consultant to RENDER their opinion about (a)
p
Consults to patients seen in an outpatient area o the hospital,
i
speci c question(s), and the ensuing recommendations must be
t
a
including observation status patients, are billed with the outpa-
l
REPORTed back in writing to the requesting provider. The essence
M
tient consult codes [99241-99245]. I a ollow-up consultation visit
o a consult in this system, is that another provider seeks the opinion
e
is needed, it would be billed using the established outpatient
d
o a consultant with expertise above and beyond his own or a par-
i
visit codes [99212-99215]. Check with individual state Medicaid
c
ticular problem or set o problems. The concept o a “REQUEST or
i
n
programs and private payors to see i they require the subsequent
e
opinion” is stressed or three reasons (1) to distinguish it rom a sel -
observation visit codes [99224-99226] instead, or observation status
a
re erral by the patient, which cannot be reimbursed as a consult (2)
n
patients.
d
to distinguish it rom a trans er o care, and (3) to prevent improper
S
payment or “protocol” care that may not have adequate medical
y
■ EMERGENCY DEPARTMENT PATIENTS
s
necessity documented in the chart. For example, i a hospital man-
t
e
Just as or Medicare patients, the physical location o the patient is
m
dates that 100% o patients admitted or surgery or to a psychiatry
not what determines the selection o codes to use or billing con-
s
ward must have a medicine consult, this could lead to nonbillable
o
work. In these situations, a preprinted admission orm request- sults. It is the patient’s “status.”
f
C
ing a medicine consult per “hospital policy” without documented Admission to an inpatient service. For patients located in the
a
r
medical necessity or each individual patient situation, would lead emergency department who have already been admitted to
e
to nonbillable work. I a 25-year-old healthy emale is admitted to an inpatient service, use the inpatient consult codes [99251-
general surgery or complications a ter an elective cholecystectomy, 99255]. I the consultant admits the patient to their inpatient
but has no other medical issues, there is no medical necessity or service on the same date as the consult, only one CPT code
a medicine consultant to see the patient. The surgeon does not or the work o both the consult and the admission should be
consciously request this service or have a particular question or the submitted—an initial inpatient admission code [99221-99223].
medicine consultant to answer, and does not need the internist’s Observation status. I the patient in the emergency department is
expertise to manage the postoperative issues. A hospital mandated assigned observation status on a di erent service rom the con-
task is there ore not reimbursable through the patient’s insurance. sultant’s, the consult will be billed using the outpatient consult
Because the consult codes reimburse higher level per level, than codes [99241-99245]. I the consultant places the patient into
their subsequent ollow-up counterparts, there is a big push to observation status on his own service on the same calendar date,
ensure that an opinion was requested, and not a trans er o care. only one CPT code should be submitted or all o this work—an
This is the essence o a consult or those insurers that still ollow initial observation visit [99218-99220, 99234-99236].
these guidelines. For example, medical consultation request or Discharge to home. I the patient is only under the care o an
an opinion regarding selection o anti-hypertensive medication emergency room physician, and is eventually discharged home,
in a pregnant inpatient would meet criteria or allowing a consult report this work with outpatient consult codes [99241-99245].
[99251-99255] to be billed, even i the recommendations include
continued concurrent care o the hypertension by the internist.
■ CRITICAL ILLNESS
On the other hand, i the obstetrician is not com ortable managing
hypertension at all, and wants the care o that problem to be totally I the patient is critically ill during the initial consult visit and
managed by the internist, the essence o the request or help is no >30 minutes o critical care is provided by the consultant, only
longer or an “opinion” about management. The request is or the critical care [99291, 99292] and no other consult code set would be
internal medicine consultant to take over the care o the hyperten- reported or that day. In the event that a consult is completed and
sion while the obstetrician continues to manage the other aspects at a later time on that calendar date, the consultant provides >30
o the patient’s care is re erred to as a “trans er o care.” I the patient minutes o critical care to the patient, both the consult and critical
remains on the obstetrician’s service and the internist provides con- care service can be paid. The consult work would be reported using
current care, managing hypertension, this work cannot be reported the rules above to select the appropriate code set.
with the “trans er o the patient to another service” consult codes
[99251-99255]. In this concurrent care situation, the internal medi- ■ SPLIT/SHARED VISITS
cine consultant would report this work with a subsequent inpatient Although previous guidance rom Medicare has stated that a con-
visit code [99231-99233]. For those payors who do not ollow the sult cannot be provided as a split/shared service, those rules only
184
applied to consults reported with CPT codes 99241-99255. With the billing provider’s time, not time spent on patient care by the
Medicare consults now reported using di erent code sets, the resident, student, social worker, or nursing sta . For hospital based
C
rules or split/shared visits or Medicare patients now de ault to the services, in addition to ace-to- ace time, the majority o E/M services
H
rules or whichever set o codes are being reported or the consult. also allow inclusion o time spent by the provider on the patient’s
A
P
Thus, or inpatient consults reported with initial inpatient visit codes f oor or in their unit (“unit/f oor time”) providing patient care, toward
T
[99221-99223], these services can be provided and billed as split/ billable time. In general, time is added up rom 12:01 a m to 11:59 pm,
E
shared visits. Check with individual nonmedicare payors or their not over a 24-hour period, and does not need to be continuous. One
R
rules regarding this situation. exception to this is in the event that a time-based service is started
2
be ore midnight, continuously provided, but not completed until
8
a ter midnight. The CPT handbook advises i this should occur, to
COMANAGEMENT
add up time that the service was provided continuously, and to bill
C
The terms “comanagement” and “concurrent care” are o ten used or it on the date the service began.
o
n
interchangeably. Concurrent care is the situation when two phy-
s
u
sicians are managing di erent aspects o the patient’s care on the
l
t
same calendar date on a more extensive basis than a one-time ■ COUNSELING AND COORDINATION OF CARE
a
t
i
consultation, usually discussed in re erence to two physicians
o
When counseling and/or coordinating care (CCC) is the dominant
n
submitting claims or inpatient ollow-up services [99231-99233]. eature o certain E/M visits, that is, when more than 50% o the total
,
C
Two initial inpatient services [99221-99223] submitted on the visit time ( ace to ace and/or unit/f oor), is spent counseling the
o
m
same calendar date will be interpreted as consultation care and patient and or coordinating their care, the level o service provided
a
not all under this discussion. The Medicare manual clearly states may be determined by the total visit time rather than by quanti -
n
a
that the work o both physicians is reimbursable as long as the cation o the documented history, physical, and medical decision
g
documentation ref ects the medical necessity or each physician
e
making (MDM). The typical total visit times associated with these
m
to provide their service. Although the two physicians are usually services are shown below.
e
n
rom di erent subspecialties, in the event that both are rom the For example, i a patient is seen or inpatient ollow up o a
t
,
same specialty but one has documented expertise above and pulmonary embolism (PE) and a compliant level 2 [99232] visit is
T
i
beyond the other, both services may be paid. An example o
m
documented based on history, physical, and MDM, but it is also
e
this could be an internal medicine hospitalist who is managing a documented that >50% o a 35-minute ace-to- ace visit is spent
-
B
patient with a COPD exacerbation, but elicits the help o another counseling the patient on risk actors or a PE, evaluation, and treat-
a
s
internal medicine hospitalist with expertise in pain management, ment, time is now the controlling actor and a level 3 [99233] visit
e
d
to take over the daily care o the patient’s severe re ractory chest can be billed even though the history, physical, and MDM only
,
a
pain rom lung cancer. The medical necessity o each physician’s amount to a level 2. Medicare will only allow time-based billing or
n
d
care should not only be ref ected in the medical record, but also by CCC or inpatient admissions and ollow-ups [99221-99233], SNF
P
each claim being submitted with di erent ICD-10 codes, showing admissions and ollow-ups [99304-99318], new outpatient visits
a
l
the di erent aspects o care each provider is managing on that
l
[99201-99205], established outpatient visits [99212-99215], and
i
a
particular date. I both providers submit their service with the same
t
other home services [99324-99350]. Medicaid and other insurers
i
v
primary ICD-10 code, only the rst claim received by Medicare will
e
who ollow the CPT handbook may also allow time-based bill-
C
be reimbursed. The second one will be denied. ing or CCC or observation services [99218-99236] and consults
a
Comanagement, de ned by the Society o Hospital Medicine
r
[99241-99255] as well.
e
(SHM) and its Advisory Panel, is the “shared responsibility, author-
B
The Medicare Claims Processing Manual Chapter 12 states that
i
l
ity, and accountability or the care o a hospitalized patient.” By
l
the amount o CCC time may be estimated. In addition, the CPT
i
n
convention, this term is used when a more ormal arrangement is manual instructs the provider to round the total visit time to the
g
made between two physicians providing concurrent care. In the closest “average” total visit time. Thus in the example above, i 32
common situation involving a surgeon and medicine provider, the minutes were spent with the patient, 32 is closer to 35 minutes
surgeon manages surgery related issues and the medicine hospital- (average visit time or a level 3 ollow-up [99233]) than it is to 25
ist manages the patient’s medical conditions on an ongoing basis, minutes (average visit time or a level 2 ollow-up [ 99232]), and a
o ten or the entire duration o the admission, and usually with level 3 service would be billed. This di ers rom the rules governing
the ormal arrangement outlining each specialty’s responsibilities. prolonged service total time determination below.
From the start, the patient will be admitted to one service, with an Tasks that count toward counseling the patient include but
order or the other subspecialty physician to see the patient. Just are not limited to discussions o the plan, evaluation, procedures,
as in the example in the consult section above, the documenta- prognosis, treatment options, risk actor reduction, and patient and
tion must support the medical necessity or the second physician amily education.
to be involved, not just or the initial visit, but also or each daily In order to bill or these services, two dif erent amounts o time
visit. Situations will arise where each physician in a comanagement must be documented: the total visit time, and the portion o that
arrangement may not need to see the patient every single day o total visit time that was spent CCC. Check with insurers and local
the admission. No billing should be submitted or these days, even i compliance o ces to see i the term “>50% o the total visit time was
the patient is physically seen by the physician, based on a mandated spent counseling the patient” is acceptable or i the speci c number
hospital protocol. Without documented medical necessity, this work o minutes spent CCC needs to be explicitly documented. Neither
is not reimbursable. the Medicare manual nor the CPT handbook list this as a documen-
tation requirement. The most common documentation pit all in
TIME-BASED BILLING success ully getting reimbursed or CCC time, is reporting only one
Several E/M services can be billed based on time (Table 28-1), with time amount, thus making it impossible to determine that >50% o
di erent billing nuances or each code set. In certain situations, the the total visit time was spent CCC. In addition, lack o a brie descrip-
rules allowing reimbursement or these services di er between the tion o what was discussed during that time will cause a denial; the
CPT and Medicare manuals. When billing based on time, the only medical necessity or investing that time into the visit may not be
time that can be counted toward reimbursement o a service, is obvious.
185
TABLE 28-1 Average Visit Times or E/M Services Billed as Counseling/Coordination Services and Threshold Times
or Prolonged Services
P
Medicare allows billing or counseling/coordination time and prolonged service only or services in bold below. The CPTmanual allows both or all
A
services in the table.
R
Threshold to (Minutes) Average Threshold to
T
CPT Code Visit Type Level Visit Time Bill 99356 Bill 99357
I
99218 105 Initial observation (1) 30 60
99219 (2) 50 80 125
99220 (3) 70 100 145
T
h
99221 Initial inpatient (1) 30 60 105
e
99222 (2) 50 80 125
S
p
99223 (3) 70 100 145
e
c
i
99224 Observation (1) 15 45 90
a
l
ollow-up
t
y
o
99225 (2) 25 55 100
f
H
99226 (3) 35 65 110
o
99231 Inpatient (1) 15 45 90
s
p
ollow-up
i
t
a
99232 (2) 25 55 100
l
M
99233 (3) 35 65 110
e
d
99234 Observation same (1) 40 70 115
i
c
i
day discharge
n
e
99235 (2) 50 80 125
a
n
99236 (3) 55 85 130
d
99251 Inpatient consult (1) 20 50 95
S
y
s
99252 (2) 40 70 115
t
e
m
99253 (3) 55 85 130
s
99254 (4) 80 110 155
o
f
99255 (5) 110 140 185
C
a
99304 Initial SNF (1) 25 55 100
r
e
99305 (2) 35 65 110
99306 (3) 45 75 120
99307 SNF ollow-up (1) 10 40 85
99308 (2) 15 45 90
99309 (3) 25 55 100
99310 (4) 35 65 110
99318 Annual NF 30 60 105
assessment
(Minutes)
Threshold to Threshold to
CPT Code Visit Type Level Average Visit Time Bill 99354 Bill 99355
99201 New outpatient (1) 10 40 85
99202 (2) 20 50 95
99203 (3) 30 60 105
99204 (4) 45 75 120
99205 (5) 60 90 135
99212 Established (2) 10 40 85
outpatient
99213 (3) 15 45 90
99214 (4) 25 55 100
99215 (5) 40 70 115
99241 Outpatient consult (1) 10 40 85
99242 (2) 20 50 95
99243 (3) 30 60 105
99244 (4) 45 75 120
99245 (5) 60 90 135
CPT codes 99324-99350 not included in the chart above as not used in hospital medicine.
186
Examples o minimally acceptable reporting o time or billing the There are several di erences between Medicare and CPT manu-
visit based on CCC: als’ rules governing how and when prolonged services may be
C
provided and how they must be documented. Medicare only allows
H
1. I spent 20 minutes out o a 25-minute visit ace-to- ace and
A
unit/f oor time, counseling the patient on the sa e use o home prolonged services to be billed when the primary E/M visit is an
P
oxygen, and coordinating care with social work, setting up inpatient admission or ollow-up [99221-99233], SNF admission or
T
home oxygen, home nursing, and pulmonary rehab (level 2 ollow-up [99304-99318], new outpatient visit [99201-99205], estab-
E
lished outpatient visit [99212-99215], or home or domiciliary visit
R
inpatient ollow-up visit [99232]).
2. I spent 35 minutes ace to ace with the patient today. Twenty- [99324-99350]. Medicare does not allow prolonged services to be
2
billed with observation admissions and ollow-ups [99218-99236],
8
ve o those minutes were spent counseling the patient on her
new diagnosis o breast cancer, the next steps in work-up, and however, the CPT manual does allow prolonged service billing
possible treatment options including chemotherapy and XRT with these observation services as well as with the consults codes
C
[99241-99255]. For all payors including Medicare, prolonged services
o
(level 3 inpatient ollow-up visit [99233]).
n
can never be billed with discharge management services [99238-
s
3. I spent 75 minutes with Mr X ace-to- ace and unit/f oor time.
u
99239, 99217], critical care [99291-99292], or procedures. In general,
l
More than ty percent o that time was spent counseling the
t
a
patient. The patient has very low health literacy and needed the direct patient contact services 99356-99357 are added on to
t
i
o
extensive counseling regarding acute renal ailure, necessary inpatient and observation services, and 99354-99355 are added
n
on to outpatient services (Table 28-1). 99358-99359 are prolonged
,
inpatient work-up, and possible need or dialysis (level 3 inpa-
C
services without direct patient contact, that is, without a ace to ace,
o
tient admission [99223]).
m
4. I spent 40 minutes ace to ace with the patient. One hundred and are not reimbursed by Medicare. These codes are usually used
a
when extensive time is spent in chart review. Payment or these
n
percent o the time was spent counseling the patient on hos-
a
services varies among private insurers.
g
pice and the grave prognosis or his metastatic lung cancer
e
(level 3 inpatient ollow-up visit [99233]). 99354, 99356, and 99358 are billed or the “ rst hour” o pro-
m
longed service provided. These services can be billed once 30 min-
e
Examples o unacceptable reporting o CCC time or billing:
n
utes o prolonged service has been provided; the rules governing
t
,
1. I spent 25 minutes counseling the patient on new diagnosis o prolonged service care allow or each unit o service to be billed
T
i
lung cancer.
m
once 50% o the service time in the description has been met.
e
2. I spent 15 minutes ace to ace with the patient discussing the 99355, 99357, and 99359 are billed or each additional “hal hour”
-
B
hospital course or his pneumonia. o prolonged service a ter the rst hour o prolonged care. These
a
s
3. I spent 10 minutes out o a 30-minute visit counseling the services cannot be billed unless 99354, 99356, or 99358 has already
e
d
patient on what CHF is, low salt diet, and decreased f uid intake been billed with the primary E/M. Again, once hal o the time in the
,
a
(10 minutes is less than 50% o the total visit time). This visit description is met, that is, 15 minutes, these services can be billed.
n
d
cannot be billed based on time; the documentation o his- Thus, in the inpatient setting, i 105 minutes o prolonged service
P
tory, physical, and MDM would be used to determine the level has been provided beyond the typical visit time or the primary
a
l
o service. Speci cally, even though 10 minutes is >50% o a
l
E/M visit, the rst hour o prolonged care may be billed, PLUS the
i
a
t
level 1 inpatient ollow-up average visit time (15 minutes), this next hal hour o care, PLUS another hal hour o care. In addition to
i
v
statement cannot be used in lieu o documenting the history,
e
the primary E/M service, 99356 and 99357 × 2 would be submitted
C
physical, and MDM to bill or the service. in this example. No modi ers are needed or billing these services.
a
r
4. I spent today’s entire visit counseling the patient on the di - Medicare requires the service to “meet or exceed” the threshold
e
erential diagnosis o his diarrhea and the planned work-up.
B
time to allow reimbursement or prolonged services (Table 28-1).
i
l
(An auditor cannot try to determine how much time you spent
l
Note that this is di erent rom the instruction to “round” to the near-
i
n
with the patient based on the time orders were written, or the
g
est average visit time when billing or CCC. Thus i a level 2 inpatient
note was entered in the EMR. This service cannot be billed visit is per ormed (typical visit time 25 minutes) and 52 minutes o
based on time.) ace-to- ace care was provided, the 55-minute threshold needed to
bill or the rst hour o prolonged service [99356] has not been met,
and the additional 27 minutes o care provided in this example is not
■ PROLONGED SERVICES reimbursable; only the level 2 inpatient ollow-up [99232] would be
Prolonged service codes [99354-99359] are add on codes used submitted or this care. Table 28-2 shows the cut o times needed
when more than 30 minutes o care is provided beyond the typical to bill each unit o prolonged service.
or average visit time or the E/M service. They can never be billed The rest o the discussion will ocus only on services with direct
alone, without their companion code—the primary E/M service. patient contact [99354-99357]. Medicare only allows ace-to- ace
Time can be added up over the course o the calendar date not time with the patient to be counted toward prolonged service
only by the provider, but also by their covering partners, including billing, both in the outpatient as well as the inpatient setting
involvement by a nocturnist; the time each provider spent over the [99354-99357]. Notice that this is di erent rom Medicare’s rules
calendar date can be aggregated to determine to level o service or CCC, discharge management services, and critical care which all
provided. In this situation, the prolonged service would be reported allow unit/f oor time to be included as well. The CPT manual de nes
under the NPI o the provider who is billing or the primary E/M ser- “direct patient contact” to mean only ace-to- ace time in the outpa-
vice. Relative value units (RVUs) between the two providers would tient setting [99354-99355] but also time spent on the patient’s f oor
need to be adjusted internally or productivity purposes. Both physi- or in their unit or inpatient services [99356-99357]. This has a large
cians and nonphysician providers (NPPs) can bill or prolonged ser- impact on how prolonged services may be billed.
vices. Finally, prolonged services may be provided and billed or as Prolonged services may be provided in addition to any level o
a split/shared visit. No restrictions are identi ed in either the CPT or service within the code sets above as long as >50% o the prolonged
Medicare manuals on how much time the physician must spend on service time was not spent counseling the patient or coordinating
the service relative to the NPP, in order to bill a split/shared visit. The their care. In the situation where CCC is the key eature o the visit,
documentation must clearly ref ect that each provider per ormed a the highest level o service in the E/M code set must be used as the
substantive portion o the service. primary E/M be ore determining i any prolonged service time is
187
TABLE 28-2 Prolonged Service Billing
CPT Codes Submitted

P
Minutes of Prolonged Service Outpatient Inpatient and Observation No Direct Patient Care
A
0-29 Only primary E/M Only primary E/M Only primary E/M
R
T
30-74 99354 99356 99358
I
75-104 99354 + 99355 99356 + 99357 99358 + 99359
105-134 99354 + 99355 × 2 99356 + 99357 × 2 99358 + 99359 × 2
135-164 99354 + 99355 × 3 99356 + 99357 × 3 99358 + 99359 × 3
165-194 99354 + 99355 × 4 99356 + 99357 × 4 99358 + 99359 × 4
T
h
195+ - - - - - - - - - - - - - - - - - - - ollow the pattern above - - - - - - - - - - - - - - - - - - -
e
S
p
e
c
i
a
l
t
y
billable. I less than 30 minutes o prolonged services are provided, ■ CRITICAL CARE
o
that work is not separately reimbursable.
f
Critical care is de ned in the Medicare Claims Processing Manual,
H
Example 1: A patient is seen with a new diagnosis o colon can- Chapter 12, Section 30.6.12, as a physician’s direct delivery o medi-
o
cer and a ully compliant level 2 inpatient ollow-up visit [99232] is
s
cal care o a high-complexity MDM, to a critically ill or injured patient
p
documented, based on history, physical, and MDM. Seventy- ve
i
t
to prevent or stop imminent or active acute organ ailure. A critical
a
minutes o ace-to- ace plus unit/f oor time care (total visit time) is
l
illness or injury acutely impairs one or more vital organ systems so
M
provided to the patient. Only 60 minutes o that time was ace-to- that there is high probability o imminent or li e threatening dete-
e
d
ace time, and >50% o the total visit time was spent counseling the rioration in the patient’s condition. Some examples o critical illness
i
c
patient regarding the new diagnosis, urther necessary testing, treat-
i
rom the Medicare manual include circulatory ailure, shock; renal,
n
ment options, and plans or consultants. Although only a level 2 visit
e
hepatic, or respiratory ailure. Chronic organ ailure management
a
is documented, since CCC is the key eature o the service, the visit does not meet criteria or critical care services. Examples o this
n
d
can be billed based on time. The highest level o inpatient ollow-up include chronic vent management and routine hemodialysis.
S
service is a level 3 [99233] which is associated with an average visit For a service to quali y as critical care, all our criteria must be met
y
s
time o 35 minutes. In Medicare patients, only ace-to- ace time can
t
and adequately documented:
e
be used when determining i prolonged service time is separately
m
1. The medical necessity re ers to a statement about the patient’s
s
reimbursable. Only 60 minutes o the total visit time was ace to
o
ace. This means that 25 minutes o prolonged service care was pro- illness. The Medicare manual states “…The ailure to initiate
f
these interventions on an urgent basis would likely result in sudden,
C
vided (60 minutes total ace-to- ace time minus 35 minutes typical
a
clinically signif cant or li e threatening deterioration in the patient’s
r
visit time or a level 3 ollow-up). In this scenario, those 25 minutes
e
are not separately reimbursable. This service or a Medicare patient condition….”
would be reported as a level 3 inpatient ollow-up [99233]. 2. The service meets criteria or high-complexity medical decision
Reimbursement would be di erent i this patient does not have making, and the high-complexity MDM is adequately docu-
Medicare, and the insurer ollows the CPT manual. The total visit mented in the chart.
time would now be 75 minutes, adding in time spent providing care 3. The care is directly urnished by the physician, not by the resi-
on the patient’s f oor or in their unit. Now in addition to the level 3 dent, and not by a NPP in a split/shared situation. Critical care
inpatient ollow-up visit, there are 40 minutes o prolonged service cannot be per ormed and billed as a split/shared visit.
(75 minutes direct patient care time minus 35 minutes typical visit 4. The patient has a critical condition as de ned above. Since
time or a level 3 inpatient ollow-up). This service would now be critical care encompasses not only the “treatment o organ
reported as a level 3 inpatient ollow-up [99233] PLUS the rst hour ailure” but also the “prevention o urther deterioration in the
o prolonged service [99356] which would result on average in ~$95 patient’s condition,” it is not a requirement that the person
additional income. have an emergency or crisis situation to bill or critical care. The
Example 2: A non-English-speaking patient with an asthma physician’s documented clinical judgment that the patient is at
exacerbation and lots o anxiety continues to call nursing several high risk o impending organ ailure, or urther deterioration in
times during the day with shortness o breath, requiring repeated clinical condition, will support this criteria.
history taking, physical exam, and reassessment a ter each albuterol Seeing a critically ill patient does not automatically allow critical
nebulizer treatment. In addition, each interaction with the patient care billing. For example, an ophthalmologist seeing a septic patient
requires the assistance o an interpreter. A compliant level 2 inpa- on pressors, or glaucoma treatment, cannot bill or critical care.
tient ollow-up visit [99232] is documented. Sixty minutes o ace- There is no restriction to where critical care can be provided. It
to- ace care is provided intermittently over the course o the day. is not uncommon to provide critical care to a patient in a regular
The typical visit time or 99232 is 25 minutes. Thirty- ve minutes o medical ward with an acute change in clinical status, while awaiting
prolonged service that is not dominated by counseling or coordina- trans er to an intensive care unit. A patient is in the intensive care
tion o care is best ref ected by 99232 PLUS 99356. unit does not automatically allow or a critical care service to be
For Medicare, start and stop times or prolonged services and the billed. For example, a patient admitted to the ICU or acute respira-
indication or the prolonged time need to be documented along tory ailure 7 days ago, unable to wean o the ventilator, but stable
with a brie summary o what was done during that time in addition or 3 days on their current vent settings, awaiting tracheostomy, no
to the total ace-to- ace time providing care on that calendar date. longer meets critical illness criteria. Neither does a patient who is in
The CPT handbook does not include a requirement or document- the ICU due to increased nursing needs such as a patient in diabetic
ing start and stop times to bill or this service. There may be variable ketoacidosis who needs glucose nger checks every hour, and the
requirements among state Medicaid and private payors. high nurse to patient ratio on the general ward, makes this level o
188
care impossible or unsa e. Critical care billing is not warranted on a
patient in an ICU bed only because the hospital policy requires it or TABLE 28-4 Procedures Bundled into Critical Care
C
their treatment, such as an insulin drip.
H
Procedure CPT
Critical care is a time-based service that ollows the general rules
A
IVplacement 36000
P
o time-based coding. Both ace-to- ace and unit/f oor time can be
T
counted toward billable critical care time as long as the physician Arterial blood gas 36600
E
is in close proximity to the patient to immediately intervene or the Blood draws, physician skill needed 36410, 36415, 36591
R
minutes o reported critical care. I the physician is reviewing lab data NG tube placement (+ luoro, +lavage) 43752, 91105
2
on the patient’s f oor but the patient is on a di erent f oor getting
8
Chest x-ray, pro essional component 71010-26, 71015-26,
a test done, in this scenario, the unit/f oor time cannot be counted 71020-26
toward critical care time. For time reported as critical care time, the Temporary transcutaneous pacing 92953
C
physician must provide their entire attention to management o the
o
Interpretation o cardiac output indices 93561, 93562
n
critically ill patient. For example, i the physician is at the critically ill
s
Ventilator management 94002-94004, 94660,
u
patient’s bedside while an ECG is being per ormed, and while wait-
l
t
94662
a
ing, he is multitasking and looking up labs on another patient, this
t
i
Pulse-oximetry 94760, 94761, 94762
o
time cannot be counted toward critical care time. A minimum o
n
30 minutes o critical care must be provided to bill or critical care Data analysis rom a computer 99090
,
C
services. Twenty nine minutes or less o critical care would be billed
o
m
with an E/M code that best re lects the service provided, such
a
as an admission code [99221-99223] or inpatient ollow-up code
n
a
[99231-99233]. Once 30 minutes o critical care is provided, the rst Certain procedures when per ormed on the same date as a criti-
g
cal care service and by the same provider, are bundled into the
e
hour o critical care [99291] can be billed. Each additional hal hour
m
o critical care [99292] can be billed or therea ter (Table 28-3). reimbursement or critical care and cannot be billed or separately.
e
n
In addition to time spent ace to ace with the patient taking a (Table 28-4). Any medically necessary procedure not included in
t
,
history and per orming a physical exam, other activities that count this table can be billed separately rom critical care, such as central
T
i
line placement, cardiopulmonary resuscitation, and intubation to
m
toward critical care time include the ollowing:
e
name a ew. The documentation should clearly state that the time
-
1. Time spent reviewing test results while on the patient’s f oor or
B
per orming these unbundled procedures is not included in the
a
in their unit, and does not all into the pit all situation above.
s
reported critical care time.
e
Time spent personally viewing a chest x-ray in the radiology
d
When critical care is provided by the same physician/billing group
,
suite on a di erent f oor or looking at a peripheral smear in the
a
on the same date that another E/M service such as an inpatient
n
hematology lab cannot be counted toward critical care time.
d
admission, ollow-up, or consult was provided earlier in the day
2. Time spent discussing the care plan with nursing sta or other
P
when the patient was not critically ill, both the earlier E/M and the
a
consultants.
l
critical care service can be reimbursed. The exception to this rule
l
i
3. Discussions with amily members or surrogates when both o
a
is when the other E/M service is an emergency room visit [99281-
t
i
these criteria are met (a) the patient lacks capacity or unable
v
99285]. An emergency room service cannot be paid on the same
e
to participate in their own care and (b) the discussion directly
C
date as a critical care service by the same physician/billing group.
a
impacts decision making. These discussions will count toward
In a coverage situation, such as a change in shi ts where a partner
r
e
critical care time even i they occur via phone on the patient’s
in the same billing group continues to provide critical care on the
B
f oor or in their unit. However, time spent on routine updates
i
l
same calendar date, critical care time can be aggregated as long as
l
i
to amily cannot be counted toward critical care services, and
n
the initial physician provided at least 30 minutes o critical care. For
g
is not separately reimbursable.
example, i Dr A provides 40 minutes o critical care at 6 pm and then
4. Time spent per orming procedures that are bundled into criti-
Dr B takes over and provides an additional 35 minutes o critical
cal care (Table 28-4).
care at 7 pm, the critical care time can be aggregated. A total o 75
Activities that cannot count toward billable critical care time minutes o critical care was provided; the rst hour o critical care
include work that does not directly contribute to the patient’s care, [99291] would be submitted by Dr A and an additional hal hour o
even i per ormed at the patient’s bedside: critical care [99292] would be submitted by Dr B. However, i Dr A
1. Time teaching had only provided 25 minutes o critical care and Dr B provided 50
2. Time looking up literature minutes, although they still provided a total o 75 minutes o criti-
3. Family updates, even i at the request o the patient cal care, Dr A did not provide at least 30 minutes be ore going o
4. Time per orming separately reimbursed procedures shi t. They cannot aggregate their critical care time. Dr A would bill
an appropriate E/M code that best represents their work such as an
inpatient admission or ollow-up, and Dr B would bill or the rst
hour o critical care [99291].
TABLE 28-3 Critical Care Billing
Minutes of Critical Care CPT Codes Submitted ■ CONCURRENT CARE

0-29 Submit appropriate E/M Two di erent physicians rom di erent specialties can both receive
30-74 99291 payment or critical care services on the same calendar date i there
is medical necessity or each o them to provide this level o care;
75-104 99291 + 99292
however, two physicians cannot get paid or the same exact same
105-134 99291 + 99292 × 2 minutes o critical care. For example, i Dr A, an internal medicine
135-164 99291 + 99292 × 3 hospitalist, provides 30 minutes o critical care rom 9 to 10 a m
165-194 99291 + 99292 × 4 stabilizing a patient with impending respiratory ailure, and Dr B,
195+ - - - - ollow the pattern above - - - - the intensive care physician, provides critical care between 10
and 11 a m, each physician would submit a claim or the rst hour o
189
critical care [99291]. In the event that both physicians provide care reimbursement or the higher level o care [99239]. No time amount
at the same time, only one physician will get paid or critical care. needs to be documented when billing the lower level o care
The other physician should submit an E/M code that best represents [99238]. I extensive amounts o time over 30 minutes are spent on
their work. discharge management, there is no additional reimbursement or
P
A
Having di erent subspecialty designations does not automati- this care. Prolonged services cannot be billed with any discharge
R
cally allow each physician to submit a claim or the rst hour o management code. There is only one level o discharge manage-
T
critical care. In the event that a member cross covering a patient ment or observation services regardless o the amount o time
I
happens to have a di erent subspecialty designation, but their role spent on care.
is to continue the care started by the previous physician, the work o One o the most common documentation errors preventing
the two physicians would be aggregated. For example, in a hospital reimbursement or the higher level o inpatient discharge is the
medicine group, i Dr A is an internist and provides critical care or statement “I spent 30 minutes on discharge management.” As
T
h
30 minutes and then their partner Dr B, a cardiologist, comes on to opposed to other time-based services where 30 minutes is the mini-
e
cover the night shi t and provides an additional 30 minutes o criti- mum amount needed to bill a unit o service (ie, prolonged service,
S
p
cal care, even i the patient’s critical illness happened to be a cardiac critical care), more than 30 minutes must be spent in discharge
e
problem, in this situation, the two physicians would aggregate their
c
management to allow reimbursement or the 99239. This author
i
a
work. A total o 60 minutes o critical care was provided, which is suggests documenting the actual amount o time spent, in minutes.
l
t
y
only enough to bill or the rst hour o critical care. Only 99291 can
o
be submitted in this scenario.
f
PALLIATIVE CARE
H
When providing critical care in tandem with a resident, the only
o
Although palliative care has had its own subspecialty designation
s
time that the physician can count toward critical care billing, is the
p
(17) since October 2009, many internal medicine hospitalists (sub-
i
time they personally provide critical care, or time they are physically
t
specialty 11) have developed expertise in this eld without ormally
a
l
present while the resident provides critical care. For documentation,
changing their subspecialty designation rom 11 to 17. When unc-
M
the resident’s note can help support the high complexity o MDM.
e
tioning as a consultant providing concurrent care to patients under
d
In addition to meeting all teaching physician (TP) presence and
i
the care o a colleague in the same subspecialty, there are several
c
documentation criteria, and writing a valid attestation statement
i
n
“best practices” to be aware o to minimize claims denials.
e
linking the TP’s note to the resident’s note, the TP must personally
a
document the ollowing to bill or critical care.
n
■ SUBSPECIALTY DESIGNATION
d
1. The patient was critically ill at the time the TP was physically
S
When providing concurrent care, one o the rst pieces o evidence
y
present and provided the service.
s
that medical necessity exists or two physicians to manage a patient
t
2. What the critical illness is. ICD-10 codes or a critical illness must
e
on the same calendar date, is di erent subspecialty designations. A
m
be used, or the claim will be at high risk o initial denial.
physician can go into the Medicare Provider Enrollment, Chain and
s
3. The nature o the treatment and management they personally
o
Ownership System (PECOS) at any time and change their primary
f
provided.
C
and secondary subspecialty designations. The physician should
4. The number o minutes o critical care that they personally
a
have evidence o expertise in their primary subspecialty designation;
r
e
provided.
this does not have to be completion o an accredited ellowship
Documentation o the above by the resident is insu cient or bill- program in that eld. Recognition at the local, regional, or national
ing. An example rom the Medicare manual o minimally acceptable level o expertise via activities such as publications, public speaking
TP documentation or critical care: “Patient developed hypotension engagements, or development o institutional protocols or cur-
and hypoxia; I spent 45 minutes while the patient was in this con- ricula, would satis y this criteria. Documentation o constant and
dition, providing f uids, pressor drugs, and oxygen. I reviewed the updated Continued Medical Education (CME) credit in that eld
resident’s documentation and I agree with the resident’s assessment would also count. Many physicians list their primary subspecialty
and plan o care.” designation as the eld that they provide >50% o their services
in. For internal medicine palliative care hospitalists, the percent o
■ DISCHARGE MANAGEMENT SERVICES services provided as an internist may o ten outweigh that o their
Discharge management services can only be billed by the primary palliative care services; symptom-driven ICD-10 coding, complete
attending o record, or a partner covering that day. All other con- descriptive documentation, and vigilant claims tracking are key to
sultants should bill an inpatient ollow-up visit [99231-99233] or an ensure appropriate reimbursement or these services.
established outpatient visit [99212-99215] based on the patient’s
inpatient versus observation status as well as payor rules or obser- ■ SYMPTOM-DRIVEN ICD-10 CODING
vation services. For inpatient discharge management, there are two Palliative care providers are o ten consulted to assist with maximiz-
levels o service: <30 minutes [99238] or >30 minutes [99239]. Both ing quality o li e and minimizing su ering or the time a patient
ace-to- ace and unit/f oor time spent in discharge activities count, has le t. In doing so, the goals o care o ten center around symptom
including but not limited to taking a history, the nal physical exam, control and not disease management. This may di er rom the
counseling the patient and amily, reviewing data, writing prescrip- primary attending’s care plan, which may employ palliative services,
tions, and ordering ollow-up appointments. but which are still overall managing a disease state. Use o ICD-10
The only required documentation to bill or discharge manage- codes or symptoms being managed, not only better ref ects the
ment services is a statement attesting to having a ace-to- ace visit work o the palliative care provider, but also ensures that di erent
with the patient on the date o discharge management services. ICD-10 codes will be submitted rom the primary physician’s claim.
For legal, ethical, and high-quality patient care reasons, urther For example, a patient with end-stage COPD is admitted or the
documentation o the details o that visit would be prudent. When eighth time this year with a COPD exacerbation, contemplating hos-
providing the higher level o inpatient discharge management pice care, and most bothered by dyspnea and atigue. The palliative
[99239], documentation must explicitly state how much time was care provider’s documentation and plan will ocus on control o the
spent on discharge management as well as a brie summary o symptoms o dyspnea and atigue, while the primary hospitalist’s
what was done during that time. Lack o this statement will prevent note will ocus on appropriate management o COPD. Claims or
190
these services will be submitted with the ICD-10 codes or dyspnea One way to help track these services internally would be to
[R06.0] and atigue [R53.83], and COPD [J44.1] respectively. remember to submit ICD-10 code [Z51.5] or “encounter or pal-
C
liative care services” with 100% o these services. This code should
H
A
■ COMPLETE AND DESCRIPTIVE DOCUMENTATION never be the primary diagnosis. Its use will not alter the amount
P
Palliative care providers o ten spend enormous amounts o time o reimbursement or the visit; however its presence on a claim
T
counseling patients, and may bill a signi cant percent o their visits may alert an insurer that this service is separate and distinct rom
E
another E/M submitted by the same billing group, and prevent an
R
based on CCC. More detailed descriptions o the clinical situation
will better support the medical necessity to spend these larger up ront denial. At minimum, it will allow a physician to more easily
2
search or these encounters in their data base to monitor payment
8
amounts o time with the patient. Adjectives may make a huge di -
erence. A statement such as “patient extremely distraught over their or these services.
terminal condition and required intense counseling regarding goals o
C
o
end o li e care” paints a much better picture or the need or 90 min- SUGGESTED READINGS
n
s
utes o counseling than a statement like “Patient upset over terminal
u
CPT 2015: Current Procedural Terminology. Cpt/Current Procedural
l
diagnosis. I counseled the patient on goals o care.”
t
a
Terminology Pro essional Edition. American Medical Association
t
i
o
■ TRACK CLAIMS AND CLAIM DENIALS Press.
n
,
Lustbader Dr, Nelson JE, Weissman DE, et al. Physician reimburse-
C
The best way to prevent uture denials is to determine the reason
o
or denial o a medically necessary visit, and use in ormation learned ment or critical care services integrating palliative care or
m
patients who are critically Ill. Chest. 2012;141(3):787-792.
a
rom that encounter to prevent a denial or similar services. For
n
example, a private insurer who sees two inpatient ollow-up ser-
a
g
vices [99231-99233] rom the same subspecialty group on the same ONLINE RESOURCE
e
m
calendar date, even with di erent ICD-10’s, may deny it up ront
e
simply because they require modi er –25 to be attached to one o Centers or Medicare and Medicaid Services. Medicare Claims
n
t
the services. Another provider may require documentation to be
,
Processing Manual: Section 30.6.1.C; Section 30.6.10, Section
T
i
submitted up ront or all situations where two inpatient ollow-up 30.6.12, Section 30.6.15.1 (Chapter 12). CMS Web Site. Available at
m
visits are submitted on the same date. Lessons learned rom these www.cms.hhsgov/manuals/downloads/clm104c12.pd . Accessed
e
-
B
two denials alone may prevent hundreds more. December 14, 2015.
a
s
e
d
,
a
n
d
P
a
l
l
i
a
t
i
v
e
C
a
r
e
B
i
l
l
i
n
g
191
CHAP TER
29 PROCEDURES
■ INTRODUCTION
Procedures are commonplace in hospital medicine; more and more
hospitalists have invested additional educational time in becoming
pro cient in per orming them. Understanding the documentation
requirements necessary to ensure proper payment o procedures,
including adequate documentation o the medical necessity to
per orm them, is crucial. In addition, making sure that the Cur-
Billing for Procedures rent Procedural Terminology (CPT) code is correctly linked to the
proper International Classi cation o Diseases, 10th Revision (ICD-10)
code(s) will urther prevent claim denials.
and Use of Modifiers Details regarding the service requirements o each procedure are
outlined in the CPT Handbook, published by the American Medi-
in Inpatient Practice cal Association. The Medicare Claims Processing Manuel discusses
nuances in Medicare guidelines regarding situational and docu-
mentation requirements or reimbursement. Procedures may be
described as “minor” or “major.” These terms have caused signi cant
con usion in the literature. There are actually two di erent ways in
Yvette M. Cua, MD
which the terms “minor” and “major” procedures are used within the
Medicare manual.
■ MINOR VERSUS MAJOR PROCEDURES

Surgeries and global packages
Within the Medicare Processing Manual, procedures are re erred to
as “surgeries” or “surgical procedures.” When discussing payment
or the medically necessary care provided immediately be ore and
a ter the procedure, one must be aware o the global period associ-
ated with the service. Services with global periods o 90 days are
de ned as “major surgeries” or “major procedures.” These terms are
used interchangeably in the manual. Services with global periods
o 0 or 10 days are described as “minor surgeries” or “minor proce-
dures.” Endoscopies are included in this category. Most procedures
per ormed by nonsurgical specialties all into the “minor” procedure
category.
One example o the importance o this distinction is use o
modi er –57 (decision to per orm surgery). Modi er –57 is only
used when deciding to per orm major and not minor procedures.
Reimbursement or all care related to the surgery, by the surgeon
per orming the procedure, occurring 1 day be ore the procedure
and or the ensuing 90-day global period, is bundled into the reim-
bursement or the surgery’s CPT code. One exception is the reim-
bursement or the E/M visit at which time the decision to operate
is made i it occurs 1 day prior to surgery or on the day o surgery.
Modi er –57 attached to that E/M code will allow or its reimburse-
ment in addition to the reimbursement or the major procedure.
For the situation where the decision to per orm a minor procedure
is made at an E/M visit the same day o the procedure, modi er –25
should be attached to that E/M code to show that it was medically
necessary work, separate rom the usual history and physical needed
to sa ely per orm the procedure.
Teaching physician procedural billing

With regard to teaching physician (TP) rules or procedural billing,
a “minor” procedure is de ned as taking <5 minutes to per orm,
and the TP must be present or the entire procedure to bill or it. A
“major” procedure takes >5 minutes. The TP must be present or the
critical or key portion(s) o the procedure and must be immediately
192
available or the entire procedure, in order to bill or it. The de ni- terminate in either the subclavian, brachiocephalic, innominate or
tion o “major” and “minor” in this context is based on the usual total iliac veins, or where one o these joins with the superior or in erior
C
procedure time. Examples o minor procedures include drainage o vena cava. Medical documentation should include these key pieces
H
a subungal hematoma [11740], or a skin punch biopsy [11100]. Most o in ormation not only to select the correct CPT code, but also to
A
P
inpatient procedures hospitalists per orm, are considered “major” adequately support reimbursement or these services. This in orma-
T
procedures, such as central line placement [36555-36556], thoracen- tion is usually ound in a procedure note.
E
tesis [32554, 32555], or lumbar puncture [66270, 66272]. “Placing a central line” at bedside, by convention, has become
R
synonymous with “placing a nontunneled catheter in a central
2
■ DESCRIPTION OF SELECTED PROCEDURES location” such as the internal jugular vein or the subclavian vein.
9
Below are nuances o some common procedures per ormed in This corresponds to CPT codes [36555, 36556] which di er only
hospital medicine. by the age o the patient. In this example, to ensure accurate CPT
B
code selection, at minimum, the body o the procedure note should
i
l
l
i
Central venous access procedures
n
include re erence to insertion o a nontunneled catheter (such as
g
Central venous access can be achieved through various locations a triple lumen) into a central vessel (such as the right subclavian).
o
and processes, and with various apparatuses. Di erent CPT codes In addition, postprocedure chest x-ray results veri ying tip location
r
P
exist based on the ollowing key pieces o in ormation: (1) location should be documented. The CPT codes or peripherally inserted
r
o
o access—central, peripheral; (2) method o placement—tunneled, central catheter (PICC) placement [36568, 36569] only di er by age
c
e
nontunneled; (3) apparatus—catheter, port, pump; (4) number o o the patient.
d
u
access sites in the apparatus; and (5) age o the patient. Each o Use o ultrasound guidance [76937] or uoroscopy [77001] or
r
e
these actors plays a role in determining the usual amount o work placement, is reimbursed through an additional code and is not
s
needed to per orm the service, which directly ties into graded reim- bundled into the reimbursement or central line placement. Billing
a
n
bursement amounts. For example, tunneled catheters require more or imaging used to guide line placement requires very speci c
d
U
expertise and take longer to place than nontunneled lines; centrally documentation. Ultrasound guidance or central vascular access
s
placed lines are more challenging than their peripherally placed [76937] is only reimbursable when used with a “dynamic” technique
e
o
counterparts; and a device with a port takes more work to place and not a “static” technique. A dynamic technique implies that the
M
than a catheter without one. There are separate codes or repair, ultrasound is used throughout the entire procedure, not only to
o
replacement, and removal o these apparatuses as well (Table 29-1). identi y the target vessel, but to watch success ul entrance o the
d
i
These central venous catheters (CVC), also called “central lines,” and needle into it. “Static” use o the ultrasound purely identi es the
i
e
central venous access devices (CVAD), like “ports,” have tips that vessel and its patency, and is not reimbursable.
r
s
i
n
I
n
p
a
t
i
e
TABLE 29-1 CPT Codes for Central Venous Access Procedures
n
t
P
Centrally Placed
r
a
Method Type Age Insert Repair Partial Replace Total Replace Remove
c
t
i
c
Nontunneled CVC <5 y old 36555 36575 * 36580 *
e
≥5 y old 36556 36575 * 36580 *
Tunneled CVC <5 y old 36557 36575 * 36581 36589
≥5 y old 36558 36575 * 36581 36589
CVAD (port) <5 y old 36560 36576 36578 36582 36590
≥5 y old 36561 36576 36578 36582 36590
CVAD (pump) 36563 36576 36578 36583
2-Tunneled cath CVC 36565 36575 × 2 * 36581 x 2 36589
CVAD (port) 36566 36576 × 2 36578 × 2 36582 × 2 36590 × 2
Peripherally Placed
Method Type Age Insert Repair Partial Replace Total Replace Remove
Nontunneled PICC <5 y old 36568 36575 * 36584 *
≥5 y old 36569 36575 * 36584 *
Tunneled CVAD (port) <5 y old 36570 36576 36578 36585 36590
≥5 y old 36571 36576 36578 36585 36590
Imaging guidance CPT
Ultrasound guidance or 76937
vascular access
Fluoroscopic guidance or 77001
vascular access
* Not separately reimbursable.

Insertion = placement or catheter in newly established site.
Repair = ixing device without replacing any component o it.
Partial replacement = replacing only the catheter portion o a CVAD, not a port or pump.
Total replacement = replacing entire device through the same access site. I a new access site is established, submit CPT or insertion, not replacement.
193
TABLE 29-2 CPT Codes for Pleural Access Procedures
Service Description Imaging Guidance CPT

P
Chest tube placement Tube thoracostomy with or without connection to add 75989 32551
A
waterseal
R
T
Thoracentesis Needle/catheter aspiration o pleural space without imaging 32554
I
with imaging 32555
Pigtail catheter placement Percutaneous pleural drainage with without imaging 32556
placement o indwelling catheter with imaging 32557
T
h
e
S
p
e
At minimum, documentation needs to include a description o dissection. The procedure entails moderate sedation which is not
c
i
potential access sites, patency o the vessel selected, and success ul separately reimbursable. Chest tube insertion is not separately
a
l
t
entry o the seeker needle into the target vessel in order to bill or reimbursable i placed intraoperatively during another cardiac or
y
ultrasound guidance or central vascular access [76937]. In addition, thoracic procedure.
o
f
permanent visual recording in the orm o a printed image, digital There is no separate reimbursement or removal o chest tubes or
H
o
image, or video must be captured and available in an audit. The CPT pigtail catheters. I any o these procedures is done bilaterally, modi-
s
p
manual implies that at least an image o the patent target vessel is er –50 should be used.
i
t
a
suf cient. Attempting to shi t ocus away rom the procedure to try
l
M
to capture an image o the actual seeker needle in the vessel may Paracentesis
e
pose a risk to the patient. A postprocedural image o the catheter Just as or a thoracentesis, paracentesis codes [49082, 49083] were
d
i
sitting in the target vessel would be a sa er and acceptable way to
c
revised in 2012 to re ect the increased use o imaging guidance.
i
n
meet these criteria. When submitting the claim, remember to attach The intent to do a diagnostic versus therapeutic procedure, the
e
modi er –26 to the CPT code since the technical component o
a
di erence between small and large volume taps, and per orming
n
the service is not separately reimbursable in a acility setting. Place- the initial versus subsequent tap, do not play a role in determining
d
ment o a nontunneled Quinton triple lumen catheter into the right
S
the level o reimbursement. For the higher level o reimbursement,
y
subclavian vein o a 45-year-old patient using dynamic real-time
s
documentation must clearly state that imaging was used, otherwise
t
e
ultrasound guidance, would be reported using CPT codes 36556 the reimbursement will de ault to the lower reimbursing service
m
PLUS 76937-26. done without imaging [49082]. [49083] is a bundled code reim-
s
There is no reimbursement or repositioning or removing a non-
o
bursing or both the work o the procedure as well as the imaging.
f
tunneled catheter at bedside. However, changing a nontunneled
C
A separate CPT code such as 76942 (ultrasound guidance) should
a
central line over a wire (central location [36580], PICC [36584]) is not be used with it (Table 29-3).
r
e
separately reimbursable. I a second line needs to be placed on the
same day as the rst (ie, the patient pulls the rst one out com- Incision and drainage
pletely, or the patient had a dialysis catheter placed on the same
The selection o the correct CPT code or incision and drainage
day as a triple lumen CVC), the CPT code or the procedure would be
(I&D) procedures is based on the location o the lesion, on the
reported twice with the appropriate modi er (–76, –77) appended
type o lesion being treated, the complexity o the procedure, and
to the second procedure.
i it occurred in the postoperative period. Certain locations have
unique CPT codes or I&D such as an abscess o the eyelid [67700]
Thoracentesis and chest tube placement or external ear [69000, 69005]. I&D o the majority o abscesses are
The CPT codes or accessing the pleural space [32554-32557] reported using CPT codes 10060 (simple) and 10061 (complex or
were updated in 2013 to better re ect a change in practice with multiple). There is no universal de nition o “complex”; the CPT
improved technology. At that time, over 75% o thoracenteses with manual states that it is le t up to the clinical judgment o the
or without catheter placement were being done with ultrasound provider and substantiated by their documentation. Complex
guidance. This new set o CPT codes contains bundled codes, such I&D procedures include (1) large lesions, though “large” is also
that a separate CPT code or the use o imaging (ultrasound [76942], not universally de ned; (2) signi cant packing or placement o a
uoroscopy [77002]) to identi y a pocket o uid, should no longer drain; (3) disintegration o loculations; (4) incision and drainage o
be submitted with the appropriate pleural access code; reimburse- multiple lesions. I&Ds o hematomas are reported with CPT code
ment or the use o imaging is already included in the payment or 10040 (Table 29-4).
the primary procedure code [32555, 32557]. Documentation or a
thoracentesis (“puncture and aspiration o the pleural cavity”) needs
to speci cally identi y i ultrasound guidance was used [32555] or
not [32554]. No adjustment in reimbursement is made or a diag- TABLE 29-3 CPT Codes for Paracentesis and Peritoneal
nostic versus therapeutic procedure. Similarly, placement o a pigtail Lavage
catheter (“percutaneous pleural drainage with insertion o indwell-
Service Imaging Guidance CPT
ing catheter”) with ultrasound guidance [32557] reimburses higher
than the blind placement [32556] (Table 29-2). Paracentesis Without imaging 49082
For chest tube placement [32551], i imaging guidance is used With imaging 49083
(ultrasound, uoroscopy, computed tomography), a separate CPT Peritoneal lavage Includes reimbursement or 49084
code [75989] must still be submitted or reimbursement o imag- imaging
ing. The reason or higher reimbursement than pigtail catheter Even i not used
placement is that this open procedure requires cutdown and
194
and necessity o per orming the procedure. A consultant is asked or
TABLE 29-4 CPT Codes for Incision and Drainage Procedures and provides an opinion on diagnosis, evaluation, and treatment.
C
A consultant can be reimbursed or both the consult that generated
H
CPT Code Description
A
the decision to per orm an LP plus the procedure.
10060 I&D o cutaneous or subcutaneous abscess, cyst,
P
or paronychia; simple or single
T
Two E/M services
E
10061 I&D o cutaneous or subcutaneous abscess, cyst,
R
or paronychia; complicated or multiple A modi er –25 may be needed when two medically necessary E/M
services are provided on the same calendar date. Only one primary
2
10080 I&D o pilonidal cyst; simple
9
E/M code should be submitted or all E/M work done rom 12:01 a m
10081 I&D o pilonidal cyst; complicated to 11:59 pm by one physician and/or multiple physicians in a billing
10120 Incision and removal o oreign body, subcutaneous group unctioning as coverage/continuance o care or each other.
B
tissues; simple
i
l
However, a ew exceptions to this rule exist. One exception is when
l
i
n
10121 Incision and removal o oreign body, subcutaneous an E/M service such as an inpatient admission or inpatient ollow-
g
tissues; complicated up visit is completed and ollowed later that same day by a critical
o
10140 I&D o hematoma, seroma or luid collection
r
care service by the same provider and/or billing group. The provider
P
10160 Puncture aspiration o abscess, hematoma, bulla, may be compensated or both services. Some insurers may require
r
o
or cyst modi er –25 to be attached to the CPT code or the noncritical care
c
e
service to show that it was separate and distinct rom the critical
d
10180 I&D, complex, post-op wound in ection
u
care service. Medical necessity and complete documentation or
r
e
both services must exist in the medical record.
s
a
n
d
MODIFIERS ■ MODIFIER –26: BILLING FOR THE PROFESSIONAL
U
COMPONENT OF A PROCEDURE
s
A modi er is a two character alpha numeric code, attached to a
e
Many procedures such as chest x-rays and ultrasounds have reim-
o
CPT code, to urther describe a service or to signi y that a service
has been altered by a speci c circumstance without changing bursement that is divided into a technical component (TC) and a
M
pro essional component (PC). Reimbursement or the TC covers the
o
the description/de nition o the service. Some situations require
d
more than one modi er to be used with a CPT code. There are two cost o purchasing or leasing equipment, and or maintenance. The
i
i
reimbursement or the PC covers the supervision/per ormance o a
e
levels o modi ers. Level 1 modi ers, also called CPT modi ers, are
r
test, and its ormal written interpretation. When billing or the pro-
s
numeric, and updated by the American Medical Association. Level 2
i
n
modi ers, also called HCPCS modi ers, begin with a letter ollowed essional component o a service, attach modi er –26.
I
n
by either a number or another letter, and are updated by CMS. For example, many hospitalists read their own chest x-rays (CXR)
p
to make immediate management decisions or their patients, and
a
There are hundreds o modi ers. Some directly impact reim-
t
document their ndings in their progress notes, ollowed by a
i
e
bursement amounts; some directly enable payment or a service,
n
while others are or in ormational purposes alone. Below are some ormal or of cial radiology report some time later. This work counts
t
P
o the more common ones needed or proper claims processing in toward their level o medical decision making or the E/M service
r
a
Hospital Medicine. and is not separately reimbursable. However, some institutions do
c
t
not have radiologists immediately available a ter hours and may
i
c
e
have a system in place whereby hospitalists or emergency depart-
■ MODIFIER –25: SIGNIFICANT AND SEPARATELY ment providers are responsible or producing the of cial written
IDENTIFIABLE E/M SERVICE BY THE SAME PROVIDER reports, not a radiologist the next day. In this situation, with a valid
ON THE SAME DAY AS A PROCEDURE OR OTHER written report generated, this work is separately reimbursable by
SERVICE the hospitalist or emergency department provider. The TC o the
This may only be attached to an E/M code and not to a procedure global ee (total reimbursement) or the procedure, is not separately
code. On occasion, it is necessary or a provider and/or covering reimbursable in this setting. Modi er –26 should be added to the
member o their billing group, to provide more than one medically procedure’s CPT code or reimbursement o the PC. For the CXR
necessary service to a patient on the same calendar date. above, 71020-26 would be submitted or reimbursement or the
work o viewing and interpreting the lm, and creating the of cial
E//M and procedure report. The provider cannot count that work toward the complexity
o his medical decision making or any associated E/M service to the
The most common scenario would be a medically necessary E/M
patient i receiving separate reimbursement or CXR interpretation.
visit plus a procedure on the same calendar date. Modi er –25
Minus one situation, in the erroneous event that both the hospitalist
should be attached to the E/M code to signi y that it is a separately
and a radiologist submit 71020-26 or reimbursement or the same
identi able service, distinct rom the routine pre- and postwork
lm, only the rst bill received with be reimbursed. A system needs
needed or the procedure. For example, a hospitalist admits a
to be in place to ensure that this situation does not occur.
patient with altered mental status and ever, determines that a
There are some procedures that do not have a global ee divided
lumbar puncture (LP) is needed, and per orms it on the same date
into a TC and PC. Instead, there are separate CPT codes or billing or
as the admission. The hospital may receive payment or BOTH the
TC + PC, TC alone, or PC alone. For example, EKGs: CPT 93000 is or
initial inpatient visit [99221-99223] PLUS the LP [62270]. Modi er –25
TC + PC, 93005 is or TC alone, and 93010 is or the PC alone.
should be attached to the initial inpatient visit CPT code [99221-
99223] to show that the cognitive work o that visit is separate and
above and beyond that necessary or just sa ely per orming the LP. ■ MODIFIER –76: SAME EXACT PROCEDURE
In contrast, a neurologist may only bill or the LP and not or an E/M PERFORMED BY THE SAME PROVIDER ON
visit, i his or her role is one o proceduralist and not consultant. THE SAME DATE
A proceduralist should take and appropriately document a brie his- A procedural CPT code and never an E/M code is used when the
tory and per orm a brie physical preprocedure to ensure the sa ety same exact procedure is done twice on a patient on the same
195
calendar date and/or during the procedure’s global period. For ■ MODIFIER –Q5: RECIPROCAL BILLING
example, a hospitalist places a central line on a patient in the morn- Attach modi er –Q5 to services provided in a reciprocal billing
ing, the patient pulls it out, and the hospitalist places another one arrangement. A reciprocal billing situation exists when two or more
later that day. The central line CPT code [36556] would be reported
P
physicians rom di erent billing entities agree to cover or each
A
twice, attaching modi er –76 to the second one. Likewise, you other in one’s absence. This agreement does not need to be in writ-
R
would report cardiopulmonary resuscitation CPR [92950] twice ing, but clearly understood between parties. The criteria or recipro-
T
attaching modi er –76 to the second CPR billing code i a patient cal billing are (1) the regular physician is unavailable; (2) the patient
I
requires CPR at two di erent times in the day. Reimbursement or is seeking care rom the regular physician; (3) the substitute physi-
two procedures requires appropriate documentation o each proce- cian does not cover or more than 60 consecutive days. The substi-
dure (ie, two notes with the time each procedure was per ormed). tute physician, with one exception, is not a member o the regular
A le t knee arthrocentesis and a right shoulder arthrocentesis physician’s billing group. These services are actually billed under the
T
h
would not be coded using this modi er. Although an arthocentesis regular physician’s National Provider Identi er (NPI) number, attach-
e
was per ormed in both locations, it is not the exact same proce- ing modi er –Q5 to each CPT service code. The reimbursement will
S
p
dure—one is o the knee and the other is o the shoulder. This would go to the regular physician. Ideally, when the substitute physician
e
be coded using modi er –XE instead.
c
needs time o , the regular physician will return the avor, and in the
i
a
l
long run, reimbursement evens out. This arrangement is commonly
t
y
■ MODIFIER –77: SAME EXACT PROCEDURE used by solo practitioners, but may also be used by group practices
o
PERFORMED ON THE SAME DATE BY A DIFFERENT
f
who acquire coverage rom a physician outside their billing group.
H
PROVIDER Notice that the usual scenario: partners in a billing group and sub-
o
s
This modi er is used exactly like modi er –76 except when the same specialty cover or each other on weekends, holidays, and vacations,
p
i
t
exact procedure is per ormed by a di erent provider on the same does not t the de nition o a reciprocal billing arrangement. In this
a
l
date. scenario, the physician providing the service would simply submit
M
charges under their own NPI number.
e
d
■ MODIFIER –AI: PRINCIPAL ATTENDING The hospice setting is an exception rom the situation where the
i
c
i
PHYSICIAN OF RECORD substitute physician is actually rom the same billing group. Hospice
n
e
This modi er is used to identi y the principal physician o record patients select or are assigned a primary attending who oversees
a
their care. Unlike the role o principal attending o record or other
n
or the patient, ie the main physician overseeing care. Medicare
d
stopped paying or inpatient consult codes [99251-99255] on Janu- hospital inpatients, the role o primary hospice attending is speci c
S
y
ary 1, 2010. All consultants seeing inpatients now bill their consult to an individual provider, not to a billing group/subspecialty. In the
s
t
event that the hospice attendings require coverage, such as week-
e
with the same initial inpatient visit codes [99221-99223] and initial
m
skilled nursing acility codes [99304-99306] that the admitting end coverage, and one o their partners covers these patients, this is
s
also considered a reciprocal billing arrangement. These services, pro-
o
attending uses. The –AI modi er (letter I, not number 1) will distin-
f
guish the admitting attending rom all other consultants, but will vided by a partner in the same billing group, will be billed under the
C
regular hospice attending’s NPI number with modi er –Q5 attached
a
not change the amount o reimbursement or the visit. By attaching
r
e
it to the initial inpatient or skilled nursing visit code, it also aids CMS to each CPT code. Thus in a group practice, Dr Sub may be covering
with identi ying which physician will bill or discharge management or Dr Hospice or the weekend. All services provided by Dr Sub to
services. Only attach it to the initial inpatient [99221-99223] or initial Dr Hospice’s regular inpatients will be billed under Dr Sub’s own NPI;
skilled nursing acility [99304-99306] visit codes. Do not attach it to however, those services provided by Dr Sub to hospice patients will
any other CPT codes such as subsequent visits [99231-99233, 99307- be billed under Dr Hospice’s NPI, attaching modi er –Q5.
99310], discharges [99238-99239, 99315-99316], or observation
services [99217-99220; 99224-99226]. ■ MODIFIER –Q6
Modi er –Q6 is attached to all services per ormed by a locum tenens
■ MODIFIER –GC: CARE PROVIDED IN PART BY A provider (LT), with claims submitted under the regular physician’s
RESIDENT, SUPERVISED BY A TP
NPI. Reimbursement will go to the regular physician. A LT has no
Attached modi er –GC to all services per ormed in conjunction with practice o their own, and unctions as an independent contractor,
a resident (including interns and ellows) in a teaching situation, receiving payment on a xed per diem or ee- or-time basis. They
provided that the TP was present and personally participated in the can only provide coverage or any individual physician or a con-
critical or key component(s) o the encounter. It is not appropriate secutive not cumulative 60-day period, even i not providing care
to attach to services per ormed in supervision o medical students. on all o those days. The 60-day period starts to the rst day the LT
provides services. I a physician returns to work or a period o time
■ MODIFIER –GV: CARE RELATED TO THE and needs to leave again, the locum tenens may provide coverage
TERMINAL ILLNESS or another 60-day period. The only exception to the 60-day time
The –GV modi er can only be attached to services rendered by period or Medicare patients is or coverage o a physician called to
a hospice patient’s primary attending i (1) the attending is not active duty—coverage time is unlimited. I a second physician in a
employed or paid by the hospice under any agreement and is not group practice needs time o , the LT can provide a separate 60 days
a volunteer, and (2) the care is or the patient’s terminal illness. This o coverage or that physician as well, regardless o the number o
modi er is used to submit claims to Medicare Part B or reimburse- days o coverage they already served or the rst physician.
ment. It cannot be used by consultants, nor can it be used by the Common reasons a physician may not be available, include
patient’s primary attending or care unrelated to the terminal illness. pregnancy, illness, continuing medical education, and vacation. In a
group practice setting, a physician may need to leave the practice,
■ MODIFIER –GW: CARE UNRELATED TO THE and the practice may employ a LT or up to 60 days while nding a
TERMINAL ILLNESS replacement. The purpose o a LT is to provide temporary coverage
Attach modi er –GW to services provided to a hospice patient when or an existing physician. They cannot be used as temporary help to
care is unrelated to the terminal illness. ll a newly created position while growing a practice.
196
LT is an attractive short term coverage option. They do not have ONLINE RESOURCES
to be enrolled in Medicare to provide services since the billing
C
is under the regular physician’s NPI. LT can only be employed to Centers or Medicare and Medicaid Services. Medicare Claims Pro-
H
substitute or an absent physician. They cannot be hired to cover an
A
cessing Manual: Section 40 Chap 12. CMS Web site. www.cms.hhs.
P
absent nonphysician provider (NPP). Also, they cannot be employed gov/manuals/downloads/clm104c12.pd . Accessed December
T
to cover an absent physician i the reason or absence is death. NPI 14, 2015.
E
numbers are deactivated upon death.
R
Centers or Medicare and Medicaid Services. Medicare Claims Pro-
LT guidelines were created by Medicare, but private insurers and cessing Manual: Section 100.1.2. Chap 12. CMS Web site. www.cms.
2
Medicaid may di er. For example, Texas Medicaid and Healthcare
9
hhs.gov/manuals/downloads/clm104c12.pd . Accessed Decem-
Partners has a 90 consecutive day LT limit. Some private insurers ber 14, 2015.
do not recognize LT services and require LTs to be credentialed
B
2012 Coding Updates, American Gastroenterology Association.
through the normal new member process. It is imperative or this
i
l
l
http:/ / www.gastro.org/ journals-publications/ aga-edigest/ Tri-
i
n
and all billing situations, to check with individual payors or devia-
g
Society_Coding_Sheet.pd .
tions rom the Medicare guidelines.
o
Noridian healthcare Solutions. https://med.noridianmedicare.com/
r
web/jeb/topics/modi ers.
P
r
o
c
e
d
u
r
e
s
a
n
d
U
s
e
o
M
o
d
i
i
e
r
s
i
n
I
n
p
a
t
i
e
n
t
P
r
a
c
t
i
c
e
197
30
CHAP TER TEACHING PHYSICIAN GUIDELINES
■ INTRODUCTION
Teaching physician (TP) guidelines date back to 1969 with the pub-
lication o Intermediary Letter 372, but did not come under serious
scrutiny until 1995 when the results o the O ce o the Inspector
General’s (OIG) rst physicians at teaching hospitals (PATH) audit
resulted in a $30 million dollar settlement by the University o
Pennsylvania. The audit revealed that there was insu cient docu-
Billing in the mentation in the medical record that the TP was physically pres-
ent and actively participated in the critical portions o the services
billed. This index case prompted a nationwide PATH audit initiative,
Teaching Setting and on December 8, 1995, the Health Care Financing Administra-
tion (HCFA), now the Centers or Medicare and Medicaid Services
and Billing with (CMS), published in the Final Rule, more detailed instructions or
documenting and billing or services in a teaching setting. The
Advanced Practice
rules o cially went into e ect July 1, 1996 and were added to the
Medicare Claims Processing Manual, Chapter 12, Section 100. These
rules have had modi cations and clari cations over the years, last
Providers in 2011. Although state Medicaid programs and private insurers
o ten ollow Medicare, there are individual di erences among these
programs, especially in regards to requirements or TP presence
and TP documentation. This author suggests using this chapter as
Yvette M. Cua, MD a starting point, but checking with local payers or deviation rom
these guidelines.
DEFINITIONS
The term “resident” in the Medicare Processing Manual re ers to
“… an individual who participates in an approved graduate medical
education (GME) program or a physician who is not in an approved
GME program but who is authorized to practice only in a hospital
setting ….” O note, receiving a aculty or sta appointment, or
being included in a hospital’s ull time equivalency (FTE) count, in
and o itsel does not alter a resident’s status. For billing purposes,
and or this chapter, the term “resident” includes interns and ellows,
but not students.
A student is someone in an accredited educational program
like a medical school, dental school, or school o osteopathy that is
not an approved GME program. Medicare does not pay or student
services.
A TP is a physician who is not another resident, who provides
care to patients in conjunction with a resident. A TP does not need
to have a aculty appointment, thus these regulations are also
applicable to private practice physicians who work with residents.
Finally, these guidelines are speci cally or physicians, and are not
applicable to nonphysician providers (NPP), discussed separately in
the second part o this chapter.
PAYMENT FOR PHYSICIAN SERVICES PROVIDED IN A

TEACHING SETTING
The reason that medical and surgical services provided only by
residents, within the scope o their training program, cannot be
reimbursed through the Medicare Physician Fee Schedule (MPFS), is
because Medicare Part A already pays or resident services up ront
through direct graduate medical education (DGME) and indirect
medical education (IME) payments. In a teaching setting, these
services may be reimbursed through the MPFS i the TP person-
ally per orms or is physically present while a resident per orms the
198
critical or key portions o the service, assuming that the documenta- 2. The TP personally per ormed the service or was present or the
tion clearly ref ects this. critical or key portions per ormed by a resident.
C
3. The TP participated in the management o the patient. In a teach-
H
A
ing setting, the TP’s management o the patient is o ten not
PRACTICE POINT
P
the direct writing o orders or calling o consultants. Instead,
T
• In a teaching setting, pro essional clinical services may be care plan development and supervision o proper execution or
E
reimbursed through the Medicare Physician Fee Schedule orders are commonplace: review o the resident’s documented
R
(MPFS) i the teaching physician (TP) personally per orms or assessment and plan or accuracy, discussion with the resident
3
o the details o the plan, including any deviances rom their
0
is physically present while a resident per orms the critical or
key portions o the service, assuming that the documentation documentation, and vigilant ollow-up to ensure timely and
clearly ref ects this. correct ollow-through o the plan. Documentation o these
B
activities will satis y these criteria.
i
l
l
i
n
These three criteria are o ten documented in a sentence called
g
■ EVALUATIONS AND MANAGEMENT SERVICES an “attestation statement.” This statement must always be written
i
n
by the TP. The resident’s documentation o the presence and participa-
t
Teaching physician presence
h
tion o the TP is insu cient to support reimbursement or the service.
e
There are several common scenarios in which the TP provides ser-
T
An acceptable attestation correctly linking the TP’s note to the
e
vices in tandem with a resident. The key eature in all o these is that
a
resident’s note will allow or the combination o the resident’s and
c
the TP is physically present and has a ace to ace encounter with
h
the TP’s documentation to determine the level o service provided.
i
the patient on the date o service, personally per orming all o the
n
Medicare manual examples o unacceptable attestation state-
g
service or at minimum, the critical or key portions to allow or billing, ments that do not meet the above criteria include the ollowing
S
regardless o the extent o care the resident provided.
e
statements which do not make it possible to determine i the TP
t
t
1. The TP personally provides the service. There may be situations
i
was actually present, evaluated the patient, and/or participated in
n
g
when a resident has a day o , or the TP has completed a service the plan o care.
a
be ore the resident has seen a patient and/or has discussed it
n
1. “Agree with above.” ollowed by valid authentication (ie, legible
d
with the TP. The TP would document the service as they would
signature or electronic password protected signature).
B
in a nonteaching setting as there is no resident note to link to.
i
l
2. “Rounded, Reviewed, Agree.” ollowed by valid authentication.
l
i
In absence o a resident’s note, attempting to write an attesta-
n
3. “Discussed with resident. Agree.” ollowed by valid
g
tion alone without personally documenting elements o the
authentication.
w
history, physical, and medical decision making, is insu cient to
i
4. “Seen and agree.” ollowed by valid authentication.
t
h
support reimbursement or the service provided.
5. “Patient seen and evaluated.” ollowed by valid authentication.
A
2. The TP personally provides the service at a di erent time on the
d
6. Valid authentication alone.
v
same calendar date as the resident. An example o this situation
a
In addition, a TP’s attestation statement re erencing a resident’s
n
would be when a resident sees a patient and provides a service
c
like a subsequent inpatient visit [99231-99233], and later on note but written and signed be ore the resident’s note is written
e
d
the same date, the TP sees the patient and personally per orms will not allow linkage to the resident’s documentation, and thus not
P
allow the resident’s documentation to be used to determine the
r
the service. Notice that the personal per ormance and not the
a
level o service provided. You cannot attest to reviewing and agreeing
c
personal documentation o the entire service is the main eature
t
i
with documentation that has yet to be written.
c
that allows the TP to bill or this service. I the documentation
e
by the resident earlier in the day contains the same in ormation Examples rom the Medicare manual o minimally acceptable
P
r
attestations include:
o
that the TP personally obtained later, the TP does not have to
v
redocument any o it. The TP needs to document a valid attes-
i
1. “Hospital Day #3. I saw and evaluated the patient. I agree with
d
e
tation statement linked to the resident’s note, plus any addi- the ndings and the plan o care as documented in the resi-
r
s
tional ndings, discrepancies, or changes in clinical condition. dent’s note.”
3. The resident per orms the entire service, or at minimum the critical 2. “I saw the patient with the resident and agree with the resi-
or key components in the physical presence o the TP, and the TP dent’s ndings and plan.”
personally documents that he or she was present or the criti- 3. “See resident’s note or details. I saw and evaluated the patient
cal or key portions o the service. The clinical judgment o the and agree with the resident’s ndings and plan as written.”
TP determines the critical or key portion(s) o a service, based
Although not speci cally stated in the Medicare manual as a require-
on standards o care, speci cs o the service, and individual
ment, the use o the word “I” at the beginning o the attestation is the
patients’ situational actors.
best way to document that you personally per ormed the service.
In addition to the attestation statement, the TP should document
■ DOCUMENTATION REQUIREMENTS any in ormation that the resident orgot to document, new in orma-
For E/M services (excluding outpatient Primary Care Exception situation or changes in clinical picture since the time o the resident’s
tions), at a minimum, TPs must personally document: note, as well as any discrepancies or disagreements with the resi-
dent’s ndings, thoughts, or plans. This not only helps support bill-
1. The TP was present. All E/M services in the context o hospital
ing, but more importantly, high-quality patient care as well. Below
medicine require a ace to ace encounter with the billing TP on
are a couple examples o this rom the Medicare manual:
the date o service. I something prevents the TP rom seeing
the patient, such as the patient being in the operating room all 1. “Hospital Day #5. I saw and examined the patient. I agree with
day, or at a procedure during the only time rame that the TP is the resident’s note except the heart murmur is louder, so I will
available that day, even though the TP may have invested the obtain an echo to evaluate.”
time in complete data review, discussion with consultants, and 2. “I saw and evaluated the patient. I reviewed the resident’s note
care plan development with the resident, the service cannot be and agree, except that picture is more consistent with pericar-
billed or reimbursed. ditis than myocardial ischemia. Will begin NSAIDs.”
199
In act, this additional documentation, individualizing the docu- TP is present or the entire procedure, the resident or operating
mentation to each patient, is better veri cation o the TP personal room nurse may document the TP’s presence without need or the
participation in the case. TP to personally document it. The manual does not elaborate on
the TP documentation requirements i the TP is only present or
P
A
the critical or key components o a single surgery. In this situation,
PRACTICE POINT
R
check with institutional compliance o ce regulations regarding i
T
• Keep in mind that the documentation must ref ect the medical the TP needs to personally document their presence versus having
I
necessity o the service being provided; the pure volume the resident or operating nurse’s documentation o their presence
o documentation alone will not determine the level o satis y billing criteria.
reimbursement o the service.
Two overlapping surgeries
T
h
In order to bill or each o two overlapping surgeries, the TP must
e
■ STUDENT E/M DOCUMENTATION be physically present or the critical or key components o each
S
p
The only parts o a student’s documentation that may be re erred surgery. Thus, the critical or key components cannot overlap. They
e
c
to and used to determine the level o E/M service provided, are the must also be immediately available to intervene or the noncritical
i
a
l
review o systems (ROS) and past medical history, amily history, components o each surgery. They must arrange or another quali-
t
y
and social history (PFSH). The TP note must document their review ed physician to be available or the noncritical components o the
o
f
and agreement with the in ormation. No matter how accurate and other surgery i they are not available or the noncritical compo-
H
complete a student’s documentation o the history o present illness nents o one surgery (ie, they are in the second surgery during its
o
s
(HPI), physical exam (PE), and medical decision making (MDM), re er- critical or key components). The TP must personally document that
p
i
t
ence to it without the TP personally veri ying and redocumenting it, they were present or the critical or key components o each surgery
a
l
will not allow it to contribute to a billable service. to bill or them.
M
e
d
■ LATE NIGHT ADMISSIONS Three or more overlapping surgeries
i
c
i
n
In all o the above examples, the understanding is that the TP and I a TP is involved in three or more overlapping surgeries, they
e
resident per ormed a service jointly on the same calendar date. The cannot bill or any o those services.
a
n
TP can only link their attestation to the resident’s documentation
d
rom the same calendar date, and thus use it to determine their level Minor versus major
S
y
o service. There is one special exception to this rule. In the teaching
s
In the context o discussions about extent o TP participation in
t
e
setting, it is not uncommon or a resident to admit a patient late at a procedure, one must determine i the procedure is “minor” or
m
night when there is no TP in the acility to immediately provide the “major.” In this context, a “minor” procedure is de ined as taking
s
o
initial inpatient service. In 2010, the author o this chapter worked <5 minutes to per orm. By de ault, a major procedure takes >5 min-
f
directly with CMS to augment the Medicare manual to include this
C
utes to per orm. The vast majority o Hospital Medicine procedures
a
scenario. In 2011, this scenario was added to the Medicare manual
r
are major. For minor procedures, the TP must be present or the
e
allowing a TP to link their attestation statement to the resident’s entire procedure to allow or reimbursement. For major procedures,
documentation o the initial inpatient visit the night be ore, allow- the TP needs to be present or the critical or key portions to bill or
ing that documentation to be used to determine the level o service, the service. Although the procedure note may be written by the
provided that the TP also documents any additional in ormation or resident, the TP must personally attest to their presence or the
signi cant changes since the time o the resident’s note. The date o entire procedure or or the critical or key portions o the procedure,
the service or claim submission should be the date that the TP actually or minor and major procedures respectively. The resident’s re er-
had their initial ace to ace visit, even i it is di erent rom the patient’s ence to their presence is insu cient.
admission date.
■ PROCEDURES ■ ENDOSCOPIES
The TP must be physically present or the entire viewing time
There is some con usion and controversy over TP documentation
(insertion o endoscope until removal o endoscope) to bill or
requirements or procedures due to the way the Medicare manual
endoscopic procedures. Viewing the procedure rom a monitor in
is worded. In addition, with the wide variability in state Medicaid
a di erent room does not meet criteria or TP physical presence.
and private insurer regulations, many compliance o ces opt or
Because endoscopies are discussed in the Medicare manual in the
more de endable approaches than what seems to be required in
same section as single surgeries, it is unclear i they ollow the guide-
the Medicare manual, to ensure documentation meets any entity’s
lines or single surgeries, meaning that it is not explicitly stated that
strictest regulations. The Medicare manual uses the terms “proce-
the resident or the operating room nurse may document the TP’s
dure” and “surgery” interchangeably at times, making the guidelines
presence in lieu o the TP personally documenting it. This is another
or TP documentation in certain circumstances, le t up to interpre-
situation to clari y with the institution’s compliance o ce.
tation. For all “surgical procedures,” the manual is clear that the TP
must be present or the critical or key components o the procedure,
but no blanket statement is made regarding TP documentation ■ TIME-BASED SERVICES
requirements. When a resident is involved in a service that is billed based on time,
to meet minimum time requirements, the TP can only count their
Single surgeries personal time spent, not the resident’s time spent. For example, i a
In order to bill or these services, a TP must be physically present resident spends 25 minutes on discharge management services or
or the critical or key components, plus or the noncritical portions, an inpatient, and the TP spends another 10 minutes, the TP cannot
the TP must be immediately available to return i needed. I the TP aggregate their documentation o time with the resident’s time to
is unavailable or the noncritical components, they need to make achieve 35 minutes o time, enough or the higher level o discharge
arrangements or another quali ed physician to be available. I the service [99239]. The TP can only bill or the 10 minutes o time
200
they personally spent on the encounter and thus only bill or the physician’s license, and allowing 100% MPFS reimbursement, in two
<30 minutes discharge service [99238]. very speci c ways.
C
H
■ CRITICAL CARE
A
NPP REQUIREMENTS
P
When providing critical care services in conjunction with a resident, In addition to meeting state licensing and credentialing require-
T
just as in other time-based services, the TP may only count personal
E
ments, in order to allow NPPs to bill under the supervising physi-
time spent providing critical care to the patient, and not time spent
R
cian’s license, these two criteria need to be ully satis ed.
by the resident. Time spent rounding and teaching, but not in direct
3
1. The NPP must be a nancial responsibility to the billing physi-
0
patient care also cannot be counted toward critical care time. In
addition, the TP can only bill or critical care services i the patient cian and/or their group. O ten the NPPs and the physicians are
was critically ill during the time the TP was physically present and employed and paid by the same nancial entity; they are mem-
B
bers o the same group practice. Sometimes a physician group
i
managing the patient. For example, i the patient is critically ill,
l
l
i
may directly hire an NPP or NPP group to work with them as
n
going into impending respiratory ailure, and the TP is not physi-
g
cally at the hospital, but assists the resident over the telephone in an employee or as an independent contractor. In the event that
i
n
deciding to initiate noninvasive ventilation, order blood gases, and the NPP is not the nancial responsibility o the physician they are
t
h
consult the intensive care unit, this time cannot be counted toward working with, this criteria will not be satis ed. An example o this
e
critical care, as the TP was not physically present providing the criti- type o relationship would be a private hospitalist group work-
T
e
cal care service. More importantly, i the patient is o noninvasive ing with an NPP directly employed by the hospital. The hospital
a
c
ventilation and oxygen, breathing normally, and clinically no longer would usually report the NPP’s salary and bene ts on their cost
h
i
report. In this situation, the work o the NPP is not separately
n
at risk o impending organ ailure at the time the TP nally has their
g
ace to ace visit, the TP still cannot bill or critical care since the reimbursable, and thus any contribution by the NPP to a ser-
S
vice per ormed with a physician, cannot be aggregated with
e
patient did not require critical cares services at the time o the TP’s
t
the physician’s work to determine the level o service provided.
t
i
physical presence.
n
2. The service being provided by the NPP must be within the scope o
g
Although the resident’s documentation o the patient’s critical
their practice as de ned by laws o the state they are practicing
a
condition and their details o the history, physical, and high com-
n
in, and in accordance with the Collaborate Billing Agreement
d
plexity o medical decision making all greatly support the medical
B
necessity or critical care billing, in addition to a valid TP attestation with their supervising physician(s). These agreements are usu-
i
l
l
ally reviewed and signed annually, based on state regulations.
i
statement, the TP must personally document their physical pres-
n
g
ence while the patient was critically ill, the amount o time they
w
personally spent providing critical care, and a brie summary o their E/M SERVICES PROVIDED IN CONJUNCTION
i
t
h
activities. The resident’s documentation o the TP’s presence and partici- WITH A PHYSICIAN
A
pation is insu cient to support billing or critical care services. ■ SPLIT/SHARED BILLING
d
v
a
Service requirements
n
MODIFIER –GC
c
When an NPP and a physician each provide a “substantive” ace to
e
Attach modif er –GC to all services’ CPT codes when per ormed
d
ace portion o an E/M service on the same calendar date, the service
in conjunction with a trainee. By attaching modi er –GC, the TP
P
may be billed under either the NPP or the physician’s NPI. Medicare
r
a
is attesting to meeting all physical presence and documentation
c
does not de ne what exactly constitutes a “substantive” portion
t
requirements to bill or these services.
i
c
o the visit. It is expected however that a “substantive” portion will
e
include all or some portion o history, physical, and/or medical deci-
P
MIDLEVEL PROVIDERS
r
sion making. There is no required percentage o the service that has
o
■ INTRODUCTION
v
to be provided and documented by the physician versus the NPP
i
d
Advanced practice providers (APPs) including physicians are trained, in order to bill under the physician’s NPI. The NPP and the physician
e
r
s
licensed clinicians who can provide medical services. Some nonphy- must each personally document their portion o the E/M service in
sician providers (NPPs) may work and bill independently, such as the medical record, and sign it. The level o service will be deter-
nurse practitioners (NP) and physical therapists (PT). Others may only mined by the aggregate o their documented work.
work under the direct supervision o a physician, such as physician
assistants (PA). Other health care workers that all under the term Documentation requirements
NPP include: certi ed registered nurse anesthetists (CRNA), certi ed
nurse midwives (CNM), clinical nurse specialists (CNS), clinical psy- The medical record needs to clearly ref ect that both the NPP and
chologists (PhD), nonclinical psychologists, occupational therapists, the physician each had a ace to ace encounter with the patient
and speech pathologists. Medicare has very speci c rules that gov- on the date o service. Both NPP and physician must personally
ern how services involving an NPP can be provided and how they document their work in the medical record, both must be identi ed
can be reimbursed. However, each state’s Medicaid program, and in the chart and provide valid signatures, and the two documenta-
private insurers have their own regulations which may di er rom tion entries must be linked to show that the aggregate o the docu-
Medicare. For example, some private insurers such as Aetna will only mentation will be used to determine the level o service provided.
allow PAs to bill incident to a physician, not under their own license. There are two key eatures to the billing physician’s documenta-
This chapter will discuss NPP billing with regard to Medicare guide- tion to ensure proper support or billing under their NPI in a split/
lines. Clinicians should check with individual state Medicaid pro- shared situation.
grams and private payers or deviations rom the in ormation below. 1. The billing physician’s documentation should ref ect that the phy-
NPPs must have their own National Provider Identi cation (NPI) sician had a ace to ace visit with the patient on the date o service,
number in order to bill Medicare. When NPPs bill under their own unlike “incident to” services which may be billed without the
license, Medicare reimbursement is at 85% o the MPFS rate. Certain physician having a ace to ace visit. The physician’s personal
NPPs, including NPs, PAs, and CNSs, can provide evaluation and man- entry in the medical record o their work, plus a valid signature
agement (E/M) services in tandem with a physician, billing under the constitute evidence o the ace to ace visit.
201
2. The billing physician’s documentation should ref ect the medically
necessary and unique contribution to the patient’s evaluation TABLE 30-1 Split/Shared Billing
and care plan development. In contradistinction to visits in con-
Services Eligible for Split/Shared Billing CPT Code
junction with a resident, where the billing TP must personally
P
Initial inpatient visits (admissions) 99221-99223
A
per orm the critical and key portions o the service, in a split/
R
shared service, the billing physician does not need to per orm Subsequent inpatient visits ( ollow-up) 99231-99233
T
or attest to per orming the critical and key portions o the ser- Inpatient discharge management services 99238-99239
I
vice. Instead, the billing physician needs to personally document Observation same day discharge services 99234-99236
exactly what they obtained or history and physical, and their Initial observation visits 99218-99220
thoughts related to medical decision making. Examples o insu -
Subsequent observation visits ( ollow-up) 99224-99226
cient documentation o a substantive portion o work being
T
Observation discharge management services 99217
h
done by the physician include:
e
a. A notation by the NPP o the physician’s involvement. Emergency room services 99281-99285
S
p
b. A simple cosignature o the NPP’s note. Hospital-based clinic: new and established 99201-99215
e
outpatient visits*
c
c. The billing physician’s documentation should not look like
i
a
a TP attestation statement. Although the billing physician Nonhospital based (private) o ice: established 99211-99215
l
t
y
does play a supervisory role in their relationship with the outpatient visits*
o
NPP, and rom an institutional policy, may need to make Prolonged services 99254-99357
f
H
comment about reviewing the NPP’s work or legal pur- Services Ineligible for Split/Shared Billing CPT Code
o
s
poses, the billing physician’s documentation should not look
p
Nonhospital based (private) o ice: 99201-99205
i
like they did nothing more than review and agree with the
t
new outpatient visits
a
l
NPP’s work and documentation.
Consults 99241-99255
M
In the above three documentation scenarios, the service may only
e
Prolonged services 99291-99292
d
be billed under the NPP’s NPI.
i
c
Skilled nursing acility and nursing home 99304-99316
i
The roles o the NPP and physician should be complementary; the
n
services Procedures
e
relationship between a billing physician and an NPP in a split/shared
a
visit is similar to two colleagues covering or each other, seeing a
n
* Only i “incident to” criteria are met. I not, then bill under NPP.
d
patient sequentially on a given day. There will inevitably be overlap
S
y
in the historical, physical exam, and decision-making in ormation
s
physician’s work the next morning will go toward determining the
t
obtained and documented; however, the documentation needs to
e
m
clearly ref ect that the NPP and physician each contributed unique level o E/M service on day 2, such as a subsequent inpatient visit
s
and medically necessary care. Speci cally, the physician documen- [99231-99233].
o
f
tation needs to ref ect their independent contribution to the care
■ CONSULTS
C
o the patient outside the work o the NPP. When the physician
a
r
Although Medicare’s guidelines do not allow consult services to
e
sees the patient earlier in the day and documents their encounter
be ore the NPP, their contribution to the service is more easily be billed as split/shared visits, since January 1, 2010 Medicare no
discerned. In the scenario where the NPP sees the patient and longer recognizes or reimburses or consult codes [99241-99255].
documents their encounter rst, by remembering to document Medicare however still pays or consult services, just through di -
additional historical in ormation, physical exam ndings, new in or- erent code sets [99221-99223; 99231-99232; 99201-99215; 99281-
mation since the NPP’s visit, changes in clinical status, and additional 99285]. All o these code sets may be billed as split/shared services.
medical decision making not ref ected in the NPP’s note, the medi- So now or Medicare patients, even though the intent o the service
cal necessity o the physician’s role in the service is better supported. is to provide a consult, when submitted with one o these services
Split/shared services can only be provided in the hospital inpa- codes, the care may now be provided as a split/shared visit. Clini-
tient, hospital outpatient (Place o Service, POS 22), hospital obser- cians should check with Medicaid and private insurers or how these
vation, emergency room, and nonhospital based (private) o ce services should be billed.
settings. In order or split/shared services to be per ormed in a
private o ce setting (POS 11), “incident to” criteria must be met,
■ INCIDENT TO
thus new patient visits [99201-99205] in the private o ce setting “Incident to” services are de ned as medically necessary services
can never be billed as split/shared visits (see section “Incident to”). that are provided incident to a physician’s pro essional services, but
Note that new patient visits can be billed in a hospital based clinic are an integral part o the plan o care. “Incident to” services can
since “incident to” rules do not apply to hospital settings and are only be provided in a nonhospital based (private) o ce (POS 11) or
thus not a requirement or these services in that setting. Below is a patient’s home, but never in a hospital setting. Several criteria must
summary o services that can and cannot be per ormed with split/ hold true to allow or this billing.
shared billing (see Table 30-1). 1. The care by the NPP is provided under the direct supervision
The aggregate work in a split/shared service ollows CPT E/M rules o the physician. The physician must have had an initial ace
such that work is added up rom 12:01 a m to 11:59 pm, and not over to ace visit with the patient at some previous point in time to
a 24-hour period. I the NPP sees a patient at 10 pm, or example and establish a plan o care, with the intent o the NPP carrying out
per orms the work o admission, but the physician does not see the the plan o care at (a) subsequent visit(s). Thus “incident to” ser-
patient until 8 a m the next morning, the physician cannot aggre- vices can never be provided to new patients or to established
gate his work the next calendar date with the NPP’s rom the night patients with a new problem. I a patient presents or planned
be ore to submit a claim as a split/shared service. The TP rules or late ollow-up with the NPP, and in the course o the visit, in addi-
night admissions do not hold true or split/shared services. The NPP tion to the expected problem with an established plan o care,
would bill the admission (initial inpatient visit [99221-99223]) under the NPP uncovers a new problem, the NPP must now bill the
her or his own NPI, not the physician’s NPI, on calendar date 1, and entire encounter under their own NPI.
202
2. The physician must stay active in the patient’s care. Usually this documentation to improve billing, nor can they add their opinions
is satis ed by the physician seeing the patient at intervals to or personal observations. When using a scribe, the documenta-
C
review, revise, or update the plan o care. The interval o time tion must be very clear that the scribe “A” is unctioning purely to
H
between physician visits is le t up to the clinical judgment o document physician “B’s” ndings and plans. Practices that employ
A
P
the physician, the speci c patient needs, and standards o care. NPPs as scribes need to be very care ul to have protocols in place to
T
3. The physician must be physically present in the o ce or suite ensure that the documentation clearly ref ects the physician’s work,
E
and be immediately available during the time the NPP provides and not the clinical work o the NPP while serving as scribe. Physi-
R
the service. The physician does not have to be physically pres- cian B needs to attest that the documentation o history, physical
3
ent inside the individual treatment room with the NPP and and decision making ref ects his own personal work.
0
patient. It is help ul to keep records o physician schedules
in the event o an audit to aid in supporting their presence. I ONLINE RESOURCES
B
however the physician must leave the o ce while the NPP is
i
l
l
i
n
providing a planned “incident to” service, i the physician’s part- Centers or Medicare and Medicaid Services. Medicare Claims Pro-
g
ner is physically present in the o ce and immediately available cessing Manual: Section 100. Chap 12. CMS Web site. Avail-
i
n
to cover, the visit can still be submitted under the physician able at: www.cms.hhs.gov/manuals/downloads/clm104c12.pd .
t
h
as an “incident to” service. Otherwise, the visit must be billed Accessed December 14, 2015.
e
under the NPP’s NPI.
T
Centers or Medicare and Medicaid Services. Medicare Claims Pro-
e
a
Incident to services do not require the billing physician to have cessing Manual: Section 30.6.1.B. Chap 12. CMS Web site. Avail-
c
h
a ace to ace visit with the patient on the date o service. In addi- able at: www.cms.hhs.gov/manuals/downloads/clm104c12.pd .
i
n
tion, or billing purposes, “incident to” services do not require the Accessed December 14, 2015.
g
supervising physician to document anything in the medical record.
S
Medicare Learning Network. Available at: https://www.cms.gov/
e
However, based on physician com ort level, and certain state or
t
Outreach-and-Education/ Medicare-Learning-Network-MLN/
t
i
individual institution internal policies, the supervising physician
n
MLNProducts/ Downloads/ Teaching-Physicians-Fact-Sheet-
g
may need to sign each note. No modi er is necessary or submitting
ICN006437.pd .
a
these claims.
n
d
B
SCRIBES
i
l
l
i
n
Occasionally practices employ NPPs as scribes. A scribe’s role is to
g
document an encounter or a physician, in the medical record. Scrib-
w
i
ing is not reimbursable. Scribes are not allowed to enhance or alter
t
h
A
d
v
a
n
c
e
d
P
r
a
c
t
i
c
e
P
r
o
v
i
d
e
r
s
203
CHAP TER
31 INTRODUCTION
The documentation o the inpatient episode o care ollows the
general principles outlined in Chapter 27: Pro essional Billing and
Coding, and is primarily a record o the patient’s clinical experience.
A number o regulatory changes in the payment or hospital acility
care have occurred over the past decades which subtly change the
documentation requirements and may lead to increased interac-
tions o hospitalists with others in the hospital systems such as clini-
cal documentation improvement (CDI) and utilization review (UR)
Hospital-Driven personnel. A better understanding o these regulatory changes and
their impact on both acility reimbursement and patient nancial
Documentation liability will help the hospitalist document care more completely. In
addition, there will be ewer interruptions and queries, less back-end
work or acilities and ewer surprises or the patients when their
hospital bills arrive.
Jeannine Z. Engel, MD
■ DIAGNOSTIC RELATED GROUPS AND PAYMENT
Diana L. Snow, MA, CCS, CPC, CHC FOR HOSPITAL SERVICES
Predetermined payment based on diagnosis
Medicare pays hospitals or inpatient services using the inpatient
prospective payment system (IPPS). Although IPPS is only applicable
to Medicare, the concept o prospective payment or inpatient stays
has been adopted by most third party payers. The primary driver o
payment in the inpatient setting is a grouping methodology, called
a diagnostic related group (DRG).
The basic premise o a DRG system is prospective payment, or
a predetermined, xed amount. The original DRG system began in
the early 1980s in some states, with congress implementing the IPPS
or Medicare services in 1983. Prior to the introduction o the DRG
system, hospital reimbursement was made on a percentage o costs,
and hospitals billed or the actual costs or an inpatient admission
and received a true “ ee or service.”
Each DRG payment is based upon an analysis o prior claims
data in regard to how much it costs, and how many resources are
required on average, to treat patients o that type. There are modi -
cations to this payment amount or hospitals based on location, the
percentage o low income patients treated, teaching hospital status,
and whether or not the speci c case was “unusually costly.” DRGs
are three digit numerical assignments which are divided into cat-
egories, called major diagnostic categories (MDCs) based upon the
patient’s symptoms, comorbidities, whether the care provided was
medical or surgical, the organ system a ected, age, and discharge
status. Each DRG was intended to be a re ection o the resources
required to care or patients in speci c categories, and is designed
to make one single payment to cover all o the services provided by
the acility during the inpatient stay.
Modification of the DRG system

There were initially many limitations to the DRG system, as it was
developed or use with a Medicare population and had limited
applicability to other populations o patients, most notably children
and expectant mothers. As a result, many private payers began
using a modi cation o the DRG system called the All Patient
Re ned DRG system, or an APR-DRG. This was the rst introduc-
tion o adjustments or patient severity o illness and the intensity
o resources needed to care or speci c patients who had mul-
tiple conditions which represent signi cant comorbidities and thus
intensi ed resource requirements.
204
On October 1, 2008, the Centers or Medicare and Medicaid Ser- ■ OBSERVATION STATUS
vices (CMS) ollowed suit and introduced the MS-DRG system: the
C
Short hospital stays
Medicare Severity DRG. Upon initial implementation, the MS-DRG
H
Observation status applies to conditions that do not meet criteria
system changed the number o available code assignments rom
A
as an inpatient, but it is otherwise inappropriate to send the patient
P
500 to 746. The increase in the number o possible code assign-
home. Observation is considered an outpatient stay in the hospital
T
ments was a re ection o the introduction o major complications
E
and is reimbursed under Medicare Part B. This distinction means that
or comorbidities (MCCs), complications or comorbidities (CCs), or
R
a patient will have to pay a coinsurance, which is generally 20% or
noncomplications or comorbidities (non-CCs).
3
Medicare patients. Out o pocket costs can go up signi cantly when
1
ICD 9/ICD 10 codes they are paying a percentage o Medicare approved charges. Addi-
tionally, patients who are classi ed as outpatients are not eligible or
Each year, CMS produces lists o ICD-9/ICD-10 codes which are des-
H
coverage or a Skilled Nursing Facility (SNF) by Medicare.
o
ignated as major comorbid conditions (MCCs) or comorbid condi- Hospitals began using observation services almost immediately
s
p
tions (CCs) which impact the severity and increase the payment o a a ter the IPPS was introduced in 1983. A problem with a capitated
i
t
DRG. Frequently, physician education is centered around these con-
a
system o inpatient hospital payment began to sur ace, speci cally
l
-
cepts to ensure that physicians are adequately documenting these
D
what to do with cases which were short stays and did not approach
r
diagnostic categories so the hospital can be paid appropriately or
i
the median length o stay or a particular DRG, but still ell into that
v
e
services they have provided to patients and the hospital is able to DRG. Patients and their hospital care began to all into the gap
n
report the most appropriate DRG on the claim.
D
between what was expected as an inpatient admission and what
Although the concept o DRG coding sounds simple, it may be
o
was expected as an outpatient service. As part o this, there became
c
very tricky or patients with multiple, complicated conditions. Docu-
u
three distinct categories or patients: Inpatient, Outpatient, and
m
mentation by the treating physicians is the key to understanding Observation.
e
the principal diagnosis. The medical record must contain suf cient
n
t
clarity or a coder to determine which diagnosis is principal. Fre-
a
Review of observation status
t
i
quently, a physician may not know what the principal diagnosis is,
o
In the late 1980s and early 1990s, organizations called Peer Review
n
especially at the beginning o an admission when a patient presents
with unde ned symptoms. It is also common or a patient to be dis- Organizations, or PROs, reviewed admission decisions by providers
charged without a de nitive diagnosis, in which case the symptom and questioned the validity o the provider’s decision. This resulted
treated becomes the principal diagnosis. in providers keeping patients in observation status or weeks at a
Coders determine the DRG code. The coder abstracts, or reads the time rather than risk admitting a patient who did not quite meet
medical record, and determines which diagnosis was the principal inpatient criteria. This resulted in excessive charges to Medicare
diagnosis or the hospitalization. According to the Of cial Guidelines and to the Medicare bene ciary, as outpatient stays or Medicare
or Coding and Reporting, the Principal Diagnosis is the “condition, are billed as a coinsurance (20%, copay or instance) while the
a ter study, which occasioned the admission to the hospital.” This bene ciary charges or an inpatient stay are a set amount called
principal diagnosis assists the coder in determining the category o the deductible ($1260 deductible in 2015). As a response CMS
the MS-DRG code. The other two main actors in determining the implemented a 48-hour target or observation services stating that
proper DRG assignment are (1) whether or not a surgical procedure it would be “unusual” or a patient to stay in observation longer. Cur-
was completed during the stay, and (2) whether or not there are rently, the CMS manuals state, “Observation care is a well-de ned
complicating or comorbid conditions present. When a coder is set o speci c, clinically appropriate services, which include ongoing
unclear about what the documentation is attempting to convey, short-term treatment, assessment, and reassessment, that are ur-
he/she will submit a query to the physician. Many physicians have nished while a decision is being made regarding whether patients
expressed consternation at the lack o clarity in coder queries; will require urther treatment as hospital inpatients or i they are able
speci cally that it is unclear what the coders are asking o the physi- to be discharged rom the hospital.”
cians. Because it could be nancially advantageous to the acility or Issues with observation status
a physician to document a speci c diagnosis, there are pro essional
standards regarding the wording o queries which are not “leading” Observation status has been problematic or many hospitals rom
the physician to answer a certain way. For example: both operational and payment perspectives. CMS has issued addi-
tional instruction on the use o observation services almost every
Documentation states, “Obtunded patient admitted with 3-day year or the past 20 years. Each year, they have changed the rules
history o nausea and vomiting. CXR revealed right lower lobe and the payment modeling slightly. For a period o time in the early
(RLL) pneumonia. Clindamycin ordered.” 2000s, hospitals could only receive additional observation payments
Inappropriate Query, “Is the patient’s pneumonia due to or patients who had congestive heart ailure, chest pain, or asthma,
aspiration?” as CMS had recognized that those were the categories most likely
Appropriate Query, “Can you speci y the etiology o the patient’s to require a longer assessment than an Emergency Room visit, but
pneumonia? It is noted in the admitting H&P this obtunded a shorter visit than the average inpatient admission or those condi-
patient had a history o nausea and vomiting prior to admis- tions. Several years later, CMS reversed this policy and implemented
sion. I the etiology o the pneumonia can be urther speci ed, di erent policies allowing any patient with any condition to be
please document the type/etiology o the pneumonia in the treated under observation, but they would only pay the hospital i
progress note.” the stay lasted at least 8 hours.
Hospital payment is directly tied to the quality o the clinical doc-
umentation provided by the practitioners treating the patient. In the ■ RECOVERY AUDIT CONTRACTORS
inpatient setting, there are many coding rules regarding what docu- Further changes in the payment and de nition o inpatient services
mentation can be utilized, and although there are exceptions to this or Medicare bene ciaries were largely shaped by the actions o and
rule, in general, the patient’s attending documentation carries the reactions to the Recovery Audit Contractor program.
nal weight i there are inconsistencies or contradictions regarding Congress created the Recovery Audit Contractors (RACs) in
the patient’s diagnosis rom residents, consultants or specialists. 2003 as part o the Medicare Modernization Act (MMA) to recover
205
underpayments and discover overpayments to the Medicare Trust the time o the inpatient order. For instance, time in the emergency
Fund. The RAC program began as a demonstration project in department, in observation, in the clinics, in the operating room
selected states (NY, FL, SC, CA, MA, and AZ). From 2005 to 2008, or other procedures all count toward to the 2MN. In addition, i a
the program recovered nearly $38 million in overpayments to the patient is trans erred rom an outside acility, the accepting acility
P
A
Medicare Trust und, although these numbers may not include can count the time spent in the outside acility’s ED toward the
R
claims that were appealed. In 2006, Congress wrote into law (Tax 2MNs. One example: i a patient is originally placed in observation
T
Relie and Health Care Act o 2006) a permanent and national because he is not expected to require 2MN o care, but does not
I
Recovery Audit Program to recoup overpayments associated with recover as expected and requires an additional night o medically
services made under part A or Bo title XVIII o the Social Security Act necessary hospital services, the order can be changed to inpatient
(Medicare payments). While the RACs reviewed a number o areas on hospital day #2, and the entire stay can be billed under Part A
o potential overpayments, the largest area by dollar amount was because the patient stayed in the acility or a total o 3 days and
T
h
short-stay (1-3 days) inpatient hospitalizations. The records or these 2MN. However, to quali y or skilled nursing acility care, the ben-
e
hospitalizations were reviewed, and the DRG payments denied and e ciary must still have a three midnight inpatient hospitalization,
S
p
recouped, with the rationale “inpatient stay not reasonable and nec- speci cally, three midnights a ter an inpatient admission order. In
e
essary, services could have been provided in an outpatient or obser- the example above, only one midnight was spent in the acility a ter
c
i
a
vation setting.” The ull payment amount was recouped, and the the inpatient order was written, there ore only one midnight would
l
t
y
acilities had no mechanism or recovering a portion or the entire count toward the SNF quali ying stay.
o
payment amount except to appeal inpatient denial. Even i the The 2MN rule de ned several new terms, the 2MN presumption,
f
H
acility agreed that the patient could have received the service as an the 2MN expectation and the 2MN benchmark. The presumption
o
s
outpatient, or in observation status, to agree to this meant or eiting essentially states that medical reviewers (auditors), in the absence
p
i
all o the DRG payment. This lead to an unprecedented number o o evidence o gaming, should not review claims or patients who
t
a
l
appeals by acilities around the country, and the ederal appeal sys- receive 2MN o hospital care a ter the inpatient order, as it will
M
tem was overwhelmed. The result was a backlog o appeals being be presumed that the inpatient admission was reasonable and
e
d
processed, and millions o dollars tied up in the appeal system. necessary. The 2MN expectation and 2MN benchmark are used
i
c
In March 2013, CMS issued an interim ruling which allowed acili- interchangeably. These terms de ne the premise under which the
i
n
e
ties to rebill Part A claims which were denied by a CMS contractor admitting provider makes the decision to write an admission order,
a
or the inpatient admission deemed “not reasonable and necessary,” using the in ormation available a t the time of admission. That is,
n
d
to Part B. In addition, they released a proposed rule de ning which physicians and other providers should generally admit as inpatients
S
services could be rebilled under Part B. This proposed rule increased the Medicare bene ciaries they expect will require two or more mid-
y
s
the number o services that could be rebilled or Part B payments, nights o hospital services, and should treat most other bene ciaries
t
e
m
making this “Part B rebill” an attractive option in certain circum- on an outpatient or observation basis.
s
stances. In essence, the rule allowed acilities to rebill the denied The 2MN rule was and is a departure rom the previous method
o
inpatient claim as i the services had originally been provided in an o determining inpatient status, which involved screening criteria
f
C
outpatient setting/status. Later in 2013, the IPPS (IPPS) proposed o severity o illness and intensity o service. And, this rule applies
a
r
and then Final Rules (1599-F) wrote into law a new de nition o to Medicare bene ciaries only, not other payers. In order to help
e
“inpatient” or Medicare Bene ciaries, and the “two Midnight Rule” providers and acilities understand the rule, and allow CMS to
was born. determine the need or more education, they immediately insti-
tuted a “probe and educate” period or all acilities. CMS instructed
■ TWO MIDNIGHT RULE AND PROBE AND the Medicare Administrative Contractors (MACs) to review a small
EDUCATE AUDITS sample o Medicare Part A inpatient hospital claims spanning 0 or 1
On October 1, 2013, the two midnight (2MN) rule took e ect or midnight a ter ormal inpatient admission to determine the medical
Medicare Bene ciaries, and the RAC audits or short stays were necessity o the inpatient status in accordance with the 2MN bench-
placed on temporary hold. The de nition o an inpatient changed mark. Each acility then received eedback, and i required, a second
rom a subjective one o potential risk and intensity o service, to a sampling o 1MN stays. This process required more time than origi-
mostly objective time-based payment model. Per CMS: I a Medicare nally expected, and the probe and educate period was extended,
bene ciary requires medically necessary hospital services or two or rom March 2014 to September 2014, then March 2015, April 2015,
more midnights, then the acility may bill or the services under Part and most recently to September 2015. During this period, the RAC
A, as a DRG. I the patient is expected to require only 1MN o services, audits or medical necessity related to inpatient status were placed
the provider should place the patient in observation, and the acility on hold, and all admissions during this period are prohibited rom
should bill or the stay under Part B. The only statutory exception is uture review related to inpatient status by the Recovery Audit
a patient receiving a procedure on the Medicare inpatient-only list. Contractors.
These patients must receive their care as an inpatient in order or the
acility to receive payment or the surgical procedure. ■ CERTIFICATION REQUIREMENTS
There are certainly nuances to this rule. The most straight orward In addition to the 2MN rule, the 2014 IPPS nal rule (1599-F) outlined
are the exceptions to the 2MN time rame. These include: patients new certi cation requirements or all inpatient hospitalizations.
who leave Against Medical Advice (AMA) prior to the second While certi cation was not a new concept, it had been in place pri-
MN; patients who die unexpectedly; patients who are discharged marily or long stays and payment outliers. For 14 months, until most
to hospice unexpectedly; patients requiring new, unexpected o the Certi cation requirements or all inpatient admissions were
mechanical ventilation; and patients who have an unexpected clini- removed in January 2015, many hospitals and providers scrambled
cal improvement and are discharged prior to the second MN. CMS to understand and comply with the certi cation requirements. As o
is clear that these exceptions should be rare, and that documenta- January 1, 2015, the only portion o these short-lived requirements
tion would need to support the Part A billing o 1MN stays in these which remained or all inpatient hospitalizations was the inpatient
circumstances. order requirement. CMS de ned the admission order requirements
Another change was start time or the 2MN. The clock starts at the in the Code o Federal Regulations, making this a condition o pay-
time the patient begins to receive care at the acility, regardless o ment in ederal law under Medicare Part A. The inpatient admission
206
order must be present in the medical record, and supported by physician director o CM/UR), particularly i the hospitalist does not
admission and progress notes. The order must be “ urnished by a respond, or is not moving in the direction that seems appropriate
C
quali ed and licensed practitioner who has admitting privileges at to the UR nurse.
H
the hospital as permitted by state law.” Other requirements include Because the downgrading o a Medicare patient rom inpatient
A
P
amiliarity with the patient’s case, and the timing o the order should status to outpatient status involves potential increased nancial
T
be at or be ore the time o inpatient admission. The regulations impact or the bene ciary, CMS requires that the acility ollow cer-
E
speci cally state that the admitting practitioner cannot delegate tain speci c steps i they change a patient rom IP to OP (observa-
R
the order to another individual who does not meet the criteria, but tion) while the patient is still hospitalized. This process is called CC44.
3
subregulatory guidance does address trainees and verbal admission The basic steps include noti ying the patient, and providing nancial
1
orders. The expectation is that any inpatient admission order placed counseling, in addition to actually changing the order. Upgrading
by a resident, mid-level provider (without admitting privileges) or status rom Observation to Inpatient only requires the written order.
H
verbal inpatient admission orders will be signed by an authorized The 2014 IPPS nal rule also allowed acilities to change the bill-
o
s
provider prior to the patient’s discharge rom the hospital. The ing status o inpatient hospital stays which are internally audited, as
p
i
t
requirement to have delegated inpatient orders signed prior to dis- long as the billing change occurs within 1 year o the date o service
a
l
charge has been challenging to success ully implement or acilities (timely ling). This regulatory change allowed acilities to put in
-
D
around the country, especially or short stays. Process choices have place back-end review processes or 1MN inpatient stays and decide
r
i
v
included close tracking with e-mail or other reminder noti cations, whether to leave the billing under Part A (DRG) or change the billing
e
n
hard stops on D/C orders or actual discharges, and back-end rebill- to Part B, i the documented care did not ul ll the 2MN expectation/
D
ing or IP stays that do not meet this requirement. benchmark. Facilities must ollow the Utilization Review Conditions
o
c
It may be help ul or Hospitalists to view the 2MN rule and the o Participation, which require noti cation o the patient and the
u
m
inpatient order requirements as a payment policy as opposed to attending provider within 2 days o the decision to change the bill-
e
one which de nes the care provided to the patient. CMS and its ing rom IP (Part A) to OP (Part B). The primary reason to change the
n
t
contractors are not making judgments about the need or or the billing status to Part B is that the patient received care in the acility
a
t
quality o care provided to Medicare Bene ciaries. They are de ning or only 1MN, there was a signed inpatient order, but there was no
i
o
n
the parameters under which they will pay or this care. Under certain documentation to support that 2MN o medically necessary hospital
parameters (signed admission order and greater than 2MN stay or care was expected at the time o admission.
one o the exceptions), the care will be paid under Part A, as a DRG.
Failing these parameters, the care will be paid through Medicare ■ INPATIENT DOCUMENTATION OF CARE:
Part B as an outpatient hospitalization. NUTS AND BOLTS
Adding to the con usion created by an onslaught o regulatory
■ ROLE OF UTILIZATION REVIEW
changes is the multitude o payers and their di ering rules or sta-
Getting the patient status correct is not always obvious, and it is tus decisions and payment. Clinicians could spend all o their time
a single action (the admission order) embedded in a very busy thinking about these issues, and lose sight o the patient and their
sequence o evaluating and starting the course o care or a new care. For inpatient hospital care, the documentation requirements
hospitalized patient. Most acilities have a team o nurse review- include a history and physical exam (admission note), daily progress
ers who are looking at the admission in ormation, and helping the notes and a discharge day note. These notes are used or pro es-
admitting team make this decision. These pro essionals, generally sional billing. For the hospital to bill or the inpatient care, a signed
named UR nurses, have the training and expertise to understand inpatient admission order is also required, as well as a discharge
the nuances o admission status or all payers. Some acilities have summary.
divided Case Management and UR, while other acilities may have
the same team members per orming both o these tasks. UR teams
have been trained to use screening criteria such as Milliman or Inter- PRACTICE POINT
Qual to determine i a patient meets inpatient or observation status.
Medicare patients
These criteria are based in medical literature, but also are clear that
they are guidelines, and are not meant to overrule the provider’s • Facility Inpatient Part A billing = signed inpatient order and two
medical judgment. Interestingly, the Medicare Conditions o Partici- midnight medically necessary hospital care.
pation (COP) require that acilities employ some type o screening
criteria, but do not designate which one. With the implementation
o the 2MN rule, the use o screening criteria became essentially The best documentation o hospital care tells the patient’s story,
moot or Medicare patients, as they now de ned inpatient in terms and the provider’s plan or care. There are three levels o services or
o hospital care and time, not intensity o service and severity o ill- initial hospital care, CPT codes 99221, 99222, and 99223.
ness. However, the COP or UR did not change, and screening crite- The only signi cant di erence in documentation requirements
ria are still required. In addition, many commercial payers, and some between the three service levels is the medical decision making
state Medicaids still rely on these screening criteria to determine (Table 31-1). While the lowest admission service may be billed with
inpatient versus outpatient status. a lower level history and examination, in general, when a patient is
The interaction o the hospitalist with the UR nurses o ten sick enough to require inpatient admission, there is usually medi-
involves a conversation around the expected length o stay o a cal necessity to document a comprehensive history and physical
Medicare patient on hospital day two. I the patient has an order or examination. In order to demonstrate the medical decision mak-
observation, and it appears the patient will stay an additional MN, ing and cognitive work per ormed, any personal work should be
the UR nurse will suggest that the patient have an IP order written. documented, including personal review o ECGs, x-rays, or other
I the patient has an IP order, and it appears that the patient will tests, conversations with consultants, pertinent history obtained
be discharged, the UR nurse will discuss whether or not the status rom amily members, and any review and summary o old records.
should be down-graded to Observation, which will require a pro- This work is counted in the “data” section o medical decision mak-
cess called condition-code 44 (CC44). Any o these interactions may ing, along with tests ordered and reviewed. Risk to the patient may
involve an additional physician or secondary review (usually the be indicated by presenting problems, the diagnostic procedures
207
day note or discharge summary outlining the clinical course and
TABLE 31-1 Documentation Requirements for Hospital-based unexpected recovery, i this occurred. Again, the goal is to tell the
Professional Billing (All Three Elements Needed: patient’s story: this patient looked very ill when they arrived; we
History, Exam, and MDM)
expected 2 to 3 days o IV antibiotics so we wrote an IP order, but
P
A
Inpatient Admit 99221 99222 99223 they recovered more quickly than we expected, and we were able
R
to send them home a ter 1MN o outstanding care.
Observation 99218 99219 99220
T
Inpatient acility billing requires a discharge summary (both
Observation/same 99234 99235 99236
I
Medicare Conditions o Participation, and most accreditation bodies
day D/C
such as The Joint Commission), and it is a use ul tool or summariz-
History: CC and 4 HPI 4 4 ing the story o a patient’s hospitalization. Pro essional billing allows
2 ROS 10 10 or a discharge day note (99238, 99239) which is time based. The
T
h
1 PFSH 3 3 documentation requirements are slightly di erent, but it is simple to
e
(DET) (COMP) (COMP) combine these two documents into a single note, saving time and
S
p
Exam 5 (DET) 8 (COMP) 8 (COMP) duplication o e ort. Both require a discussion o the hospital stay,
e
c
discharge instructions including medications and plans or ollow-
i
Medical decision Straight- Moderate High
a
up, as well as the attending provider’s signature. The discharge day
l
t
making (MDM) orward/low
y
note requires a ace-to- ace service rom the billing provider, a nal
o
Time (admit only) 30 50 70 examination, and some notation o time spent. The Joint Commis-
f
H
sion standards also include procedures and treatments provided
o
s
and the patient’s discharge condition. The authors suggest that
p
i
hospitalists discuss the expectations o the discharge summary with
t
a
ordered, or management options selected. The risk table which
l
their acility, and work to make a discharge day template which
M
most coders and auditors use as a re erence does not include all dis- includes all o the required elements or both the pro essional
e
d
ease processes, so the documentation o risk to the patient, whether day o discharge service and the discharge summary into a single
i
c
through literature re erences (eg, high PESI or PSI scores) or your note. To prevent scrutiny o the discharge day service payment,
i
n
expert opinion is a reasonable part o the admission and subsequent
e
the documentation needs to clearly re ect that a ace-to- ace visit
a
hospital day documentation. Notations such as “stable on pressors” occurred on discharge day. I residents are documenting this service,
n
d
can be misinterpreted by nonclinical personnel. The admission then only the billing provider’s time can be counted toward the
S
note should include all o the patient’s problems which are being discharge day time. Mid-level provider time may be combined with
y
s
addressed or treated, including chronic medical conditions, i pre- attending MD time and billed as a split-shared service. This time
t
e
scribing chronic medications, or the condition a ects the patient’s
m
may be discontinuous, and you can count time that is spent in the
s
current reason or hospitalization, include it in the assessment. days prior to discharge, as long as you only “count” this time once.
o
The authors avor problem-based notes, ollowing the pattern And, i your partner spends time on discharge on Sunday, and you
f
C
o : problem, attribution, and plan. Each problem should include are the actual discharging provider on Monday, then your partner’s
a
r
the provider’s clinical reasoning or a di erential diagnosis (attri- time is combined with your time, and all billed together on Monday,
e
bution), and some indication o the patient’s short-term risks as there can only be a single discharge day service billed per hospi-
and severity o illness. What makes this particular patient need talization. I no documentation o time is included in the discharge
to be in the hospital? Why is this illness particularly risky or this day note, then 99238 will be billed. In order or 99239 to be billed,
patient? The proposed plan o care with an expectation o length the provider must document that he or she spent greater than 30
o stay should be stated outright, or implied. Instead o stating “IV minutes coordinating the patient’s discharge.
ce epime,” include an expected length o treatment: “IV ce epime
or 2 to 3 days, or until a ebrile or 24 hours.” When cosigning or
attesting a ter residents, students or advanced practice clinicians, Observation documentation of care
the responsible physician should include additional in ormation to Observation documentation requires an admission note (initial
augment the documentation, particularly i these providers down- observation services 99218-20), updates in the patient’s clinical
play the severity o a patient’s presentation, or do not have the condition at some appropriate interval by providers or nursing
knowledge or experience to recognize the signi cance o certain sta , and a discharge note (observation care discharge, 99217). I
clinical scenarios. the observation services occur within a single calendar day, the
Daily progress notes should include all o the problems rom the services must span greater than 8 hours to be billed to Medicare
admission note until resolved, and any new problems. Hospitalists as the combined observation admission/discharge on the same
should be aware o in ormation carried orward via a macro to be date (99234-36). Documentation that re ects less than 8 hours o
sure that the documentation is accurate and makes sense or the observation services on a calendar date may only be billed as initial
date o service. Resolved problems should be deleted (eg, that observation services (99218-20) with the reimbursement or the
hypokalemia rom 3 days ago, now resolved). Daily progress notes discharge work bundled into that payment or Medicare billing;
which include pages o radiographic and laboratory data rom the however, this 8-hour rule or observation billing does not necessar-
past week will o ten be ignored by a reader due to their prohibitive ily apply to private insurance providers. These time-based require-
length. Consider re erring to laboratory test and x-rays as ordered or ments may vary among private payers and Medicaid. There is no
reviewed and only include pertinent details in your notes. Linking requirement or a discharge summary in order or the acility to bill
your orders with a statement in your notes “see my orders or any observation services. The decision to require a discharge summary
additional plans,” also adds all o your orders to the daily plan and or observation stays should be local and acility-based, and is o ten
medical decision making. related to having a concise summary o the patient’s hospitalization.
A “quality” note generally has the ollowing attributes: concise, The basic documentation requirements or initial observation care
clear, current, organized, prioritized, and trustworthy. I a Medicare are identical to those or initial inpatient care (Table 31-1). In order
patient is discharged earlier than expected, especially a ter 1MN or the acility to bill or observation services, an order or observa-
o care, the provider should include a statement in the discharge tion is also required.
208
The uncertain future (a last minute update) hospitalist to comprehensively document the patient’s story and
the provider’s medical decision making, while decreasing the back-
C
CMS has issued the Calendar Year 2016 Outpatient Prospective
end work and queries. The classic admission note, daily progress
H
Payment (OPPS) Proposed Rule (CMS-1633-P) which adds another
notes and discharge summary that we all learned in medical school
A
wrinkle to the payment o short stay inpatient hospital services. The
P
proposed changes include allowing or Part A payment on a case are still the backbone o hospital documentation.
T
by case basis based on the judgment o the admitting physician,
E
SUGGESTED READINGS
R
despite the expectation o less than 2MN stay at the time o admis-
sion. In addition, the task o reviewing short hospital stays will now
3
1
all to the Quality Improvement Organizations (QIOs) with re erral to CFR § 412.3 Admissions.
the Recovery Audit Contractors or acilities that have “high” error CMS, 2013. Medicare Learning Network: Acute Care Hospital Inpatient
rates. The details o the review process have not yet been de ned
H
Prospective Payment System; 2013. http://www.cms.gov/Outreach-
o
at this time. The CY16 OPPS Final Rule will be published around and-Education/Medicare-Learning-Network-MLN/MLNProducts/
s
p
November 1, 2015 or implementation on January 1, 2016. downloads/AcutePaymtSys ctsht.pd .
i
t
a
CMS 1455-P Proposed Rule. Medicare Program: Part B. Inpatient Bill-
l
-
CONCLUSION
D
ing in Hospitals. Q&ARelated to Patient Status Reviews. http://www.
r
i
Documentation o a patient’s hospital care has become increasingly
v
cms.gov/ Research-Statistics-Data-and-Systems/ Monitoring-
e
complex as the regulations surrounding acility billing or inpatient
n
Programs/Medicare-FFS-Compliance-Programs/Medical-Review/
D
and observation care have changed, and continue to evolve. Every Downloads/Questions_andAnswersRelatingtoPatientStatusRe-
o
year, the Center or Medicare and Medicaid Services publishes new views orPosting_31214.pd .
c
u
rules which alter the billing regulations or both pro essional and
m
CMS 1633-P Proposed Rule. https://www.cms.gov/Medicare/Medicare-
acility services. Commercial payers may have slightly di erent
e
Fee- or-Service-Payment/ Hosp italOutp atientPPS/ Hosp ital-
n
interpretations. All o these services use provider’s notes and orders
t
Outpatient-Regulations-and-Notices-Items/CMS-1633-P.html.
a
to explain and justi y the medical necessity or the services ren-
t
i
o
dered. Some knowledge o these billing regulations will allow the CMS 1455-R Interim ruling, March 13, 2013 (18 pages).
n
209
CHAP TER
32 INTRODUCTION
The transition rom ICD-9 to ICD-10 and to uture ICD revisions is
challenging but also provides unique opportunities to improve
health care and disease/procedure tracking and data. This tran-
sition will better support current technology and health care
re orm initiatives including the Centers or Medicare & Medicaid
Services (CMS) value-based purchasing and pay- or-per ormance
programs; coordinated care models such as accountable care
organizations and patient-centered medical homes; the govern-
Taming the ICD-10 ment’s Physician Quality Reporting System; and the move toward
adoption o electronic health record systems. In addition, it a ords
Monster great potential to decrease audit risk since ICD-10 codes allow the
physician’s documentation to be translated into a more accurate
clinical picture, thereby reducing the chance o misinterpretation
by third parties, auditors (eg, recovery audit contractors—RAC),
CoLette Morgan, MD, FHM, CCDS, CDIP and attorneys.
With improved speci city, ICD-10 codes help health care pro-
Yvette M. Cua, MD viders submit highly speci c codes or the care they provide, and
better re ect severity o illness and support medical necessity. This
will hope ully re ect in a physician’s pro ling on national registries,
consumer health sites, and also with health care entity’s quality
reporting.
HISTORY
Classi ying illnesses to document disease prevalence and causes
o death is not a new concept. The rst attempts were made as ar
back as the 1600s. Initially, it was nothing more than a crude and
inconsistent nomenclature with little reliable data. However, during
the 1800s, an earnest e ort began to create a uni ormed system.
Medical statisticians were commissioned to embark upon this
enormous task. They began their work by using the Bertillion Clas-
si cation o Causes o Death developed by the French statistician,
Jacques Bertillion.
In 1893, the rst international classi cation o diseases (ICD) was
adopted by the International Statistical Institute. The United States
did not utilize a ormal disease classi cation system until 1898, when
the American Public Health Association recommended all o North
America adopt this system, and recommended it be updated every
10 years.
Each revision made the ICD more detailed. In 1948, the World
Health Organization assumed responsibility or the ICD and pub-
lished the sixth version, which incorporated morbidity or the rst
time. At this time, they decided to make ICD the of cial means o
collecting international data or epidemiological surveillance and or
health management. The WHO Nomenclature Regulations, adopted
in 1967, stipulated that Member States use the most current ICD
revision or mortality and morbidity statistics. In 1979, the United
States adopted ICD-9, and in 1983 the Inpatient Prospective Pay-
ment System in the US required ICD-9 codes to be linked to diagno-
sis related groups (DRGs) or reimbursement. ICD-9 quickly became
antiquated, running out o room or new codes or new diseases
and new technology. The 43rd World Health Assembly endorsed
ICD-10 in May 1990, however, the United States did not adopt it
until October 1, 2015. ICD-10 is dramatically more robust and allows
or many more codes than any previous edition, including ICD-9
which has been used in the United States since 1979 (Table 32-1).
The 11th revision process is underway and the nal ICD-11 will be
released in 2017.
210
TABLE 32-1 Differences Between ICD-9 and ICD-10 Code Sets TABLE 32-2 Unusual ICD-10 Codes and Their Analogous
C
ICD-9 Codes
H
ICD-9-CM ICD-10 Code Sets
A
3284 Codes 71,924 Codes ICD-10 ICD-9
P
Procedure
Accidental injury due to Accident caused by paintball
T
Codes 14,025 Codes 69,823 Codes
paintball discharge W34.011 gun, E922.5
E
Diagnoses Old New
R
Hurt/ all walking into a Activities involving walking,
Diagnosis ICD-9-CM ICD-10-CM
3
lamppost, initial encounter— marching and hiking, E001.0
structure
2
• 3-5 characters • 3-7 characters W2202XA Other all, E888.8 (required two
• First character is • Character 1 is alpha codes)
numeric or alpha
T
• Character 2 is numeric Burn due to water-skis on ire, E831 Accident to watercra t
a
m
• Characters 2-5 are • Characters 3-7 may be initial encounter—V9107XA causing other injury
i
numeric
n
alpha or numeric Problems in relationship with Con lict amily V61.9
g
Procedure ICD-9-CM ICD-10-PCS (only used in-laws—Z63.1
t
h
structure in the inpatient hospital
e
• 3-4 characters
setting)
I
C
• All characters are
D
numeric • ICD-10-PCS has seven
-
any additional encounters with the provider or this same condition
1
characters
0
• All codes have would code to T75.82XD.
M
at least three • Each can be either
There are not yet ICD-10 codes or “injured while texting,” or “dis-
o
characters alpha or numeric
n
tracted by cell phone while driving.” However, with the ability ICD-
s
• Numbers 0-9; letters
t
10 has to expand, one would expect such codes to be added soon.
e
A-H, J-N, P-Z
r
SOME CLINICALLY NOTABLE CHANGES IN ICD-10
• Acute MI codes changed rom 8 weeks duration to <4 weeks
BENEFITS AND USES OF ICD
(28 days) duration.
• The ICD is the global health in ormation standard or codes to • New terminology or asthma rom worldallergy.org.
allow comparison o mortality and morbidity statistics. This sys- • Additional codes need to be used to identi y resistance to
tem helps to monitor death and disease rates worldwide and antimicrobial drugs (Z16-), whenever in ection with resistance
measure progress toward the Millennium Development Goals. is documented (V09 was used for all in ICD-9).
• About 70% o the world’s health expenditures (USD 3.5 billion) are • A new code Z91.120- was created or “under dosing due to
allocated using ICD or reimbursement and resource allocation. nancial hardships.”
• It is the diagnostic classi cation standard or all clinical and • Injuries grouped by body part, rather than category o injury
research purposes. These include monitoring o the incidence (eg, emur racture will be under emur rather than racture).
and prevalence o diseases, observing reimbursements and • Three di erent categories or pathologic ractures (due to
resource allocation trends, and keeping track o sa ety and neoplastic disease, due to osteonecrosis, and due to other
quality guidelines. It allows the counting o deaths as well as speci ed disease).
diseases, injuries, symptoms, reasons or encounter, actors that
in uence health status, and external causes o disease.
ICD-10 IMPACT ON SOME COMMON HOSPITAL
The United States was one o the last industrialized nations to adopt DIAGNOSES
ICD-10. They actually use considerably more codes, about 90,000 ■ HYPERTENSION
o those in ICD-10, than most other nations. They use ICD-10 codes
• Some o the speci city o hypertension has been eliminated—
in the outpatient setting, and are one o the ew countries that use
ICD codes or billing. These actors played a major role in the huge there is only one code (I10) or benign, accelerated, malignant,
monetary investment health care entities made in switching rom hypertensive urgency, hypertensive emergency, and hyper-
ICD-9 to ICD-10, and are speculated to have led to the numerous tensive crisis. Despite lumping o these hypertension types, it
delays in implementation. is still important to describe hypertension completely to show
severity o illness.
• There are still di erent speci c hypertension codes i it is second-
WHY DOES ICD-10 HAVE SO MANY UNUSUAL CODES? ary to heart (I11), kidney (I15), or both diseases combined (I13).
ICD was developed to track diseases and procedures or morbidity
and mortality data, not or billing. Many ICD-10 codes are quoted as ■ ACUTE RENAL FAILURE
unusual, however were present in ICD-9, just less speci c and rarely • Documenting Acute Renal Insuf ciency and “prerenal azote-
used (Table 32-2). An example is the ICD-10 code X52.XXXA which mia” are vague and nonspeci c and usually result in incorrect
is “prolonged stay in weightlessness environment.” ICD-9 has the coding.
code E928.0 which is “e ects o weightlessness in spacecra t.” • ARF and AKI both code to N17.9 and is a CC.
Another point that comes up when comparing ICD-10 to ICD-9 • ATN codes to N17.0 and is a MCC, “vasomotor nephropathy”
is that ICD-10 has codes or initial and subsequent encounters. and “toxic nephropathy” also code to ATN (N17.0).
It is o ten asked, “do we really need a code or ‘getting hit by a
spacecra t’ the rst time and then another code or ‘getting hit by
a spacecra t’ a second time?” The “initial” and “subsequent” terms ■ SEPSIS
actually re er to the care the provider gives, not the condition itsel . • ICD-10 does not o er much clari cation regarding the use o
So, the rst encounter with this patient would code to T75.82XA and the diagnosis sepsis.
211
• Sepsis replaces the term “septicemia” that was used in ICD-9. diagnosis that is most bene cial or reimbursement, but that
• SIRS diagnosis use has changed rom a coding standpoint— selection must be supported by documentation in the chart.
only use or nonin ectious origin (eg, DKA, trauma, burn) SIRS For example, i a patient is admitted with both a COPD and CHF
w/ organ dys unction (R65.11) and SIRS i w/o organ dys unc- exacerbation without any other conditions that qualify as a CC or
P
A
tion (R65.10). MCC, the provider can choose to list COPD exacerbation as the
R
• I SIRS leads to an in ection resulting in “sepsis” or “severe sep- principal diagnosis with acute on chronic LV systolic dys unc-
T
sis,” sepsis supersedes SIRS. So, a nonin ectious SIRS leading tion as the secondary diagnosis, which reimburses higher than
I
to an in ection resulting in severe sepsis is coded as “severe i the acute on chronic LV systolic dys unction was listed as
sepsis” (R65.2-), not SIRS (R65.1-). principal with COPD exacerbation as secondary.
• Sepsis must be speci cally diagnosed under ICD-10 coding COPD exacerbation with acute on chronic LV systolic CHF
since there is not a code or “SIRS due to in ection” as there is as an MCC, maps to MS-DRG 190 with a RW = 1.1743 ($5000 ×
T
h
in ICD-9. 1.1743 = $5871.50). Acute on chronic LVCHF with COPD, only a
e
CC, maps to MS-DRG 292 with a RW = 0.9824 ($5000 × 0.9824 =
S
p
■ CLINICAL IMPORTANCE OF ICD-10 TO $ 4912.00). Assuming a BR $5000, ailing to correctly label
e
c
MS-DRG SELECTION principal and secondary diagnoses would result in $1000 lost
i
a
revenue or the institution.
l
Clinical documentation is paramount to code selection not only or
t
y
the provider’s pro essional ee, but also or the hospital’s MS-DRG Notice that i the patient had a metabolic encephalopathy
o
rom concurrent hyponatremia, this is an MCC. Now listing
f
reimbursement. The medical record must clearly document the
H
acute on chronic LV systolic CHF as the principal diagnosis
o
primary diagnosis, and each actively managed condition, contribut-
s
with metabolic encephalopathy as a MCC, mapping to MS-
p
ing comorbidity, and di erential diagnosis to the highest speci city
i
DRG 291 (RW = 1.5097, $7,548.50) will reimburse higher than
t
known.
a
l
COPD exacerbation with an MCC, MS-DRG 190 (RW = 1.1743,
M
■ DEFINITIONS $5871.50). I a patient has more than one MCC, no additional
e
d
reimbursement is given. It also does not matter which condi-
i
Medical Severity Diagnosis Related Groups (MS-DRG) is a classi ca-
c
tion is counted as the MCC.
i
n
tion system that bases hospital reimbursement or inpatient admis-
e
2. Documentation should re ect the highest level o speci city.
sions on similar resource use or the identi ed principal diagnosis
a
In the example above, i COPD exacerbation is the principal
n
and contributing complications or comorbidities. The majority o
d
diagnosis but only “CHF” is listed in the chart without urther
principal admitting diagnoses have three levels o reimbursement
S
speci ying it as “acute on chronic,” this is a nonspeci c term
y
based on the presence or absence o complications or comorbidi-
s
which does not count as either a CC or MCC; the MS-DRG
t
e
ties (CC) or major complications or comorbidities (MCC). Each MS-
m
would now change to COPD without MCC/CC, which is MS-
DRG is assigned a relative weight (RW). Each acility is assigned a
s
DRG 192 (RW = 0.7190, $3595), and result in a $2276 loss in
base rate (BR) which is the dollar value o a RW o 1. The hospital’s
o
deserved hospital revenue.
f
reimbursement or an admission is calculated by multiplying the
C
3. Complete dif erentia l dia gnosis a nd listing o correctly
a
RW × BR. Thus, the MS-DRG is the key determinant o the hospital’s
r
la beled comorbid conditions. Many patients with COPD exac-
e
reimbursement or an admission.
erbations are ound to have hyponatremia. A signi cant num-
Complications or comorbidities (CC) and major complications
ber o these present with an altered mental status rom the
and comorbidities (MCC) are conditions known to increase neces-
hyponatremia, also called metabolic encephalopathy. How-
sary resources and length o stay (LOS) or a given primary admitting
ever, o ten neither a comment about the sodium, nor mention
diagnosis. This translates into increased hospital reimbursement
o a metabolic encephalopathy ever enters the chart clinical
when they are present. A list o CCs and MCCs can be ound at the
documentation by the clinician within the history and physical
CMS website: https://www.cms.gov/icd10manual/ ullcode_cms/
or progress notes. With lack o any comment on comorbid con-
P0370.html.
ditions, COPD, MS-DRG 192 would only reimburse $3595. With
■ KEY PRINCIPLES OF DOCUMENTATION hyponatremia listed, a CC, the MS-DRG would change to 191,
reimbursing at $4685, and nally with the documentation o
The selection o MS-DRG’s is based entirely on ICD-10 combinations. metabolic encephalopathy, a MCC, the MS-DRG would change
The ICD-10 codes are chosen based on documentation in the medi- to 192, worth $5871.50.
cal record. Although a more speci c diagnosis may actually exist
clinically, without adequate documentation to alert a coder or audi- The medical record should always represent the patient’s true
tor o its presence, the optimal code cannot be submitted. complete medical status. These examples are meant to illustrate
Three key eatures o ideal documentation necessary to empower common documentation pit alls resulting in unnecessary revenue
optimal hospital reimbursement are: loss, but should not be misconstrued or used to alsi y the clinical
picture to “up-code.”
1. Clear identi cation o the primary diagnosis whenever possible.
The patient’s physician, is in the best position to determine the
■ SUMMARY
principal and secondary diagnoses. Coders and auditors who
review the medical record at a later time may never appreciate The transition rom ICD-9 to ICD-10 presents challenges to clini-
the clinical picture as well as the provider. Thus, it is crucial or cians and hospitals, but also o ers unique opportunities to improve
the provider to document their medical judgment regarding health care and disease/procedure tracking and data. A ew notable
all aspects o care, including principal diagnosis. The usual possible bene ts o ICD-10 include:
scenario is that ancillary documentation will support these • Addressing technology and health care re orm initiatives—
statements, and even make it possible or a coder to come to CMS value-based purchasing, pay- or-per ormance programs;
the same conclusion as the provider; however, sometimes that coordinated care models such as accountable care organiza-
is not the case. O ten there may be more than one diagnosis tions and patient-centered medical homes; the government’s
responsible or the admission. In this circumstance, ICD rules Physician Quality Reporting System; and the move toward
allow or the provider or the coder to select the principal adoption o electronic health record systems.
212
• Decreasing audit risk—ICD-10 codes allow the physician’s processing, and public health reporting in the United States. While
documentation to be translated into a more accurate clinical the outcomes and potential bene ts (or risks) o transition to ICD-
C
picture, thereby reducing the chance o misinterpretation by 10 remain to be realized, hospitalists and all clinicians will need to
H
third parties, auditors (eg, recovery audit contractor—RAC),
A
practice within this new documentation environment to optimize
P
and attorneys. clinical care and the accurate representation o patients’ diseases
T
• Improving care rom improved disease tracking—ICD-10 codes and clinical conditions.
E
will generate more detailed health care data and a greater
R
ow o that data to improve medical communication, which
3
can potentially lead to advances in disease management and ONLINE RESOURCES
2
clinical pathways.
CMS ICD-10 in ormation guide. https://www.cms.gov/medicare/
• Improve accurate reimbursement—ICD-10 codes help health care
T
coding/icd10/downloads/032310_icd10_slides.pd .
a
providers submit highly speci c codes or the care they provide.
m
• Better determine severity o illness and prove medical neces- CMS website: History o ICD-10. https://www.cms.gov/Medicare/
i
n
sity—ICD-10 codes better describe the extent and comorbidities Me d ic a re -Co n t ra c t in g / Co n t ra c t o rLe a rn in g Re so u rc e s/
g
involved in a patient’s illness, which in turn acilitates validation downloads/ICD-10_Overview_Presentation.pd .
t
h
e
o a patient’s condition in support o utilization o goods, services Complete MCC and CC tables. https://www.cms.gov/Medicare/
I
C
and complex procedures. Correct utilization o ICD-10 should Medicare-Fee- or-Service-Payment/AcuteInpatientPPS/FY2015-
D
translate into more accurate severity o illness and risk o mortal- IPPS-Final-Rule-Home-Page-Items/FY2015-Final-Rule-Tables.html.
-
1
ity or patients, resulting in a likely more accurate re ection o a Tables 6I and 6K.
0
M
physician’s pro ling on national registries, on consumer health MS-DRG Relative weight les. https://www.cms.gov/Medicare/
o
sites and with health care entity’s quality reporting. Medicare-Fee- or-Service-Payment/AcuteInpatientPPS/FY2015-
n
s
As o October 1, 2015, the ICD-10 coding classi cation became IPPS-Final-Rule-Home-Page-Items/FY2015-Final-Rule-Tables.html.
t
e
r
the new baseline or clinical data, clinical documentation, claims Table 5.
213
SECTION 7
Principles of Medical Ethics and
Medical-Legal Concepts
215
CHAP TER
33 INTRODUCTION
Nearly 2500 years ago, the Hippocratic writers decreed in the
Epidemics, Bk. I, Sect. XI, “Declare the past, diagnose the present,
oretell the uture; practice these acts. As to diseases, make a habit
o two things—to help, or at least to do no harm.” The basic tenets
o ethics apply to all medical specialties. The hospitalist-patient
relationship depends upon a long-standing tradition o practices
and mani estations o pro essionalism in which the physician places
the interests o the patient above his or her own, and practices
Common Indications with competence, integrity, and bene cence. The context o care
determines the application o ethical principles; however the ethical
for Ethics issues encountered in Hospital Medicine share with other orms o
medical practice a rich history in bioethics, social movements, and
Consultation landmark court cases.

The nature o the doctor-patient relationship and the new dichot-
omy o the inpatient and outpatient settings continue to evolve as
specialized care becomes more localized to geographic areas such
as the emergency room, intensive care unit, most recently, general
Joseph J. Fins, MD, MACP medical units, and in the uture, the medical home. This ragmenta-
Matthew W. McCarthy, MD tion o the clinical encounter into a unit o hospitalization represents
a departure rom the time-honored, and almost mythic, longitudinal
Walter Limehouse, MD, MA, FACEP doctor-patient relationship o general practice and primary care.
Unlike the classic doctor-patient relationship, decision making in
the hospital is generally more harried and o a more critical nature.
This may jeopardize the doctor-patient relationship i patients do not
understand the role o hospitalists, perceive that their primary care
physicians have abandoned them, or have questions o trust due to
cultural di erences or other actors. Dedication to ethical practice
preserves stability in a “crisis” and promotes a culture o trust neces-
sary or advocacy and a sound doctor-patient relationship.
Ethics, expertise, and availability o the hospitalist may help bal-
ance patient-centered obligations with the need to e ciently man-
age interventions. Hospitalists should not make initial assumptions
about their patients’ priorities without rst evaluating each patient
with a resh perspective. Communication with the patient’s outpa-
tient provider, amiliarization with the medical record, and meetings
with patients and their amily or riends who may have essential
in ormation to share during the patient’s illness are both good clini-
cal care and congruent with ethical practice.
PRACTICE POINT
• Having a longitudinal perspective rom an outpatient colleague
may mitigate diagnostic and prognostic errors that may occur
when the object o one’s practice is hospitalized patients.
Consultation with an outpatient physician builds a trusting
hospitalist-patient relationship. Patients may suspect potential
conf icts o interest or dual agency between hospitalists and
the patient needs.
The ethical mandate to optimize cooperation, or comanagement,

between doctors and other members o the health care team is an
essential element o the hospitalist model. In hospitals di erent
pro essional specialties have traditionally unctioned, o ten in isola-
tion. Hampered care coordination—and the splitting o the clinical
team—potentially compromises the therapeutic relationship. How-
ever, the hospitalist’s appropriate use o consultants, awareness o
217
TABLE 33-1 Decision-making Capacity and Competence • Discharge/placement
• Do Not Resuscitate (or Allow Natural Death) orders
• Ability to communicate a choice
P
• Understand the nature and consequences o the choice • Isolated incapacitated patient
A
•
R
• Manipulate rationally the in ormation necessary to make Maternal/ etal conf ict
T
the choice
• Medical error
I
• Reason consistently with previously expressed values and
goals • Pain management
• Re usal o recommended treatment
T
• Research ethics
h
e
one’s sphere o practice, and appreciation or continuity o care sus-
•
S
tains trust. Being transparent and sharing the results o consultations Resource allocation
p
e
with patients and their amilies promotes a trusting and e ective • Surrogate decision making
c
i
doctor-patient relationship.
a
•
l
Transplant issues
t
This chapter will review key ethical concepts and standards in the
y
o
context o Hospital Medicine and explain the role o ethics consulta- • Truth telling
f
tion to acilitate patient care.
H
•
o
Withdrawal o ventilator
s
p
HOSPITAL ETHICS COMMITTEES AND ETHICS
• Withdrawal o other li e sustaining therapy
i
t
CASE CONSULTATION
a
l
• Withdrawal or withholding arti cial nutrition and hydration
M
The Joint Commission requires that hospitals develop and imple-
e
ment a process to handle ethical issues in patient care, but it •
d
Withholding o other li e-sustaining therapy
i
does not speci y how this should be done. Typically, an ethics
c
i
Contextual Issues
n
consultant—alone or as a small team—or the ull ethics commit-
e
tee may conduct consults. Ethics committees consist o physicians, •
a
Cultural/ethnic/religious
n
nurses, social workers, attorneys, theologians, and others represen-
d
tative o the immediate community that the hospital serves. Ideally, • Communication
S
y
the committee should have speci c members capable o ethics •
s
Dispute/conf ict
t
e
mediation. Importantly, the role o ethics committees is advisory.
m
Through mediation, ethics consultants usually make recommenda- • Intra amily
s
o
tions rather than prescribe solutions to resolve conf ict. Ethics com- • Intrasta
f
mittees consult on a range o issues across the li e cycle helping
C
•
a
patients, amilies, and sta grapple with challenging questions that Sta - amily
r
e
require expert assistance. • Sta -patient
COMMON INDICATIONS FOR ETHICS CONSULTATION • Patient or amily in denial

■ INFORMED CONSENT AND REFUSAL • Physician attitude toward treatment
In ormed consent is the ethical lynchpin o modern medical ethics • Quality o li e
in which the dialogue between the patient and physician preserves
the patient’s voice in directing care. Respect or persons grounds Data rom Nilson EG, Acres CA, Tamerin NG and Fins JJ. Clinical
this doctrine. In ormed consent promotes autonomy and patient Ethics and the Quality Initiative: A pilot study or the empirical
sel -determination as an interpersonal process between physicians evaluation o ethics case consultation. Am J Med Qual. 2008;23(5):
and patients to select an appropriate course o medical care, with 356-364.
the patient critically assessing his or her own values and pre erences
(Table 33-1).
Once a patient has made a choice, ongoing consent main- PRACTICE POINT
tains the moral warrant permitting clinicians to in ringe upon the
• Medical decision-making capacity does not ref ect global
patient’s zone o privacy. Patients who provide consent retain the
cognitive ability or intelligence, but rather a discrete capability
ability to revise that decision and withdraw it.
o understanding the consequences o speci c choices o ered
and made. This capability or a patient may change during the
PRACTICE POINT course o an illness. Determining medical decision-making
capacity is a clinical judgment. A court presumes competence
Indications for ethics consultation but may determine its absence. Moreover, or a person to be
Ethical Issues competent legally, he or she must possess both decision-
making capacity necessary to make a speci c choice and be o
• Advance directive age or an emancipated minor.
• Brain death
• Capacity/in ormed consent
In ormed consent imposes responsibilities on both the patient
• Con dentiality and the hospitalist but also creates opportunity to build a trusting
• Futility (demands or medically ine ective treatment) doctor-patient relationship. A properly executed in ormed consent
depends upon mutual respect, good communication (involving
218
adequate, accurate, and relevant in ormation provided in a orm decision with a progressive degree o explication. I a patient were to
and language the patient can understand), and shared agreement re use li e-sustaining therapy, he would have to more ully demon-
C
about the course o medical care. A relationship orged through the strate his reasons, rationale, and appreciation o the consequences
H
in ormed consent process can acilitate realistic patient expecta-
A
than i he were re using an elective procedure. The level o decision-
P
tions and help prevent disputes. Barriers to the in ormed consent making capacity should be in accordance with the risks and bene ts
T
process include: o the decision to be made. These reasons can be ounded on
E
personal, religious, or cultural belie s. The stringency o the standard
R
• Poor Care Coordination. Mixed communication, or even con-
tradictory in ormation to the patient and/or amily, may cause o capacity at each level correlates with the dangerousness o the
3
treatment decision. Re usal o care by a capacitated patient who
3
con usion and undermine care decisions.
• Fragmentation of Care. Lack o clarity about one’s overall is well in ormed needs to be respected, even i that re usal would
condition may also encumber the patient’s ability to make lead to serious harm. This is ethically supported by the principle o
C
autonomy and legally by the patient’s right to privacy and dominion
o
in ormed choices.
m
• Adaptation to the New Set of Potentially Limited Choices. over one’s sel .
m
Prolonged hospitalization or progressive illness may displace Every e ort should be made to discern the patient’s rationale or
o
re usal o recommended treatment and counter any misin ormation
n
the patient’s sense o autonomous decision making; it requires
I
with appropriate acts. Ethics consultation may help resolve ethical
n
both the patience and compassion o the hospitalist to help
d
the patient understand how the experience o hospitalization issues when treatment re usals are made by a surrogate, on behal
i
c
o an incapacitated patient.
a
might inf uence the response to illness.
t
i
o
A clinician claiming to have “consented the patient” mistakenly
n
■ TRUTH TELLING AND SHARING BAD NEWS
s
prizes outcome (agreement to a proposed procedure) over process
o
(a clear and in ormed decision whether that choice resulted in In our pluralistic society, some patients may pre er nondisclosure
r
o medical in ormation. In traditional Japanese culture, withholding
E
the acceptance or re usal o a treatment). The importance o the
t
h
in ormed consent process is the act o deliberation in making a o medical in ormation is the norm ounded upon ishin denshin,
i
c
the Japanese term or the nonverbal communication o the truth.
s
sound medical choice (Table 33-2).
C
Many Islamic societies operate with less doctor-patient dialogue
o
and a greater reliance on cultural and contextual clues in order to
n
s
PRACTICE POINT implicitly, versus explicitly, communicate in ormation to the patient.
u
l
t
In such contexts o overt nondisclosure, patients eventually deduce
•
a
In ormed consent is not a recitation o acts by the physician,
t
their status because they observe the nature o their treatments and
i
o
nor is it the placement o a signature on a document. Instead,
n
the reactions o those around them (Table 33-3).
in ormed consent provides the patient with decisional
Part o the art in doctor-patient communication is titrating
capacity or surrogate decision maker, either to accept or
generic advice to the particular patient’s willingness to know and
re use a medical intervention. This process requires adequate
use in ormation. The burdens and bene ts o “truth telling” and
disclosure in clear language o the nature and purpose o the
breaking bad news are weighed against the in ormation necessary
contemplated procedure and associated risks and bene ts. The
to make an in ormed treatment choice.
decision must be voluntary without coercion, and the patient
Approach the patient by stating that in this country standard
or surrogate must be o ered any available alternatives to the
practice requires sharing all medical in ormation with the patient,
proposed intervention.
unless he decides to delegate the dialogue and the implicit author-
ity to make choices using that in ormation to a surrogate as the
Concern about patient decision-making capacity typically occurs health care agent.
with re usal rather than with agreement o a proposed therapy. I a patient requires emergent treatment, physicians can provide
A clinician may equate treatment re usal with loss o decision-making care to a patient who is unable to provide consent when no sur-
capacity because the decision challenges the doctor’s expert rec- rogate is available to provide authorization. The ethical oundation
ommendation. However, under the rubric o sel -determination, or this provision o care is grounded in the legal notion o the
patients retain the right to re use treatments and physicians have “emergency presumption” rst articulated by Benjamin Cardozo
a corollary obligation to ascertain that the patient understands the when he was chie judge o the New York State Court o Appeals in
consequences o that choice. Mere re usal o a recommendation Schloendorf versus the Society o the New York Hospital in 1914. In this
does not mean a patient lacks capacity or decision making. landmark decision, Cardozo articulates and anticipates the modern
The philosopher James Drane developed a tool to assess a notion o in ormed consent and opines that urgent and necessary
patient’s capacity or making choices. Invoking a “sliding scale o treatment can occur in the absence o consent by a patient when
competence,” Drane linked the increasing gravity o a patient’s such treatment is necessary. I the patients recovers and regains
TABLE 33-2 Informed Consent and Refusal TABLE 33-3 Truth Telling and Sharing Bad News
• The physician must communicate to the patient speci ic

• In ormed consent is a patient right
in ormation necessary or making in ormed and deliberate
• Not all patients retain this right, predicated on the ability to be choices
sel -determining, which requires capacity
• No precise metric determines what patients need to know
• Patients must have decision-making capacity to participate in to make choices; however, most physicians adhere to a
in ormed consent “reasonable person” standard by providing the amount o
• Patients without capacity need a surrogate decision maker in ormation that an “average” person would need to make an
• Urgent care can be provided when appropriate under the in ormed choice
emergency presumption • Full medical disclosure is the norm or most Western people
219
capacity or i surrogates become available, physicians should dis- dynamic decision making and provides an individual who can inter-
close the rationale and nature o the emergent treatment. These pret the patient’s prior wishes in light o evolving circumstances and
discussions should not impede the urgent provision o care. the patient-proxy covenant.
All 50 states recognize an advance directive as an extension o
P
A
■ THERAPEUTIC EXCEPTION OR PRIVILEGE the patient’s voice under the Patient Sel -Determination Act (PSDA)
R
On that rare occasion when the risks associated with disclosure o 1990, which requires health care institutions that participate in
T
outweigh the bene ts, practitioners can deliberately withhold Medicare and Medicaid programs to ask patients whether they
I
in ormation counter to the patient’s sel -determination and right to have an advance directive, in orm patients o their right to complete
know. Typically, such a deviation rom standard practice, re erred to an advance directive, and incorporate advance directives into the
as “therapeutic exception” or “privilege,” would involve a severely medical record.
depressed patient who might become suicidal with grievous news. Surrogates make decisions or incapacitated patients according
T
h
We recommend that a psychiatrist and the local ethics committee to three distinct decision-making standards: patients’ expressed
e
participate in determining the need to limit disclosure. wishes, substituted judgments, and best interests. When invok-
S
p
ing substituted judgment, the surrogate places themselves in the
e
c
■ ADVANCE DIRECTIVES shoes o the patient and tries to make a decision as the patient
i
a
would. When neither knowledge o expressed wishes nor in er-
l
t
Seventy percent o seriously ill patients are unable to decide treat-
y
ence o substituted wishes exist, the surrogate makes a decision
o
ment options at the end o li e. The majority o these patients do not
based on what a reasonable person would make, balancing ben-
f
have advance directives at the time o hospitalization. By taking on
H
e ts, and burdens.
o
discussions at the onset o care, practitioners can establish a doctor-
s
Even when surrogates consider the patient’s values, the stress o
p
patient/ amily relationship and mitigate many ethical dilemmas
i
the surrogate role coupled with amily dynamics and imprecise prior
t
a
that could ensue should the patient deteriorate. These discussions
l
patient wishes can lead to morally ambiguous situations. Conf icts
M
should occur ideally within a ramework that prospective planning
can also arise between surrogates o equal standing, such as two
e
can be emotionally raught. To guide these discussions it is best
d
sisters who cannot agree on their mother’s care. A rigid hierarchical
i
to rst identi y goals o care and then determine the desirability o
c
approach to surrogate decision-making oversimpli es a process
i
n
speci c diagnostic and therapeutic interventions.
e
that is complex, dynamic, and personal. When two surrogates dis-
Advance directives allow the patient the opportunity to speci y
a
agree, ask them to set aside their own pre erences and articulate
n
pre erences in advance o incapacity through a living will and/or
d
what each believes is in the patient’s best interest. This minimizes
designate a surrogate to speak on his behal as a “durable power o
S
potential conf icts o interest and may lead to a concordance o
y
attorney” or “health care agent” or “proxy.” With an advance direc-
s
views. When this approach ails, hospitalists may give ethical—i not
t
e
tive, an incapacitated patient can be treated in accordance with his
m
legal—precedence to that surrogate who has been assuming more
prior wishes. Such advance care planning can decrease speculation
s
o the care responsibilities.
o
about what the patient would have wanted and decrease the moral
Sometimes the conf ict about goals o care arises rom an incon-
f
angst associated with the proxy role, a burden that is o ten under-
C
gruity between a written directive and an oral one. Given the objec-
a
stated (Table 33-4).
r
tive reality o documentation, de erence will more likely be given to
e
When there is no surrogate (a health care agent appointed by the
a previous written rather than verbal directive, even i more contem-
patient or a guardian appointed by a court), ethical norms and the
poraneous verbal pre erences emerge—which ethically would take
law assign standing to these surrogates:
precedence. Hence, all practitioners should clearly document, in the
1. The patient’s spouse (and in many jurisdictions to domestic medical record, any and all expressed pre erences on the part o the
partners) patient. The discharge summary should include documentation o
2. Therea ter, a relative such as an adult child, parent, sibling advance care planning as a guide to uture care.
3. Finally, in some jurisdictions, a close riend Physician Orders or Li e-Sustaining Treatment where available
Each state has its own hierarchy or this prioritization. We rec- also provide means o expressing patient near end-o -li e treatment
ommend the use o a health care agent over a living will, when choices. Patients and their clinicians complete and together sign
the patient can identi y a trusted surrogate. In the living will, an this order set when the patient has a 1-year li e expectancy with
adult with capacity sets orth directions regarding medical inter- understanding o the patient’s current clinical condition and poten-
ventions and other actions that should or should not be taken in tial complications. These medical orders are valid across treatment
circumstances i he becomes incapacitated. A living will document locations, rom home/nursing home/hospice to emergency medical
may contain inherent contradictions and ail to anticipate possible services to hospital emergency/inpatient units. The orders address
scenarios. In contrast to a living will, a designated surrogate through whether to attempt resuscitation in the event o cardiopulmonary
a health care proxy or durable power o attorney allows or more arrest, and how aggressively to manage intercurrent illness or dis-
ease progression, such as, medical interventions including intensive
care with invasive monitoring, aggressive intervention short o
intensive care (especially mechanical ventilation), or care directed
toward hospice; orders may also include whether to use eeding
TABLE 33-4 Standards for Surrogate Decision Making
tubes or antibiotics. Some states give POLST order the statutory
• Adhere to the explicit wishes o the patient authority o an advance directive. Regardless, the orders provide
• I unknown, determine what decision the patient probably direct expression o patient choices to guide treatment when a
would have made based on the patient’s values, belie s, and patient lacks capacity, until so con rmed with the surrogate or the
past decisions as interpreted by the surrogate recapacitated patient.
• When neither knowledge o expressed wishes nor in erence While POLST orders can be use ul, like living wills they can contain
o substituted wishes exists, make a “best-interests” judgment contradictions. Our pre erence or advance care planning is through
based on what a generic patient would want in a given the use o a designated surrogate decision maker, when available,
circumstance to allow maximal responsiveness to real time challenges in the care
o the decisionally incapacitated patient.
220
■ ETHICS CONSULTATION AT THE END OF LIFE containment motivates triage decisions, adherence or individual
clinicians creates ethical conf icts because o their primary duciary
C
Establishing clear goals o care and having a working awareness o
obligation to the patient. Existing data suggests that the presence
H
the inherent conf icts and biases that may arise at li e’s end can help
o a DNR order at the time o MICU consultation was signi cantly
A
prevent conf ict and enhance patient care at a time when com ort
P
and tranquility are at a premium. Decisions to accept or re use associated with the decision to re use a patient to the MICU.
T
li e-sustaining therapy are all predicated upon the a orementioned Because any arrest during surgery could be considered reversible,
E
secondary to the procedure, physicians, patients, and/or surrogates
R
principle o sel -determination as exercised through a process o
in ormed consent or re usal. Clinicians need to distinguish responsi- should discuss DNR status prior to surgery. We echo the recom-
3
mendation o the American College o Surgeons or a process o
3
bility and culpability when considering their role in helping patients
die com ortably. “required reconsideration” o the preexisting DNR order as part o
the in ormed consent process or surgery. I the patient or surro-
C
gate rescinds the DNR order perioperatively, a decision is made to
o
■ WITHHOLDING LIFE-SUSTAINING THERAPY: DO NOT
m
reinstitute it upon arrival in the recovery room or at a speci ed time
m
RESUSCITATE ORDERS OR ALLOW NATURAL DEATH interval a ter surgery.
o
Causality is least complex in cases where a decision is made to with- I the patient, or surrogate, wants to maintain a DNR status dur-
n
I
hold li e-sustaining therapy (LST). By withholding LST, we mean a ing the procedure, this must be documented in the preoperative
n
d
decision not to institute an intervention that could prevent death or consent. I the patient dies in the OR, it is considered an “expected
i
c
prolong a dying process. The prototypic example o withholding LST death” under the rubric that DNR situations result in the patient’s
a
t
i
is a do-not-resuscitate (DNR) order, which, in the hospital setting, demise. In some institutions, the physicians may nd honoring an
o
n
means the orgoing o cardiopulmonary resuscitation (CPR) or basic intraoperative DNR order to be against their conscience. In these
s
cardiac li e support (BCLS) as well as advanced cardiac li e support cases, there should be a provision o conscientious objection and
o
r
(ACLS) to patients who have sustained a cardiopulmonary arrest. removal rom the case, so long as the primary physician responsibil-
E
t
When a patient, health care agent, or other surrogate consents to ity o nonabandonment is not breached.
h
i
a DNR order, an intervention will not interrupt the natural course Both decisions to treat patients with a DNR order in the ICU or in
c
s
o events. In such cases, the cause o death is clearly the underlying the surgical suite ultimately hinges on achieving clarity about the
C
o
disease process. goals o care. In each care decision, a conf ict may exist between
n
s
Unlike other interventions, to withhold CPR requires consent, a negative right to be le t alone (the DNR order not to resuscitate)
u
l
based on the emergency presumption o providing care i consent and the positive right to needed care. This balance o negative and
t
a
cannot be obtained. Traditionally, physicians are obligated to per- positive engagement makes sense when the goals cohere, such as
t
i
o
orm CPR unless the patient has agreed to or requested a DNR order, in the example o a palliative colostomy or an obstructing colon
n
which constitutes an in ormed re usal. cancer. In that case, the surgical diversion is meant to provide com-
Essentially, all patients who undergo cardiopulmonary arrest ort to a dying patient who had a DNR order, an ethically balanced
receive CPR unless the patient or their surrogate consents to a DNR plan o care.
order. When a patient or surrogate provides consent, the order In summary, the choice to orgo cardiopulmonary resuscitation
should be clearly placed into the medical record and the medical supports the patients’ right to re use medical care even i this re usal
and nursing teams in ormed in a standard manner. The patient’s leads to death. Notably, surrogate decision makers make approxi-
DNR status should travel with the patient when he or she goes o mately 80% o DNR requests. Hospitalists and primary care physi-
the f oor and be readily available or consultation should an event cians are encouraged to initiate DNR discussion with the patient or
occur. A DNR order should be reviewed periodically and may be proxy as soon as possible, pre erably not when patients are imme-
reversed by the patient at any time and by the surrogate decision diately aced with cardiopulmonary cessation and imminent death.
maker i the decision does not undermine a patient’s decision while
capacitated. ■ THE DO-NOT-INTUBATE (DNI) CONUNDRUM
Seventy to 80% o deaths occur with DNR orders in place or Do-Not-Intubate (DNI) or partial DNR orders compromise the
hospitalized dying patients. In the modern hospital, DNR orders take integrity o practitioners because they imply resuscitation without
many orms, ensconcing the patient’s negative right to be le t alone. intubation as a medically e cacious intervention despite the mar-
Despite the prevalence and resonance o a dying patient with DNR ginal e ectiveness o comprehensive cardiopulmonary resuscitation
orders on a general medical ward, DNRorders do not preclude other itsel . Most patients who have asystole or a ventricular tachyarrhyth-
treatments or interventions. DNR solely applies to decisions about mia requiring cardioversion and/or chest compressions will also
cardiopulmonary arrest. Ethically, patients with DNR orders may need intubation. Restoration o a viable cardiac rhythm places clini-
receive care in the intensive care unit (ICU) or in the operating room. cians in the untenable position o being unable to ully complete
DNR status should be honored across specialties and not impede resuscitation e orts.
access to appropriate palliative care i it can only be o ered through In our experience, a DNI decision suggests ambivalence about
an intensive care or operative intervention. Indeed, some assert that goals o care, that is, a desire to survive without remaining on a
to condition appropriate care upon the presence or absence o a ventilator or a protracted period o time. We recommend that these
DNR order is patently unethical. patients be ully resuscitated and also complete an advance direc-
In practice, institutions vary in their triage o patients with a DNR tive that would allow a withdrawal o the ventilator i they were to
order regarding ICU or surgical care or other interventions requiring linger beyond an a orementioned time limit.
intubation such as endoscopy. The Task Force o the American Col- Some patients or their surrogates would choose to be DNR but
lege o Critical Care Medicine and the Society o Critical Care Medi- desire intubation in nonarrest situations in order to “pull through an
cine argues against ICU admissions, or example, noting that: “ICU illness.” This might occur in the management o a COPD exacerba-
admissions should be reserved or patients with reversible medical tion, acute congestive heart ailure, or pneumonia. Analogous to
conditions who have a reasonable prospect o substantial recovery.” patients who are DNR in the OR, such patients want to be palliated
Signi cant variation may relate to how to precisely identi y (as per the a orementioned, diverting colostomy or obstructive
patients who have a “reasonable prospect o substantial recovery.” colon cancer) or treated or potentially reversible conditions (sepsis)
Moreover, i allocation or scarce resources (eg, ICU beds) or cost while setting limits on resuscitation should they deteriorate and
221
sustain a cardiac arrest or complete respiratory ailure. Intubation o patient or surrogate consents to a DNR, given the similarity between
patients with DNR orders might be regarded as a time trial. internal and external de brillation.
P
■ WITHDRAWING LIFE-SUSTAINING THERAPY ■ RELIEF OF SUFFERING AND PALLIATIVE SEDATION
A
R
Accepted as a norm, dating rom the blue-ribbon President’s Com- Su ering is de ned as an existential threat to the sel and distinct
T
mission or the Study o Ethical Problems in Medicine and Biomedi- rom pain. Palliative sedation is de ned as the use o speci c sedat-
I
cal and Behavioral Research, no ethical distinction exists between ing medications to relieve intolerable pain and su ering rom re rac-
withholding and withdrawing treatment. The statement that the tory symptoms, even at the risk o death. Palliative sedation aims
removal o LST device causes a patient to die, especially i the death to control symptoms rather than to end li e; archaic, misleading
is closely related temporally, is a misconstrual o causality. With- terminology such as “terminal sedation” should be avoided. Pallia-
T
drawal o LST simply removes an impediment to death. The intent tive sedation provides di erent levels:
h
e
is reedom rom interventions that are perceived as burdensome. 1. Ordinary sedation ( or relie o heightened anxiety or stress
S
Death a ter re usal or withdrawal o an intervention results rom the
p
without reduction o consciousness)
e
underlying disease. A decision to withdraw li e-sustaining therapies
c
2. Proportionate sedation ( or reduction o patient’s awareness
i
a
may be challenging due to: o distressing symptoms with the minimum dose necessary to
l
t
y
• Trans erence and counter trans erence o ten embedded in promote the patient’s ability to engage with his amily and his
o
end-o -li e decisions immediate environment)
f
H
• Physicians reluctance due to a misconstrued view that there 3. Palliative sedation to unconsciousness (when less extreme
o
s
is an ethical, and certainly psychological, di erence between measures have not relieved su ering)
p
i
withholding and withdrawal o LST
t
The initiation o palliative sedation to unconsciousness o ten
a
l
• A sense o ailure or sense o culpability invokes ethical dilemmas due to con usion about physician-assisted
M
• Uncertainty about prognostication
e
suicide or euthanasia. The doctrine o double e ect, originating
d
• Inadequate communication with patients and/or surrogates rom Catholic theology, re ers to the doctrine where a physician
i
c
about goals o care
i
uses a treatment, or gives a medication, or an intended e ect
n
e
• Di erences between how physicians and lay people view these where the potential outcome is good (eg, relie o a symptom),
a
decisions
n
knowing that there could be an undesired secondary e ect (such
d
State and ederal law regulate who is entitled to authorize the as death). Double e ect distinguishes between the ethically man-
S
y
plan when the patient cannot speak or himsel regarding a decision dated goal o treating intolerable patient su ering rom hastening
s
t
to withdraw LST. In the wake o the US Supreme Court decision in death by engaging in physician-assisted suicide or euthanasia. In all
e
m
the Cruzan case and the ederalism o the a orementioned Patient circumstances, the degree o sedation must be proportional to the
s
Sel -Determination Act, each state can set an evidentiary standard severity o su ering and is given only a ter the patient and or sur-
o
f
about the amount o evidence rom the patient’s prior wishes or rogate have completed the in ormed consent process and agree.
C
values, i known, necessary to permit a surrogate to authorize with-
a
Having noted the importance o disclosure and shared decision
r
e
drawal o LST. Although the Supreme Court in Cruzan observed that making, we do not believe that surrogates have the moral authority
there is no di erence between the withdrawal o arti cial nutrition to limit the provision o analgesia, including the provision o pallia-
and hydration and the withdrawal o a ventilator, some religious tive sedation or general anesthesia, i they are the only means by
traditions view the provision o ood and water as normative obliga- which to ensure the patient’s com ort. We assert that the right to
tions that require a higher degree o oreknowledge o the patient’s pain management is a basic human right that can not be abridged
wishes. by surrogate intervention. When there is a disagreement on the
scope and level o pain management it is best to involve the exper-
■ MEDICAL DEVICES AT LIFE’S END tise o the ethics committee and palliative care consultants.
The philosophical and legal status o implantable medical devices

has been a source o con usion in clinical practice amongst hos- ■ MEDICAL INEFFECTIVENESS (FUTILITY)
pitalists. These devices, which include automated implanted car- Broadly de ned, medical utility can be broken down into several
dioverter de brillators (AICDs) and permanent pacemakers (PPMs) domains:
may engender ambivalence in the clinician. They can be viewed as
• Physiologic utility (when it is absolutely—or to a reasonable
a treatment, comparable with external medical devices like dialysis
degree o medical certainty—impossible to achieve a physi-
machines or ventilators, or they may considered similar to a biologic
ologic e ect such as CPR in the setting o persistent acidosis)
transplant, like a liver or lung transplant. The distinction, i not prop-
• Qualitative utility (when the patient’s physiology may improve,
erly understood can lead to the perception that neither physician
but there is no patient-centered bene t)
nor patient has the ability to withhold or withdraw the e ects o the
• Quantitative utility (when the intervention has not worked in
device. It is critical to appreciate that the patient—or in the setting
similar patients within an accepted con dence interval)
o decision incapacity, the surrogate—retains the ability to deacti-
vate or remove these devices under the rubric o in ormed re usal. Disagreement about the e ectiveness o ongoing care may
When AICDs and PPMs only provide standby interventions, we evoke strong emotions on the part o patients, amilies, and physi-
view deactivation o their resuscitative role as withholding o care. cians entrusted to provide care. Multiple prior admissions when the
In contrast, when a patient is paced continuously by a permanent patient “pulled through” despite negative odds, many clinicians
pacemaker, either in a PPM or as a unction o the AICD, or uses a le t with disparate views about aggressiveness o care at the end o li e,
ventricular assist device, deactivation is de ned as a withdrawal o communication ailures, and cultural di erences all contribute to a
LST. Deactivation decisions should adhere to evidentiary decision- amily’s view o the patient’s overall condition, prognosis, and how
making standards as consistent with applicable state law, and ethics they would want him to spend the end o his li e. Clinicians should
consultation may provide expert guidance. In our view, consider try to prevent these disputes through ongoing communication dur-
disabling the AICD and/or the episodic unctions o the PPM when a ing the course o the illness, to be ref ective about the implicit orce
222
o one’s countertrans erence and avoid mixed messages rom di er- ACKNOWLEDGMENT
ent physicians by ensuring coherent comanagement.
C
The authors acknowledge the coauthorship o Heather X. Cereste,
The best way to overcome ragmentation is to have a meeting
H
MD, FACP o this chapter in the prior edition o this text.
including all key clinicians involved (physicians, social work, nurs-
A
P
ing, etc), signi cant others designated by the patient, and amily
T
members. The aim o the discussion is to create a comprehensive SUGGESTED READINGS
E
actual understanding o patient’s condition and prognosis. Then,
R
a ter achieving a broad understanding o the medical acts, a ruit- Berger JT, DeRenzo EG, Schwartz J. Surrogate decision making:
3
ul discussion regarding both amily and clinician expectations can reconciling ethical theory and clinical practice. Ann Intern Med.
3
ensue over the course o multiple meetings. 2008;149(1):48-53.
Ultimately, the imperative or members o the health care team Bosslet GT, Pope TM, Ruben eld GD, et al. An o cial ATS/AACN/
C
rests on exploring the intricacy o their patient’s history and values
o
ACCP/ESICM/SCCM policy statement: responding to requests or
m
and appreciate that many surrogates may be reluctant to imme- potentially inappropriate treatments in intensive care units. Am J
m
diately accept a physician’s prediction o medical utility. In these Resp Crit Care Med. 2015;191(11):1318-1330.
o
circumstances, clinicians should ask the surrogate to make judg-
n
Drane JF. Competency to give an in ormed consent. A model or
I
ments believed to be in the patient’s best interest and to articulate
n
making clinical assessments. J Am Med Assoc. 1984;252(7):925-927.
d
goals o care. On many occasions a surrogate may desire something
i
c
that is unachievable through the provision o care. Having the sur- Etchells E. Bioethics or clinicians. 3. Capacity. Can Med Assoc J.
a
t
1996;155:657-661.
i
rogate articulate these goals provides an opportunity or reality
o
n
testing and an occasion to redirect a bene cent impulse, so long Fins JJ. APalliative Ethic o Care: Clinical Wisdom at Li e’s End. Sudbury,
s
as practitioners appreciate why they are so potentially distressed by MA: Jones &Bartlett; 2006.
o
r
amily demands or interventions perceived by the treating team as
E
Guidelines or intensive care unit admission, discharge, and triage.
t
medically ine ective.
h
Task Force o the American College o Critical Care Medicine, Soci-
i
The increasing complexity o medical decision making o ten
c
ety o Critical Care Medicine. Crit Care Med. 1999;27(3):633-638.
s
necessitates the assistance o an ethical consultant (or consulting
C
Institute o Medicine. Committee on approaching death: addressing
o
committee). These consultations tend to be utilized by physicians
n
key end o li e issues. Model advance care planning initiatives. In:
s
who believe in shared decision making. Providers who do not utilize
u
Dying in America: Improving Quality & Honoring Individual Pre er-
l
these services tend to believe that it is their responsibility to resolve
t
a
issues with patients and amilies and that they are already pro cient ences Near the End o Li e; 2015:172-192.
t
i
o
in ethics. These sel -assessments may not ref ect the physician’s Limehouse WE, Feeser VR, Bookman KJ, et al. A model or emergency
n
knowledge, skills or ability to mediate ethical disputes in clinical department end-o -li e communications a ter acute devastating
practice and thus practitioners need to cultivate a ref ective stance events. Part I. Decision-making capacity, surrogates, and advance
toward the utility o ethics consults in order to best serve patients directives. Acad Emerg Med. 2012;19(9):E1068-E1072.
and their amilies. Orlowski JP, Hein S, Christensen JA, et al. Why doctors use or do not
use ethics consultation. J Med Ethics. 2006;32(9):499-502.
CONCLUSION
Rosenbloom AH, Jotkowitz A. The ethics o the hospitalist model. J
The hospitalist model was ounded on the premise that it could Hosp Med. 2010;5(3):183-188.
improve inpatient medical care. Ensuring that a model ounded on
Satyanarayana Rao KH. In ormed consent: an ethical obligation or
e ciency and cost-e ectiveness does not in any way compromise
legal compulsion? J Cutan Aesthet Surg. 2008;1(1):33-35.
the primacy o the patient requires a amiliarity with bioethics and
legal precedent as well as a willingness to call upon clinical ethics Truog RD, Campbell ML, Curtis JR, et al. Recommendations or end-
consultants and hospital ethics committees when their role in the o -li e care in the intensive care unit: a consensus statement by
institution places them in a potentially integrity-compromising the American College o Critical Care Medicine. Crit Care Med.
situation. These conf icts can be best understood and managed 2008;36:953-963.
through interdisciplinary consultation and collaboration.
At the epicenter o inpatient care, hospitalists should maintain
clarity, mediate misunderstandings about the diagnosis, prognosis, ONLINE RESOURCE
or goals o care, and minimize the opportunity or conf ict to arise
Aid to Capacity Evaluation (ACE). http://jcb.utoronto.ca/tools/
by organizing the medical team, without dominating it—all in the
documents/ace.pd . Accessed May 31, 2016.
service o patient-centered care. Hospital ethics consultants and
committees stand ready to provide assistance and remind us o the
centrality o patient bene cence in all its many orms.
223
CHAP TER
34 INTRODUCTION
Advance directives, such as living wills, health care powers o attor-
ney, do-not-resuscitate orders, and mental health care directives,
are written legal documents that o er patients the opportunity
to re ect upon and provide direction or their uture medical care.
They are important and use ul tools to acilitate discussions between
patients, amily members, and physicians about end-o -li e care
choices, and they provide guidance and legal protection in those
situations when a patient is no longer capable o declaring care pre -
Medical-Legal erences, when critical and di cult treatment decisions need to be
made. Following an advance directive acilitates making end-o -li e
Concepts: Advance care decisions at the patient’s bedside rather than through conten-
tious court proceedings. More importantly, advance directives are
Directives and legal mechanisms that rein orce the undamental pro essional and
moral responsibility o health care providers and institutions to pro-
mote and protect patient autonomy, wel are, and dignity.
Surrogate Decision
Making •
PRACTICE POINT
Advance directives are legal mechanisms that rein orce
undamental pro essional and moral responsibilities o health
care providers and institutions to promote and protect patient
Kelly Armstrong, PhD autonomy, wel are, and dignity. Advance directives are best
thought o as the result and documentation o a patient-
Ross D. Silverman, JD, MPH centered process aimed at extending the rights o patients to
guide their medical care, even through periods when they are
no longer able to directly participate in decisions about their
own care.
This chapter o ers an introduction to advance directives, examin-

ing the general structure o the various types o advance directives,
when they may be triggered, what clinical circumstances and deci-
sions they may cover, and the relative strengths and weaknesses
o the di erent advance directive instruments. Every state has laws
describing the types o advance directives available in its jurisdic-
tion, the processes by which such documents may be created and
triggered, how and where they can be employed, and the legal
protections a orded to care providers and health care acilities that
carry out care decisions when guided by such documents. These
laws may be supplemented by policies and procedures adopted by
your local hospital or health care acility to direct the use o advance
directives in your particular setting. Given the unique idiosyncrasies
ound in di erent state laws and di erent acility policies, hospital-
ists should note that not all varieties o advance directive instru-
ments may be available locally, and the processes used to carry out
a particular advance directive may diverge rom what is described
in this chapter.
CONCEPTUAL FOUNDATIONS
The 1960s and 1970s saw a movement toward greater patient par-
ticipation in health care that resulted in new ways to ensure shared
decision making between patients and their physicians. During this
time, dramatic medical and technological advances underscored
the importance o recognizing and incorporating the goals and
values o all patients, while ethicists, the courts, and others came
to a consensus that a decisionally capable patient has the right to
accept or re use any type o medical care. However, many persons
224
eared a loss o control may occur i they became incapacitated and statute or case law, have provisions or honoring advance direc-
unable to make their own medical decisions. As a result, patients tives. Despite this, di erent jurisdictions utilize di erent standards,
C
and potential patients became increasingly aware o the need to terminology, and limitations o authority. Physicians should become
H
make provisions or their own uture medical treatment.
A
amiliar with the local requirements or completing and honoring a
P
legal advance directive.
T
■ PATIENT SELF-DETERMINATION ACT The legal requirements or a valid advance directive vary rom
E
state to state, but in general any adult or emancipated minor with
R
In 1990, the US Supreme Court heard the case o Nancy Cruzan, a
33-year-old woman diagnosed in a persistent vegetative state a ter decisional capacity can execute an advance directive. In some rare
3
instances an advance directive may have an expiration date, how-
4
a car accident in 1983. In the case o Cruzan v. Commissioner, Missouri
Department of Health, 497 US 261 (1990), Nancy’s parents sought to ever, in general an advance directive remains in e ect until such
have the eeding tube removed and allow Nancy to die. The Court time as the patient revokes it. Di erent states have di erent require-
M
ound that decisionally capable patients have the right to re use li e- ments or validly revoking an advance directive, but most recognize
e
d
sustaining therapies, including nutrition and hydration. However, a decisionally capable person may revoke an advance directive at
i
c
any time simply by making the intention to revoke clear to a lawyer
a
the Court also indicated that states could require third parties act-
l
-
or health care provider, either verbally or in writing.
L
ing on behal o patients who are no longer decisionally capable to
e
Because o the additional e ort required to complete a written
g
submit evidence o the patient’s wishes be ore granting a request to
a
withdraw li e-sustaining treatment. directive, the written directive is generally held to have more power
l
C
Partly in response to the Cruzan decision, Congress passed the than oral statements; however, this should be evaluated on a case-
o
by-case basis.
n
Patient Sel -Determination Act in 1990. This law attempts to make it
c
e
clear that patients have the right to make decisions regarding their
p
■ NONCONFORMING DOCUMENTS
t
medical care. This includes the right to accept or re use treatment
s
:
and the right to complete an advance directive as evidence o their
A
While many state statutes contain standard or recommended lan-
d
wishes. The law requires any health care provider participating in guage or advance directives, people are introduced to advance
v
a
Medicare or Medicaid to provide all persons over the age o 18 directives rom a variety o sources, including f nancial planners,
n
c
with written in ormation regarding the patient’s right to accept senior organizations, national agencies, and religious organiza-
e
or re use treatment and right to complete an advance directive.
D
tions, not to mention the Internet. Most state statutes explicitly
i
Providers include hospitals, nursing homes, home health care pro-
r
acknowledge that versions o advance directives that do not contain
e
viders, hospices, and health maintenance organizations, but not
c
the recommended statutory language, as well as some oral state-
t
i
outpatient-service providers or emergency medical personnel. The
v
ments, may also be valid. Such noncon orming directives cannot be
e
Patient Sel -Determination Act also requires health care providers to
s
dismissed just because they do not contain the language recom-
a
document whether patients have advance directives, establish poli-
n
mended by statute. Evaluating directives that do not con orm to the
d
cies to implement advance directives, and educate their sta and recommended language ound in state statutes requires physicians
S
the community about advance directives. Patients should also be
u
to determine, with the assistance o the ethics committee or legal
r
in ormed that having an advance directive is not required to receive
r
counsel i necessary, whether a noncon orming directive meets
o
medical care.
g
the state’s standard o reliability. In other words, is it clear that the
a
t
patient intended to document his or her wishes with the intention
e
■ UNIFORM HEALTH CARE DECISIONS ACT
D
that such documentation would be relied upon and ollowed by
e
A ter the Cruzan decision, signif cant changes occurred in state the health care team? Such seriousness o purpose can be demon-
c
i
s
laws across the United States regarding health care decision mak- strated by a variety o means including i the document contains the
i
o
ing. While every state passed legislation authorizing the use o at patient’s signature, i the document is witnessed, i the document
n
M
least one orm o advance directive, there was very little uni ormity was presented to the physician by the patient as an advance direc-
a
between the laws. By 1993, state laws regarding health care deci- tive, or i the patient discussed the document with amily members
k
i
sion making were o ten ragmented, incomplete, and sometimes as evidence o the patient’s wishes.
n
g
inconsistent. Statutes in one state requently con icted both with
other statutes rom the same state, and with statutes rom other ■ ADVANCE CARE PLANNING
states. With this con usion in mind, the Uni orm Health care Deci-
When advance directives f rst came into existence, they were
sions Act (UHCDA) was dra ted in 1993. Under the UHCDA, any adult
viewed as legal documents o ering legal protection rom unwanted
or emancipated minor may execute or provide an “advance health
treatment at the end o li e. As the practice o medicine has become
care directive,” which re ers to either a “power o attorney or health
more patient centered, a greater ocus has emerged on ensuring
care” or other “individual instruction.” I an individual ails to execute
all medical decisions are not only clinically sound, but also based
a power o attorney or health care or i the agent is not available,
on the patient’s personal goals or care. Advance directives are best
the UHCDA authorizes health care decisions to be made by a “sur-
thought o as the result and documentation o a patient-centered
rogate” to be selected rom a priority list. The Act also recognizes
process aimed at extending the rights o patients to guide their
an individual’s authority to def ne the scope o any instruction or
medical care, even through periods when they are no longer able to
agency as broadly or as narrowly as the individual chooses.
directly participate in decisions about their own care.
Optimally, be ore completing an advance directive, individu-
ADVANCE DIRECTIVES als will have the opportunity to have a structured discussion with
Advance directives empower individuals to make their goals, val- their physician or other clinician about their health care wishes
ues, and treatment pre erences known be ore a loss o decisional and goals. Ideally, these conversations are held be ore the patient
capacity may occur. The term advance directives re ers to oral or becomes ill, even though physicians cannot discuss with specif c-
written instructions about a person’s medical care that provide ity every scenario that may occur. Advance directives may simply
guidance regarding medical treatment to health care pro essionals contain general pre erences, or they may contain specif c instruc-
i and when the patient becomes unconscious or otherwise unable tions about particular treatments. Those that contain only general
to make his or her own decisions. All 50 states, either through pre erences can be less help ul in guiding care than those with
225
specif c instructions because general pre erences can require more would be the proximate cause o death (rather than the patient’s
interpretation in light o the current medical evidence. For example, terminal condition). Most states also have provisions invalidating
a patient’s statement “I don’t want to live hooked up to machines,” the document i the patient is pregnant, and some states dictate
may mean the patient’s does not want CPR, or it may mean the that the power granted to a designated agent such as a power o
P
A
patient wants all aggressive therapies until such time as those thera- attorney or health care takes precedence over the power provided
R
pies ail to restore the patient to an acceptable quality o li e. A com- in the living will. Physicians should re er to the specif c statutes in
T
prehensive process o advance care planning includes a discussion their state or guidance.
I
o possible or likely scenarios based on the patient’s unique medical As the f rst widely known advance directive, the living will is
situation, a discussion o the patient’s values and goals, documenta- the most well known and physicians should be aware that many
tion o the patient’s values and goals, and a way to ensure that this patients re er to all advance directives by the name “living will,” and
in ormation is available to present and uture care providers. The others may con use the living will with a last will and testament.
T
h
advance directive as a document provides a legally recognized way Care should be taken to veri y the exact type o directive(s) the
e
to record this discussion. patient may have.
S
p
e
TYPES OF ADVANCE DIRECTIVES
c
Do not resuscitate orders
i
a
l
Although state statutes vary, there are three basic types o advance
t
Do-not-resuscitate (DNR) orders have become increasingly com-
y
directives: instructional, such as living wills; proxy, such as a power o
o
mon in the care o the dying patient. The decision whether to resus-
f
attorney or health care; and combination directives, which provide citate a patient who su ers a cardiac or respiratory arrest involves
H
both instructions or care and name a proxy to make decisions.
o
consideration not only o the potential clinical outcomes, but also
s
p
the patient’s pre erences regarding the intervention and i the likely
i
t
■ INSTRUCTIONAL ADVANCE DIRECTIVES
a
outcome is one that would be desired by the patient. In order to
l
M
Instructional directives provide consent or re usal or specif c treat- streamline care rom the inpatient to outpatient setting, a majority
e
ments that may need to be utilized when the patient is unable to o states now have legislation permitting the use o out-o -hospital
d
i
make the decision, traditionally at the end o li e (Table 34-1). DNRorders. These DNRorders are physician orders that direct health
c
i
n
care pro essionals across all emergency, living, or health care set-
e
Living wills tings to withhold or withdraw (i treatment has already begun in
a
n
Living wills are the most common instructional directive. A living will absence o the orm) specif c types o li e-sustaining treatment, such
d
as CPR or def brillation, in the event o a respiratory or cardiac arrest.
S
directs physicians to withhold or withdraw li e-sustaining measures
y
s
and to provide only com ort care i the patient has a terminal illness
t
e
and li e-sustaining measures are prolonging the dying process with- Physician’s orders for life sustaining treatment (POLST)
m
out a chance or recovery. The living will thus addresses only a small
s
POLST stands or physician’s orders or li e-sustaining treatment,
o
subset o medical situations and critics cite the typically vague lan- and the POLST orm has been adopted by several states (some
f
C
guage and di culty physicians have predicting with certainty when states such as New York use the acronym MOLST or medical orders
a
r
a patient is at the end o li e as primary reasons that the document or li e-sustaining treatment). POLST addresses the desire some per-
e
has allen out o avor in terms o utility. Nevertheless, the document sons may have, particularly those persons who are experiencing a
can be help ul as a starting point or urther discussion, and it serves chronic or li e-limiting illness, to avoid unwanted emergency medi-
as evidence o a patient’s values, specif cally that there are certain cal care like CPR or a trans er to the hospital. POLST takes advance
outcomes the patient would not wish to pursue. The living will is directives a step urther by not only documenting a patient’s treat-
also a use ul document or persons who do not have a designated ment pre erences, but also providing emergency and other medi-
agent or proxy since it speaks directly to the physician and does not cal personnel with clear physician orders to ollow in the case o an
require the consent o a designated agent or proxy in order or the emergency. Depending on the state, the POLST orm has three or
physician to take action. our sections outlining the patient’s desire to have or re use CPR,
Some jurisdictions have placed restrictions on how the living will whether the patient would like to be taken to a hospital, and the
may be invoked. For example, Illinois and Missouri do not permit types o medical interventions desired by the patient, including the
the withdrawal o artif cial nutrition and hydration i the withdrawal provision o com ort care, antibiotics, or artif cially administered
TABLE 34-1 Types of Advance Directives
Advance Directive Type Covered Activity

Living will Instructional End-o -li e document or patients with a terminal illness who wish to
orgo death-delaying procedures
Physician’s orders or li e-sustaining Instructional Portable physician orders or li e-sustaining treatment that apply across
treatment/medical orders or li e- all emergency and care settings
sustaining treatment
Power o attorney or health care Proxy Appoints a person over the age o 18 to make decisions on behal o
the patient
Psychiatric advance directive Combination Allows patient to preauthorize certain types o mental health
treatment in the event o acute psychiatric illness; also allows the
patient to appoint someone to make mental health treatment
decisions
“Five wishes” Combination Combines a living will, power o attorney or health care, and
instructions or com ort care and personal matters
226
nutrition. In those states that have adopted the POLST paradigm, to re use treatment, as long as they do not pose a serious risk o
the orders are valid in all emergency, living, and health care set- harm to themselves or others. However, during an acute episode
C
tings. I a patient presents the POLST document in a state that o psychiatric illness an individual may become unable to make or
H
has not adopted the POLST paradigm, the document should be
A
communicate decisions about treatment. A psychiatric advance
P
interpreted as strong and reliable evidence o the patient’s known directive allows currently competent patients who may experience
T
wishes regarding treatment. an acute episode o psychiatric illness in the uture, to agree in
E
advance to treatment they may re use later when ill. Unlike many
R
Other types of instructional directives other types o advance directives that ocus on empowering the
3
patient to re use unwanted interventions, psychiatric advance
4
Other types o instructional directives allow individuals to re use
directives are generally “opt-in” documents where the patient
specif c therapies, such as blood trans usion or dialysis, or per-
may preconsent to inpatient hospitalization, medication, or other
sons who have specif c desires to re use those types o therapy.
M
help ul treatment modalities. In addition to o ering a clear written
e
For example, many Jehovah’s Witnesses do want blood or blood
d
statement o an individual’s treatment pre erences, the psychiat-
i
products administered under any circumstances, even though
c
ric advance directive can also be used to assign decision-making
a
death may be an outcome o the re usal. The instructional direc-
l
-
authority to another person while the individual is incapacitated.
L
tive provides evidence o the patient’s wishes, and many will
e
Correctly executed and implemented directives not only promote
g
specif cally release the physician rom any liability or ollowing
a
individual autonomy and empowerment, they can eliminate the
l
the directive.
C
need or court involvement and assist in recovery by communicat-
o
ing to the physician and others the types o treatments that have or
n
c
■ PROXY ADVANCE DIRECTIVES have not worked or the person in the past.
e
p
Proxy advance directives such as the power o attorney or health
t
s
:
care allow a patient to appoint another person over the age o 18 “Five wishes”
A
to make health care decisions in circumstances when the patient
d
“Five wishes” is a combination advance directive introduced in 1996
v
is unable to make those decisions or himsel or hersel . The proxy
a
that has become popular because o its easy-to-read language. It
n
advance directive is not limited to end-o -li e decisions and is there-
c
introduces the subjects o a living will, power o attorney, com ort
e
ore a more use ul document or the vast number o decisions that care, spirituality, and other personal matters such as orgiveness and
D
can be encountered throughout the li espan. The only requirement
i
memorial plans in the orm o f ve distinct but overlapping “wishes.”
r
e
be ore the instrument takes e ect is the patient must lack decisional
c
The f rst two wishes—”the person I want to make care decisions
t
capacity regarding the decision at hand. Physicians should note that
i
v
or me when I can’t” and “the kind o medical treatment I want or
e
because decisional capacity is decision specif c, a patient may have
s
don’t want”—are intended to serve as legal documentation ulf lling
the capacity to make some decisions, but require assistance to make
a
the requirements o many states’ living will and power o attorney
n
other types o decisions.
d
statutes.
S
The person appointed by a power o attorney or health care is
u
called a power o attorney or an “agent.” Because a power o attor-
r
r
o
ney has been directly appointed by the patient, a power o attorney HONORING ADVANCE DIRECTIVES
g
a
typically has broad authority to make the same kinds o decisions Today’s health care providers, and physicians in particular, have the
t
e
as the patient unless that authority has been limited by the patient simultaneous responsibility o respecting patients’ autonomous
D
or by statute. Thus, in most states, a power o attorney can accept choices while protecting rom harm those patients who are incapa-
e
c
or re use any type o treatment. Because the power o attorney can ble o making an authentic or in ormed decision. For patients with
i
s
i
execute a broad range o powers, health care pro essionals and
o
variable or uctuating capacity, circumstances may arise where the
n
others should ensure the appointment o a power o attorney is patient is capable o being involved in some, but not all, decisions.
M
executed without coercion or duress. Some states require the power The minimum standard or decisional capacity is directly related to
a
k
o attorney to accept the agency in writing, while others restrict the risk involved in the decision. Thus, a patient may legitimately
i
n
agency to a single individual in order to minimize the chances
g
consent to a low-risk procedure such as a blood draw or CT scan,
that disagreement or di erences o opinion between two or more but need assistance with more complex decisions such as whether
agents stall medical decisions. Physicians should re er to the specif c to have surgery. The presumption is that a coherent care plan
statutes in their state or guidance. should be established around the patient’s known wishes whenever
possible.
■ COMBINATION ADVANCE DIRECTIVES
There is a growing trend in advance directives toward allowing ■ INFORMED CONSENT
persons to combine instructions regarding treatment pre erences In ormed consent is, with rare exceptions, required be ore treatment
with appointing a decision maker to make proxy decisions. Many can be provided to a patient. In order or consent to be valid, it must
state orms now allow persons completing an instructional orm be given voluntarily in light o accurate and relevant in ormation—in
to also speci y or appoint a power o attorney at the same time. It a language intelligible to the recipient—regarding the risks, ben-
is important to note that while many combination orms grant the ef ts, and alternatives o the proposed interventions. In addition to
proxy decision maker broad powers o decision making, the docu- providing in ormation that the average reasonable person might
ment may also contain a limitations or exclusions section that limits desire, it is important that patients receive su cient in ormation
the types o decisions that the proxy decision maker can make on within the priorities that have the most meaning or them. For
behal o the patient. instance, some religious belie systems require or preclude certain
orms o medical treatment. Other patients may want to know how
Psychiatric advance directives incapacitated they may be a ter a surgery, or how long they must
Psychiatric advance directives allow persons with mental illnesses wait be ore driving. Valid consent also requires the person giving
to engage in advance planning with their physicians regarding consent to have su cient decisional capacity to make the decision
potential uture care. Persons with mental illnesses retain the right at hand.
227
PRACTICE POINT
TABLE 34-2 Standards for Assessing Decisional Capacity
• Most advance directives are not applicable when the patient
has decisional capacity and is capable o exercising his or her Assessing Decisional Capacity
P
right to accept or re use medical treatment. Thus, a patient’s 1. The patient can make and communicate a choice
A
oral statements made while the patient is decisionally capable
R
2. The patient demonstrates understanding o his or her
T
should be ollowed even i those statements con ict with a medical condition, prognosis, and the risks and bene its o
written directive.
I
the available treatment options
3. The patient has the capacity or reasoned decision making
■ DECISIONAL CAPACITY 4. The patient is able to apply his or her own values to the
decision at hand
T
There are relatively ew guidelines in orming physicians and oth-
h
5. The patient’s decision remains stable over time (i the
e
ers when to conduct and document an explicit assessment o an situation allows), or the patient can make a reasoned
S
individual’s decisional capacity. However, assessments o decisional
p
explanation why the decision is not consistent with the
e
capacity should be ongoing and not restricted to instances where patient’s previously known wishes
c
i
the patient disagrees with a physician’s recommendation or treat-
a
l
t
ment. Assessment and documentation may be warranted whenever
y
o
the individual consents to complex or high-risk interventions, when
f
incompetent or have been appointed a guardian. On the contrary,
H
the patient makes choices that do not appear prudent or when the
o
decision is outside the patient’s norms, or when the patient has there is signif cant variability in the decisional capacity among these
s
p
marked cognitive def cits or other risk actors or impaired decisional patients. The examining clinician should evaluate each patient’s
i
t
capacity in light o the particular decision at hand.
a
capacity.
l
M
e
PROXY DECISION MAKING
d
PRACTICE POINT
i
c
Physicians respect a patient’s autonomy by recognizing that the
i
n
• The presumption is that patients can make their own decisions patient’s right to accept or re use treatment remains even a ter
e
he or she loses decision-making capacity. Patients who have lost
a
until they demonstrate otherwise. Although some complex
n
decisional capacity may continue to communicate through their
d
situations may or call or a specialized assessment by a
advance directives, or through a power o attorney or surrogate
S
psychiatrist or ethicist, most capacity assessments can be
y
who interprets what the patient would choose in a given situation.
s
completed by the physician obtaining consent to treat.
t
e
When medical decisions need to be made or a patient who
m
lacks decisional capacity, physicians should f rst inquire whether the
s
o
It is important to note that because capacity may uctuate over patient has documented any wishes in a written advance directive
f
C
time, physicians may be dealing with a proxy decision maker or such as a POLST orm or a durable power o attorney or health
a
some aspects o treatment, and directly with the patient or other care. Optimally, the patient will have documented pre erences that
r
e
aspects. When reasonable, important decisions should wait until directly relate to the proposed treatment. However, this rarely hap-
the patient regains decisional capacity (i expected), and decisions pens. More commonly, a third party is necessary to represent the
made by others should be reviewed with the patient i and when he patient’s interests and interpret the patient’s known wishes and
or she regains capacity. The presumption is that patients can make values in light o the current medical situation. I the patient has des-
their own decisions until they demonstrate otherwise. There ore, ignated a decision maker in an advance directive such as a power o
each decision should be put be ore the patient and i the patient attorney or health care, the designated agent should be relied upon
concretely demonstrates that he or she is unable to make the par- to make decisions. I the patient does not have a power o attorney
ticular decision at hand, this should be care ully documented in the or some other document that directs care, such as a POLST orm or
patient’s medical record. Many jurisdictions also require documen- living will (or those documents exist but do not apply to the situa-
tation o the reason or the patient’s incapacity and its expected tion at hand), a surrogate should be appointed.
duration.
Decisional capacity is unctionally def ned in light o a specif c ■ SURROGATE DECISION MAKING IN THE ABSENCE
decision. As the risk o a proposed intervention increases, so too OF ADVANCE DIRECTIVES
does the threshold standard or decisional capacity. For instance, Many states have established protocols or identi ying a legal third-
the capacity needed to consent to routine labs would be lower than party decision maker in the absence o a documented advance
the capacity required or consent to open heart surgery. Grisso and directive. This third-party decision maker may be called a proxy
Appelbaum have put orward the most requently used model or or a surrogate, depending on the state. Most patients have not
assessing decisional capacity standards 1 to 4 in Table 34-2. completed an advance directive and there should be no presump-
Some commentators have also indicated that under certain cir- tion in erred rom the act that an advance directive has not been
cumstances it may be prudent to add the f th criterion, particularly executed. The guiding principle in appointing any surrogate is to
i the decision being made by the patient represents a radical or f nd the person, or group o persons, who best know the patient’s
abrupt change in treatment goals, or does not con orm to what is values and health care goals. This person or persons should be able
known about the patient values or other decisions. to e ectively communicate with the health care team, and be will-
ing to make choices the patient would most likely make i he or she
■ COMPETENCE AND DECISIONAL CAPACITY could speak or himsel or hersel .
Decisional capacity and competence are not synonymous. Compe- Where they exist, state statutes generally indicate the amily o
tence is a legal designation that must be made by a court, whereas the patient should be responsible or making medical decisions. The
decisional capacity is determined by a physician or other clinician. order o priority or appointing a surrogate is listed in Table 34-3.
Not all patients with a developmental, cognitive, or mental illness While “ amily” is generally recognized as a biological or legal relation-
have impaired decisional capacity, even i they have been declared ship, most states have not specif cally addressed more complicated
228
Scope of authority
TABLE 34-3 Priority Hierarchy for Appointing a Surrogate
C
The exact scope o a surrogate’s decision-making authority varies by
H
Typical Order of Priority for Appointing a Surrogate in state. Patients may also document certain limits on the kinds o deci-
A
Absence of Applicable Advance Directive (Check Local sions that may be made by their surrogate. For example, some living
P
Jurisdiction) wills state the surrogate decision maker cannot override specif c
T
1. Court-appointed guardian instructions such as a request not to receive CPR. Some state laws
E
R
also suggest the patient must assent to, or minimally not re use,
2. Spouse or domestic partner (where legally recognized)
decisions made by the surrogate. This o ers a layer o protection
3
3. Adult children
4
to those patients who may not reach the threshold or decisional
4. Either parent capacity with respect to the decision at hand, but who are still aware
5. Adult siblings o and engaged with what is happening to them. When there is
M
disagreement between the patient and surrogate, and a clinically
e
6. Other amily members
d
appropriate compromise that is acceptable to both parties cannot
i
7. A close riend o the patient
c
a
be ound, it is advisable to contact the hospital’s ethics committee.
l
-
L
Because the surrogate is o ten an individual whom the patient has
e
g
not explicitly appointed to the role o decision maker, many states
a
relational ties such as when the patient has ull siblings, hal siblings, place restrictions on the types o end-o -li e decisions that can be
l
C
and step siblings who all consider themselves to be close relatives made by the surrogate. Broadly speaking, these restrictions may
o
n
o the patient on the same level o the surrogate hierarchy. I more require the attending physician and one additional physician to docu-
c
e
than one person has the same level o priority (such as several adult ment that the patient has a “quali ying condition,” such as a terminal
p
children), consensus is pre erred, but many states allow or a majority
t
illness or permanent unconsciousness, be ore honoring a request by
s
:
decision when consensus cannot be reached. The disagreeing party the surrogate to withhold or withdraw li e-sustaining treatment.
A
d
then has the option o turning to the court to assist in resolving the Generally, a proxy decision maker or a decisionally incapable
v
a
dispute. In states that do not appoint surrogates by statute, case law patient may conduct routine medical a airs or the patient includ-
n
may o er guidance, or the physician or ethics committee can nomi-
c
ing consulting with the patient’s health care providers, providing
e
nate the decision maker according to the standards o the institution. verbal or written consent or medical procedures, applying or pub-
D
i
lic benef ts such as Medicare or Medicaid, authorizing the release
r
e
Guiding standards
o in ormation and clinical records needed or continued care, and
c
t
Two basic principles should guide treatment decisions or decisionally
i
authorizing the trans er o the patient to or rom health care acili-
v
e
incapable patients: respecting and promoting the patient’s auton- ties. Care should be taken that such activities are warranted by the
s
a
omy, and ostering the patient’s well-being. All surrogates and care patient’s clinical condition, do not con ict with the patient’s known
n
providers have an obligation to ollow the in ormed verbal or written
d
wishes, and that they are undertaken by the proxy only when it
S
wishes o the patient and to act in the person’s best interests. They is not expected that the patient will be returned to a decisionally
u
r
should also take into account the person’s values and goals i those capable state in the time necessary to assure continuity o care.
r
o
are known. Three legal and ethical standards have been established
g
a
to guide such decisions: (1) The highest standard is a directly relevant Conflicts
t
e
autonomous directive where the patient’s wishes in regard to the
D
Occasionally, situations arise in which a surrogate makes a decision
decision at hand are known, either through documentation or discus-
e
or requests an intervention that con icts with either the patient’s
c
sion. (2) The most common standard used is substituted judgment
i
s
advance directive or other instructions the patient provided to
i
o
where the proxy decision maker is tasked with making the decision he
the care team while the patient was still decisionally capable. For
n
or she believes the patient would have made in this situation based
M
instance, it is not uncommon or a patient’s child to request CPR be
on what is known about the patient’s wishes, personal values and
a
provided even when the patient has specif cally requested a DNR
k
pre erence, and goals. (3) In the best interest standard, the patient’s
i
order. Plans and treatment goals previously established between
n
wishes are unknown or have never been known (such as cases involv-
g
the patient and the physician should not be changed without
ing in ants), and the proxy decision maker must weigh the risks and
concrete evidence that those decisions were made due to actual,
benef ts o all o the alternatives and make a decision that achieves
conceptual, or clinical error. The primary responsibility o the physi-
the greatest net benef t rom the perspective o the patient.
cian is to the patient, and decisions made a ter the patient loses
Confidentiality and HIPAA decisional capacity should urther the continued interests o the
patient. Additionally, surrogates are morally and legally required
Although medical care has always included the need to keep
to make decisions that con orm as closely as possible to those the
patients’ medical in ormation conf dential, the Health Insurance
patient would choose or himsel or hersel . Physicians thus are not
Portability and Accountability Act (HIPAA) has urther specif ed and
obligated to ollow requests by surrogates that do not comply with
codif ed the responsibility o health care providers. In order to make
the patient’s known wishes. I the physician is unable to resolve the
e ective decisions, proxy decision makers need in ormation about
con ict, it may be help ul to involve a third party to help mediate
the patient’s medical history and care. HIPAA regulations recognize
the situation such as an ethics consultant or committee, or hospital
this and entitle duly documented proxy decision makers to the
legal counsel. I an ethically, legally, and clinically sound agreement
same medical in ormation as the patient in regard to the decision
cannot be reached, court review may be necessary.
at hand. However, physicians should disclose only that in ormation
needed by the proxy to make an in ormed choice regarding the
■ SPECIAL CASES
decision at hand. I possible and/or when directed by the patient,
physicians should avoid discussing highly personal in ormation such Implied consent
as sexually transmitted diseases, HIV status, chemical dependency, In health care, the principle o implied consent is sometimes col-
mental illness, or any history o sexual or physical abuse, unless such loquially re erred to as “emergency consent” since it is most com-
in ormation is absolutely necessary in order or the proxy to make monly invoked when the ollowing conditions are met: (1) there
appropriate and in ormed decisions. is an emergency circumstance where the patient is unable to
229
participate in the in ormed consent process (usually because the administration o ANH be ore a proxy request may be honored.
patient is unconscious), and there is no available evidence o the Clear and convincing evidence would require either a written direc-
patient’s wishes not to receive the therapy; (2) no other proxy deci- tive or evidence o a serious, re ective discussion with the patient
sion maker is available to make decisions or the patient; and (3) the on the subject. Other states have lesser evidentiary standards, and
P
A
physician is compelled to immediately provide necessary treatment some include ANH within the parameters o li e-sustaining treat-
R
without which serious or irreversible harm to the patient’s li e or ment that may be terminated upon request o a proxy decision
T
health may result. Thus, implied consent is presumed when a person maker whenever the standards are met or withdrawal o li e-
I
needs help and cannot explicitly provide consent. Implied consent sustaining treatment. Physicians should re er to the specif c statutes
also re ers to situations where the patient does not expressly state, in their state or guidance.
either verbally or in writing that a procedure may be done, but his or
her actions imply consent. A common example is when the patient Physician assisted suicide
T
h
extends an arm a ter being told that the physician wants to draw The US Supreme Court ruled in 1997 that states may maintain laws
e
blood or laboratory analysis. Implied consent is legally accepted that prohibit or allow euthanasia and assisted suicide. A ew states
S
p
and provides a de ense against claims o battery, but not against now allow terminally ill adult patients, who are decisionally capable
e
claims o negligence. Physicians should ollow their institutions’
c
and able to communicate their wishes, to end their lives through
i
a
documentation standards when relying on implied consent in the
l
the voluntary sel -administration o a lethal dose o medication pre-
t
y
context o a medical emergency, and clearly identi y why emer- scribed by a licensed physician expressly or that purpose. In Oregon
o
gency treatment was necessary as well as the nature and immediacy
f
and Washington, the statutes are called the Death with Dignity Act.
H
o the threat. Patients must ulf ll several requirements be ore receiving the medi-
o
s
cation, including initiating verbal and written requests, undergoing
p
i
Minors
t
a second opinion consultation, receiving psychiatric intervention i
a
l
The rights o minors to make decisions about their medical care the patient is perceived to be depressed, and undergoing a 15-day
M
have expanded over the past decade. While in general, a parent waiting period. This process is sometimes called “physician-assisted
e
d
or legal guardian o a minor has to provide consent or treatment, suicide” because patients sel -administer the medication at a time
i
c
i
some jurisdictions are granting older minors, especially those over o their choosing with the intention o ending their li e. Because the
n
e
the age o 16, broad leeway to make decisions about their medical patient actively takes steps to end his or her li e, this is di erent than
a
treatment. Nearly all jurisdictions allow legally emancipated minors withholding or withdrawing medical treatment where barriers to
n
d
and minors who are pregnant or parents themselves to make medi- the dying process are removed. This is also di erent than voluntary
S
y
cal decisions or their care and their own minor children. However, active euthanasia where the physician acts upon the voluntary
s
t
some states now have statutes or case law that allow mature minors request o a decisionally capable patient and the physician inten-
e
m
to provide consent to procedures i they can demonstrate that tionally administers medications or other interventions to cause the
s
they are mature enough to understand and appreciate the nature patients death. In states that allow it, patients must request and sel -
o
f
and consequences o a proposed medical procedure or treatment. administer the lethal medication. No state allows the medications
C
Rather than reliance on an “all-or-none” phenomenon, the mature to be administered or requested by a proxy decision maker, even
a
r
e
minor doctrine allows or individual assessment o the stability o where there is clear evidence o the patient’s wishes.
the minor patient’s value system along with their emotional and
intellectual development. This approach recognizes that deci- SUGGESTED READINGS
sional capacity in a minor is a gradual process a ected by personal
characteristics and environment. The laws concerning the extent Appelbaum PS. Assessment o patients’ competence to consent to
minors are allowed to make their own medical decisions vary rom treatment. N Engl J Med. 2007;357:1834-1840.
jurisdiction to jurisdiction, and most jurisdictions are reluctant to
Burns JP, Edwards J, Johnson J, et al. Do-not-resuscitate order a ter
allow minors to re use potentially li e-sustaining treatment without
25 years. Crit Care Med. 2003;31(5):1543-1550.
involvement o the court. It is there ore advisable that each physi-
cian be amiliar with the local statutes. Johnstone MJ, Kanitsaki O. Ethics and advance care planning in a
culturally diverse society. J Transcult Nurs. 2009;20(4):405-416.
Artificial nutrition and hydration Jonsen AR, Siegler M, Winslade WJ. Clinical Ethics: APractical Approach
Medical and legal issues abound when it comes to the use o artif - to Ethical Decisions in Clinical Medicine, 7th ed. New York: McGraw-
cial nutrition and hydration (ANH), particularly in the dying or per- Hill; 2010.
manently unconscious patient. The US Supreme Court made it clear Scheyett AM, Kim MM, Swanson JW, et al. Psychiatric advance direc-
in its 1990 Cruzan decision that ANH is a medical treatment and a tives: a tool or consumer empowerment and recovery. Psychiatr
decisionally capable patient may re use any and all types o medical Rehabil J. 2007;31(1):70-75.
treatment, including ANH and other types o li e-sustaining treat- Siegel MD. End-o -li e decision-making in the ICU. Clin Chest Med.
ment. Patients may also execute a specif c instructional directive 2009;30:181-194.
that re erences ANH and clearly communicates the patient’s re usal
Valvano TJ. Legal issues in sexual and reproductive health care or
o ANH. However, state laws are highly variable in regard to proxy
adolescents. Clin Pediatr Emerg Med. 2009;10:60-65.
decision-maker requests to withhold or withdraw ANH, and many
living will laws expressly orbid the withholding or withdrawal o Von Gunten CF, Ferris FD, Emanuel LL. The patient-physician rela-
ANH as a means to shorten the dying process. Some states require tionship. Ensuring competency in end-o -li e care: communica-
“clear and convincing” evidence o the patient’s wishes to orgo tion and relational skills. JAMA 2000;284(23):3051-3057.
230
35
Medical malpractice is a serious concern or many physicians and
a topic that o ten prompts intense debate. In this chapter, we
review the elements o medical malpractice, as well as data about
the requency o both negligent medical care and actual claims o
medical malpractice. We will also review how well the malpractice
system achieves its purpose o deterring negligent medical care
and compensating patients who are harmed by such negligence.
We also discuss malpractice issues that are o particular concern to
Medical Malpractice hospitalists, and what can be done to reduce the risk o being the
subject o a medical malpractice claim.
THE ELEMENTS OF MEDICAL MALPRACTICE

Adam C. Schaf er, MD
Medical malpractice is a orm o negligence that applies to health
care providers including doctors, nurses, and institutional medical
care providers like hospitals. At the core o negligence-based liability
is the notion that individuals committing unintentional but reason-
ably avoidable acts that cause injury should be required to compen-
sate the victims o those acts. To determine whether negligence is
present in a given situation, courts require plainti s to prove our
elements through a preponderance o the evidence: duty, breach,
causation, and harm.
PRACTICE POINT
• To determine whether negligence is present in a given
situation, courts require plainti s to prove our elements
through a preponderance o the evidence: duty, breach,
causation, and harm.
• In the hospital setting, physicians have a duty to provide care
with the same skill and diligence as a reasonably competent
physician in the same specialty or eld o practice would under
similar circumstances.
• The question o whether a physician breached the duty o care,
then, o ten hinges on competing testimony provided by expert
witnesses as to the applicable standard o care and whether
the conduct in question ailed to meet that standard.
• To establish legal causation, the plainti must show that the
breach was both the “cause in act”and the “proximate cause”o
the injury.
The duty o care in negligence claims is a hypothetical stan-

dard by which the court judges the conduct o the de endant to
determine whether he or she had an obligation to act di erently.
In the hospital setting, physicians have a duty to provide care with
the same skill and diligence as a reasonably competent physician
in the same specialty or eld o practice would under similar cir-
cumstances. Failure to meet this standard constitutes a breach o
the physician’s duty o care. In most cases, or this duty to exist, a
physician-patient relationship must have been established.
In order to determine whether a physician has breached the duty
o care, an expert witness must testi y as to the applicable standard
in court. In the majority o states, physicians are judged by a national
standard o care that all physicians in the same specialty would be
expected to ollow. However, in a signi cant number o states, phy-
sicians are judged by what other physicians in the same specialty
231
and in the same geographic area would have done in a particular data rom a medical liability insurer, covering physicians in the New
situation. In either case, the relevant testimony must come rom England region. This study showed that the claims rate against hos-
expert witnesses who have the education, training, or other creden- pitalists (0.52 claims per 100 physician coverage years [PCYs]) was
tials that would make them amiliar with the applicable standard signi cantly lower than that or nonhospitalist internal medicine
P
A
o care. The question o whether a physician breached the duty o physicians (1.91 claims per 100 PCYs), and emergency medicine
R
care then, o ten hinges on competing testimony provided by expert physicians (3.50 claims per 100 PCYs). Among the claims led
T
witnesses as to the applicable standard o care and whether the against hospitalists, 32% resulted in payment, with a mean payment
I
conduct in question ailed to meet that standard. o $384,617. The severity o injury to the patient in the claims against
Even i a physician breaches this duty by ailing to adhere to hospitalists was high, with 50% o the claims involving the death o
the standard o care, the plainti in a case cannot establish liability the patient.
unless that breach is the actual cause o the injury. To establish legal Some studies have analyzed the epidemiology o medical injury
T
h
causation, the plainti must show that the breach was both the in speci c states. Examining more than 30,000 records o patients
e
“cause in act” and the “proximate cause” o the injury. As a breach hospitalized in New York State in 1984, the Harvard Medical Practice
S
p
o the duty o care is the “cause in act” o damages i the plainti Study is the largest study to assess the rate o medical malpractice
e
can establish that the presence o the breach was the “deciding ac- injuries and claims. This study showed that adverse events occurred
c
i
a
tor” in determining whether the damage would have occurred. Put in 3.7% o hospitalizations; o these adverse events, 27.6% were
l
t
y
di erently, a breach o the duty o care would not be the “cause in determined to be due to negligence. In a urther analysis o the
o
act” o harm i the harm would have occurred despite the negligent Harvard Medical Practice Study by Localio et al, the overall rate o
f
H
care o a physician. In addition to being the cause in act o harm, a malpractice claims per discharge was 0.13%. In this study, the vast
o
s
breach o the duty o care must also be the proximate cause in order majority o adverse events did not result in a malpractice claim. O
p
i
to satis y the causation element o negligence. To be the proximate the adverse events due to negligence that were identi ed, remark-
t
a
ably only about 2% resulted in malpractice claims. The estimated
l
cause o harm, the harm must be, by its nature, a oreseeable or
M
direct consequence o a breach o the duty o care. Some courts ratio o negligence to claims was 7.6 to 1.
e
d
require an additional or alternative nding that the breach was a Testing the generalizability o the results o the Harvard Medi-
i
c
“substantial actor” in causing the injury, especially when two or cal Practice Study, a subsequent, methodologically similar study
i
n
e
more parties may be responsible. by Thomas et al, examined 15,000 hospital records rom Utah and
a
I the court or jury nds that a physician has breached a duty Colorado. In this study, which yielded comparable results to the
n
d
by ailing to adhere to the applicable standard o care and that Harvard Medical Practice Study, 2.9% o hospitalizations in each
S
the breach is the cause in act and proximate cause o a patient’s state involved adverse events. O these adverse events, 32.6% were
y
s
injury, then the physician will be liable or damages. The measure a result o negligence in Utah, and 27.4% were a result o negligence
t
e
m
o such damages is o ten highly dependent on the acts and cir- in Colorado. Additional analysis o these data by Studdert et al in
s
cumstances surrounding the particular incident. Generally, a claim- 2000 showed that only about 3% o those patients who su ered
o
ant can recover compensatory damages or both economic and a negligent injury led a malpractice claim. Characteristics more
f
C
noneconomic harm. Economic damages include the speci c costs common among patients who su ered negligence but did not
a
r
associated with treating the injury, such as medical bills and drug le a malpractice claim include low income, uninsured, insured
e
expenses, as well as current and uture loss o earnings. Economic by Medicare or Medicaid, and age ≥75 years. O those malpractice
damages also include the costs o living with the injury such as claims identi ed during the study period, 78% were made despite
modi cations to the home to accommodate a wheelchair. Noneco- the absence o negligence and 56% were made despite the absence
nomic damages most o ten include pain and su ering (physical and o an adverse event. The ratio o negligent adverse events to claims
emotional) rom the injury. In addition, courts may order punitive was 5.1 to 1 in Utah and 6.7 to 1 in Colorado.
damages or injuries that are the result o malicious conduct or a The two main purposes o the medical malpractice system are
will ul disregard o patient sa ety. However, such instances are rare. to compensate patients who su ered injuries resulting rom neg-
Medical malpractice claims usually involve numerous medical ligence, and to deter negligent behavior by imposing costs on
personnel involved in every stage o the patient’s care. Malpractice physicians who practice negligently. These data call into question
plainti s cast a wide net when ling suit or a number o reasons. whether the medical malpractice system is achieving these objec-
First, it is o ten cost prohibitive to le an individual suit against each tives. Given the large number o adverse events due to negligence
de endant because o the increased costs o legal discovery. Second, not leading to a malpractice claim, the medical malpractice system
because most states employ a comparative negligence standard is not e cient at holding negligent physicians accountable, and
(meaning that total damages are calculated and then allocated many patients who have been injured as a result o malpractice are
to de endants based on their percentage o contribution o ault), not receiving compensation. One implication o these data is that
it is di cult to allocate damages among separate claims. It is also the rate o claims is a problematic metric to use in assessing quality
more likely that naming multiple de endants will help the plainti o care, since most episodes o negligence do not lead to malprac-
narrow down which o the de endants was actually at ault (i any). tice claims, and a signi cant number o malpractice claims are led
Third, joining multiple de endants in the same suit allows plainti s in the absence o negligence or injury.
to use the de endants’ own knowledge and testimony to establish A somewhat di erent picture emerges when the outcomes o
standards o care, cutting down on the costs o hiring independent claims are analyzed, rather than simply the ling o claims. Studdert
expert witnesses. Finally, there may be jurisdictional rules that pro- et al in 2006 evaluated 1452 closed malpractice claims in which
hibit separate claims and require naming all the responsible parties objective assessments were made by reviewers as to whether there
in a single claim i a ailure to do so would result in an un air out- were medical errors resulting in injury. O those claims led involv-
come or an increased burden on the judicial system. ing injuries, 63% were determined to be a result o error. In cases in
which there was injury due to error, compensation was paid 73% o
THE EPIDEMIOLOGY OF MEDICAL MALPRACTICE the time. In cases in which there were no errors, no compensation
Up until relatively recently, little data existed about liability environ- was paid 72% o the time. The authors o the study concluded that,
ment or hospitalists speci cally. However, data recently published although the malpractice system does a reasonable job o provid-
by Scha er et al, looked at claims rates or hospitalists, based on ing compensation only when there is injury as a result o a medical
232
error, the process has signi cant shortcomings. Namely, cases take a
long time to come to resolution (5 years, on average, rom injury to
CASE 35-1
C
disposition) and the monetary costs o litigating the claims are steep
H
(54% o the compensation paid). FAILURE TO FOLLOW UP ON AN INCIDENTAL FINDING
A
P
Thus the data show a very limited correlation between mal- A 62-year-old male with a signi cant smoking history presented
T
practice claims made and acts o actual malpractice. Based on the to the emergency department (ED) in November 1999 a ter a all
E
2006 data rom Studdert et al looking at the outcomes o claims, it resulting in a le t shoulder injury. The ED physician took x-rays o
R
appears that the majority o claims with merit result in compensa- the chest and le t shoulder, read them as showing no racture,
3
tion and the majority o meritless claims are denied compensation. and discharged the patient home. Four days later, the attending
5
However, the system o determining which claims have merit is radiologist read the x-ray as showing a le t lung nodule, and a
protracted and expensive. report o the x-ray was sent to the ED physician and primary care
M
physician (PCP). The radiologist did not call either the ED physician
e
AREAS OF MEDICAL MALPRACTICE OF SPECIAL
d
or the PCP. The patient saw his PCP twice in December 2000 or back
i
CONCERN TO HOSPITALISTS
c
and shoulder pain and was sent or physical therapy. The patient
a
l
■ INTRODUCTION presented to the ED in August 2001 with chest and shoulder
M
a
One analysis o 272 malpractice claims made against hospitalists pain. A chest x-ray was obtained, which the ED attending read as
l
p
ound that the most common contributing actor underlying the normal, but the radiologist noted a large mass in the le t lung. This
r
a
in ormation was not conveyed to the patient’s PCP. A ter another
c
cases was a problem with clinical judgment, such as ailing to order
t
i
an indicated diagnostic test or having too narrow a diagnostic ocus. visit to his PCP in September 2001, the patient presented to the ED
c
e
The second most common contributing actor was a breakdown in again in October 2001 with back and chest pain, and a chest x-ray
communication—either between the providers and the patient/ showed a mass occupying the majority o his le t lung. The patient
amily or among providers (Table 35-1). died o metastatic disease soon therea ter. The patient’s children
Many o the areas o malpractice risk o speci c concern to hos- led suit against the PCP, ED physician, and radiologist, and the
pitalists relate to communication issues. The discontinuity between suit was settled or more than $500,000.
inpatient and outpatient care that is inherent to hospital medicine, Adapted rom Wright J, McCormack P. Failure to act on
as well as the multiple hando s o patient care that can occur when incidental nding. CRICO Forum 2007;25:6-7.
hospitalists work shi ts, both increase the risk o a communications
breakdown that could result in injury due to negligence. Examples
o areas o liability concern or hospitalists related to inadequate The preceding case illustrates the liability pit alls that can result
communication include ailure to ollow up on incidental ndings rom inadequate communication among the physician ordering a
(Case 35-1) and appropriately addressing test results that may be radiologic study (the ED physician), the physician interpreting the
pending at the time o discharge. study (the radiologist), and the physician who is best suited to ollow
TABLE 35-1 Contributing Factors in Hospitalist Medical Malpractice Cases (n = 272)*
Contributing Factor # o Cases % o Cases (95% CI) De inition or Example

Clinical judgment 148 54.4% (48.3%-60.4%) Problems with patient assessment or choice o
therapy; ailure/delay in obtaining consult/re erral
Failure or delay in ordering a diagnostic test 36 13.2% (9.4%-17.8%)
Failure or delay in obtaining a consult or 35 12.9% (9.1%-17.4%)
re erral
Having too narrow a diagnostic ocus 34 12.5% (8.8%-17.0%)
Communication 99 36.4% (30.7%-42.4%) Issues with communication among clinicians or
between the clinicians and the patient or amily
Inadequate communication among 61 22.4% (17.6%-27.9%)
providers regarding the patient’s condition
Poor rapport with/lack o sympathy toward 15 5.5% (3.1%-8.9%)
and patient and/or amily
Insu icient education o the patient and/or 9 3.3% (1.5%-6.2%)
amily regarding the risks o medications
Documentation 53 19.5% (14.9%-24.7%) Insu icient or lack o documentation
Administrative 47 17.3% (13.0%-22.3%) Problems with sta ing or hospital policies and
protocols
Clinical systems 44 16.2% (12.0%-21.1%) Failure or delay in scheduling a recommended
test or ailure to identi y the provider
coordinating care
Behavior related 28 10.3% (7.0%-14.5%) Patient not ollowing provider recommendations;
seeking other providers due to dissatis action
with care
*An individual case may have multiple contributing actors. Categories including <10% o cases are not reported. Nonsubstantive categories, such as
inadequate in ormation available, are excluded. Where subcategories are speci ied, only the top three subcategories are reported.
Adapted rom Scha er AC, Puopolo AL, Raman S, Kachalia A. Liability impact o the hospitalist model o care. J Hosp Med. 2014;9(12):750 755.
233
up on the abnormal results (the primary care physician). This case the ef ects o the atenolol. A partner o the patient’s cardiologist
had eatures that are common in ED cases leading to malpractice was contacted and advised to put an external pacemaker on the
claims, including the misreading o plain radiographs, the involve- patient, but the cardiologist did not see the patient. The patient
ment o multiple individual ailures, and process breakdowns. was admitted to the ICU by a hospitalist. In anticipation that an
P
A
internal pacemaker might be needed, the hospitalist reversed
R
■ PENDING TESTS AND INCIDENTAL FINDINGS the patient’s war arin with resh rozen plasma. The ICU nurses
T
Hospitalists requently nd themselves in the same position as called the cardiologist to report that the patient was bradycardic
I
the ED physician in the above case, ordering a study o which and eeling unwell. The cardiologist never placed an internal
the nal results may not come back until a ter the patient has pacemaker. The hospitalist was next contacted by the ICU nurses
been discharged. The same problem applies to laboratory tests. once the patient was in cardiac arrest. The patient’s amily led
One study by Roy et al encompassing the hospitalist services at suit against both the cardiologist and the hospitalist. The court
T
h
two academic medical centers ound that 41% o patients had ultimately ound in avor o the hospitalist.
e
laboratory or radiology results pending at the time o discharge Domby v. Moritz (2008 Cal. App. Unpub. LEXIS 1856) illustrates
S
p
and in 9.4% o cases the results o these studies were considered this risk. In ling suit against both the cardiologist and the
e
c
potentially actionable. Seventy percent o the inpatient physicians hospitalist, the amily o the patient asserted that the hospitalist
i
a
should have ensured that the cardiologist physically came in to
l
and 45.8% o the outpatient physicians were unaware o these
t
y
potentially actionable results. evaluate the patient. Although the court ultimately ound in avor
o
o the hospitalist, this case shows that hospitalists have to take an
f
The problem o pending test results at the time o discharge is
H
best addressed at a systems level— or example, through a mecha- active role in discussing the treatment plan with consultants and
o
s
nism that automatically noti es the ordering provider o the nal in clearly delineating who has responsibility or which aspects
p
i
o the patient’s care. It may be legally hazardous to consider a
t
results o such tests. However, such systems are not widely in place
a
l
and even when they are, physicians still o ten ail to ollow up on clinical decision “not my call” and exclusively within the purview
M
clinically signi cant results. Consequently, physicians need to take o a specialist, because the hospitalist, as the attending physician
e
d
responsibility or ollowing up on the nal results o the tests that o record, may ace litigation based on the decisions made by the
i
c
consulting specialists.
i
they order.
n
e
Physicians may also be held responsible or responding to test
a
results ordered by another physician when these results come
n
d
back while that physician is on duty, as was held in Siggers v. Barlow
S
RISK FACTORS FOR MEDICAL MALPRACTICE CLAIMS
y
(906 F.2d 241). Responding to these test results o ten means com- AND STRATEGIES TO REDUCE THIS RISK
s
t
municating with the patient’s PCP about what additional ollow-up
e
In considering ways to reduce the risk o acing a medical mal-
m
needs to occur, such as serial imaging or an incidentally discovered
s
pulmonary nodule. Discharge summaries, while important, are gen- practice claim, a key question to ask is why patients decide to le
o
claims, given that the vast majority o patients who are injured
f
erally not adequate as the only means o communicating important
C
ndings that need to be ollowed up by the PCP. Kripalani et al and due to medical errors do not initiate a malpractice action. One
a
r
study, by Beckman et al examined 45 plainti depositions in
e
Pantilat et al identi ed a number o potential de ciencies in the
discharge summary as the sole means o communicating with the PCP. medical malpractice cases and ound that in 71% o cases there
These de ciencies include the possibility that the discharge sum- were signi cant relationship issues between the plainti and the
mary does not reach the correct PCP (occurring 25% o the time), de endant physician. The most common issue was the eeling by
ailure to include tests pending at discharge (occurring 65% o the the patient o having been deserted by the physician. Examples
time), and the PCP not receiving the discharge summary prior to include abandonment, and the physician being unavailable and
ollow-up (occurring 67% o the time). There ore, hospitalists should sending associates such as residents in the place o the attend-
contact PCPs directly regarding important test results or other mat- ing physician. Other relationship issues that were present in the
ters that need to be ollowed up, by phone and/or letter, and this examined depositions included: devaluing the patient (such as by
communication should be documented in the patient’s chart. discounting the patients’ illness or pain); delivering in ormation
poorly (including ailure to explain what was occurring); and ailing
to understand the patient’s or amily’s perspective (such as by not
■ COORDINATION OF CONSULTANT CARE asking or the patients’ opinion).
Another potential area o malpractice liability is the use and coor- The behavior o consulting specialists who are brought in a ter
dination o consultants. Hospitalists list active coordination o an adverse event has occurred may also inf uence whether a mal-
consulting specialists as one o the bene ts they bring to patient practice claim is led. In 54.8% o cases, health care pro essionals
care. However, with this responsibility or coordination o specialists, raised questions about the care the patient had received, and in
and in their role o the attending physician o record or the patient, 70.6% o these cases, the health care pro essional who cast doubt
hospitalists are at risk o incurring malpractice liability based on the on the quality o the care that had been provided by the de endant
actions o the consulting specialists (see Case 35-2). physician was a consultant who saw the patient a ter the adverse
event. In a couple o cases, it was an acquaintance—who happened
to be a health care pro essional but was not directly involved in
CASE 35-2 the case—who suggested that the care received was substandard.
There ore, consultants seeing a patient a ter an adverse event need
DOMBY V. MORITZ (2008 CAL. APP. UNPUB. LEXIS 1856) to be mind ul that even an o hand remark on the care the patient
A 67-year-old emale with a history o hypertension checked her has received may a ect whether the patient pursues a malpractice
own blood pressure, ound that it was elevated, and contacted claim.
her PCP. As instructed by her PCP, the patient took an extra dose These data suggest speci c measures that may be taken to
o atenolol, a ter which she had an episode o syncope. She reduce the risk o a malpractice claim being led. It is important
presented to the hospital, where she was bradycardic and so the to avoid those physician behaviors, such as creating conditions in
ED physician gave the patient atropine and glucagon to reverse which the patient may eel abandoned and not ully acknowledging
234
the patient’s concerns or discom ort. Given the possibility that
having associates such as residents or physician assistants see the
CASE 35-3
C
patient runs the risk o the patient eeling abandoned, the attending
H
COMPLAINTS AGAINST DOCTOR A
physician should explain the expected involvement o associates up
A
Dr A is a physician who joined the hospitalist service 4 years ago.
P
ront. It may also be help ul to rame the care to the patient as being
T
provided by a team, so the patient does not eel connected only to During his time as a member o the hospitalist group, his scores
E
the attending physician. It is also important to ensure that patient on patient satis action surveys have been in the lowest decile
R
expectations about the outcome o a procedure or treatment are o physicians at the hospital. As the director o the hospitalist
3
realistic. The in ormed consent process is an opportune occasion to service, you receive a call rom a manager in the patient relations
5
address the patient’s expectations. department saying that Dr Ahas been the subject o two complaints
Strong communication skills are also important. Supporting the within the past 6 months. The patient relations manager says both
M
bene t o good communication skills in reducing litigation, another complaints are very similar, and the complaints describe Dr A as
e
being unwilling to ully discuss his patients’ medical conditions.
d
study by Lester et al ound that physicians, who exhibited “positive
i
c
communication behaviors” such as making eye contact, acknowl- The complaints urther state that it seems like Dr A is always trying
a
l
edging what the patient says, and spending more time with the to get out o the patients’ rooms as quickly as possible and that
M
patient, elicited reduced litigious eelings in observers. Hickson et he appears annoyed when the patients ask questions. One o the
a
l
p
al in 1994 showed that the patients o obstetricians with a high patients who complained wrote: “Dr A just did not seem like he
r
a
requency o malpractice claims complained about these physi- really cared about me or my many medical problems. I would not
c
t
want him to take care o me again or any members o my amily.”
i
cians’ communication skills, including these physicians not listening
c
e
and not providing in ormation. The rate o these complaints about When you, as the director o the hospitalist service, meet with
poor communication was signi cantly higher or physicians with a Dr A about these issues, he seems irritated and explains that every
history o a high requency o medical malpractice claims than or physician has at least a ew disgruntled patients and that is part
physicians with a better claims record. o practicing medicine. Dr A says he wants to make sure he sees
On a systems level, it may be possible to identi y physicians all his patients and completes his billing orms promptly, and so
within an organization who are at increased risk o a malpractice he cannot be expected to linger in patients’ rooms. Dr A urther
claim. Physicians with an increased number o patient complaints explains that i a patient asks a question that he deems important,
have more risk management episodes, de ned as both malpractice then he makes sure to answer that question ully.
claims that are led and incidents reported by sta members to
the risk management department. One study by Hickson et al rom
I it is possible to recognize physicians at increased risk o mal-
2002 retrospectively examined a cohort o 645 physicians, look-
practice suits, such as Dr A in the preceding case, then potentially
ing or an association between the number o unsolicited patient
actions may be taken to mitigate this risk. For example, physicians
complaints and the number o risk management episodes. A small
who receive a high number o complaints or with particularly low
number o physicians generated a markedly disproportionate
satis action ratings could undergo educational programs aimed
number o patient complaints, with 9% o the physicians garnering
at enhancing their patient-communication skills. One approach,
more than 50% o the complaints. There was a signi cant positive
advocated by Moore et al consists o a tiered intervention system.
correlation between the number o complaints received and both
Initially, a physician who has been identi ed as being at high risk is
the total number o risk management episodes and the number
approached by a peer to discuss the issue. I that is not e ective in
o lawsuits.
improving the physician’s complaint rate, then a plan or improve-
Similarly, another study by Stel ox et al ound that scores rom
ment is developed in conjunction with someone in authority, such
a commonly used hospital satis action survey were signi cantly
as the department chair. I these e orts are still unsuccess ul, meet-
associated with risk management episodes, which included both
ings can be held with senior o cials in hospital management, with
malpractice lawsuits and incidents identi ed by risk management
the possibility o discipline or dismissal. Components o the plan to
as having the potential to result in a malpractice claim. The survey
reduce the complaint rate (and so potentially also the risk o acing
instrument included ve questions asking patients to rank their
a malpractice action) can include enhancements to the manage-
inpatient attending physician in di erent areas, using a scale o 1 to
ment o the physician’s practice, continuing medical education on
5 or each question, with a score o 5 denoting the highest rating.
the physician-patient relationship, and/or mental health evaluation.
Each 1-point decrement on the survey correlated with a 5% increase
in the rate o risk management episodes. The speci c questions on
the survey that had the strongest correlation with risk management DISCLOSING ERRORS TO PATIENTS
episodes were those regarding the time the physician spent with Physicians understandably are o ten conf icted about whether to
the patient and the concern the physician showed or the patient’s disclose medical errors. Historically, physicians have been hesitant to
qualms. No signi cant correlation was ound between the responses disclose mistakes or ear o inviting litigation over an error that may
to questions on how satis ed the patient was with the physician’s otherwise have gone unnoticed by the patient, and in order to avoid
skill and the rate o risk management episodes. There was also a possible censure over having made a mistake. An opposing view
positive correlation between the rate o patient complaints and the holds that disclosing errors will help avoid the strain in the patient-
rate o risk management episodes. Notably, a breakdown o com- physician relationship and the breakdown in communication that
plaints against physicians again suggested the crucial importance may occur a ter a mistake, and so may decrease the risk o litigation,
o good communication with the patient. O the 483 complaints or at least lead to smaller awards.
analyzed in the study, 75% o them concerned communication The empirical evidence is inadequate to clearly answer what
issues and 25% o them related to patient care matters. These two e ect disclosure o medical errors will have on the likelihood o mal-
studies by Hickson et al and Stel ox et al show that by using data practice litigation. Policy changes regarding noti cation o patients
that are commonly collected by hospitals—number o complaints about medical errors implemented by some medical centers do
and the results o patient satis action surveys—it may be possible provide examples o the possible consequences o disclosure poli-
to identi y physicians who are at elevated risk o being named in a cies. In 1987, the Lexington, Kentucky, Veterans A airs Medical Cen-
malpractice action. ter (VAMC), in reaction to two large malpractice payouts, decided to
235
put into place a policy o proactively identi ying and investigating are increasingly implementing standards requiring error disclosure.
cases o possible medical negligence. I medical negligence was Error disclosure, with the accompanying ability to gather data on
ound, the representatives rom the Lexington VAMC would have what types o mistakes are recurring, also supports the public policy
a ace-to- ace meeting with the patient or next o kin. At this meet- goal o improving systems so as to reduce uture errors.
P
A
ing, hospital representatives would explain the situation, answer Disclosure o medical errors is generally considered the ethi-
R
any questions, and o er restitution—the amount o which was cally appropriate course. Honesty is necessary to maintain a strong
T
based on a determination o actual loss. Claims assistance was also physician-patient relationship, and in ormed consent requires that
I
o ered. Reviewing 15 years o experience with this ull disclosure patients be ully aware o the circumstances surrounding their treat-
policy at the Lexington VAMC, Kraman et al concluded that this ment so they can decide about urther care. Demonstrating this
approach appeared to reduce the amount o overall malpractice trend toward increasing disclosure o medical errors, a consensus
payouts. Although the Lexington VAMC had an increased number o statement rom the Harvard-a liated hospitals in 2006 expressed a
T
h
payouts, the average amount o these payouts was relatively small, commitment to ull disclosure o medical errors in order “to change
e
at $14,500. This compares to a mean pretrial settlement amount our systems to prevent uture error” and because “it is the right thing
S
p
o $98,000 or all medical centers in the VA system. Despite the to do.”
e
Lexington VAMC being in the top quarter o all VAMCs in the num-
c
i
a
ber o tort claims led, it was in the bottom quarter o all VAMCs
l
t
PRACTICE POINT
y
in terms o total malpractice payouts. This VAMC experience has
o
major limitations regarding its generalizability, because physicians in •
f
Disclosure o medical errors is generally considered the
H
VAMCs do not pay individual malpractice premiums and, as ederal ethically appropriate course. Honesty is necessary to maintain
o
s
government entities, VAMCs are not subject to punitive damages. a strong physician-patient relationship, and in ormed consent
p
i
Several hospital systems and liability insurers have instituted
t
requires that patients be ully aware o the circumstances
a
l
programs that couple disclosure o unanticipated care outcomes surrounding their treatment so they can decide about urther
M
with rapid o ers o compensation in appropriate cases. The most care.
e
d
widely published program is the one implemented by the Univer-
i
c
sity o Michigan Health System (UMHS) in 2002. In this program,
i
n
e
unanticipated outcomes are promptly disclosed and investigated.
DEFENSIVE MEDICINE
a
The three principles the UMHS cites as de ning their risk manage-
n
d
ment approach are: (1) rapid o ers o compensation when “unrea- De ensive medicine, as de ned by a 1994 O ce o Technology
S
sonable” care was the cause o the injury; (2) orce ul de ense o Assessment report, is “when doctors order tests, procedures, or
y
s
claims in which the care provided was reasonable; and (3) use o visits, or avoid high-risk patients or procedures, primarily (but not
t
e
necessarily solely) to reduce their exposure to malpractice liability.”
m
knowledge gained rom the incidents to prevent uture injuries
s
and claims. With this policy in e ect, the UMHS saw a decrease in De ensive medicine may be categorized by whether it is positive,
o
monthly liability costs, in the rates o new claims, and in the time such as ordering o extra tests to try to orestall a malpractice claim,
f
C
to resolution o claims. As with the VAMC program, questions exist or negative, such as avoiding patients perceived as representing an
a
r
about the generalizability o the UMHS program, especially since increased malpractice risk. Some authors pre er the term “assurance
e
UMHS as an institution could assume legal responsibility when the behavior” in place o positive de ensive medicine, and “avoidance
outcome was due to systems-level problem. As a result o how tort behavior” in place o negative de ensive medicine, so as to avoid the
laws are structured in some states, claims in those states are usually suggestion o approval or disapproval about de ensive medicine.
led against individual physicians, rather than institutions, which Particularly in environments o high-liability stress, de ensive
might make some physicians hesitant to accept settlements, with medicine appears to be very common. A 2005 study by Studdert
the accompanying requirement o reporting the payment to the et al surveyed physicians in litigation-prone specialties (emer-
National Practitioner Data Bank. gency medicine, general surgery, orthopedic surgery, neurosurgery,
Despite the encouraging reports rom organizations implement- obstetrics/gynecology, and radiology) in Pennsylvania, which had
ing disclosure-and-o er programs, some uncertainty remains about experienced rapidly increasing malpractice premiums. O the physi-
disclosure as a risk management strategy, particularly when dis- cians who responded to the survey, 93% had engaged in de ensive
closures are not made in the context o compensation programs. medicine and 42% were limiting the scope o their practice because
A major legal concern about disclosure in the absence o some o ear o liability. The most common type o de ensive medicine
mechanism or awarding rapid and modest compensation is that, in the survey was ordering extra tests, which 59% o physicians
because most medical errors do not result in malpractice claims, reported doing. This was especially common among emergency
aggressive disclosure o medical errors may prompt claims that physicians, 70% o whom reported ordering extra tests. Physicians
would otherwise not have been led. A theoretical modeling o this concerned about whether their malpractice insurance coverage
issue by Studdert et al in 2007 concluded that routine disclosure was adequate and those who elt their insurance premiums were
would have a 94% likelihood o increasing malpractice compensa- particularly onerous were especially likely to engage in de ensive
tion costs. Regulatory protections that exist, such as state “apology medicine. Common negative de ensive medicine practices included
laws” designed to allow physicians to apologize without having it avoiding high-risk patients, reported by 39% o physicians, and
used against them, may provide only very limited protection. These avoiding high-risk procedures, reported by 32% o physicians.
laws may prevent expressions o regret rom being used against the Positive and negative de ensive medicine practices have di ering
physician, but not ancillary in ormation surrounding that expression implications or the health care system. Positive de ensive medicine
o regret, such as in ormation about causation or ault. has the potential to increase costs while o ering modest, i any,
Ultimately, one may expect to see progressively wider implemen- bene ts to patients. In contrast, negative de ensive medicine may
tation o policies encouraging or even requiring error disclosure. limit patients’ access to certain medical services viewed as high risk,
The basis or this expectation is independent o the e ect o error such as obstetrics.
disclosure policies on malpractice costs, but is instead based on Not only is de ensive medicine common, but it is also expensive.
regulatory, public policy, and ethical considerations. Some states Estimates o the costs o de ensive medicine vary and are raught
and accreditation organizations, such as The Joint Commission, with methodological limitations. One estimate is that approximately
236
5% to 9% o health care spending can be labeled as de ensive. A summary can have standardized sections dedicated to tests pend-
concern is that i de ensive medical practice becomes common ing at the time o discharge and issues requiring outpatient ollow-
C
enough, it may become the standard o care, which could orce all up. Having these sections in all discharge summaries ensures that
H
physicians to practice in a de ensive manner. the person preparing the discharge summary addresses these areas
A
P
Overall, there is no clear empirical evidence that de ensive and also gets PCPs accustomed to looking or this in ormation in
T
medicine a ects patient outcomes. There are some theoretical the discharge summaries. Even with standardized discharge orms,
E
arguments against the practice o de ensive medicine. Patients who important issues requiring outpatient ollow-up should still be
R
perceive that their physician is ordering a test or procedure or a directly communicated to the PCP, so as to minimize the chance
3
de ensive reason may react negatively to this and be more likely that these matters get overlooked.
5
to le a claim in the event o an adverse outcome. Some orms Checklists have been ound to reduce complications and mortal-
o de ensive medicine involve physical risk to the patient— or ity in the surgical setting, and the bene ts o checklists also extend
M
example, ordering unnecessary biopsies and other invasive proce- to the medical setting. Checklists in Hospital Medicine have the
e
d
dures. Particularly or these cases, services ordered primarily to serve potential to reduce errors that could give rise to a malpractice claim,
i
c
the desire o the physician or minimizing risk and not the medical such as the ailure to use appropriate deep vein thrombosis (DVT)
a
l
needs o the patient are ethically suspect. However, i ear o mal- prophylaxis in a patient who subsequently develops a pulmonary
M
practice causes physicians to lower their tolerance or the possibility embolus while in the hospital, or leaving a central venous catheter
a
l
p
that a signi cant nding, such as a cancer, could be missed, then in a patient who then develops a catheter-related bloodstream
r
a
this e ect is not necessarily deleterious. Indeed, some tests ordered in ection. Checklists could also improve e ciency, such as by mak-
c
t
i
primarily or solely to bene t the physician (by reducing medicolegal ing sure a patient who needs a physical therapy evaluation receives
c
e
risk) end up having clear bene t to the patient. In the aggregate, one promptly. To enhance the e ectiveness o a checklist in Hospi-
though, de ensive medical practices are likely cost ine ective. tal Medicine, other members o the care team, such as the nurses,
should be involved in ensuring the components o the checklist
ADDITIONAL STRATEGIES TO REDUCE THE RISK OF have been met, and are empowered to raise the issue when the
A MALPRACTICE CLAIM components o the checklist have not been met.
A number o di erent strategies can be employed to potentially In ormal “curbside” consultations is a potentially legally perilous
reduce the risk o a malpractice action. As discussed above, good practice. Questions to consultants about a speci c patient should
communication practices with patients and their amilies are cru- generally be made as a ormal request or consultation, not an in or-
cially important. Feelings on the part o patients that the physician mal “curbside” consultation. When consultants provide “curbside”
is unavailable or dismissive o the patients’ concerns may increase consultations, they are usually not seeing the patient and evaluat-
the risk o a malpractice claim. Delegating important communica- ing all the data, so their assessment may be based on incomplete
tion tasks should be avoided. Residents and other trainees may not in ormation. Moreover, a “curbside” consultation does not result in
provide complete in ormation to patients, may not convey in orma- a note rom the consultant in the chart, so the basis or the consul-
tion in a sensitive manner, and may not care ully document the tant’s recommendations will not be part o the medical record. A
communications they do have with patients. consultant who ormally sees the patient will also usually be able
A 2007 case decided by the Massachusetts Supreme Judicial to continue to ollow the patient as an outpatient, which can help
Court highlighted the importance o in orming patients about the with the transition o care to the outpatient setting and provide a
potential side e ects o their medications. The case, Coombes v. resource to whom the patient’s PCP can turn or assistance. I the
Florio (450 Mass. 182), concerned a 72-year-old patient on multiple name o a consultant who provides a “curbside” consultation is
medications (including oxycodone, tamsulosin, and oxazepam) who placed in the chart, then i a malpractice claim arises, it is likely that
was driving and atally struck a 10-year-old boy. The boy’s mother the consultant will be named in the claim. Table 35-2 summarizes
sued both the driver o the car and the driver’s physician, Dr Roland strategies designed to reduce the risk o a malpractice claim.
J. Florio. The Massachusetts court ruled that this was not a medical
malpractice case, because the boy who was killed and his mother COPING WITH A MALPRACTICE CLAIM
had no physician-patient relationship with Dr Florio. Nonetheless, Being the subject o a malpractice claim is usually intensely stress ul.
the court held that Dr Florio could still be subject to a negligence Common reactions to being sued include anger, depressed mood,
claim, because he did have a duty to make the patient aware o rustration, irritability, and insomnia. Samko and Gable have even
the side e ects o the medications the patient was taking so that the compared physicians’ reaction to a lawsuit with the ve Kübler-
patient could make an in ormed decision about whether it was sa e Ross stages o grie : denial, anger, bargaining, depression, and then
to drive. The court reasoned that i it was not sa e or the patient to acceptance. Physicians are at risk o personalizing the claim and
drive, then an accident, which might result in harm to parties other considering it an attack on their competence and character. The
than the patient, was a oreseeable consequence. The court drew process o adjudicating is commonly protracted, o ten taking 4 to
an analogy with a bar owner being ound negligent when some- 5 years rom the time o the adverse event to resolution o the case,
one becomes inebriated at the bar and then drives and becomes thereby adding to the stress o a malpractice claim.
involved in a atal collision. Approaches that may help physicians cope with the stress o
Inadequate communication among physicians, both between a malpractice claim include discussing the stress with trusted
hospitalists, and between hospitalists and PCPs, is a signi cant riends, amily, and colleagues. Discussions o speci c details o the
liability concern or hospitalists. These communications should be case should occur only in settings where privilege applies, such as
standardized whenever possible. Hando s o patient care between with one’s lawyer or with a therapist with whom one has a ormal
hospitalists should use a standardized orm so that crucial in orma- patient-clinician relationship. Open discussion with amily about the
tion, such as diagnostic uncertainties and the status o commu- accompanying stress may be especially help ul, since the stress o
nication with the PCP, is not overlooked. Although the discharge the malpractice claim is likely to a ect amily members. Colleagues
summary is not in and o itsel adequate as the only means to should express support when they know an associate is acing a
communicate important in ormation to a patient’s PCP, it can be malpractice action. Some pro essional societies also o er speci c
designed to help make sure the PCP receives important in orma- counseling resources or re errals or physicians trying to deal with
tion arising rom the hospitalization. For instance, the discharge the stress o a malpractice action.
237
TABLE 35-2 Strategies to Reduce the Risk o a Medical Malpractice Claim
Strategy Explanation
P
Maintain open and empathetic Inadequate or insensitive communication rom the physician is commonly cited as a reason that
A
communication with patients patients ile a malpractice claim. Good communication with the patient may reduce the likelihood that
R
and their amilies a malpractice claim is iled, especially in the event o an unexpected outcome
T
Be care ul about delegating Delegating important communication tasks to trainees runs the risk o the in ormation being conveyed
I
communication in an insensitive manner, the communication not being well documented in the chart, and the patient
taking o ense that the attending physician did not care enough to come in person
Standardize hando s Standardizing hando s, such as by having predesigned hando orms, helps ensure that important
in ormation, such as pending tests requiring ollow-up, are communicated to the incoming physician
T
h
Standardize discharge One way to make sure that the discharge summary contains crucial items, such as tests pending at
e
S
summaries the time o discharge and issues that require outpatient ollow-up, is to have a standardized discharge
p
template with sections prompting the inclusion o this in ormation
e
c
i
Directly communicate with Important items requiring outpatient ollow-up, such as an incidentally discovered lung nodule, should
a
l
t
PCPs be communicated directly to the PCP by letter or telephone call, and this communication should be
y
documented in the chart. A discharge summary alone is not adequate to communicate important
o
f
ollow-up matters
H
o
Use checklists Checklists can ensure that routine measures required or most patients, such as DVT prophylaxis, are
s
p
not overlooked. Implementation o checklists should involve the entire care team
i
t
a
Avoid “curbside” consultations Consultants who provide “curbside” consultations make recommendations based on what may be
l
incomplete in ormation and there is no record o the consultation in the chart
M
e
Recognize that hospitalists When a consultant is negligent, the hospitalist, as the attending o record, is likely to be named in the
d
i
can be held responsible or claim. When a hospitalist has concerns about the decisions o a consultant, this should be discussed
c
i
consultants’ decisions with the consultant. The responsibilities o the consultant should be clearly de ined
n
e
Collect and provide eedback Negative eedback rom patients about a physician, particularly about the physician’s communication
a
n
to physicians skills, can signal that this physician is at elevated risk o a malpractice claim. This eedback should be
d
conveyed to the physician, and a plan to remedy the identi ied de iciencies should be made
S
y
s
t
e
m
s
o
One o the reasons malpractice claims can be so stress ul or (NPDB), which was established by the Health Care Quality Improve-
f
C
physicians is that so much o their own identity revolves around ment Act o 1986. The Act was intended to improve the quality o
a
r
their pro ession. Realizing this, the physician should attempt to medical care, in part by requiring the submission o malpractice
e
depersonalize the claim. Most claimants have as their primary objec- payments to the NPDB, which can then be queried by health care
tive obtaining compensation, not vili ying the physician. Physicians institutions when making hiring decisions. Patients and individuals
dealing with a malpractice suit should use it as an occasion to assess do not have access to reports o physician malpractice payments
whether they have appropriate balance between their pro essional made to the NPDB, although physicians can request their own NPDB
lives and their leisure time. Spending time engaged in vocational les. Some states, however, have web sites that allow patients to
pursuits, such as hobbies and time with riends, is important. Physi- look up individual physicians and nd out in ormation about their
cians should also have the lawyer representing them explain what malpractice histories.
the process o adjudicating the claim will entail, so that the process
is demysti ed and surprises are minimized. The acts surrounding
the case should be examined to see i there is a systems-level issue CONCLUSION
that can be addressed to help prevent uture claims— or example,
Popular perceptions notwithstanding, the medical malpractice
designing a system or reviewing incoming radiology studies i a
system appears to do a reasonable job o awarding compensation
radiographic nding was missed.
primarily in cases that actually involve an injury due to negligence.
There are some speci c pit alls that must be avoided during
Nevertheless, the system is ine cient and expensive. In addition,
the stress o malpractice litigation. Physicians acing a malpractice
most adverse events resulting rom negligence never lead to claims
claim who do not have a ormal PCP should obtain one, as a PCP
or compensation, and meritless malpractice claims also remain a
can be help ul with medication or symptoms and re errals or
problem.
counseling. Sel -medication should be avoided. Insomnia is a com-
mon symptom arising rom the stress o a malpractice claim, and
physicians who eel medication is needed to treat insomnia should
discuss this with their own physician, and should not sel -prescribe PRACTICE POINT
or obtain medication rom a colleague in ormally. One action that
should never be taken is going back and altering any documents • In seeking to avoid malpractice claims, physicians need to be
conscientious about communicating with the patient, so that
in an e ort to assist one’s de ense. Not only is this unethical and
the patient does not eel abandoned or devalued. Although
potentially criminal, but also by the time a physician is aware that a
medical liability experience involving hospitalists speci cally
malpractice claim may be led, the ling party almost certainly has
copies o the medical records and related documents. is limited, issues hospitalists need to be care ul about include
When a claim results in payment on behal o an individual coordinating the actions o consulting specialists, ollowing up
physician, there are reporting requirements. In ormation about this on pending tests, and communicating with PCPs about issues
payment must be reported to the National Practitioner Data Bank that require outpatient ollow-up.
238
Legal citations SUGGESTED READINGS
C
Domby v. Moritz, 2008 Cal. App. Unpub. LEXIS 1856
Beckman HB, Markakis KM, Suchman AL, et al. The doctor-patient
H
Coombes v. Florio, 450 Mass. 182 (2007)
A
relationship and malpractice. Lessons rom plainti depositions.
Siggers v. Barlow, 906 F.2d 241 (1990)
P
Arch Intern Med. 1994;154:1365-1370.
Beilke v. Coryell, 524 N.W.2d 607, 610 (N.D. 1994)
T
Brennan TA, Leape LL, Laird NM, et al. Incidence o adverse events
E
Hill v. Medlantic Health Care Group, 933 A.2d 314, 325 (D.C. App. 2007)
R
Kent v. Pioneer Valley Hospital, 930 P.2d 904, 906 (Ut. App. 1997) and negligence in hospitalized patients. Results o the Harvard
Medical Practice Study I. N Engl J Med. 1991;324:370-376.
3
Palandjian v. Foster, 842 N.E.2d 916, 921-22 (Mass. 2006)
5
Polozie v. United States, 835 F. Supp. 68, 72-74 (D. Conn. 1993) Scha er AC, Puopolo AL, Raman S, et al. Liability impact o the hos-
Health Care Quality Improvement Act of 1986, Pub. L. No. 99-660, 100 pitalist model o care. J Hosp Med. 2014;9(12):750-755.
M
Stat. 3743 Stel ox HT, Gandhi TK, Orav EJ, et al. The relation o patient satis ac-
e
(codi ed as amended in scattered sections o 42 U.S.C.). tion with complaints against physicians and malpractice lawsuits.
d
i
c
Am J Med. 2005;118:1126-1133.
a
l
Studdert DM, Mello MM, Brennan TA. Medical malpractice. N Engl J
M
ACKNOWLEDGMENT
a
Med. 2004;350(3):283-292.
l
The author would like to thank Pro . Michelle M. Mello or her review
p
r
Studdert DM, Mello MM, Gawande AA, et al. Claims, errors, and
a
o the manuscript and Nicholas Beshara, JD, MPH or his contribution
c
compensation payments in medical malpractice litigation. N Engl
t
to the previous edition’s chapter.
i
c
J Med. 2006;354:2024-2033.
e
239
SECTION 8
Professional Development
241
CHAP TER
36 INTRODUCTION
Most physicians learn through in ormal approaches, including read-
ing, point-o -care learning, and consulting colleagues. More ormal
adult learning occurs mainly through continuing medical education
(CME). The approach to physician learning through CME is changing
and hospitalists are poised to play a crucial role in its development.
Hospitalists are increasingly the primary teachers in the hospital
setting and play a major role in per ormance improvement. Modern
CME integrates these two processes. In this chapter, we discuss
Principles o Adult principles o adult learning, the changing landscape o CME, and the
resultant responsibilities and opportunities or hospitalists.
Learning and ADULT LEARNING

Continuing Medical Adult learning is complex. Understanding the ramework o adult
learning theory can help in orm curricular design, teaching, and
Education evaluation. O the many theories o adult learning, three are most
in uential: the behaviorist, cognitivist, and constructivist theories.
No single theory ts the learning style o all adult learners, and most
educators use elements rom each.
Jef rey A. Tabas, MD • Behaviorism, popularized by B.F. Skinner, ocuses on using con-
sequences to shape behavior. A desired behavior is rewarded
Robert B. Baron, MD, MS with positive rein orcement, while undesired behavior is dis-
couraged with negative rein orcement. This theory emphasizes
that eedback is critical to learning.
• Cognitivism tries to explain learning through in ormation-
processing models and minimizes the ocus on the behavioral
response. It highlights the importance o in ormation that is
appropriately organized by the educator and the development
o problem-solving skills by the learner.
• Constructivism, popularized by Jean Piaget, teaches that learn-
ers construct new knowledge rom experiences they integrate
into their own existing ramework o understanding when
the experience is consistent with that ramework. When the
experience is inconsistent with that ramework, they either
change their perceptions o the experience or re rame their
internal model o understanding. This theory emphasizes the
educator’s role as acilitator instead o didactic teacher and the
learner’s need or a social and active learning process.
Together, these theories suggest that adults learn most ef ectively
when they (1) perceive the relevance o educational material, (2) are
actively engaged, (3) have input into choosing educational experi-
ences and directing their own learning, and (4) have the chance to
step back and re ect on their learning.
Moore has proposed that adult learning involves a ve-stage
process. These are: (1) recognizing an opportunity or learning; (2)
searching or resources or learning; (3) engaging in learning to
address an opportunity or improvement; (4) trying out what was
learned; and (5) incorporating what was learned. Learning occurs
not as a linear progression through these stages, but as a dynamic
process with complex interactions that include revisiting and con-
currence o various stages. CME may stimulate the rst stage by pro-
viding the recognition that the opportunity or learning exists (“I did
not know that continuous positive airway pressure [CPAP] decreases
intubation in patients with congestive heart ailure [CHF]”) or the
third stage by providing the learning needed to address the oppor-
tunity or improvement (“I developed the competence to appro-
priately select candidates or CPAP and the steps to implement it”).
243
As described above, learners most readily progress through stages through auditing individual physicians or practice groups, or
when they see relevance (“ICU beds are limited and I can save hos- rom reported hospital, regional, or national data involving
pital resources by avoiding intubation”), are actively engaged (“The registries or national quality measures. Per ormance can re ect
CME presentation used dynamic learning approaches such as case patient outcomes, such as mortality or readmission rates, or
P
A
presentations, question and answer, or audience response systems”), process measures, such as counseling or smoking cessation
R
have chosen the subject (“I want to learn how to implement CPAP”), in patients with pneumonia. Optimal per ormance can be
T
and re ect on their learning experience (“Is this something that determined through assessing practice guidelines, medical
I
would work in my institution and do I need more learning to ef ec- literature, national benchmarks, and the like. The dif erence
tively implement this?”). between the two is the practice gap. Pro essional practice gaps
are not limited to patient care but can also involve other areas,
CME EFFECTIVENESS such as research or administrative practice.
T
h
The best available evidence suggests that CME is ef ective in achiev- 2. Identi y the educational needs—improved knowledge, com-
e
ing and maintaining knowledge, competence, and procedural skills, petence, or per ormance—that should be addressed to close
S
p
as well as improving physician per ormance and patient health the practice gap. These can be determined by surveying
e
learners, interviewing thought leaders in the content area, or
c
outcomes, i the activity is planned and implemented according to
i
a
reviewing literature.
l
recommended approaches. Assessments show that interventions
t
y
using live educational strategies are more ef ective than print, that 3. Identi y which outcomes the activity is designed to improve—
o
competence (strategies to apply knowledge), per ormance
f
multimedia are more ef ective than single media, and that multiple
H
exposures are more ef ective than a single exposure. Simulation (what is done in practice), or patient outcomes—and how
o
s
methods in medical education seem e ective in disseminating they will be measured ( or example by looking at changes in
p
i
hospital quality measures). Improved knowledge alone, which
t
psychomotor and procedural skills.
a
l
may be an educational need, is not an adequate outcome. For
M
THE ROLE OF THE HOSPITALIST IN CONTINUING example, the educational need may be a lack o knowledge o
e
d
EDUCATION the indications or CPAP in patients with CHF, but the outcome
i
c
should be the ability to apply that knowledge to appropriately
i
Research reveals signi cant gaps between the medical care that
n
e
patients actually receive and the care they should be getting. Hos- select the patients that might bene t rom the therapy when
a
pitals that have hospitalists provide care closer to the ideal—their presented with several case scenarios.
n
d
care is associated with better per ormance measures, such as 4. Select an appropriate educational ormat to encourage this
S
improved diagnosis, treatment, counseling, and prevention, as well change ( or example, a case-based lecture with spaced learn-
y
s
ing using ollow-up e-mail reminders).
t
as decreased length o stays and hospital costs. The hospitalist plays
e
m
an essential role in closing the quality gaps through implement- 5. Identi y the potential barriers to change, and describe how to
s
ing guidelines and reporting and implementing quality measures. address them.
o
6. Identi y ways to cooperate and collaborate outside o the CME
f
Much o that role involves changing physician behaviors, which
C
is the essence o CME. Given that many nationally reported activity to help acilitate change, such as interaction with the
a
r
quality department or other organizations.
e
quality measures—heart ailure, acute myocardial in arction, pneu-
monia, asthma, venous thromboembolism, and stroke care, among Educators who thoroughly understand and apply this approach
others— all within the purview o the hospitalist, the hospitalist will be better able to provide the quality learning needed to ef ect
must be versed in the theory o CME and be prepared to deliver the desired improvements.
such activities.
OTHER FORMS OF CME
THE CHANGING LANDSCAPE OF CME
Point o Care CME (PoCCME) is CME developed by an accredited
Continuing medical education has traditionally used didactic lec- provider that includes sel -directed online learning. A physician
tures and reading ollowed by testing to con rm knowledge, with answers a clinical patient question in real time using an evidence-
little heed paid to the importance o physician practice behaviors. based source and then documents the question, the sources
The ocus o modern CME has evolved rom increasing knowledge consulted, and the resultant application to practice. For example, a
to improving physician competence, physician per ormance, and physician needs to know the parameters used to determine when a
patient outcomes. This is re ected in the drive to incorporate prac- patient with CHF can be weaned of CPAP. Working with an accred-
tice-based learning and improvement into all aspects o continuing ited provider, she goes online to determine the criteria or weaning,
education and accreditation. Driven by the link to quality in patient and then documents her question, sources, and how she will use
sa ety, the American Board o Medical Specialties has mandated that the in ormation. That physician can then receive PoCCME credit.
all specialty boards adopt practice per ormance assessment as the Per ormance Improvement CME (PICME) is awarded when a
ourth component in maintaining certi cation. The American Board physician or group o physicians per orms a three-step process in
o Internal Medicine was one o the rst specialty boards to adopt which they (1) learn how to measure per ormance and then assess
this requirement. The Federation o State Medical Boards has also their practice, (2) develop an intervention based on best practice,
discussed the value o per ormance improvement in maintaining and (3) remeasure per ormance and then re ect on the impact o
licensure. As a result o this changing ocus, the Accreditation Coun- the intervention. PICME typically involves longer-term interven-
cil or Continuing Medical Education (ACCME) has incorporated sel - tions and activities that require chart review or data collection.
directed physician improvement and change as a desired outcome These activities may address the structure, process, or outcome
o CME activities. ACCME has developed the ollowing systematic o a physician’s practice with direct implications or patient care.
approach to ensure appropriate planning, implementation, and The goal is or CME to be an active process that occurs within the
assessment o each CME activity: clinical care setting, as opposed to a more passive process in a non-
1. Identi y the pro essional practice gap—the dif erence between clinical setting. It acknowledges that some process changes require
current practice and optimal per ormance—appropriate to systems-level change. Objectives include learning the per ormance
learners o this activity. Current practice can be identi ed improvement process, taking an active role in learning and change,
244
and attempting to directly improve patient care processes and out- CONCLUSION
comes through an educational activity.
C
Hospitalists are poised to play a major role in the uture o CME.
For example, a hospitalist group identi es a gap between the
H
This will include use o interactive case-based learning, simulation,
observed Central Line-Associated Bloodstream In ection (CLASBI)
A
and other learning ormats, in the same way we teach students
P
rate and the desired rate, based on national guidelines. They
and residents. It will also include de ning practice gaps, providing
T
develop a tool to sel -report compliance with the central catheter
process and outcomes data as a basis or sel -assessment, teaching
E
insertion “bundle” and measure the rate o compliance. The group
R
per ormance improvement principles as a template or educational
then implements a training activity and monitors the sel -reported
3
interventions, and incorporating point-o -care learning and teach-
rate o compliance or subsequent CLASBI in ection rates. I rates
6
ing into their practice.
have not improved, they analyze their approach again, looking or
other needs that can be addressed to improve these outcomes.
P
While drawbacks to this approach are the prolonged ef ort and time SUGGESTED READINGS
r
i
n
required, these may not be any greater than the ef ort, time, and
c
Accreditation Council or Continuing Medical Education (ACCME).
i
p
cost o traveling to a several-day live meeting, and may yield sig-
l
Proposal or New Criteria or Accreditation with Commendation.
e
ni cantly greater improvements in care. With their established track
s
http://www.accme.org/requirements/accreditation-requirements-
o
record in patient sa ety and quality and the advent o PICME, hos-
pitalists are a natural resource or implementing this orm o CME. cme-providers/proposal-new-criteria-accreditation-commendation.
A
April 23, 2014. Accessed October 1, 2015.
d
The American Board o Medical Specialties (ABMS) released
u
l
updated standards in 2015 or maintenance o certi cation part Cervero RM, Gaines J. The impact o CME on physician per or-
t
L
4 activities—Improvement in Medical Practice. These require the mance and patient health outcomes. J Contin Educ Health Prof.
e
a
physician to demonstrate competence in systematic measurement 2015;35(2):131-137.
r
n
and improvement in patient care. Such activities can also meet the
i
Combes JR, Arespacochaga E. Continuing Medical Education as
n
g
requirements or PICME, thereby allowing multiple orms o credit a Strategic Resource. http://www.ahaphysician orum.org/
a
or activities that are integrated into the normal work ow o practice resources/leadership-development/CME/index.shtml. September
n
d
to improve patient care. 2014. Accessed October 1, 2015.
C
o
Iglehart JK, Baron RB. Maintenance o certi cation. N Engl J Med.
n
PRACTICAL IMPACT OF THE HOSPITALIST ON CME
2013;368(13):1262-1263.
t
i
n
The role o hospitalists in CME continues to expand. Programs have
u
Institute o Medicine (IOM). Measuring the Impact o Interpro es-
i
been developed or hospitalists as agents o change in venues rang-
n
sional Education and Collaborative Practice and Patient Out-
g
ing rom university CME programs, to the Society o Hospital Medi-
comes. http://www.iom.edu/Reports/2015/Impact-o -IPE.aspx.
M
cine, to the Agency or Healthcare Research and Quality (AHRQ).
April 22, 2015. Accessed October 1, 2015.
e
Hospitalists should incorporate the principles described above
d
i
McMahon GT. Advancing continuing medical education. JAMA.
c
into the CME planning and education process. This would include
a
2015;314(6):561-562.
l
close collaboration and communication between educators and
E
d
the institutional staf responsible or quality and sa ety initiatives. O’Leary KJ, A sar-Manesh N, Budnitz T, et al. Hospital quality and
u
c
Approaches to planning internal CME such as grand rounds might patient sa ety competencies: development, description, and rec-
a
t
include comparing hospital-quality data to national guidelines ommendations or use. J Hosp Med. 2011;6(9):530-536.
i
o
to identi y gaps and then surveying physicians to identi y needs
n
Rosenbluth G, Tabas JA, Baron RB. What’s in it or me? Maintenance
related to those gaps. An example o a planned activity a ter identi - o certi cation as an incentive or aculty supervision o resident
cation o a practice gap and learning needs, such as or CLASBI rates, quality improvement projects. Acad Med. 2016;91(1):56-59.
might include hands-on skills workshops with trainers to improve
Tabas JA, Baron RB. Commercial unding o accredited continuing
techniques (per ormance) in central line insertion practices. CME
medical education. BMJ. 2012;344:e810.
activity assessment could include be ore-and-a ter surveys to deter-
mine whether the activity’s content relates to their practice and how
it has changed that practice. More robust assessment would include
measuring changes in process or patient outcomes in conjunction
with the hospital quality department.
245
CHAP TER
37 INTRODUCTION AND NEEDS ASSESSMENT
During the last 25 years cultural diversity has increased dramatically
due to global migration. The 2013 International Migration Report
rom the Department o Economic and Social A airs o the United
Nations revealed that the United States has six times as many immi-
grants as all o Latin America. The United States also resettles the
largest number o migrants in the world and provides more bene ts
and wel are than any other nation. The United States hosts about
20% o the world’s global migrants.
Cultural Competence In 2013, about 39% o the US population identi ed themselves
as members o minority groups. Hispanic and A rican American
groups were the largest minority groups, accounting or 17% and
13% o the population, respectively. By 2050, it is projected that
Germán E. Giese, MD minority groups will account or almost hal o the US population
(Figure 37-1).
Sarahi Rodríguez-Pérez, MD
While A rican Americans and Hispanics are the largest minority
Efrén Manjarrez, MD, SFHM groups in the United States, they are also the most underrepre-
sented minorities in medicine. Only 6% o practicing physicians
come rom these groups (Figure 37-2). These physicians o ten
carry the responsibility o providing health care or these minority
communities. Although the number o minority students entering
medical school is increasing, it is not increasing at a rate to ensure
a culturally competent physician work orce (Table 37-1). The racial
and ethnic disparities are an ongoing phenomenon in both medi-
cal school enrollment and graduating physicians rom US medical
schools (Table 37-2). The inclusion o International Medical Gradu-
ates (IMGs) into residency programs in di erent specialties might
mitigate some o these disparities. However, the proportion o
residency positions lled with IMGs is still small compared to the
number lled by US graduates (Table 37-3).
IMPACT OF CULTURAL COMPETENCE AND DISPARITIES

ON HEALTH CARE QUALITY
The National Quality Forum (NQF) de nes cultural competency as
the “ongoing capacity o health care systems, organizations, and
pro essionals to provide or diverse patient populations high-quality
care that is sa e, patient and amily centered, evidence based, and
equitable.” Cultural competency may be achieved through “poli-
cies, learning processes, and structures by which organizations and
individuals develop the attitudes, behaviors, and systems that are
needed or e ective cross-cultural interactions.” When caring or
di erent cultural groups, physicians should ideally have awareness
and appreciation o the speci c traditions, and cultural and religious
values o their patients, as these actors greatly in uence human
behavior and decision making.
Di erences in ethnicity and traditions may impact the patient-
provider relationship, speci cally when there is discordance
between provider and patient ethnicity or language, and impact
e ective delivery o quality health care. Health care systems need to
be able to understand and accommodate multicultural population
needs. In addition to this being “the right thing to do,” understand-
ing and accommodating diverse patient populations will also a ect
health care system reimbursement. For example, the Physician
Quality Reporting System (PQRS, ormerly known as the Physician
Quality Reporting Initiative), is a health care quality improvement
incentive program that was initiated by the Centers or Medicare
and Medicaid Services (CMS) in 2006; it constitutes an example o
a “pay or per ormance” program that nancially rewards providers
246
(Middle -s e rie s proje ctions )
C
1990
H
75.7 2000
71.6
A
2025
P
62.0 2050
T
E
52.5
R
3
7
22.5
C
15.7 16.8
12.3 12.8 14.2
u
7.5 10.3 11.3
l
9.0
t
u
4.4
3.0
r
0.8 0.9 1.0 1.1
a
l
C
White , Bla ck Ame rica n India n, As ia n a nd His pa nic Origin
o
Non His pa nic Es kimo, a nd Ale ut Pa cific Is la nde r (of a ny ra ce )
m
p
Figure 37 1 Percent of US population by race and Hispanic origin: 1990, 2000, 2025, 2050.
e
t
e
n
c
or reporting health care quality data to CMS. Beginning in 2015, care people should receive, and the quality o care they actually
e
the program will apply a negative payment adjustment to practices receive. Another in uential IOM report, Unequal Treatment: Confront-
which do not satis actorily report data on quality measures. The ing Racial and Ethnic Disparities in Health Care (2002), concluded that
PQRS program includes 254 quality measures urther grouped into people o color o ten receive lower-quality care than their white
six main health care quality domains, one o which is related to counterparts—even when insurance and socioeconomic status,
patient satis action. The same is true o the hospital Value-Based comorbidities, stage o presentation, and other actors are taken
Purchasing program; hospitals can gain or lose up to 2% o their into account.
CMS unding based on their per ormance in the program, o which Health disparities at any level pose a signi cant threat to patient
25% rests on patient satis action scores. Meeting a patient’s needs sa ety and there ore represent an opportunity to implement insti-
with cultural competence (and there ore ensuring their satis action) tutional quality improvement strategies to mitigate the dispari-
is now necessary to be nancially viable in many value-based pur- ties. For example, improving patient–provider communication via
chasing programs. interventions such as the appropriate use o interpreters should
The 2001 Institute o Medicine (IOM) report, Crossing the Quality decrease the risk o misdiagnosis, decrease unnecessary procedures
Chasm, identi ed a signi cant gap between the quality o health and diagnostic testing, and increase the participation o patients
Bla ck, 3.3%

His pa nic/La tino, 2.8% (30,598) Na tive Ame rica n, 0.3%
(26,094) (2457)
As ia n, 5.7%
IMG,*** 23.6%
(53,799)
(221,633)
White ,* 36.7%
(344,821)
Unknown,** 27.6%
(258,950)
Note : As ia n include s Chine s e , Filipino, Kore a n, Ja pa ne s e , Vie tna me s e , India n/Pa kis ta ni, Othe r As ia n;
His pa nic/La tino Include s Mexica n Ame rica n, C’We a lth Pue rto Rica n, Ma inla nd Pue rto Rica n, Othe r His pa nic;
Na tive Ame rica n include s Ame rica n India n/Ala s ka Na tive , a nd a s of 1996, Na tive Hawa iia ns.
*The s e da ta ma inly re pre s e nt White phys icia ns who gra dua te d from U.S. a llopa thic me dica l s chools from 1978 to 2004.
**Include s phys icia ns who gra dua te d from Ca na dia n me dica l s chools, doctors of os te opa thic me dicine , a nd mos t
like ly White s who gra dua te d prior to 1978.
***Inte rna tiona l Me dica l Gra dua te s .
Da ta S ource : AAMC Da ta Wa re hous e : Minority P hys icia n Da ta ba s e , Applica nt-Ma tricula nt file, a nd
AMA P hys icia n Ma s te rfile , a s of Ma rch 16, 2006.
Figure 37 2 2004 US physicians by race/ethnicity.
247
TABLE 37-1 Total US Medical School Enrollment by Race, Ethnicity, and Sex
Student Race/Ethnicity Responses1 2013-2014 2014-2015

P
Women2 American Indian or Alaska native 94 97
A
R
Asian 8,329 8,511
T
Black or A rican American 3,185 3,188
I
Hispanic, Latino, or o Spanish Origin 11,527 1,615
Native Hawaiian or Other Paci ic Islander 47 56
White 20,623 20,895
T
Other 402 529
h
e
Multiple Race/Ethnicity 3,447 3,560
S
p
Unknown Race/Ethnicity 494 648
e
c
Non-U.S. Citizen and Non-Permanent Resident3 762 789
i
a
l
Total For Women 38,910 39,888
t
y
Men2 American Indian or Alaska Native 89 108
o
f
Asian 8,818 8,885
H
o
s
p
i
t
a
l
M
White 26,371 26,497
e
d
Other 507 635
i
c
i
n
e
Unkown Race/Ethnicity 655 799
a
n
d
Non-U.S. Citizen and Non-Permanent Resident3 853 852
S
Total for Men 44,481 45,372
y
s
t
All American Indian or Alaska Native 183 205
e
m
Asian 17,147 17,396
s
o
f
C
a
r
e
White 46,994 47,392
Other 909 1,164
Unknown Race/Ethnicity 1,149 1,447
Non-U.S. Citizen and Non-Permanent Resident3 1,615 1,641
Total 83,391 85,260
1
In 2013, the methodology or acquiring race/ethnicity in ormation was updated. Rather than one question asking an applicant’s Hispanic origin and a sec-
ond question asking the applicant’s race, the Hispanich origin and race response options are now listed together under a single question about how appli-
cants sel -identity, Applicants could select multiple response options.
2
Four applicants in 2013 and six applicants in 2014 who declined to resport gender are not re lected.
3
The “Non-U.S. Citizen and Non-Permanent Resident” Category may include students with unknown citizenship.
and amilies in clinical decision making. Utilizing providers that are o physicians and health care providers are unable to articulate
amiliar with the social and cultural traditions o patients is an e ec- approaches or culturally sensitive strategies to practice in a culturally
tive strategy to improve their health care outcomes. For example, competent way, and there is an historical lack o standardized train-
a study per ormed at a rural clinic in Canada treating an aboriginal ing nationally. This is consistent with the literature, which suggests
community ound that the introduction o culturally sensitive that although “cultural competence” is broadly supported, consen-
programs was associated with increased appointment attendance sus de nition, detailed understanding, and standardized training
( rom 25% be ore the program to 85% a ter the program). Involve- rameworks are lacking.
ment o aboriginal nurses, inclusion o culturally sensitive activities, Facing this reality, the US Department o Health and Human Ser-
and participation in spiritual ceremonies were also associated with vices’ Of ce o Minority Health published in 2000 the rst National
higher patient satis action, higher trust toward the health care team, Standards for Culturally and Linguistically Appropriate Services in Health
and better communication amongst participants. Care (National CLAS Standards), which provided a ramework or all
health care organizations to best serve the nation’s increasingly
STANDARDS FOR CULTURAL COMPETENCE diverse communities (later updated between 2010 and 2013). The
Un ortunately, cultural competence is a vague concept or ront line CLAS Standards are divided into a “principal standard,” plus 14 other
health care providers. That is one o the reasons why speci c strate- standards grouped into three domains which are (1) Governance,
gies to improve it are still lacking across the board. The majority Leadership, and Work orce, (2) Communication and Language
248
TABLE 37-2 Total US Medical School Graduate Race/Ethnicity and Sex, 2012-2013 and 2013-2014
C
H
Graduate Race/Ethnicity Responses1 Class of 2013 Class of 2014
A
Women2 American indian or Alaska Native 11 16
P
T
Asian 1,837 1,829
E
Black or A rican American 656 675
R
Hispanic, Latino, or o Spanish Origin 446 479
3
7
Native Hawaiian or Other paci ic Islander 3 3
White 4,816 4,674
C
Other 132 126
u
l
Multiple Race/Ethnicity 634 587
t
u
r
a
l
Non-U.S. Citizen and Non-permanent Resident3 180 172
C
o
Total for Women 8,720 8,579
m
p
Men2 American Indian or Alaska Native 14 11
e
t
Asian 1,809 1,872
e
n
Black or A rican American 336 377
c
e
Hispanic, Latino, or o Spanish Origin 429 449
White 5,947 5,784
Other 166 174
Multiple Race/Ethnicity 556 581
Total for Men 9,434 9,499
All American Indian or Alaska Native 25 27
Asian 3,646 3,701
Black or A rican American 992 1,052
Hispanic, Latino, or o Spanish origin 875 928
White 10,763 10,458
Other 298 300
Total 18,154 18,078
1
In 2013, the methodology or acquiring race/ethnicity in ormation was updated. Rather than one question asking an applicant’s Hispanic origin and a
second question asking the applicant’s race, the Hispanich origin and race response options are now listed together under a single question about how
applicants sel -identity, Applicants could select multiple response options.
2
Four applicants in 2013 and six applicants in 2014 who declined to resport gender are not re lected.
3
The “Non-U.S. Citizen and Non-Permanent Resident” Category may include students with unknown citizenship.
Assistance, and (3) Engagement, Continuous Improvement, and associated with reluctance to participate in medical research, and
Accountability. The principal standard is “to provide e ective, equi- may be associated with low rates o trust (and subsequent adher-
table, understandable and respect ul quality care, and services that ence to treatment). Patterns o trust in di erent components o our
are responsive to diverse cultural health belie s and practices, pre- health care system (physicians vs hospitals vs insurance companies)
erred languages, health literacy, and other communication needs.” also di er by race. Di erences in trust may re ect di erent cultural
experiences and expectations or care.
IDENTIFYING AND OVERCOMING BARRIERS OF One major barrier in mutual cultural understanding is the lack o
CULTURAL COMPETENCE intercultural care training within the health care work orce. Hiring
Cultural and social di erences represent imaginary lines separat- translators or translation services can easily overcome a language
ing multicultural health care providers and patients, predisposing barrier, but it will not overcome cultural and social barriers. Most
and contributing to stereotyping, xenophobic, and discriminatory health care organizations have not implemented ormal training
behaviors. For example, the 1932 US Public Health Service Tuskegee programs to enhance culture competence among their work orce.
Syphilis Study on Untreated Syphilis in the Negro Male was a ed- Health care institutions and provider groups should promote hir-
erally unded trial that withheld available treatment rom A rican ing ethnically diverse physicians and pro essional sta , as working in
American men with syphilis. Research has demonstrated that A rican a culturally diverse environment allows prejudice and stereotypes to
Americans’ knowledge o this history o racial discrimination is be contrasted with reality through awareness and understanding.
249
TABLE 37-3 Matches by Applicant Type, 2015
Number of Number U.S. U.S. 5th U.S. Non-U.S. Number

P
Specialty Positions Filled Senior Grad Osteo Canadian Pathway IMG IMG Unfilled
A
PGY-1 Positions
R
T
Anesthesiology 1094 1,066 799 19 137 1 0 58 52 28
I
Child Neurology 104 95 73 0 10 0 0 5 7 9
Dermatology 22 22 20 2 0 0 0 0 0 0
Emergency Medicine 1,821 1,813 1,438 60 203 1 0 75 36 8
Emergency Med-Family Med 4 4 1 0 1 0 0 2 0 0
T
h
Family Medicine 3,195 3,039 1,405 139 446 3 0 680 366 156
e
S
Family Med-Preventive Med 5 5 5 0 0 0 0 0 0 0
p
e
Iternal Medicine (Categorical) 66,770 6,698 3,317 101 511 5 1 894 1,869 72
c
i
a
Medicine-Anesthesiology 6 6 4 1 0 0 0 1 0 0
l
t
y
Medicine-Dermatology 6 6 6 0 0 0 0 0 0 0
o
f
Medicine-Emergency Med 28 27 23 0 3 0 0 1 0 1
H
o
Medicine-Family Medicine 2 2 1 0 0 0 0 1 0 0
s
p
Medicine-Medical Genetics 4 4 2 0 0 0 0 1 1 0
i
t
a
Medicine-Neurology 1 0 0 0 0 0 0 0 0 1
l
M
Medicine-Pediatrics 380 379 319 2 34 0 0 15 9 1
e
d
Medicine-Preliminary (PGY-1 Only) 1,928 1,805 1,388 27 119 0 0 125 146 123
i
c
i
Medicine-Preventive Med 7 7 4 0 0 0 0 0 3 0
n
e
Medicine-Primary 341 339 206 3 16 0 0 38 76 2
a
n
Medicine-Psychiatry 21 21 13 0 1 0 0 2 5 0
d
S
Neurodevelopmental Disabilities 1 1 0 0 0 0 0 0 1 0
y
s
Neurological Surgery 210 208 188 4 0 1 0 1 14 2
t
e
m
Neurology 404 396 222 7 37 0 0 27 103 8
s
Obstetrics-Gynecology 1,255 1,255 1,002 17 121 0 0 65 50 0
o
f
OB/GYN-Preliminary (PGY-1 Only) 21 14 3 0 1 0 0 3 7 7
C
a
Orthopedic Surgery 703 703 663 24 3 0 0 10 3 0
r
e
Otolaryngology 299 298 283 13 2 0 0 0 0 1
Pathology 605 568 282 23 44 1 0 55 163 37
Pediatrics (Categorical) 2,668 2,654 1,889 32 303 2 0 174 254 14
Pediatrics-Anesthesiology 9 9 9 0 0 0 0 0 0 0
Pediatrics-Emergency Med 9 9 8 0 0 0 0 1 0 0
Pediatrics-Medical Genetics 15 15 12 0 0 0 0 0 3 0
Pediatrics-PM&R 1 1 0 0 1 0 0 0 0 0
Pediatrics-Prelim (PGY-1 Only) 36 30 21 0 2 0 0 3 4 6
Pediatrics-Primary 74 74 26 2 8 0 0 17 21 0
Peds/Psych/Child Psych 19 19 18 0 0 0 0 1 0 0
Physical Medicine &Rehab 107 107 65 4 29 0 0 7 2 0
Plastic Surgery (Integrated) 148 144 136 3 1 0 0 1 3 4
Psychiatry (Categorical) 1,353 1,339 774 39 183 1 0 178 164 14
Psychiatry-Family Medicine 10 10 10 0 0 0 0 0 0 0
Psychiatry-Neurology 3 2 0 0 1 0 0 0 1 1
Radiation Oncology 17 15 14 1 0 0 0 0 0 2
Radiology-Diagnostic 133 120 67 7 19 0 0 12 15 13
Surgery (Categorical) 1,224 1,222 979 86 36 0 0 71 50 2
Surgery-Preliminary (PGY-1 Only) 1,296 821 476 37 24 1 0 104 179 475
Thoracic Surgery 35 35 28 3 2 0 0 0 2 0
Transitional (PGY-1 Only) 842 790 685 5 39 1 0 30 30 52
Vascular Surgery 57 55 48 1 2 0 0 2 2 2
Total PGY-1 27,293 26,252 16,932 662 2,339 17 1 2,660 3,641 1,041
250
Another major barrier in mutual cultural understanding is the provider to be able to encourage the patient to provide acts
lack o ethnically diverse health care leaders. Health care institutions o his/her understanding o their illnesses, their rationale or
C
should also prioritize mentoring and promoting physician and pro- decision making, and the role and in uence o their amily
H
essional sta leaders rom underrepresented ethnic and minority and their opinions about available treatments and comple-
A
P
groups, to ensure uture diversity in leadership to drive health care mentary medicine. Also, the provider should be able to create
T
policy. an environment that is com ortable or the patient to express
E
opinions, concerns, sources o mistrust, prejudice, etc.
R
Multiple teaching and evaluation strategies have been evaluated
3
EDUCATING HEALTH CARE PROVIDERS IN CULTURAL
7
COMPETENCE in di erent settings, including standard surveys, structured inter-
Despite increased awareness about health care disparities and views, presentation o clinical cases, and OSCEs (Objective Struc-
tures Clinical Exam). O those, the OSCE is the only one that is able to
C
the need or a diverse work orce, there is still a ubiquitous lack o
u
appropriate cultural competence training or providers. Most medi- e ectively teach and evaluate the three conceptual approaches to
l
t
u
cal schools have cultural competency lectures embedded within cross cultural education. OSCEs provide a hands-on, real-time prac-
r
a
larger courses in the preclinical years, but those usually occur when tical training that enhances participation in a case-based learning
l
C
medical students do not get to rein orce these concepts within their environment. Trainees can be exposed to multiple clinical scenarios
o
and situations rom a wide array o patients with very diverse cul-
m
clinical training. Some also have cultural immersion programs either
p
locally or in oreign locations, but most student do not get this tural, social, racial, economic and religious backgrounds.
e
The AAMC developed the Cultural Competence Education and
t
exposure. Most residency programs completely lack structured or
e
n
ormal training programs in cultural competence, and most teach- Training Assessment Inventory (CCETAI). The CEETAI is a pool o 35
c
peer-reviewed research studies evaluating the impact o cultural
e
ing aculty do not have the skills to teach cultural competence, due
to lack o training themselves and/or having trained in an era when competence education and training in the medical pro ession.
the patient population was much less diverse. They present the educational goals addressed in each study, the
Improving intercultural communication skills o health care pro- educational activity/tool, and the outcome assessment method
viders is key to a culturally competent health care system. One way used. Each also presents the target learners (eg, residents, medical
o improving these communication skills among physicians is to students o di erent years) and allows or comparisons o the stud-
implement courses during the medical school clinical years and/or ies by curricular goals as categorized by the TACCT.
residency training, and continue throughout practice; these courses Ideally, a cultural competence curriculum should be an integral
should ocus on the cognitive, emotional and behavioral compo- part o any medical school and residency curriculum, and should
nents o communication, targeted to the ethno cultural groups involve case-based, hands-on learning with real li e scenarios. For
most prominent in the local community. This way, providers can such a curriculum to be success ul, there must be ull support rom
become more knowledgeable about local cultures and alternative institution leadership, aculty, and students; there also much be
ways they may experience health and disease processes. adequate institutional and community resources committed, with a
The Association o American Medical Colleges (AAMC), supported clearly de ned evaluation process.
by a Commonwealth Fund grant, has developed a sel -administered
assessment tool, the TACCT (Tool or Assessing Cultural Compe- POTENTIAL STRATEGIES TO IMPROVE CULTURAL
tence Training). The aim o the TACCT is to help medical schools in COMPETENCE
meeting two o the accreditation standards rom the LCME (Liaison
In addition to structured and longitudinal education, here are some
Committee on Medical Education) (ED-21 and ED-22). It also can be
strategies that hospital medicine groups can employ to improve
used as a guide to develop cultural competence curriculum, as well
cultural competence. The American Hospital Association suggests
as to examine all components o the entire curriculum to identi y
the ollowing:
the extent to which cultural competent care is currently taught. The
AAMC has identi ed six core domains needed or e ective cultural • Conduct an assessment to understand sta knowledge on
competence training: health disparities, bias and stereotyping, com- local cultures
munity strategies, cross-cultural communication skills, working with • Use the resources above to create a curriculum with multiple
interpreters, and the culture o medicine. Hospitalists and trainees methods o training, including case studies (and OSCE) with
should ensure they are receiving some training in these six domains, real patients, which should be repeated at regular intervals
and they should strive to improve sel -directed learning i a core cur- • Track data rom patient satis action scores to detect areas o
riculum does not currently exist. concern
Betancourt and colleagues have outlined three di erent but com- • Track your local health care disparities rom your organization
plementary conceptual approaches or cross-cultural education: by race, ethnicity or language pro ciency
1. “Cultural sensitivity-awareness approach” is ocused on We also suggest the ollowing to enhance culture competence
improving attitudes central to pro essionalism such as humil- within hospital medicine groups:
ity, empathy, curiosity, respect, sensitivity and awareness o all • Actively recruit physicians that are o the same ethnicity and/or
outside in uences on the patient. This approach uses exercises speak the same language as the local population.
designed to promote sel -re ection and to detect biases and • Educate your sta on mistrust in the health care system rom
tendencies to stereotype or appreciate the values o other A rican American patients, and rein orce the legitimate histori-
cultures. cal basis or it; actively employ shared decision-making strate-
2. “Multicultural-categorical approach” is ocused on improving gies to gain their trust.
knowledge about the attitudes, values, belie s and behaviors o • Create more e ective patient-physician communication prac-
di erent cultural groups. tices or your Limited English Pro ciency (LEP) patients (eg, by
3. “Cross-cultural approach” is ocused on improving commu- a better use o interpreters).
nication skills, training learners to be able to detect certain • Target patients with limited health literacy and/or alternative
cross-cutting cultural/social/health issues, and improving their health belie s by the appropriate use o resources such as pas-
skills to deal with issues, once detected. The aim is or the toral care, social workers, and case managers.
251
• Create and use instructional materials provided in multiple Boulware LE, Cooper LA, Ratner LE, et al. Race and trust in the health
languages speci c to the community being served, at least or care system. Public Health Rep. 2003;118(4):358-365.
the most common admission diagnoses. Lie D, Boker J, Cleveland E. Using the Tool or Assessing Cultural
• Create interventions to reduce such disparities where they cur-
P
Competence Training (TACCT) to measure aculty and medical
A
rent exist in your organization. student perceptions o cultural competence instruction in the
R
The ultimate goals in enhancing cultural competence are to rst three years o the curriculum. Acad Med. 2006;81:557-564.
T
avoid misdiagnosis, enhance patient sa ety, improve adherence to Weech-Maldonado R. Moving towards culturally competent
I
treatment plans, reduce inpatient and post discharge mortality and health systems: organizational and market actors. Soc Sci Med.
readmissions, improve outpatient ollow-up, and decrease medico- 2012;75(5):815-822.
legal liabilities or all o your patients.
T
ONLINE RESOURCES
h
e
CONCLUSION
S
American Hospital Association. Becoming a Culturally Competent
p
The United States has rapidly become a racially diverse country. Di -
e
Health Care Organization. http://www.hpoe.org/Reports HPOE/
c
erent ethnic groups have di erent belie s when it comes to expe-
i
a
riencing health care, and health care disparities continue to exist. becoming_culturally_competent_health_care_organization.PDF.
l
t
y
Government accrediting agencies and unders have taken notice. Association o American Medical Colleges Enrollment, Graduates,
o
With the advent o value-based payment models, there is a nancial and MD/PhD Data. Available online at https://www.aamc.org/
f
H
imperative or health care organizations to provide patient-cen- data/ acts/enrollmentgraduate/.
o
s
tered, high-quality care to all patients. The need to develop cultural Association o American o Medical Colleges. Assessing change:
p
i
competence o health care practitioners is now more important
t
evaluating cultural competence education and training. Position
a
l
than ever. More research is needed to evaluate the impact that pro- paper, March 2015. https://members.aamc.org/eweb/upload/
M
vider cultural competence has on patient outcomes, but improving Assessing%20Change%20-%20Evaluating%20Cultural%20Com-
e
d
cultural competence should improve patient satis action, as well as petence%20Education%20and%20Training.pd .
i
c
promote social harmony and cultural respect.
i
n
National Resident Matching Program, Results and Data: 2015
e
Main Residency Match. National Resident Matching Program,
a
n
SUGGESTED READINGS Washington, DC. 2015. http://www.nrmp.org/wp-content/
d
uploads/2015/05/Main-Match-Results-and-Data-2015_ nal.pd .
S
y
Berger G, Conroy S, Peerson A, et al. Clinical supervisors and cultural
s
US Department o Health and Human Services’ Of ce o Minor-
t
e
competence. Clin Teach. 2014;11:370-374. ity Health: National Standards or Culturally and Linguistically
m
Betancourt JR. Cross-cultural medical education: conceptual Appropriate Services in Health and Health Care (The National
s
o
approaches and rameworks or evaluation. Acad Med. 2003;78: CLAS Standards). http://minorityhealth.hhs.gov/omh/browse.
f
C
560-569. aspx?lvl=2&lvlid=53.
a
r
e
252
CHAP TER
38 INTRODUCTION
The rapid growth o the hospitalist movement has dramatically
changed the delivery o inpatient care in the United States over the
last two decades. Hospital medicine has an increasing presence within
general internal medicine in academic and community hospitals.
Historically, the hospitalist movement attracted a number o
recent internal medicine residency graduates who were seeking a
transitional period be ore entering subspecialty ellowship training.
With the growing nancial burden o medical school education, sev-
Career Design eral physicians also entered into short-term hospitalist practices to
help repay debt prior to urther training. Consequently, in its in ancy,
and Development a substantial percentage o hospital medicine was populated with

this transient physician population, leading to a paucity o invest-
in Academic and ment in hospitalist career development and leadership. Excessive

patient workloads, long hours, a shortage o mentorship and other
actors have led to high attrition rates and signi cant physician
Community Settings burnout. See Chapter 41: For the Individual: Career Sustainability and
Avoiding Burnout.
The site where hospitalists practice de nes their area o expertise.
Distinguishing between academic and community hospitalists
Amulya Nagarur, MD is o ten based on the type o institution in which they practice
(academic medical centers vs community hospitals). This de ni-
Andrew Z. Fenves, MD, FACP, FASN tion may lead to an oversimpli cation o the di erences between
Alberto Puig, MD, PhD, FACP these physicians, as some hospitalists in the community do have
academic responsibilities in addition to clinical care. Academic hos-
pitalists typically hold a aculty appointment con erred by a hospital-
a liated university. When community hospitals become a liated
with universities, or merge with academic medical centers, the role
o their sta may change accordingly to include an academic com-
ponent. In addition to community and academic hospitals provid-
ing salary support, companies have been developed to outsource
hospitalists and related services to multiple hospitals. In any o these
settings, hospitalists have opportunities to acquire additional skills
and leadership that will provide them with the option o transition-
ing rom one setting into another. This chapter o ers strategies or
pro essional development or early-career hospitalists.
THE ACADEMIC SETTING

Achieving traditional academic success in hospital medicine has
many challenges. Increasing demands o clinical work is due to a
number o actors, including:
• Expansion o patient care into specialty areas
• Increased on-site coverage (nights, weekends)
• Lack o unding or protected time to pursue scholarly activities
• Restricted resident work hours and availability
Many academic clinicians build careers that ollow a classic triad
o clinical care, research, and teaching in the inpatient setting. In the
inpatient setting career-hospitalists may specialize in acute general
medicine, one o the medical specialties, general medical consultation,
or comanage a specialty such as orthopedic surgery and neurosurgery.
Each o these clinical areas provides opportunities or hospitalists,
including acquiring new skills, leading a program, per orming qual-
ity improvement, and education. In addition, some hospitalists may
expand their clinical duties to involve some outpatient care.
Many hospitalists cite higher job satis action when able to
complement inpatient clinical duties with other nonclinical activities
(Table 38-1).
253
health IT sector, there may be unding through their local hospital.
TABLE 38-1 Nonclinical Roles in the Hospital Setting Regional and national societies are supportive o both health IT
training and innovation.
Research • Clinical studies o commonly
Physicians with academic careers re er to the guidance o a men-
P
encountered medical problems
A
tor as having the most positive inf uence on academic productivity
Teaching Primary teaching attending
R
and pro essional satis action. Mentored aculty at all levels in their
responsible or
T
careers are more success ul at producing peer-reviewed publica-
• Conducting rounds
I
tions and procuring grant support, and have higher con dence
• Observing clinical skills in their career trajectory. Institutions also bene t rom their aculty
• Supervising procedures having rich mentoring opportunities. Mentorship has been most
Mentorship • Trainees and junior hospitalists success ully associated with clinician-scientists as this group o
T
h
• Nurse practitioners and physician physicians has a more objective means or measuring academic
e
assistants productivity. Many junior aculty choosing careers as clinician edu-
S
p
Scholarship Local Level cators have had exposure to master clinicians; however, mentorship
e
c
is o ten di cult to nd or identi y. Typically, clinician-educators have
i
• Discussants at trainee con erences
a
a career trajectory that ollows a natural pattern: at the start o their
l
t
• Development o new curricula
y
careers, educators are entry-level teachers who are still re ning their
o
or trainees clinical and teaching skills. Over time, they develop more robust
f
H
• Interdepartmental educational content and teaching skills. Some clinician-educators move on to
o
e orts (medical consultation,
s
direct clerkships and residency programs. The highest level o ac-
p
comanagement)
i
ulty participation in education involves those who study the process
t
a
Regional, national, and international
l
and outcomes o teaching and learning or delve into curriculum
M
level development. See Chapter 39: Mentorship o Peers and Trainees.
e
d
• Speaking engagement at society
i
c
meetings
i
n
THE COMMUNITY SETTING
e
• Publication: articles, chapters, books,
a
studies o the process o teaching The early hospitalist movement originated in the community set-
n
and learning
d
ting, and many hospitalists choose to practice exclusively in this
S
• Curriculum development setting. The lack o structured pathways or the early-career
y
s
hospitalists has made career development di cult to navigate.
t
Leadership Leadership o :
e
m
• Hospitalist programs
Community hospitalists cite numerous reasons or requesting
s
ormal mentorship. Institution-based mentoring programs or hos-
o
• Medicine divisions or departments
pitalists engaged in ull-time community practice have shown to
f
C
• Administrative o icers o medical strengthen pro essional satis action.
a
centers
r
The clinical responsibilities o hospitalists who practice in a com-
e
• Medical consultative service munity setting overlap with those o academic hospitalists. Com-
• General medical service munity hospitalists provide medical ward care, medical consult
• Comanagement services services, emergency medicine triage, intensive care unit and surgical
• Hospital committees
unit care, and at times coordinated subspecialty care. These physi-
cians require expertise and skill managing the breadth and depth o
• Medical student core clerkships
resources available at their institutions and must also sa ely negoti-
• Residency programs ate trans ers to higher-care acilities when needed. Operationally,
Hospital systems • VTE prophylaxis community hospitals have expanded the roles o their hospitalists as
improvement • Management o diabetes they have become essential to the productivity and mission o local
(patient sa ety health centers. Some community hospitalists oversee treatment in
• Transitions and hand-o s
and quality postacute care acilities such as long-term acute care, rehabilitation,
improvement) • Utilization review
skilled nursing, and hospice centers.
Health in ormation • Design and implementation o Regardless o hospital size and type, hospital medicine programs
technology electronic health records bene t medical centers and communities by virtue o expertise in
• Physician order entry hospital medicine and quality improvement, availability or admis-
• Clinical decision support sions and e ciency associated with cost-savings. Many community
• Databases or research hospitals develop academic a liations over time, and their hospital-
Community • Community education ist aculty may assume academic roles teaching medical students
outreach and residents as well as serving as discussants at resident reports,
• Re erral networks
lectures and con erences. Hospitalists in the community collaborate,
educate and mentor nurse practitioners and physician’s assistants,
and supervise many o these advanced practice clinicians on medi-
Federal unding sources such as K-series and R-series grants cal services.
remain limited or the type o quality and sa ety research that many Increasingly, community hospitalists are accountable or imple-
academic hospitalists pursue; there ore, many hospitalists will menting community support systems or their patients that include
need to seek out alternative sources o unding within their institu- robust social work, case management, pharmacy, and educa-
tions. Fellowship opportunities and grants are available on local tional supports in order to decrease readmissions and improve
and national levels. Improving processes and outcomes in a cost- compliance postdischarge. Most hospital medicine groups in the
e ective manner requires meaning ul use o data and in ormation community settings are actively engaged in quality and process
via health in ormation technology (IT). For clinicians interested in the improvement and ocus on patient sa ety, resource allocation, and
254
hospital e ciency. Multidisciplinary hospital committees o ten including teaching e ectively, producing scholarly work, and under-
include or are led by sta hospitalists. Hospitalist groups encourage standing the promotion process and the business o health care.
C
their sta to take on operational and administrative duties such as Attending such meetings osters collaboration, allows or network-
H
utilization review, as these are locally valuable to the hospital sys-
A
ing, and in many cases ignites rich mentoring relationships.
P
tem. Ranging rom hospital medicine group administrators to chie
T
operating o cers and chie executive o cers, hospitalists may par- Participation and leadership in local hospital initiatives
E
ticipate or lead initiatives in hospital operations, recruitment, policy
R
From an individual and departmental perspective, hospitalists
development, nancial a airs, and compliance. should take advantage o opportunities to increase their presence
3
In addition to hospital work, community hospitalists may assume
8
by leading hospital-based committees, quality improvement initia-
expanded roles as community educators on local, regional, and tives, and comanagement. SHM, ACP, SGIM, and ACLGIM o er train-
national levels. Some partake in grassroots activities, educating their ing opportunities that teach leadership, evaluation, communication
C
local communities on various health care issues. Others may direct
a
skills, and hospital metric derivations that drive leadership decisions.
r
e
e orts to educate local or national policymakers to a ect more SHM o ers a Certi cate o Leadership in Hospital Medicine (CLHM)
e
r
widespread legislative change in hospital medicine, community through which candidates complete a leadership project and build
D
health, or hospital administration.
e
skills or success ul executive leadership positions. In addition, many
s
i
academic centers o er training in leadership through their Centers
g
n
PROFESSIONAL DEVELOPMENT PLAN or Faculty Development.
a
Most medical centers have strong ties to local organizations,
n
■ SELF-EVALUATION
d
including health advocacy groups, legislative entities, outpatient
D
Achieving success in both settings requires developing a personal community health centers, and schools. Both early and late career
e
pro essional plan. This process begins with sel -evaluation:
v
community hospitalists with speci c areas o interest (clinical or
e
l
• Strengths: What are your key strengths and how may you high- nonclinical) may advance their careers by developing durable rela-
o
p
light your strengths? tionships with such groups.
m
• Development needs: What are one or two areas that need
e
n
improvement?
t
Promotion process
i
• Career aspirations: What do you like to do? What are your goals
n
Many hospitalists nd the promotions process di cult to navi-
A
over the next one to 2 years? What are your goals over the next
c
3 to 5 years (ie, longer term milestones)? gate due to their heavy clinical workload, lack o mentorship, and
a
d
• Periodic reassessment o goals: Do you eel that you are reach- participation in nontraditional academic activities (such as QI and
e
m
ing your goals? Do you need to revise your goals? Do you need education). Criteria or promotion vary widely across various institu-
i
tions, and hospitalists should become amiliar with this process. In
c
to take additional steps to achieve your goals?
a
a cross-sectional survey published in the Journal of Hospital Medi-
n
d
■ SELECTING A HOSPITALIST POSITION cine, success ully promoted academic hospitalists considered peer-
C
reviewed publications to be the most important activity in achieving
o
Your selection o a hospitalist position hinges on your core values
m
and priorities. Your priorities may change over time depending on promotion. Promotion within community hospitals may involve an
m
your stage o li e and needs: opportunity to manage a hospital committee, lead the local hospital
u
unit, or serve as the clinical director o a speci c clinical activity.
n
• Financial priorities: Nocturnist salaries are usually higher than
i
Because many hospitalists engage in QI projects that do not
t
y
daytime work, and community hospitals may be more compet- t the traditional triad o academia (clinical care, teaching, and
S
itive than academic hospitals. There are also regional variations.
e
research), it is important to establish a port olio to showcase QI
t
t
• Work-li e balance: Is the hospitalist service chronically under-
i
endeavors. The SGIM Academic Hospital task orce created the Qual-
n
sta ed with high turnover? Is it possible to work part time?
g
ity Port olio, a structured accompaniment to a promotions package
s
• Pro essional development: Is there a dedicated aculty devel- that documents and organizes QI work. Similarly, the Educator Port-
opment program with a robust structured mentor ocus? Is olio documents teaching, awards, and educational scholarship or
there protected time or pro essional development? clinician-educators. Templates and resources or developing Quality
Does the hospital invest in career development or is the ocus and Educator port olios are available online.
solely on clinical responsibilities?
• Clinical interests: Does the hospital provide opportunities in Strategic professional plan
your area o ocus: quality improvement, teaching, medical
A strategic pro essional plan begins with:
in ormatics, or clinical research?
• Mentorship: regular meetings with aculty mentors so that hos-
Strong aculty development programs, mentorship, and transpar-
pitalists may take advantage o leadership, research, speaking,
ent promotions criteria are critical to the success o career hospital-
and teaching opportunities that are instrumental in promotion.
ists and job satis action.
• Networking: participation in local, regional, and national
meetings.
■ EARLY MILESTONES (FIRST 1 TO 2 YEARS)
• Participation in skill building sessions: local, regional and
Networking national workshops in peer-observation, educational, leader-
Exposure to other hospitalists is crucial or career development. The ship, competency in manuscript writing, grant writing and
Society o Hospital Medicine (SHM), the American College o Physi- procurement, oral presentation, study design and execution.
cians (ACP), the Society o General Internal Medicine (SGIM), and the • Additional training: ormal degree-granting programs in public
Association o Chie s and Leaders o General Medicine (ACLGIM) health, public policy, nance and operations management,
sponsor leadership and mentorship retreats as well as interest- health care administration, QI, education, health care in ormat-
speci c workshops. The annual Academic Hospital Academy is ics, courses in biostatistics, study design, epidemiology.
an intensive multi-day workshop or junior academic hospitalists • Scholarship: identi cation o an area o special interest or
that provide essential skills or succeeding in academic medicine, expertise with eventual goal o becoming an institutional,
255
regional, and national expert and pursue oundational oppor- ACKNOWLEDGMENT
tunities with a built-in curriculum, such as the various Robert
We are indebted to Dr. Sue Williams rom Baylor University Medical
Wood Johnson Foundation Scholars programs.
Center, Dallas, Texas, or her revision o the manuscript, input on sec-
• Documentation o port olio o progress: case reports, evalua-
P
tions regarding Community Hospital Careers and invaluable advice
A
tion o QI projects, development and assessment o education
on the chapter.
R
curricula, and participation in inpatient clinical studies. This
T
documentation will be required or academic promotion.
I
• Education about the promotion process rom the beginning o SUGGESTED READINGS
career: ocus on the best avenue to achieve goals.
• Review the process or achieving Society o Hospital Medicine Farrell SE, Digioia NM, Broderick KB, et al. Mentoring or clinician-
ellowship and senior ellowship or planning purposes as well educators. Acad Emerg Med. 2004;11(12):1346-1350.
T
h
as that o other pro essional societies such as the American Flanders SA, Centor B, Weber V, et al. Challenges and opportunities
e
College o Physicians. in academic hospital medicine: report rom the Academic Hospital
S
p
Medicine Summit. J Hosp Med. 2009;4(4):240-246.
e
c
Hospital medicine as a specialty: future directions
i
Harrison R, Hunter AJ, Sharpe B, et al. Survey o US academic hospital
a
l
t
Currently, the American Board o Family Medicine and the American leaders about mentorship and academic activities in hospital
y
o
Board o Internal Medicine (ABIM) o er a Recognition o Focused groups. J Hosp Med. 2011;6(1):5-9.
f
Practice in Hospital Medicine (RFPHM) program. As hospital medi-
H
Leykum LK, Parekh VI, Sharpe B, et al. Tried and true: a sur-
o
cine continues to expand and leaders rise in both the academic and vey o success ully promoted academic hospitals. J Hosp Med.
s
p
community sectors, the eld will likely develop into an independent 2011;6(7):411-415.
i
t
a
subspecialty with its own board certi cation.
l
Mookherjee S, Monash B, Wentworth KL, et al. Faculty development
M
Several clinical training programs have introduced careers in hos-
or hospitals: structured peer observation o teaching. J Hosp Med.
e
pital medicine as part o medical student and residency education.
d
2014;9(4):244-250.
i
The SHM has a comprehensive guide to residency hospitalist tracks
c
i
Reid MB, Misky GJ, Harrison RA, et al. Mentorship, productivity,
n
as well as hospital medicine ellowships or those seeking more
e
ormal training early in their careers. Building such pathways and and promotion among academic hospitals. J Gen Intern Med.
a
n
standardizing curricula or hospitalist career development will need 2012;27(1):23-27.
d
to become a ocus in hospital medicine. Creating a ramework or Sehgal NL, Sharpe BA, Auerbach AA, et al. Investing in the uture:
S
y
assessment and evaluation o the hospitalist will allow or special- building an academic hospital aculty development program.
s
t
e
ized tracks or promotion purposes within academic and commu- J Hosp Med. 2011;6(3):161-166.
m
nity settings.
s
Tietjen P, Griner PF. Mentoring o physicians at a community-based
o
In 2001, the Institute o Medicine called or an overhaul o the health system:preliminary ndings. J Hosp Med. 2013;8(11):642-643.
f
US health care system or the 21st century that prioritizes e cient
C
a
patient-centered teams. In response to this call or redesign, the
r
e
Accreditation Council or Graduate Medical Education included ONLINE RESOURCES
aptitude or working in multidisciplinary teams as part o its core
competencies. Medical students and residents will increasingly Niebuhr V, Johnson R, Mendias E, et al. Educator Portfolios. MedEdPORTAL
be expected to learn and work collaboratively alongside other Publications; 2013. Available at: https://www.mededportal.org/
health care providers such as nurses, physical therapists, and occu- publication/9355.
pational therapists. The cadre o career-hospitalists will remain Society o General Internal Medicince (SGIM) Quality Port olio—
invaluable participants and leaders in these movements, and will Template and Instructions. Available at: http://www.sgim.
need to acquire and teach the skills or success ul multidisciplinary org / File %20Lib rary/ ACLGIM/ Tools%20an d %20Re source s/
collaboration. QualityPort olioTemplate.pd .
256
CHAP TER
39 INTRODUCTION
The word “mentor” comes rom Homer’s Odyssey, in which Odysseus
entrusts his young son to the care o his close riend, Mentor. A tran-
sitional gure in the youth’s growth, Mentor acts as the son’s guard-
ian and wise advisor, and through their mutual relationship the son
develops his own identity. Good mentors have played key roles in
the history o medicine and discovery, in the development o young
doctors, and in the institutions that train physicians.
Today’s health care leaders underscore the importance o men-
Mentorship of Peers toring on career choice as well as on career advancement and
productivity. Yet, the available evidence shows that a minority o
and Trainees medical students and aculty have mentors. Because Hospital Medi-
cine is a young specialty, peer mentorship is crucial to the success
o the specialty.
Tom Baudendistel, MD, FACP PRACTICE POINT

Aubrey Ingraham, MD Bene ts o mentorship include:
• Mentoring is a power ul predictor o academic advancement
• Academic advancement and productivity promote the
specialty o Hospital Medicine
• Mentoring o medical students, trainees, and junior aculty
acilitates recruitment and retention o hospitalists
• Faculty members derive personal and pro essional satis action
rom mentoring trainees
THE VALUE OF MENTORING IN MEDICINE

Surveys o aculty and health care leaders and a systematic review
identi ed several potential bene ts o mentoring in medicine. Men-
toring in uences career choice, including medical students’ spe-
cialty selections; promotes career advancement; increases scholarly
productivity; develops physicians’ leadership skills; shapes pro es-
sional ethics; osters development o academic departments, insti-
tutions, and pro essional societies; and increases career satis action.
Clinician-educators view mentoring as an important determinant
o promotion and development, and are more likely to remain in
academia i they are mentored.
While most studies have ocused on the bene ts to the mentee,
potential bene ts to the mentor should not be underestimated.
Faculty members derive personal and pro essional satis action rom
mentoring residents, and mentoring may acilitate promotion, result
in special awards, and increase scholarly productivity (Table 39-1).
BARRIERS TO MENTORING FOR HOSPITALISTS

The absence o mentoring is a power ul predictor o delayed aca-
demic advancement. Despite consistent reports o the bene ts o
mentoring—and the disadvantages aced by those without
mentors— ewer than 20% o aculty members have a mentor.
Clinician-educators are less likely than clinician-scientists to iden-
ti y a mentor and also less likely to serve as mentors. Faculty cite
competing time pressures, inadequate aculty development around
mentoring, and lack o recognition o mentoring by promotions
committees as actors dampening their willingness to mentor.
257
OVERCOMING THE BARRIERS: STRATEGIES
TABLE 39-1 Benefits of Mentoring FOR SUCCESSFULLY MENTORING
Benefits to the Mentee To become more promotable, hospitalist clinician-educators must
P
• Increased advocacy or career development ramp up their productivity during nonclinical time. The traditional
A
• High-level career advice rom experienced senior person clinician-educator paradigm—periods o intense clinical work inter-
R
• Enhanced access to opportunities beyond the current level o spersed with more relaxing nonclinical time—must be rethought.
T
the mentee, including: To paraphrase one department chair, just as diastole is now recog-
I
research (grants, editing, publications, collaborations) nized as an active time o the cardiac cycle, so too must clinician-
promotions educators view their nonclinical periods.
new job openings (within or outside the institution) Overarching goals in mentoring clinician-educators include per-
committee membership
sonal and career satis action and pro essional advancement. To
T
roles in pro essional organizations
h
achieve this, success ul mentors can help mentees optimize their
e
networking with key thought leaders
nonclinical “diastolic” time. In some instances, clinical and non-
S
• Valuable, nonthreatening, eedback rom “third party” who is
p
separate rom the employer-employee relationship clinical duties may overlap and provide a “two- or-one” opportunity.
e
c
• Access to a role model o pro essionalism, ethics, and values During a ward attending block, or example, a hospitalist may be
i
a
l
• More directed and ormal timeline and ramework or career encouraged to pilot a quality improvement initiative or a novel
t
y
success patient-centered multidisciplinary bedside rounding structure. A
o
f
Benefits to the Mentor mentor might help identi y the opportunities, suggest local leaders
H
• Renewed sense o excitement or career brought about by to contact, determine what metrics to analyze to demonstrate suc-
o
s
revisiting own history o pro essional growth; participating in cess, and recommend a orum or disseminating the results locally
p
i
the pro essional and personal development o a colleague;
t
or extramurally.
a
and continuing pro essional legacy o shaping “the next
l
Ultimately, without protected time or unding clinician-educators
M
generation” o physicians
may struggle to advance. Resources to support the nonclinical work
e
• Enhanced recognition among peers and junior sta
d
o junior hospitalists is o ten limited, but mentors can steer mentees
i
• Cultivation o speci ic interpersonal skills (active listening,
c
i
to education research grants, administrative roles with associated
n
e ective communication, modeling)
e
• Exposure to new ideas and opportunities, o ten leading to support, or medical school work which includes remuneration.
a
When these resources all short, mentors and mentees o ten turn
n
increased creativity and productivity (including publications
d
and projects) to their division leaders or support. A strong hospitalist leader suc-
S
y
• Pride in a mentee’s successes cess ully lobbies leadership to support nonclinical hospitalist work.
s
t
• Enhanced personal growth In these negotiations, hospitalist leaders can highlight avorable
e
m
• Broadened network via mentee’s collaboration and return-on-investment by demonstrating the impact o hospitalists
s
connection to other personnel, including cross-discipline and on: length o stay; readmission rates; bed availability or more pro t-
o
cross-department interactions
f
able surgical cases; patient satis action scores; hospital-acquired
C
• Discussion and exploration o mentor’s values with others
in ection rates; satis action o re erring primary care, emergency, and
a
• An accurate perspective on barriers experienced by current
r
e
junior sta consulting physicians; physician retention; teaching evaluations;
• Credit toward career advancement as a result o mentoring and committee contributions.
Benefits to the Department Creative mentors can also acilitate career advancement and
• Bolstered sta morale, motivation, institutional dedication, and promotion by developing ocused areas o mastery within mentees.
career satis action Hospitalist duties have expanded ar beyond general medicine
• Enhancement o productivity and creativity wards to include comanagement o patients under the care o other
• Discovery and development o personnel talent specialties and protected time or peer review, quality improvement
• Insurances o department’s uture survival through and sa ety, teamwork, and leadership. Each o these areas represents
development o leadership a potential locus o expertise and leadership or a junior hospitalist.
• Communication and demonstration o the department’s A success ul mentor can guide the mentee to become the local
values, goals, and expected personal and pro essional institutional champion, and then map a strategy or demonstrating
standards
quantitatively the impact o the mentee’s work in this area. Eventu-
• Reinvigoration o senior aculty
• Development o cross-departmental, national, and
ally, the mentor can suggest strategies or disseminating the men-
international networks tee’s successes at pro essional society meetings and via publication
in peer-reviewed journals, society newsletters or blogs, and other
social media.
Speci c behaviors associated with success ul mentoring or
clinician-educators dif er at each step in the mentoring relationship:
Junior hospitalist clinician-educators o ten spend their early preparing or mentoring, approaching a mentor or the rst time,
aculty years learning to master the many aspects o the inpatient ongoing mentoring, and ending the mentorship. Novice mentors
arena, including patient sa ety, quality improvement, transitions o and junior aculty should receive structured aculty development to
care, surgical comanagement, and ward teaching. Such jack-o -all- learn a core set o mentoring skills. Table 39-2 elaborates urther on
trades hospitalists, while valuable to their groups, can be perceived key elements or success at each phase o mentorship.
by promotions committees as “masters o none.” A second tension
inherent in hospitalist careers exacerbates the problem. Hospitalist
clinician-educators o ten cannot generate enough revenue rom
clinical duties to support un unded nonclinical work, making it hard PRACTICE POINT
to secure protected time and nancial compensation to develop • Success ul mentoring requires sel -re ection, a key means by
expertise in education, scholarly activity, or administration. Without which both participants initiate personal, relationship, and
time or expertise, hospitalists nd it challenging to advance aca- practice improvements.
demically and eel underquali ed to serve as mentors.
258
TABLE 39-2 Best-Practice Behaviors for Successful Mentoring
C
H
Mentee Mentor
A
Preparing or • Assess your competencies in the roles you • Make a realistic assessment o your availability to commit to a new
P
mentoring currently hold (preparation o clinician- mentee
T
educator port olio may acilitate this) • Consider requesting and reviewing an updated CV rom your
E
R
• De ine speci ic career goals and steps prospective mentee prior to meeting
required to achieve them • Enumerate the particular areas in which you might be o value to a
3
• List speci ic activities and experiences you mentee*
9
seek*
• Identi y speci ic questions pertaining to
M
the kinds o help you think you need
e
• Determine the key personal and
n
t
pro essional qualities you desire or value
o
r
in a mentor
s
h
• Consider seeking a mentor 1 level above
i
p
your current career level and a more senior
o
mentor
P
e
First meeting • Explain your current academic role, how • Be riend the mentee to di use the power dynamic
e
your goals may be aligned with the mentor’s • Start with open-ended questions (eg, “What do you hope to gain
r
s
work, and what you think you might need in rom our work together?”)
a
n
terms o advice and guidance • Recognize your limitations and provide alternative resources to those
d
• Recognize a potential mentor’s time and whom you cannot e ectively mentor
T
r
energy • Summarize and con irm (eg, “It sounds like I could best help you now
a
i
• Ensure you have the mentor’s updated by…Is this true?”)
n
e
contact in ormation and ask which method • Ensure you have the mentee’s updated contact in ormation and ask
e
s
o communication he or she pre ers which method o communication he or she pre ers
• Consider sending a thank you note a ter
your irst meeting
Ongoing • Determine the structure o meetings (eg, • Assist mentee in establishing goals
mentoring e-mail, in person, telephone) • Listen to your mentee actively and patiently
• Take the initiative, when in doubt • Re ine mentee’s speci ic goals, and push mentee or his or her
• Prepare and set the agenda or mentoring “dreams” (what may seem unobtainable to mentee may seem
meetings achievable to you)
• Schedule meetings at regular intervals • Hold mentee to high but obtainable standards
• In orm your mentor o changes in goals, • Advocate or mentee
barriers, and progress in reaching your • In orm the mentee about new opportunities and suggest alternate
goals resources or in ormation about academic opportunities, political
• Seek and accept challenges and eedback culture, and networking
• Clari y realistic expectations o the • Protect the mentee rom possible threats
mentee-mentor relationship • Use your experience, clout, and in luence to serve as a champion or
• Consider creating and ollowing a written the mentee
checklist or timeline to track progress • When you are unable to meet mentee’s needs, re er mentee to
another mentor
• Foster mentee development
• Commit your time and energy on a regular, ongoing, and lexible basis
• Recognize di erent mentee learning styles and tailor your approach:
some mentees may need direct, task-oriented assistance, while others
need help with problem solving or articulating ideas
• Assist in the mentee’s identity development; consider how you will
“wean” the mentee
• Collaborate with mentee i this helps promote mentee’s agenda, not
yours
• Credit the mentee or his or her diligence and creative output
(includes authorship or grants)
• Provide honest eedback in a constructive and caring manner
• Serve as role model and con idante
• Exhibit high pro essional and moral character
• Be responsive and available
• Follow through on promises
• Maintain con identiality (except in rare circumstances as required
by law, such as harassment, danger to sel /others, pro essional
misconduct)
• Seek out eedback rom mentee
• Accept personal di erences with sensitivity and do not judge gender,
culture, or age-related di erences
• Share personal knowledge (medical and nonmedical), including
ailures, so that the mentee eels com ortable seeking guidance
(Continued)
259
TABLE 39-2 Best-Practice Behaviors for Successful Mentoring (Continued)
Mentee Mentor
Ending the • Talk about when the relationship should • Talk about when the relationship should end
P
A
mentoring end (eg, a certain time point, or once • O er suggestions or uture mentors or directions o interest
R
relationship certain goals are achieved) • Thank your mentee
T
• Ask or advice on uture advisors or
mentors
I
• Thank your mentor
*Common mentoring goals include: assistance with grant-writing or manuscript preparation, advice or promotion and advancement (including advice
unique to women and minorities), building a network, research collaboration, teaching skills, negotiation, enhancing educator port olio, improving under-
T
h
standing o organization politics, oral presentation skills, and work-li e balance.
e
S
p
e
c
development and role modeling; includes emotional and psycho-
•
i
The goal o sel -re ection is to identi y speci c short- and
a
logical support; involves direct personal direct interaction; utilizes
l
t
long-term goals to aid uture career development. To achieve
y
the mentor’s greater experience and in uence within an organiza-
o
this, the mentor needs to guide the mentee so the mentee
tion or eld; and is reciprocal, designed to enrich the pro essional
f
can articulate personal strengths and weaknesses and then
H
and personal lives o both mentor and mentee. While most inter-
o
trans orm the sel -assessments into concrete goals or uture
s
actions occur within the workplace, many nd that meals, social
p
career development.
i
events, shared hobbies, and pro essional meetings provide addi-
t
•
a
Role modeling sel -re ection and actively promoting sel -
l
tional opportunities or career guidance. The ef ective mentoring
M
re ection may cause the mentee to internalize an important
relationship assumes an always pro essional and nonsexual ocus.
e
skill o a sel -directed li elong learner.
d
Failures in mentoring derive rom inadequate communication, com-
i
c
mitment, or experience; personality dif erences, and perceived (or
i
n
e
A mentor would rst lead the mentee to recognize these areas real) competition. Table 39-3 illustrates the key qualities o a good
a
o relative strength and weakness. Then, the mentor might suggest mentor.
n
d
speci c ways or the mentee to improve didactic teaching skill, or The best way to pair mentors with mentees is unknown. Manda-
S
may encourage the mentee to seek experiences that maximize the tory mentor assignments of er the advantage o engaging all men-
y
s
mentee’s strengths and minimize exposure to areas o weakness. tees in some orm o advising. On the other hand, independently
t
e
m
Importantly, the mentor should periodically request the mentee to sought out mentee-mentor relationships may be more ef ective
s
evaluate the quality and ef ectiveness o the mentoring relationship. and enduring.
o
f
C
a
PRACTICE POINT
r
e
Key questions or junior hospitalists to consider include:
TABLE 39-3 A Good Mentor
• What do I perceive as my strengths and weaknesses?
• Maintains con identiality
• How do these strengths and weaknesses compare with how
• Is knowledgeable and respected in his or her ield
others have evaluated me?
• Encourages independent behavior with an approachable,
• What activities provide me the greatest source o career nonthreatening, accessible, acilitative, empowering style o
satis action? communication
• What activities would I like to do more or less o ? • Challenges and debates mentee in a constructive way
• Which senior aculty member’s career path do I want to • Employs a care ul and dynamic balance between compassion
emulate? and empathy, and impartiality and honesty
• Shows genuine interest and investment in a mentee’s
• What speci c roles do I see mysel in within the next 3 to concerns, well-being, and uture
5 years? 5 to 10 years?
• Asks questions that provoke critical thinking, reasoning,
• What obstacles stand in the way o my career goals? analysis, and contemplation
• What skills or knowledge do I need to acquire to address my • Recognizes and admits limitations, then guides mentee to
weaknesses or to be able to realize my career goals? appropriate resource
• What are the goals and vision o my supervisor (eg, • Acknowledges importance o work-li e balance in pro essional
chairperson) and do my interests and strengths align with success
these goals? I not, how can I demonstrate my value to the • Demonstrates con idence in a mentee
division or group? • Possesses strong interpersonal and negotiation skills
• Listens actively and communicates clearly
• Avoids abuse o his or her in luence or position
CHOOSING WISELY: SELECTING THE RIGHT • Provides requent and detailed eedback
MENTOR-MENTEE RELATIONSHIP • Seeks opportunities or a mentee to assist with his or her own
Mentoring is a protected relationship occurring between a more projects (i appropriate and relevant to mentee’s goals)
advanced career incumbent (mentor) and a younger novice (men- • Expects and tolerates expressions o emotion rom mentees at
tee). An ef ective mentoring relationship: ocuses on achievement times o signi icant anxiety or rustration
or acquisition o expertise through direct assistance with career
260
■ MENTORING MODELS DOCUMENTING ACCOMPLISHMENTS: THE
CLINICIAN-EDUCATOR PORTFOLIO
C
Although most discussions ocus on the traditional model o
H
mentorship as a one-to-one relationship at one institution, junior The clinician-educator “port olio” enumerates and organizes a
A
hospitalists should be aware o other models which may enhance aculty member’s educational activities and achievements or the
P
their career development, such as collaborative group mentoring, purposes o academic promotion and sel -re ection. Traditionally,
T
mentorship between one mentee and multiple mentors, integrated decisions regarding academic promotion were based on in orma-
E
R
peer mentoring, networking at annual society meetings, or rom tion summarized in a curriculum vitae (CV) and letters o recom-
a distance (telementoring), and combinations thereo . Table 39-4 mendation. The educator port olio adds several elements: a novel
3
9
reviews mentoring models and the strengths and weaknesses o structure to capture all clinician-educator activities, including those
each. elements not traditionally captured in a CV; demonstration o the
Peer mentoring can serve important advising and networking impact o the educator’s activities on trainees, peers, the organiza-
M
unctions. The nonhierarchical nature o peer mentoring addresses tion, and the eld as a whole; and sel -re ection.
e
n
problematic issues in senior-junior mentoring relationships such as The concept o the clinician-educator “port olio” is analogous to
t
o
power, dominance, dependency, and trans erence. Peer mentors an architect’s or artist’s port olio. The educator port olio includes
r
s
h
also may be more readily available and provide a way to gain dif er- aculty member’s educational achievements, which have expanded
i
p
ent perspectives and current in ormation on diverse opportunities, rom classic “teaching” duties to include a broader array o hospital-
o
especially in a young specialty such as Hospital Medicine, which has ist activities: clinical duties, teaching and curriculum development,
P
relatively ew practicing senior hospitalists. mentoring, scholarly productivity, leadership, and administration
e
e
r
s
a
n
TABLE 39-4 Mentoring Models
d
T
r
Description Advantages Disadvantages Best Uses
a
i
n
Closed Models
e
e
Assignment model Assigns mentors to Guarantees mentee Risks pairing participants Provides mentorship
s
new trainees based on engagement with a who have little natural or within the irst several
pro essional interests, senior person mutual a inity months o early in career
gender, minority status Mentee receives only one
mentor’s perspective
Choice model Requires new resident or Allows some sel - Risks leaving a mentee Facilitates seeking a
junior aculty to choose a selection between without a mentor during an con idante or a mentee
more senior aculty to serve participants, which may o ten stress ul transition with issues o a sensitive
as a mentor within several be especially important or con idential nature
months o beginning or women and minority
employment mentees
Open Models
Multiple mentor Encourages new resident Provides lexibility and Risks creating many Provides advice rom
model or junior aculty to actively broad expertise or the “diluted” relationships with senior aculty in multiple
seek multiple mentors rom mentee less continuity, no central domains/departments
di erent places, career senior overseer
levels, or career paths Relies heavily on the
mentee’s level o
sel -motivation and
management
Layered model Assigns new trainees Allows some guided Provides less individually Provides networking
to a group o similar- peer-to-peer mentoring tailored and private counsel opportunities and
experienced peers, all o and provides a lexible Risks diminishing mentor general advice
whom share one mentor and broad network o advocacy or individual to mentee with
Schedules periodic potential senior advisors mentee undi erentiated career
meetings (attended by all Sa e environment or ocus
Relies heavily on the
mentees) idea sharing mentee’s level o Involves mentee in
Provides re erence to sel -motivation and peer groups and
individual aculty or speci ic management projects spanning many
advising disciplines
Facilitated group/ Assigns trainees to Facilitates relevant Provides less individually Provides opportunity to
Collaborative structured group orums education across broad tailored and private counsel junior aculty seeking to
mentoring model acilitated by senior content areas overseeing mentee, strengthen undamental
aculty (such as skill Provides a broad and lacks content and skills in teaching,
development, career network o potential scheduling lexibility research, and publication
planning, scholarly writing, senior advisors Relies heavily on the
role play, videotaping, mentee’s level o
group discussion, peer sel -motivation and
and acilitator eedback, management
narrative writing, and sel -
re lection)
261
TABLE 39-5 Elements of a Clinician-Educator Portfolio
Category Examples to Support This

P
Clinical activities List o clinical duties and average time per week spent on each
A
R
Direct teaching List o teaching activities (ward or ICU teaching, lectures, small group acilitation, resident report, clinic
T
precepting, visiting pro essorships)
I
Summary o evaluations rom ward rotation, ideally with comparison to peer averages
Teaching awards
Curricular design Revision or implementation o new course, rotation, or program (eg, procedural simulation training,
• Instructional development journal club curriculum, homeless clinic rotation) and how participation in course was evaluated
T
h
Outcome assessments (eg, evaluations rom course participants; end-o -course test scores or observed
e
procedural competence)
S
p
Mentoring List o mentee names by academic year, with hours spent mentoring each and outcomes o mentoring
e
c
• Interns and residents (eg, implemented global health rotation in residency; nominated mentee or regional award; helped
i
a
l
• Medical students determine career choice, etc)
t
y
• Junior aculty Collaborations with mentees: manuscripts, curricular change, research
o
• Peers
f
Participation in ormal mentoring program locally, regionally, or nationally
H
o
Formal evaluations by mentees, o the clinician-educator’s mentoring abilities
s
p
i
Educational administration Admissions committees ( or residency or medical school), including interviewing activity
t
a
and leadership
l
Course or clerkship director
M
Education or curriculum design and oversight committees
e
d
i
Dean’s o ice or student a airs position
c
i
n
Evaluation committee
e
a
Program directing
n
d
Elected positions and committee involvement in pro essional societies or medical education activity
S
(eg, Association o Program Directors or Clerkship Directors in Internal Medicine; education committee
y
s
work within pro essional society, such as Society o Hospital Medicine)
t
e
m
Faculty development leader (eg, taught small group o aculty on ultrasound-guided central venous
s
catheter insertion)
o
Educational scholarly activity Grants or education research or education activity (eg, unding to design new curriculum)
f
C
a
Publications (peer-reviewed publications, book chapters, editorials, opinion pieces, letters)
r
e
Editing and peer reviewing o grants or manuscripts related to education research
Test question writer (eg, or American Board o Internal Medicine or In Training Exam)
Posters or oral presentations involving educational topics at society meetings
Workshop moderator or presenter at society meeting (eg, promoting quality improvement activity in
residency)
Hospital administration Committee work: role on committee (eg, member, chairperson) and speci ic activities or that
committee
Quality improvement activities
Patient sa ety activities (eg, peer review committee, root cause analyses)
IT development
including quality improvement and patient sa ety (Table 39-5). on the aculty member’s medical education philosophy and career
Importantly, the port olio o ten serves as the basis or pro essional goals is included. Continuous maintenance o a port olio ensures
review, academic promotion, as a “body o work” when meeting that important activities are captured as they occur. To emphasize
prospective employers, and as the basis or sel -re ection. Ideally, the most signi cant accomplishments, we suggest a limit o 10 pages
the port olio’s ormat aligns directly with institutional promotion without attachments.
criteria, thus allowing chairpersons and promotion committees
(members o which may not be clinicians) to appreciate the value
o distinct activities.
The port olio is intended to complement, not replace, the curriculum SPECIAL MENTORING RELATIONSHIPS IN ACADEMIC
MEDICINE: TRAINEES, WOMEN AND MINORITIES
vitae. Whereas the curriculum vitae serves as a detailed outline, the
port olio goes urther to exhibit the quality and breadth o accom- ■ MENTORING TRAINEES
plishments. For example, an ef ective port olio o ten contains the From the perspective o trainees, ideal mentor attributes include
actual curricula that a clinician-educator developed, manuscripts maintaining con dentiality; being approachable, accessible, and
and miniaturized poster presentations that a aculty coauthored nonjudgmental; getting to know the mentee on a personal level;
with mentees, a summary o numerical evaluations o the aculty promoting the mentee’s goals over the mentor’s; opening doors or
by trainees, and detailed in ormation on mentee outcomes result- the mentee; and encouraging sel -re ection. See Table 39-6 ingre-
ing rom the aculty member’s mentorship. Commonly, a narrative dients or success ul resident mentoring.
262
TABLE 39-6 Strategies to Promote Successful Mentoring TABLE 39-7 Strategies to Promote Successful Mentoring of
C
of Trainees Underrepresented Minorities
H
A
• Formal assignments o aculty mentors that takes into URM mentoring programs should emphasize the ollowing
P
consideration the needs o women and minorities areas in addition to characteristics common to any success ul
T
• Dedicated time or trainees and mentors to meet: ormal mentoring relationship:
E
meetings at least twice annually with readily available access
R
• Skills development in:
to mentor at other times Cross-cultural communication
3
• Buy-in rom department chairperson:
9
Socialization into the ield o medicine
Ensure protected aculty time or mentoring Networking
Establish bonuses or awards or success ul mentoring Career advancement: speci ic guidance in aculty promotion,
M
Explicitly recognize mentoring as criteria or aculty especially as URM aculty are more likely to engage in
e
promotion
n
activities such as community-based endeavors and pipeline
t
• Faculty development to augment mentoring skills around the
o
development which have less clearly established path or
r
ollowing common topics or house sta mentees:
s
promotion
h
Helping residents apply or jobs a ter residency: Negotiating
i
p
Preparing a curriculum vitae Research: grant writing, navigating the IRB, manuscript
o
Focusing the job search preparation, editing
P
Interviewing skills • Leadership development
e
e
Communication with potential employers • Creation o a welcoming environment to reduce eelings o
r
s
Negotiating strategies isolation
a
Timeline
n
• Assignment o mentees with mentors who share race, culture,
d
Assisting residents with the ellowship application process: ethnicity, language, or gender
T
Advantages to applying during second- vs third-year o • Introduction o mentees to national URM leaders
r
a
residency
i
• Partnership with the a iliated medical center to ensure the
n
Fellowship match through ERAS (how the process works,
e
mission o the medical center includes promotion o diversity
e
key deadlines) and reduction o disparities in order to achieve a success ul
s
Selecting aculty to write letters o recommendation and sustainable URM mentoring program
Preparing personal statements and curriculum vitae
Selecting programs to apply to (reputation, location,
unique program aspects, when to consider programs
outside the match)
Elements sought by ellowship directors their activities and achievements into an educator port olio, a tool
Interviewing skills designed to enhance their chances o promotion at academic
Li estyle considerations in the specialty medical centers. Resident mentoring should ocus on career plan-
Mentoring resident scholarly activity:
ning and preparedness, networking, identi ying opportunities or
Preparing a clinical vignette abstract, poster, or case
report scholarly activity, role modeling o work-li e balance, and supporting
Giving oral presentations and speaking publicly the resident.
Establishing a quality improvement project
Writing e ective review articles
Providing potential networking opportunities locally, SUGGESTED READINGS
regionally, nationally
Involving residents in pro essional societies Beech BM, Calles-Escandon J, Hairston KG, et al. Mentoring
Submitting to the institutional review board programs or underrepresented minority aculty in academic
Providing resources or statistical support medical centers: a systematic review o the literature. Acad Med.
Optimizing clinical and pro essional development: 2013;88:541-549.
Selection o elective rotations Daley SP, Palermo AG, Nivet M, et al. Diversity in academic medicine
Ethics, integrity, and pro essionalism
no. 6. Success ul programs in minority aculty development:
Interpersonal skills with physicians, nurses, and sta
Time management, organization, and e iciency ingredients o success. Mt Sinai J Med. 2008;75:533-551.
• Providing resources or handling di icult situations (burnout, Farrell SE, Digioia NM, Broderick KB, et al. Mentoring or clinical-
psychiatric illness, substance abuse, interpersonal con licts, educators. Acad Emerg Med. 2004;11:1346-1350.
amily/social stressors, career angst, inancial stress, and
Ramanan R, Phillips R, Davis RB, et al. Mentoring in medicine: keys to
medical errors)
satis action. Am J Med. 2002;112:336-341.
Ramani S, Gruppen L, Kachur EK. Twelve tips or developing ef ective
mentors. Med Teacher. 2006;28:404-408.
MENTORING WOMEN AND MINORITIES
Sambunjak D, Straus SE, Marusic A. Mentoring in academic medi-
A recent systematic review highlights the existing evidence gap in cine: a systematic review. JAMA. 2006;296:1103-1115.
mentoring programs or women and minorities and also serves
Straus SE, Johnson MO, Marquez C, et al. Characteristics o success ul
as a call to action to reduce disparities in student and aculty
and ailed mentoring relationships: a qualitative study across two
recruitment, retention, and promotion o women and minorities.
academic health centers. Acad Med. 2013;88:82-89.
Tab le 39-7 highlights strategies speci c to this problem.
Tsen LC, Borus JF, Nadelson CC, et al. The development, implemen-
CONCLUSION tation, and assessment o an innovative aculty mentoring leader-
ship program. Acad Med. 2012;87:1757-1761.
No single mentoring model is superior, so the mentoring pair
should use one that best meets the mentee’s speci ic goals. Zerzan JT, Hess R, Schur E, et al. Making the most o mentors: a guide
Clinician-educator mentees should be encouraged to assemble or mentees. Acad Med. 2009;84:140-144.
263
CHAP TER
40 INTRODUCTION
Hospitalists are o ten asked to participate in or lead quality improve-
ment (QI) and research initiatives, locally and nationally. Because
data collection and eedback are part o any QI e ort, and because
the results o these e orts are o ten published, the hospitalists who
lead these e orts o ten ask (or are asked by others) the question: “Is
this research?” The short answer is that QI research is di erent rom
standard QI e orts in many respects. In this chapter, we will address
the di erences between standard QI e orts and research, some
Research in the reasons to do QI research, the appropriate time to do QI research
( or you and or the scienti c question at hand), how “rigorously” to
Hospital conduct QI research, getting started with the process, the ingredi-
ents or a success ul project, and issues related to study design and
methods that are either unique to or are particularly relevant to QI
research.
Jef rey L. Schnipper, MD, MPH, FHM This chapter will address both “quality improvement” and “patient
sa ety” research. The two terms are o ten used interchangeably, and
o ten the line between them is gray. For example, is an e ort to
increase β-blocker use to prevent a second myocardial in arction an
issue o QI or sa ety? That said, “sa ety” is o ten used in the context o
rare incidents where there is a strong link between an error and its
associated outcome (eg, wrong-site surgery). The issues regarding
both types o research are o ten the same. This chapter will review
additional issues unique to patient sa ety research that take account
o the rarity o many sa ety events.
PRACTICE POINT
• Collecting, analyzing, and reporting data does not turn a
QI project into research. The important characteristic o QI
research, as opposed to standard QI e orts, is that the question
to be answered is not “can we improve care here?”but “does this
intervention work in general?”I the goal is to design and test
a novel intervention to improve care (or to test an established
intervention in a novel setting), to establish whether a
particular intervention works in a wide variety o settings such
that it might become a new standard o care, and/or to learn
generalizable lessons about how to success ully implement
such an intervention, then it is research.
OVERVIEW OF QI RESEARCH
■ QI VERSUS QI RESEARCH
QI research is not just writing up the results o a QI project. In act,
writing up the results should be part o almost all QI e orts so that
other institutions may learn rom your experience and you may earn
“academic credit” or having done the work. The SQUIRE Guide-
lines (http://www.squire-statement.org/) provide detailed advice on
how to write up such results. Content unique to these reports (as
opposed to conventional research manuscripts) include:
• Introduction: description o the local problem and the intended
improvement
• Methods: discussion o any ethical issues, planning the inter-
vention, and planning the study o the intervention
• Results: description o the environmental context, a timeline
o the intervention, degree o success in implementation,
how and why the plan evolved, and lessons learned rom that
evolution
264
• Discussion: issues o maintaining improvement over time, causal • What is the magnitude o bene t?
mechanisms regarding the speci c components o the inter- • Is it cost-e ective?
C
vention, and how environmental context played a role in the • Should this intervention be spread widely?
H
success (or ailure) o the intervention and its implementation
A
The time to do rigorous QI research is during phase 2 o an inter-
P
In a standard QI project, results over time may be displayed using vention, analogous to phase III drug trials (o ten randomized con-
T
run charts with statistical process control limits, with results plotted
E
trolled trials) needed or FDA approval. The intervention should be
R
over time, a central line at the mean, and limit lines at three standard studied as rigorously as possible, especially more novel, expensive,
deviations (SD) above and below the mean. In a chart with 25 data
4
or risky interventions, in order to know whether the bene ts truly
0
points, the chance o a point being outside the 3SD lines, indicat- outweigh the risks and costs.
ing “special cause variation,” is 6.5%, similar to the 0.05 threshold
or statistical signi cance in standard statistical tests. Such charts
R
e
allow participants in a QI e ort to see whether their interventions PRACTICE POINT
s
e
are working, whether the improvements seen are likely to be due to
a
• In “Phase I”o quality improvement, the question is usually “can
r
chance, and to help guide urther improvements to the intervention.
c
h
this intervention work in at least one place?”
The important characteristic o QI research, as opposed to stan-
i
•
n
dard QI e orts, is that the question to be answered is not “can we In “Phase 2,”a more rigorous evaluation is required to answer
t
the question o whether an intervention that might work is
h
improve care here?” but “does this intervention work in general?”
e
Human subject research is de ned as “a systematic investigation, ready or widespread use. This is time or QI research.
H
• In “Phase 3,”QI interventions proven e ective in Phase 2 are
o
including research, development, testing, and evaluation designed
s
disseminated widely.
p
to develop or contribute to generalizable knowledge.” “Generaliz-
i
t
able knowledge” may be urther de ned as “enduring knowledge
a
l
about the nature and unction o human beings.” I the only goal o
a QI e ort is to improve local compliance with currently recognized Once an intervention has been proven to work, the goal is to
best practices (or a sa ety e ort designed to reduce medical errors) spread the intervention widely. Phase 3 usually requires adapta-
using recognized procedures, without adding to existing knowl- tion to each local environment, ideally using lessons learned rom
edge about the general nature and unction o human beings, then prior work (eg, the most e ective components o the intervention
it is not research. On the other hand, i the goal is to design and and how to optimize implementation). Standard QI methods are
test a novel intervention to improve care (or to test an established now ocused on local adaptation o the intervention and making
intervention in a novel setting), to establish whether a particular the micro- and macro-environment more conducive to e ective
intervention works in a wide variety o settings such that it might implementation. Lessons about how, why, and where an interven-
become a new standard o care, and/or to learn generalizable les- tion works should have already been answered in prior QI research
sons about how to success ully implement such an intervention, to help guide this process.
then it is research.
■ SHOULD QI RESEARCH BE CONDUCTED RIGOROUSLY?
■ WHY AND WHEN TO CONDUCT QI RESEARCH
The best ways to think about and conduct QI research are not
Are the reasons to conduct QI research enough to motivate you to do
without controversy. There are some, such as Dr Donald Berwick,
the work required?
ormer head o the Institute or Health Care Improvement, who
Reasons to conduct QI research include general reasons, such as
would argue that all QI e orts are local. Inherently complex behav-
expanding the body o medical knowledge and helping your patients;
ioral interventions need di erent approaches than just testing a pill.
local reasons, such as answering a burning question important or
Randomized controlled trials purpose ully control or environmental
your institution or your practice; and personal reasons, such as pro-
context in the name o unbiased outcome assessment. And yet, it
essional satis action or to provide balance to your clinical duties. QI
is precisely the context and the process that are most important to
research requires a great deal o work and, especially at the beginning
evaluate, because they reveal how, why, and where an intervention
o a research career, may come as an addition to an already ull clinical
is success ul. Proponents o this approach argue or more case stud-
schedule. The rewards can be considerable i you are motivated.
ies and “ ormative evaluation” to better study these issues.
Is it the right time scientif cally to conduct QI research? However, ignoring stronger study designs may lead to gross
In phase 1 o QI, the question is usually, “Can this intervention overestimation o treatment e ects. Observational be ore-a ter
work in at least one place?” We could think about this phase using studies are inherently con ounded by temporal trends (ie, general
the analogy o drug trials, where early research and development improvement over time), cointerventions (other interventions that
work in a pharmaceutical company and phase I/II clinical trials look may a ect the outcome), and biased by the Hawthorne e ect
or sa ety and e cacy in a limited number o care ully selected (improvement that comes when people know they are being
patients. In this phase, interventions are o ten not well de ned. watched). And studies that compare those who volunteer to imple-
The best approach in this phase is to do standard QI work, iterative ment an intervention early to those who do not may be completely
re nement o the intervention using Plan-Do-Study-Act cycles, and con ounded by inherent di erences between implementation and
to monitor improvement using run charts. In other words, it is not control sites. QI research does not di er rom biomedical research
time to do QI research yet. However, some o the groundwork or in requiring some estimate about how likely an intervention will be
later research may be done at this time. In addition to optimizing success ul be ore institutions invest time, money, and resources in
the intervention itsel , measures o process and outcome may be their implementation. For example, a very success ul intervention at
developed, and measures o environmental context and interven- one hospital, but not success ul in the next 10 hospitals studied, is
tion delity may also be developed. very di erent rom a success ul intervention in 75% o the hospitals
In phase 2, the primary questions are the ollowing: in which it is evaluated.
• Does this intervention work outside o its original location? We there ore advocate or strong study designs when appropri-
• Does it require re nement? ate and possible, but we also advocate complementing this work
• How likely is it to work? with “mixed methods” (ie, quantitative and qualitative research)
265
that look care ully at contextual actors, intervention delity (ie, how investigator wants to resolve by making measurements on his or her
aith ully an intervention is implemented as designed), and barriers study population. Good research questions are indeed everywhere
and acilitators to success ul implementation. Not every study can and may be provoked by your clinical experience, by the advent o
(or should be) a randomized controlled trial. new technologies, by acknowledging the need to improve, and/or
P
A
by maintaining a healthy skepticism about prevailing belie s. Initial
R
CONCEPTUAL MODEL OF QI RESEARCH challenges to beginning a QI research project may include psycho-
T
Celia Brown and Richard Lil ord wrote a series o articles in 2008 logical (getting over the act that you are not a “researcher”), scien-
I
describing an approach or rigorously evaluating patient sa ety ti c (developing a good research question), and logistical (choosing
interventions based on the recommendations o a network spon- study designs that are easible). Developing questions is an iterative
sored by the Medical Research Council, United Kingdom that echoes process, and we recommend consulting early and o ten with advi-
this sentiment. Their conceptual ramework is shown in Figure 40-1. sors and colleagues.
T
h
It is based on the Donabedian “structure-process-outcome” model A good research question has the ollowing attributes:
e
and provides additional detail or studies o patient sa ety interven- 1. The study that can answer the question is easible: adequate
S
p
tions. They note that such interventions may be aimed at manage- number o subjects, availability o adequate technical expertise,
e
c
ment processes, such as nurse to patient ratios and time allocated to a ordable in time and money.
i
a
l
pro essional development, and/or clinical processes, such as washing 2. The question is interesting: it con rms, extends, or re utes pre-
t
y
hands between patients. The e ectiveness o an intervention should vious ndings (this implies that you have done the background
o
f
be observed at all points to the right o the intervention. Patient reading on the subject).
H
outcomes may be “hard” (such as hospital readmission or patient 3. The question is relevant: it has implications or clinical knowl-
o
s
mortality), which are clearly relevant but may be relatively insensi- edge, practice, or policy.
p
i
t
tive to change. Surrogate outcomes, process measures, and error 4. The question can be answered ethically.
a
l
rates are more sensitive to change and should complement hard
M
Research questions o ten start out vague (eg, “Are ewer nurses
outcome measurement.
e
bad or patients?”). To carry out a study, these questions need to
d
Observations at the point o intervention should be used to assess
i
be re ned so that they become measurable. The hypothesis needs
c
i
the f delity with which the intervention was implemented and how it
n
to be testable by getting speci c about three elements: the patient
e
has been adapted over time. I an intervention is not implemented population, the intervention, and the outcome. Using the above
a
with high delity, it is unlikely to cause improvements in outcomes,
n
example, a re ned research question might be “ or medical inpa-
d
even i they are observed. And when improvement is not seen, low tients, is there an association between the patient-to-nurse ratio and
S
y
delity may explain those ndings even i the intervention were in-hospital mortality?”
s
t
theoretically e cacious.
e
m
Observations to the le t o the intervention provide in orma-
s
tion about environmental context and may explain di erences PRACTICE POINT
o
f
among sites. Organizational structure may inf uence management
•
C
A good research study must be easible, relevant, con rm,
processes, which in turn a ect intervening variables such as morale,
a
extend, or re ute previous ndings, and be ethically answered.
r
e
sa ety culture, teamwork, and provider knowledge and belie s.
To carry out a study, the research questions need to be re ned
These variables then inf uence clinical processes. I an intervention,
so that they become measurable. There are two attributes that
implemented with high delity, improves all downstream processes
make a research study use ul to others: (1) internal validity,
and outcomes (even i some are not statistically signi cant), then
meaning that the results o the study ref ect reality and can be
it increases the likelihood that the intervention itsel was really the
believed, and (2) external validity, or generalizability, meaning
cause o the improvement.
that the results can be applied to other patients or settings.
GETTING STARTED
■ ASKING THE RIGHT QUESTION ■ PLANNING YOUR RESEARCH
The most important rst step is choosing the right research ques- Once a research question has been re ned, the next steps are to
tion, ie, the uncertainty about something in the population that the develop a speci c aim (or aims) and a hypothesis. The speci c aim is
Ma na ge me nt
Clinica l proce s s e s P a tie nt
S tructure proce s s e s
Ac tive e rrors outcome s
La te nt e rrors
Inte rve ning

va ria ble s
(e g, mora le )
Conte xt Fid e lity Fid e lity
Ge ne ric inte rve ntion S pe cific inte rve ntion Throughput

(e g, huma n re s ource (e g, drug inte ra ction (e g, no. of
policy) wa rning s ys te m) pa tie nts tre a te d)
Figure 40 1 Conceptual ramework or evaluating patient sa ety interventions. Observations can be made at all points in the chain to provide in or-
mation on context, idelity, and quality and sa ety outcomes. (From Brown C, Lil ord R. BMJ 2008;337:a2764, with permission rom BMJ Publishing
Group Ltd.)
266
collect yoursel is obviously under your complete control and may
TABLE 40-1 Outline o a Research Proposal be designed to answer your exact question but takes time and
C
resources to collect. Data rom collaborators is sometimes the best
H
Research Speak English Translation
A
o all worlds i such opportunities are available and you are inter-
Research question What questions will the study address?
P
ested in the questions the data can answer.
T
Background and Why are these questions important?
E
signi icance What do we already know about them?
R
■ RESEARCH TRAINING AND EXPERIENCE
Aims and What do you plan to do?
4
You can do research, but it helps to have some background. This
hypothesis
0
What do you expect to happen? background may take many orms, rom mentors and collabora-
Subjects: Who will you study? tors, short-term programs, all the way to degree programs and
R
Inclusion/exclusion How will they agree to participate? ellowships. In our experience, a summer-long program in research
e
s
criteria methods (such as those at the Harvard School o Public Health and
e
a
Consent at the University o Cali ornia, San Francisco School o Medicine),
r
c
ollowed by a research project done in close collaboration with an
h
Design How will you actually do the study?
experienced mentor, is o ten enough to pave the way or urther
i
n
What is your “protocol”?
collaborative research projects and/or small independent projects.
t
h
Data sources Where will you get your data?
e
What do you plan to collect?
H
■ TIME
o
Outcome variables Which data are key to your question
s
p
(or hypothesis)? One o the most challenging ingredients to obtain is time. Unless
i
t
you have protected time built into your schedule or you have
a
Predictor variables Which will con use (or con ound) the
l
(covariates) issue? negotiated or protected time up ront, the short-term solution is
o ten an investment o your own time. I you can prove your ability
Statistical issues How large is the study and how will it be
analyzed? to success ully conduct projects, especially QI and sa ety projects
o inherent value to the institution, then protected time can o ten
Human subjects How will you maintain ethical standards?
be negotiated the second time around. An important early step is
How will you protect patients’ rights? to estimate the time needed to complete a research project. We
recommend designing a timeline as part o your initial proposal
(Figure 40-2). You should consult early with someone else to make
what (exactly) you want to do. For example, “Determine the e ects sure the timeline is easible. Almost all studies take more time than
o an enoxaparin guideline on the appropriate use o enoxaparin.” you initially anticipate!
The hypothesis is the a priori, testable expectation or what you
think is going to happen (or has happened). For example, “imple- ■ FUNDING
menting an enoxaparin guideline will increase appropriate use o
The degree o unding required, i any, will depend on the project.
enoxaparin.”
Be sure to consider the opportunity costs o you and your col-
The next step is to develop a study proposal. Writing everything
laborators (ie, time taken away rom other activities) in addition to
down in a standardized way serves several purposes: it orces you
direct costs. Direct costs might include the paid e ort o research
to address all the issues that may come up with a study’s design
assistants and statisticians, o ce supplies and incidentals. As with
and execution; it provides a convenient way or you to explain your
a timeline, it is never too early to design a preliminary budget and
study to others, get eedback, and re ne your methods; and it is
share it with others or re nement (Table 40-2). Line items include
necessary or Institutional Review Board (IRB) approval and to obtain
personnel costs (salary and ringe or employee e ort, hourly or daily
unding i needed (but we recommend developing a study proposal
ees or consultants), equipment, travel, and miscellaneous costs like
even i you do not intend to apply or unding). Table 40-1 provides
o ce supplies, so tware, computer hardware, and photocopying.
a list o the elements o a research proposal, both in “research speak”
Potential sources o unding depend on the scale o the project
and their English translations.
and the purpose o the study. For most QI and sa ety projects, begin
with your institution, especially i the costs are modest ($50,000-
INGREDIENTS FOR A SUCCESSFUL RESEARCH PROJECT
$100,000). I you can link your research question to nancial and
Once you have a good research idea, what else do you need to turn per ormance priorities o your hospital, internal unding becomes
it into a success ul project? The basic components are data, research more likely. Internal sources include division or department unds,
training or experience, time, unding, a research team, IRB approval, hospital-wide research grants, unding rom the risk management
and a plan or dissemination o the results. organization at the hospital, and charitable giving. For questions
less closely linked to hospital priorities or or larger projects, external
■ DATA SOURCES unding is o ten required. Sources include oundations such as the
Data may come rom a wide variety o sources, including institu- National Patient Sa ety Foundation, the Commonwealth Fund (US),
tional data, publicly available data, data rom collaborators, and and the Robert Wood Johnson Foundation; governmental sources
data you collect yoursel . Every institution collects data or billing, such as Veterans A airs, Agency or Health Care Research and Qual-
public reporting, and other purposes. Find out who leads this e ort, ity, and National Institutes o Health; statewide or national QI or
whether and how you can access these data, and how clean they sa ety initiatives; the Patient-Centered Outcomes Research Institute
are (eg, whether the owners have already taken care o issues like (PCORI, a nonpro t nongovernmental organization authorized by
missing or erroneous data, misspellings, and other eatures that Congress in 2010); and industry, including pharmaceutical com-
make research quality analyses possible). Publicly available data panies and device makers. Relationships between physicians and
have obvious advantages, but it may sometimes be time consum- industry are under more scrutiny than in the past, but some compa-
ing and/or expensive to obtain, and data sets can be large and nies are still willing to und investigator-initiated projects i priorities
unwieldy, requiring a certain level o statistical expertise. Data you are closely aligned.
267
Mo nth
Ac tivity 1-3 4-6 7-9 10-12
P
A
Obta in IRB a pprova l
R
T
As s e mble re s e a rch te a m
I
De ve lop da ta colle ction
forms , pilot te s t, e tc
Colle ct da ta
T
h
e
S
Ana lyze da ta
p
e
c
i
Write up a nd pre s e nt
a
l
t
y
o
Figure 40 2 Sample timeline.
f
H
o
s
p
■ RESEARCH TEAM For our purposes, QI research (as opposed to QI e orts alone) is
i
t
human subjects research and requires IRB approval. However, in
a
The research team should be multidisciplinary, as is true o QI teams
l
many cases, it is possible to obtain expedited approval with waiver
M
in general. One economical way to nd collaborators is to work with
e
medical students and residents, as long as the work can be done o patient consent on the grounds that the studies are o minimal
d
i
risk, that patient con dentiality will be rigorously protected, that
c
around their busy schedules (as or example, retrospective chart
i
n
review). Paid research assistants are o ten required or daily, pro- the study could not be practicably done i patient consent were
e
spective data collection. It is never too early to involve a statistician required, and that the rights o patients will not be adversely
a
n
to help determine sample size (and there ore costs and easibility a ected by waiver o consent and authorization. As with your
d
research proposal, begin the IRB application process early and have
S
o the study) and identi y other methodologic issues be ore they
y
others with experience look through it (human research committees
s
become problems.
t
e
o ten have consultants who can help with this). Interview orms and
m
■ IRB APPROVAL consent orms will need to be created as part o this process. Do
s
not worry i these are not nalized—amendments can always be
o
Institutional Review Board approval is o ten the source o much
f
submitted later. Lastly, as part o the IRB approval process, study sta
C
angst and controversy. To complicate matters, there is yet to be
a
must complete training in the ethical treatment o research subjects.
r
consensus on when IRB approval is needed. Some recent clarity
e
Each institution has its own speci c requirements.
was provided by Lynn et al. in a 2007 article in Annals o Internal
Medicine, “The ethics o using QI methods in health care.” They
■ DISSEMINATION PLAN
note the characteristics o activities that are likely to be both QI and
human subjects research, such as A success ul research project should include a dissemination plan.
Research in a vacuum is not use ul. You have a responsibility to let
1. Issues that go beyond current best practice
others know the results o your study, positive or negative. Besides
2. Allocation o patients to di erent treatments
manuscripts, other study products might include detailed descrip-
3. Deliberately delayed eedback o data to avoid bias
tions o your interventions, so tware speci cations, data collection
4. Key involvement o researchers without commitment to ongo-
instruments (surveys and interview guides). Outlets or dissemina-
ing QI at that site
tion include not only peer-reviewed journals but presentations at
5. Funding by parties outside the clinical setting
your own institution and pro essional society meetings, reports to
your unders, and press releases.
STUDY DESIGN ISSUES IN QI AND SAFETY RESEARCH

TABLE 40-2 Sample Budget
■ INTERNAL AND EXTERNAL VALIDITY
Item Description % Time Amount Two attributes make a research study use ul to others: (1) internal
Principal Oversees project 20% $30,000 validity, meaning that the results o the study ref ect reality and can
investigator be believed; and (2) external validity, or generalizability, meaning
Coinvestigator Research collaborator 10% $15,000 that the results can be applied to other patients or settings. Without
Consultant Specialized assistance 5% $5000 internal validity, generalizability is a moot point, so in a sense inter-
Research Collects data 50% $15,000 nal validity is the more important o the two attributes.
assistant
Statistician Cleans, analyzes data 10% $10,000 ■ THREATS TO INTERNAL VALIDITY
Travel to SHM Presentation o $1500 Chance
indings There are three main threats to the internal validity o a study:
O ice supplies Copying, axes, etc $500 chance, con ounding, and bias. Chance means that the results
Total $77,000 were simply due to the “luck o the draw.” Type 1 error is when the
null hypothesis is rejected due to chance when in act it is true ( or
SHM, Society o Hospital Medicine. example, drug A is ound to be better than drug B when in act the
268
two are equal). This can be thought o as a alse-positive study. By ■ MANAGING THREATS TO INTERNAL VALIDITY
convention, the threshold or type 1 error in most studies (known as
C
Table 40-3 provides ways to manage the threats to internal validity
alpha) is set at 0.05. When the P value rom a statistical test is below
H
while the study is being designed (ie, be ore data are collected) and/
this threshold, we call the di erence statistically signi cant. Another
A
or while the study is being analyzed (a ter data are collected). For
P
way to think o this is to say that or every 20 studies that nd a sta- example, during the study design phase, chance is managed using
T
tistically signi cant di erence, one o those studies is wrong simply power and sample size calculations. I alpha and beta are chosen
E
due to chance.
R
and the e ect size can be estimated (ie, how bene cial the interven-
Type 2 error results in accepting the null hypothesis when in act tion will be), then the sample size can be calculated. E ect size may
4
it is alse (eg, concluding that drug A and drug B are no di erent
0
be estimated rom preliminary studies, or more conservatively, may
rom each other when in act drug A is better). The threshold, called be chosen as the smallest e ect that would be considered “clini-
beta, is typically set at 0.10 or de nitive studies and 0.20 or prelimi- cally signi cant” by clinical experts. Conversely, i the sample size
R
nary studies—two to our times higher than alpha. In other words,
e
is xed (eg, the ward has a known daily census and the study must
s
e
the medical scienti c community has implicitly decided that it is be completed in 6 weeks), then statistical power to detect di erent
a
r
more acceptable to have a alse-negative study than a alse-positive e ect sizes can be calculated. Once a study is completed, the e ect
c
h
study. This makes sense or drug trials when the consequence o a o chance is derived using tests o statistical signi cance (ie, calcula-
i
n
positive study is FDA approval. It is less clear whether this argument tion o P values).
t
holds or QI research. Statistical power is 1—β. So, a study with 90%
h
During the study design phase, con ounding can be managed
e
power means there is a 10% chance that it will not nd a di erence most e ectively through randomization (con ounding cannot exist
H
when in act one exists, simply due to chance.
o
i the con ounder is evenly distributed among those who do and
s
p
do not receive the intervention). Anything short o randomization is
i
t
Confounding going to be less e ective at managing con ounding, but attempts
a
l
A con ounder is a third actor (ie, not the exposure or the outcome), may still be made to pick patient populations that are as compa-
associated with the exposure o interest (like the QI intervention you rable as possible (eg, medical wards rom medium-sized, nonteach-
are studying) that independently causes the outcome o interest. ing community hospitals in the suburbs). During the data analysis
For example, the earliest work in the epidemiology o lung cancer phase, con ounding can be managed with strati cation (eg, looking
ound a strong association with alcohol consumption. The con- at the e ect o alcohol on lung cancer in smokers separately rom
ounder, o course, was cigarette smoking: something associated the e ect in nonsmokers). This is very e ective i there are only one
with alcohol use (people smoke in bars) that causes the outcome or two major con ounders, but is impractical when the list o con-
o interest (lung cancer). Con ounding has to do with the science o ounders is large (N con ounders means 2N subgroups to analyze,
what is being studied. Managing con ounding requires knowledge each with a very small sample size). The purpose o multivariable
o what actors could cause the outcome(s) you are studying. In (sometimes called multivariate) analysis is to simultaneously “adjust”
QI research, major con ounders include temporal trends (ie, gen- or multiple con ounders at once. But keep in mind that con-
eral improvement with time, a problem with be ore-a ter studies), ounding can still exist, either because o incomplete adjustment,
cointerventions (ie, other interventions that a ect your outcome, inaccurate measurement o the con ounder, or existence o other
implemented at the same time but apart rom your intervention), unmeasured con ounders.
and “indication” (ie, in nonrandomized studies, the reasons why sites During the study design phase, bias can be minimized by
choose to implement an intervention that may be strongly related employing principles o sound study design: blinding to interven-
to the outcomes being studied). tion status (not just “double-blind,” but as many people involved in
the study as possible: patients, research assistants, outcome asses-
sors, statisticians), prospective data collection, valid data collection
Bias
instruments such as previously validated surveys and question-
In general terms, bias is error in a study that results in an incorrect naires, and thorough ollow-up or all endpoints. Once the study
estimate o the association between the exposure and the outcome. has been conducted, bias cannot be “adjusted or” during the data
There are two broad categories: selection bias, which is an error in analysis phase, although sometimes its direction and magnitude
the process o identi ying the study populations; and, much more may be estimated. For example, to estimate the impact o loss-
common, observation or in ormation bias, which is an error in to- ollow-up, some experts recommend assuming that everyone
measurement o the exposure and/or the outcome. There are many who received the intervention and was lost to ollow-up did poorly,
types o in ormation bias, each with its own story: recall bias, inter- while everyone who received usual care and was lost to ollow-up
viewer bias, loss to ollow-up, misclassi cation. Unlike con ound-
ing, bias has more to do with human nature and how the study is
conducted. In QI research, one major bias is the Hawthorne e ect,
or changes in measurement caused by participants’ knowledge that TABLE 40-3 Managing Threats to Internal Validity
they are being observed. Another potential bias arises when the
intervention itsel a ects measurement. For example, in early studies Threat to
o rapid response teams (RRTs), patients whose status was Do Not Validity Study Design (Before) Analysis (After)
Resuscitate (DNR) were o ten excluded rom the outcome o in- Chance Power calculations Statistics
hospital mortality. When RRTs approached patients in extremis, one Con ounding Randomization Strati ication
common activity was veri ying code status. In several cases, patients Picking comparable Multivariable
and their caregivers decided to change code status to DNR. These groups regression
patients were there ore excluded rom the outcome. But be ore the
Bias Blinding, prospective Unable to manage
advent the RRTs, these patients would have been included in out- data collection, valid (although sometimes
come assessment because no one veri ed their code status at that instruments, thorough direction and
time. A be ore-a ter study could nd a di erence in mortality simply ollow-up magnitude can be
because the intervention altered who was in the denominator or estimated)
outcome assessment.
269
did well. Even large e ect sizes can crumble under the weight o intervention can be randomized; there is particularly big concern or
such assumptions i the loss to ollow-up rate is large. On the other temporal trends, cointerventions, and other con ounders that may
hand, poor data collection instruments o ten create noise and bias not be known or cannot be measured; when the costs or potential
“toward the null” (ie, nding no di erence). There ore, i a di erence risks o the intervention is high (such that they need to be balanced
P
A
is ound, it is probably not the result o such bias. Because bias can- against a precise estimate o bene ts); the target outcome is o
R
not really be adjusted or a ter the act, it is important to manage high value (such as mortality); many settings will be a ected by the
T
study design issues up ront, be ore any data are collected. results (eg, possible incorporation into a regulatory requirement);
I
and any other situation that requires a precise estimate o e ects
■ IMPROVING GENERALIZABILITY and costs.
Guidelines or improving the generalizability o a study include the
ollowing: (1) describe your patient population well, so that others Cluster randomization
T
h
can determine whether their patient populations are comparable; Cluster randomized trials are a type o RCT in which the unit o
e
(2) describe your interventions well, so that others can determine randomization is larger than the individual patient. For example, in a
S
p
what they would need to do to replicate your experience; and recent trial o a medication reconciliation intervention, we random-
e
c
(3) describe your environmental context well, so others can deter- ized by both medical team and f oor so that we had clean separa-
i
a
l
mine whether a comparable intervention could be implemented tion o both nurses and doctors in the two arms o the study. This
t
y
in their settings. A common question related to generalizability is avoids treatment group “contamination” (ie, clinicians who change
o
f
whether interventions should be maximally customized to the par- their behavior even with control patients because they have been
H
ticular site where it is being studied. Such an approach increases the exposed to or know about the intervention), it acilitates imple-
o
s
chances o success, but it may come at the price o generalizability. mentation o the intervention (eg, service-wide educational e orts),
p
i
t
The answer should depend on which aspects o the study site are and administrative convenience. The major disadvantage is loss o
a
l
unique. Ideally, during phase 1 work, exactly which eatures need to statistical power. When patients o one physician, or example, are
M
be customized (and how) has already been determined; and these treated similarly to each other but di erent rom the patients o
e
d
customizations can then be speci ed in advance. another physician, this results in “intraclass correlation.” This reduces
i
c
i
the e ective sample size, depending on the degree o the correla-
n
e
■ TYPES OF STUDIES USED IN QI RESEARCH tion and the size o the clusters (eg, the number o patients per
a
physician). This correlation there ore needs to be anticipated and
n
Study designs appropriate or QI and sa ety research include
d
estimated in advance and used when making estimates o required
S
randomized controlled trials (with randomization at either the
y
sample size. Nevertheless, this is the pre erred study design in many
s
individual patient level or “clustered” by physician, ward, service,
t
cases, especially when there are big advantages to training and
e
or hospital), be ore-a ter studies, interrupted time series, “stepped
m
implementation, the threat o contamination is high, the interven-
wedge,” and observational cohort studies (prospective and retro-
s
tion requires it, the cluster size is small, and/or when the advantages
o
spective). Each has its advantages and disadvantages. Not all QI
f
are otherwise worth the loss in power.
C
research needs to be a randomized controlled trial to provide valid
RCTs, whether clustered or not, may su er rom issues o general-
a
in ormation.
r
e
izability in that not every patient or health care setting may be will-
ing to participate in one. In general, this is a sacri ce worth making
Randomized controlled trials
in the name o internal validity, i easible and appropriate to do so.
A randomized controlled trial (RCT) minimizes con ounding by Describing your patients, intervention, and environment well goes a
ensuring that potential con ounders are equally distributed in the long way toward alleviating these concerns.
di erent arms o the study. This is true regardless o whether the
con ounders are known or can be measured. To the extent that Stepped wedge
RCTs require prospective data collection and at least allow the pos-
The stepped wedge re ers to a study design in which an interven-
sibility o some blinding, they minimize bias as well, as long as there
tion is sequentially rolled out to di erent groups at di erent times,
are valid data collection instruments and thorough ollow-up. And
such as di erent f oors o a hospital (Figure 40-3). The order o the
these studies still require adequate sample size to deal with chance.
rollout is randomized to avoid con ounding by indication (ie, those
Note that the outcome may be measured at one point in time (a ter
most ready or the intervention get it rst). Each group serves a di -
the intervention has been implemented in those sites randomized
erent amount o time in the usual care and intervention arms. This
to receive it) or as relative improvement over time (preintervention
to postintervention). The latter may be pre erred i you suspect large
variation in baseline per ormance and especially i the same patients
are going to be ollowed or the entire study period. All RCTs should E
s
r
e
be analyzed on an “intention-to-treat” basis, meaning that out-
t
s
D
u
comes are measured according to the original intent o the random-
l
c
/
ization, regardless o what treatment (i any) was actually received.
s
l
C
a
This preserves the sanctity o the randomization and is particularly
u
d
important in QI studies, where the actors that lead to success ul
vi
B
i
d
implementation or compliance with an intervention may indepen-
n
I
dently predict positive outcomes. However, RCTs may not always be A
easible or ethical. You may have limited control over who receives 1 2 3 4 5 6
a QI intervention or you may be required (eg, or regulatory reasons)
Da ta colle ction point (time pe riod)
to provide it to everyone. It is o ten not ethical to study potentially
harm ul interventions with an RCT. And or outcomes that are rare Figure 40 3 Stepped wedge study design. The intervention is rolled
or take a long time to develop, RCTs may be prohibitively expensive out to individuals or clusters sequentially over time, rom blank cells
to conduct. Rare outcomes are particularly an issue with sa ety (as (usual care) to shaded cells (intervention). (From Brown C, Lil ord R. BMJ
opposed to QI) studies. The best times to use an RCT are when an 2008;337:a2764, with permission rom BMJ Publishing Group Ltd.)
270
approach allows adjustment or temporal trends and also minimizes implementation in the intervention arm). However, even i the anal-
con ounding because each group serves as its own control. Thus, ysis incorporates baseline rates by comparing improvement over
C
the stepped wedge provides the advantages o a control group but time, these types o studies are potentially f awed. Those sites that
H
is more practical than an RCT, especially i the intervention requires a are “early adopters” are o ten very di erent rom those that are not
A
P
gradual rollout anyway. The disadvantages are practical constraints in terms o culture o quality and sa ety, leadership, organizational
T
regarding implementation, logistical challenges, and the risk o con- structure, etc. These con ounders may have large e ects on the
E
tamination (eg, colleagues hearing about the intervention occurring ability to implement and improve outcomes in response to an inter-
R
on a di erent f oor). vention. These con ounders are also pervasive, potentially unknown,
4
and di cult to measure. This con ounding may be a atal f aw to
0
Time series internal validity. A better approach is to deliver the intervention to as
Be ore-a ter studies and interrupted time series may be more ea- many groups as possible (to improve generalizability) and do an ITS.
R
sible than RCTs when you have limited control over who receives
e
s
e
the intervention and when they receive it. In a simple be ore-a ter Summary of study design issues
a
r
study, outcomes are measured at one point be ore the intervention In conclusion, the optimal study design or a QI or sa ety study
c
h
and one point a ter the intervention (eg, use o measures to prevent depends on answers to several questions.
i
n
venous thromboembolism be ore and a ter a mass e-mail reminder
1. Is it easible and ethical to have a control group?
t
is sent out). This is probably the most common QI research study
h
2. Do you have control over who gets the intervention and when?
e
design, but un ortunately is one o the worst. It is very sensitive to
H
3. Does the intervention need to be implemented all at once or
temporal trends and to cointerventions. Much pre erred is an inter-
o
gradually?
s
rupted time series (ITS), in which repeated observations (at least 3)
p
4. Can/should the intervention be delivered to individual patients,
i
t
are made prior to and a ter the intervention is implemented. The
a
or does it make more sense to deliver it to a higher level (eg,
l
e ect o the intervention can then be measured over and above
physician, f oor, service)?
temporal trends (although it still does not adjust or cointerven-
5. Are the outcomes rare or do they take a long time to develop?
tions). ITS is use ul i the intervention needs to be given to everyone,
6. Are the costs and risks o the intervention low?
all at once. Because there is no control group, one disadvantage
7. Will the results be used to promote widespread adoption o
is lack o generalizability, especially i the study site is particu-
this intervention?
larly enthusiastic and well equipped to conduct the intervention.
Another disadvantage is that the analysis assumes that the interven- Table 40-4 illustrates how the answers to these questions may
tion is implemented once and does not change over time. However, inf uence your choice o study design.
newer sophisticated analytical techniques, such as random e ects
models with nonlinear time-by-intervention e ects, can take into ■ ADVANCED TOPICS AND CONTROVERSIES
account such phenomena as continuous improvement over time IN QI RESEARCH
(eg, as the intervention is re ned) and/or reaching a plateau due Should the interventions be held xed while they are being studied?
to a ceiling e ect (inability to improve urther because quality is On the one hand, such an approach allows or better description o
already so high). the intervention, makes it clear “what” is being studied, and makes
analyses easier to conduct. On the other hand, this approach does
Cohort studies not allow or standard continuous QI methods and may there ore
A word o caution is in order about observational cohort studies, limit the e ectiveness o the intervention. The answer may depend
in which sites that choose to implement an intervention are com- on how much customization is considered necessary or a given
pared with sites that choose not to. Because o potentially large intervention. In a phase 2 study, an intervention may already be
di erences in baseline rates o quality, the analysis should never optimized and it may be appropriate to hold it xed. But even under
be a simple comparison o outcomes at one point in time (eg, a ter these circumstances, we would still recommend a long pilot period
TABLE 40-4 Questions that In luence Choice o Study Design
Preferred Study Designs

Question If Yes If No
Is it easible and ethical to have a control group? RCT or clustered RCT ITS
Stepped wedge
Do you have control over who gets the intervention and when? RCT or clustered RCT ITS
Stepped Wedge
Does the intervention need to be implemented all at once? RCT or clustered RCT Stepped wedge
ITS
Can/should the intervention be delivered to individual patients? RCT (randomization by patient) Clustered RCT
Are the outcomes rare or do they take a long time to develop? Multicenter ITS RCT or clustered RCT
Are the costs and risks o the intervention low? ITS RCT or clustered RCT
Be ore-a ter studies
Will the results be used to generate a requirement or adoption o this RCT or clustered RCT ITS
intervention? Be ore-a ter studies
ITS, interrupted time series; RCT, randomized controlled trial.
271
that allows or continuous QI methods and/or a multiphase study in time to do research, both personally and or the research question.
which improvements to an intervention are planned or at periodic Plan your study in advance and get the help and training you need.
intervals. For example, a study o a novel so tware application should Do not be a raid o the IRB: they can actually help you with your
likely be conducted as a multiphase study; changes to so tware proposal, but allow enough time or the approval process to occur.
P
A
o ten take a while and “version 1” o so tware is almost never ideal. You can do QI research, and by getting started, you may nd that
R
Can an intervention be di erent at each site? Under some circum- the rewards and the potential impact o your work on large patient
T
stances it may be desirable to standardize the goals and unctions populations are well worth the e ort.
I
o the intervention rather than the exact orm and structure. This
allows or f exibility and maximizes the chances o success at each SUGGESTED READINGS
site. Such an approach may be pre erred when the intervention is
complex and when baseline achievement and environment are very Benneyan JC, Lloyd RC, Plsek PE. Statistical process control as a tool
T
h
di erent rom site to site. Hawe et al. recommend this approach. For or research and healthcare improvement. Qual Sa Health Care.
e
a recently conducted multisite study o medication reconciliation, 2003;12(6):458-464.
S
p
we chose to standardize the components o the intervention along
e
Brown C, Ho er T, Johal A, et al. An epistemology o patient sa ety
unctional lines (eg, “improve access to sources o preadmission
c
i
research: a ramework or study design and interpretation.
a
medication in ormation”). This was necessary because each site was
l
t
Parts 1–4. Qual Sa Health Care. 2008;17(3):158-181.
y
di erent in terms o its current processes, its local strengths and
o
weaknesses, and its environment. On the other hand, it makes the Brown C, Lil ord R. Evaluating service delivery interventions to
f
H
description o the intervention and the analysis more di cult. enhance patient sa ety. BMJ. 2008;337:a2764.
o
s
One way to improve the analytic approach to both these issues Donabedian A. Explorations in quality assessment and monitor-
p
i
(continuous improvement over time and di erent interventions at ing. In: Gri th JR, ed. The Def nition o Quality and Approaches
t
a
l
each site) is to break down an intervention into its component parts to it Assessment. Washington, DC: Health Administration Press;
M
and quanti y the e ects on outcomes when each component is 1980:4-163.
e
d
implemented at each site, adjusting or site and other potential con- Dumas JE, Lynch AM, Laughlin JE, et al. Promoting intervention
i
c
ounders. Using the above multicenter medication reconciliation QI
i
delity. Conceptual issues, methods, and preliminary results
n
e
study as an example, such an approach, using time-varying covari- rom the EARLY ALLIANCE prevention trial. Am J Prev Med. 2001;
a
ates, allowed us to determine which intervention components were
n
20(1 Suppl):38-47.
d
most (and least) e ective in lowering medication discrepancy rates.
Hawe P, Shiell A, Riley T. Complex interventions: how “out o
S
y
control” can a randomised controlled trial be? BMJ.
s
CONCLUSION
t
e
2004;328(7455):1561-1563.
m
Regarding study design, chance, con ounding, and bias are the
Lynn J, Baily MA, Bottrell M, et al. The ethics o using quality improve-
s
major threats to the internal validity o a study. It is ar better to mini-
o
ment methods in health care. Ann Intern Med. 2007;146(9):
f
mize these when designing your study than to deal with them later
C
666-673.
at the analysis phase. Di erent study designs address these issues
a
r
e
to di erent degrees. More ormal study designs, such as cluster- Ogrinc G, Mooney SE, Estrada C, et al. The SQUIRE (Standards or
randomized controlled trials, are pre erred (i possible and ethical) QUality Improvement Reporting Excellence) guidelines or qual-
when the bene ts o the intervention are not sel -evident or the ity improvement reporting: explanation and elaboration. Qual Sa
intervention is costly and not without risk. But other study designs, Health Care. 2008;17(Suppl 1):i13-i32.
such as stepped wedge and interrupted time series, are also excel- Pawson R, Tilley N. Realistic Evaluation. London: Sage Publications,
lent study designs depending on the situation. Ltd; 1997.
It always makes sense to have adequate sample size to adequately Salanitro AH, Kripilani S, Resnic J, et al. Rationale and design o
answer your study question, to describe your interventions well, use the Multicenter Medication Reconciliation Quality Improvement
unbiased measurement tools to collect data, look at both processes Study (MARQUIS). BMC Health Serv Res. 2013;13:230.
and outcomes, look at potential unintended consequences o your
intervention, and examine environmental context and interven- Schnipper JL, Stein JM, Wetterneck TB, et al. What are the best ways
tion delity. Studies are most success ul when study design issues to improve medication reconciliation? Paper presented at: An on-
are anticipated and managed early, well be ore data are collected. treatment analysis o the MARQUIS study. Plenary session, Society
There ore, get the help o experts in study design and statistics as o General Internal Medicine Annual Meeting; 2015; Toronto, ON.
soon as your study question has been re ned. Stetler CB, Legro MW, Wallace CM, et al. The role o ormative evalua-
Interesting research questions are everywhere and hospitalists are tion in implementation research and the QUERI experience. J Gen
in an ideal position to recognize these questions. Choose the right Intern Med. 2006;21(Suppl 2):S1-S8.
272
CHAP TER
41 INTRODUCTION
Anywhere rom a quarter to hal o physicians worldwide are esti-
mated to experience job burnout. General internal medicine and
amily medicine—the specialties or most adult hospitalists—claim
some o the highest rates o physician burnout in the United States.
Within the adult generalist elds, several studies suggest clinic-
based primary care physicians report burnout more requently than
their inpatient peers, but hospitalists remain among the physicians
most susceptible to job burnout.
For Individuals and The unique work-li e challenges o hospital medicine practice
include demands on hospitalists to serve several contemporary
Practices: Career physician roles as boundary spanners, communicators, nonleader

team members, quality enhancers, and caregivers. Even as hospital-
Sustainability and ists have become permanent xtures in most hospitals, role con icts
remain common with the competing demands o patients, other
clinicians, administrators, and employers. Shi twork with day/night
Avoiding Burnout schedule changes and sleep deprivation causes physiologic stress.
Productivity and e ciency demands are pervasive in modern health
care organizations and a driver o high workload. Pay uncertain-
ties with health care re orm and capitated care, and the pressures
Keiki Hinami, MD, MS to standardize care which lead to loss o autonomy and intense
scrutiny o per ormance contribute to conditions under which job
Tosha B. Wetterneck, MD, MS burnout is expected to remain a concern or the discipline.
WHAT IS BURNOUT?
Burnout is a psychological syndrome leading to a worker’s erosion
o engagement with their job due to long-term exposure to emo-
tionally demanding work. It is a condition observed predominantly
among those in the helping pro essions, like health and social ser-
vices where direct, requent, and intense interactions with people
are common and where the outcomes o work are not ully depen-
dent on worker actions. Burnout is commonly conceptualized as
having three constitutive dimensions. The rst, emotional exhaustion
is a literal depletion o worker energy due to work demands. It may
mani est in hospitalists as “compassion atigue” or the tendency
to distance themselves—cognitively and emotionally— rom their
work as they realize they cannot continue to give o themselves to
patients and coworkers. In essence, it is a coping response to work
overload. The second is depersonalization, marked by a detached
emotional callousness or cynicism that mani ests as indif erence or
dys unctional attitudes and behaviors toward patients. It is o ten a
protective response to emotional exhaustion. The nal component
o burnout, diminished personal accomplishment, is the erosion o a
worker’s sense o personal ef ectiveness, which brings on a eeling
o powerlessness and the tendency to negatively evaluate onesel .
This may mani est as a hospitalist not completing assigned tasks or
as worsening pro essional sel -esteem. Emotional exhaustion is usu-
ally considered necessary or burnout to be diagnosed, the other
components may occur in parallel, sequentially, or not at all.
Burnout is distinct rom related concepts like stress, depression,
and dissatis action. The de nitions o each have been established
empirically and while they overlap signi cantly, burnout is speci c
to the context o the workplace as an ongoing emotional response
to chronic demands and interpersonal stressors. Job dissatis action
is a predictor o burnout; workers who are dissatis ed are more
likely to be burned out. However, it is not ully clear whether dis-
satis action always precedes burnout or is a result o burnout (or
other workplace conditions that also produce burnout). Individuals
273
who are depression-prone have higher rates o burnout and even improvement, and their speci c organizational mission and char-
though burnout is speci c to the workplace it can also af ect home- acter. Organizational commitment re ers to employee identi cation
li e. Job stress can be conceptualized by two models: the demand- with and involvement in a particular community. Higher levels o
control-support model and the ef ort-reward imbalance model. In commitment are re ected in lower rates o attrition and better job
P
A
the rst model, job stress is more likely when there are high job per ormance. Burnout is associated with decreased worker organi-
R
demands (workload; time pressure), low control over job (autonomy zational commitment in health care pro essionals. In addition, being
T
in decision making; power over outcomes) and low support ( rom around burned out workers and having negative interactions with
I
colleagues, supervisor, organization; inadequate resources). Given them may impact other workers’ organizational commitment.
the high job demands inherent in the medical pro ession, control Although burnout is not the only or even the dominant reason or
and support are important mediators. With ef ort-reward imbalance, physicians leaving their jobs, the 2009 Hospitalist Physician Satis ac-
there is a discrepancy between the demands and obligations o the tion Survey ound 65% o hospitalists at risk o burnout (compared
T
h
job (ef ort) and the rewards of ered like salary, career opportuni- to 31% without burnout risk) reported their intent to leave their
e
ties, esteem, and job security. For workers who are committed to current practice within 2 years. Moreover, 68% o hospitalists at risk
S
p
their jobs, this imbalance leads to job stress. Again, given the high o burnout (compared to 45% without burnout risk) reported their
e
demands associated with hospitalist jobs and the pro essional com-
c
intent to leave the hospital medicine specialty within 5 years. The
i
a
mitment displayed by most physicians, rewards are very important cost o replacing a hospitalist may be comparable to the 250,000 US
l
t
y
to mediate stress. dollars estimated turnover cost o an outpatient-based generalist.
o
The cost o physician burnout in all o Canada was estimated to be
f
H
WHY IS BURNOUT IMPORTANT TO HOSPITALISTS? over 200 million Canadian dollars, o which two-thirds were rom
o
s
Across a wide variety o pro essions, burnout has been associated early retirement. Given the high cost o turnover and recruitment, it
p
i
is essential that burnout be minimized to ensure that experienced
t
with negative work outcomes including decreasing work hours
a
l
or job turnover, decreased work ef ectiveness and productivity, hospitalists remain clinically active in the pro ession to train the
M
reduced job and organizational commitment, and stress-related uture work orce.
e
d
health outcomes such as substance abuse and depression. In addi-
i
c
■ STRESS-RELATED HEALTH PROBLEMS
i
tion, the negative attitudes and actions o burned out workers
n
e
can negatively impact others in and out o the workplace. These AMONG PHYSICIANS
a
outcomes and the impact on Hospital Medicine are considered in
n
Like all health care providers, physicians experience high levels
d
depth in the proceeding section. o job stress. A study in the Netherlands ound a correspondingly
S
y
high proportion o hospital-based physicians suf ering work-related
s
t
■ PHYSICIAN BURNOUT AND JOB PERFORMANCE atigue (42%), depression (29%), anxiety (24%), and posttraumatic
e
m
Burnout predicts poor physician job per ormance. For example, stress complaints (15%). Burnout was the term originally coined by
s
Herbert Freudenberger in the context o alcoholism and drug abuse
o
providers who are happy with their work are known to increase
f
patients’ satis action, adherence to physician advice, trust, and con- in the 1970s, and it remains a eature o the 8% to 12% o health
C
pro essionals who develop a substance-related disorder at some
a
dence. On the other hand, patients o depersonalized physicians
r
e
have been shown to take longer to recover rom their illness. Physi- point in their lives. Suicide is another disturbing problem or health
cians in the United Kingdom reported providing lower standards o care providers. According to the Psychiatric Clinics o North America,
care, being angry and sometimes abusive with patients as a result o male physicians are two times more likely to commit suicide than
chronic stress. Similarly, burned out general internists and medicine the average American, and emale physicians are three times more
residents reported engaging in suboptimal patient care such as likely. Although only a ew studies have tied job burnout to the
making errors not due to a lack o knowledge or inexperience. There pro ound psychological and spiritual dislocation that predispose
appears to be a reciprocal relationship between burnout and per- physicians to sel -destructive behavior, the traditional lack o sel -
ception o committing medical errors. In a longitudinal study, resi- care among physicians, scarcity o resources or physician support,
dents who were burned out subsequently reported more perceived and burnout are understood to be related aspects o the same
errors than those who were not burned out and those who had stress-related problem.
more perceived errors initially were more likely to report burnout in
the uture. A recent systematic review o the relevant literature dem- WHAT LEADS TO BURNOUT IN HOSPITALISTS?
onstrated burnout’s potential impact, also, on nontrainee physician The key eatures o hospitalist work and individual characteristics
productivity through sel -perceived “insu cient” work ability and that are associated with burnout are summarized in Table 41-1 and
job retention measures. discussed in detail below.
Although the mechanisms connecting burnout to poor-quality
and poor-quantity patient care have not been empirically proven, ■ WORKLOAD
one proposed causal pathway involves the providers’ emotional
Excessive workload is con rmed as the overriding source o burnout
state. One measure o emotions—positive af ect—is associated
in physicians. Cost and time pressures pose additional job demands
with enhanced decision making and problem solving as well as
on the typical hospitalist. Irregular night shi ts, extended work weeks,
higher levels o patient centeredness in health care providers. There-
and other nontraditional work patterns can predispose hospitalists
ore, burned out hospitalists may be less cognitively vigilant and less
to physical and mental exhaustion. Some hospitalists engaged in
likely to put orth the extra ef ort necessary to deliver the highest
nonclinical responsibilities spend additional time on administrative
quality, patient-centered care.
work, education, and research. The average number o daily patient
encounters by a hospitalist in 2009 was 15, but depending on prac-
■ REDUCED ORGANIZATIONAL COMMITMENT tice type this number ranged rom 13 in typical academic groups to
AND JOB TURNOVER 17 in private practice groups. Variability in workload across practice
Virtually all hospitalists work in organizations or are direct employ- types was also apparent in the number o clinical shi ts per month
ees o hospitals and health systems. There ore, employers depend or ull time hospitalists (15 shi ts in academic groups compared to
on hospitalists’ commitment to advance patient care, quality 19 shi ts in local hospitalist-only groups), and hours o nonclinical
274
variations in care, these ef orts contribute to a sense o reduced
TABLE 41-1 Characteristics Associated with Job Burnout clinical autonomy. Much o the drivers o care standardization is
C
now written into the Af ordable Care Act, and the negative impact
H
Job Characteristics Individual Characteristics
A
o quality incentives on physicians’ sense o autonomy has been
Job demands: workload, time Early career
P
empirically demonstrated.
pressure, complex patients
T
Physician autonomy is also reduced when their role is uncertain
E
Role con lict and role Male gender or unpredictable. This is o ten the case in hospitalist work where
R
ambiguity the work demands rise and all erratically. A potential or role con-
4
Lack o job control/autonomy Not married ict exists, in theory, when hospitalists are required to reconcile
1
Lack o support or good Lack o social support outside their responsibilities as advocates both or the hospital and or the
relationships with colleagues o work patient. Primary or comanagement o patients typically cared or by
F
and supervisors
o
surgeons or specialists, also introduce the potential or role ambigu-
r
Lack o reciprocity rom Personality actors: external ity and decreased autonomy.
I
n
patients locus o control, low hardiness,
d
i
low sel -esteem
v
■ INTERPERSONAL CONTACTS
i
d
Lack o resources to do the job
u
The typical hospitalist routinely interacts with many individuals at
a
Lack o organizational commitment to hospitalist groups or
l
s
work. Whether patients, colleagues, administrative leaders, or any
individuals
a
number o coworkers on an interdisciplinary team, hospitalists are
n
d
expected to communicate well and seamlessly coordinate care with
P
other stakeholders. A central proposition o social exchange theory
r
a
c
is that individuals pursue reciprocity in interpersonal relationships,
t
i
c
work per month (19 vs 17 hours in multistate hospitalist groups vs and those who nd themselves participating in an unreciprocated
e
s
academic groups, respectively). A big data study rom the Christiana relationship will become distressed. Physicians who do not receive
:
C
Care Health System demonstrated a J-shaped increase in hospital the intrinsically sought-a ter positive eedback rom patients and
a
r
length o stay and costs with increasing workload, suggesting that coworkers are more likely to mani est signs o burnout.
e
e
hospitalists are routinely stretched to manage clinical loads large A study by Bakker and colleagues examined the relationships
r
S
enough to impinge on e ciency. between general practitioners and patients over a period o 5 years,
u
s
Both qualitative and quantitative work overload contribute to and ound that emotional exhaustion evoked negative attitudes
t
a
exhaustion by depleting the capacity o physicians to meet job toward patients. In turn, physicians who attempted to gain emotional
i
n
a
demands. Numerous studies, including the Physician Workli e Study distance rom their patients as a way o coping with exhaustion
b
were ound to engender the demanding patient behaviors that they
i
(PWS) o US generalists and specialists, demonstrated the un avor-
l
i
t
sought to avoid. Additional research establishes burnout’s associa-
y
able ef ects o overload, time pressure, and resource scarcity on
a
physician stress, burnout, and dissatis action. Workload is perceived tion with a greater exposure to patients with poor prognosis. Unlike
n
d
variably by individuals and cannot be measured simply by the traditional practitioners with whom patients have longitudinal rela-
A
number o hours worked. Another study ound that work hours only tionships, clinical encounters with hospitalists are relatively brie . The
v
o
indirectly predicted quality o care through perceived overload, sug- di culty o rapidly establishing rewarding relationships with patients
i
d
gesting other considerations mitigate the ef ects o physician work may pose a unique risk or burnout among hospitalists. Hof and col-
i
n
leagues ound that avorable patient-related interactions, more than
g
hours. One consideration is the protective ef ect o recovery time.
B
Acute atigue resulting rom an especially demanding event at work peer relations, predicted hospitalists’ intent to stay in the career.
u
Coworker support augments the personal accomplishment
r
does not necessarily lead to burnout, given recovery during rest ul
n
dimension o burnout, re ecting the value physicians place on
o
periods at work or at home. When work overload is persistent, as
u
opposed to sporadic, there is little opportunity to rest and restore expert evaluation by peers. Hinami and colleagues demonstrated
t
balance. A sustainable workload or work-li e balance, in contrast, hospitalist job longevity proportionate to more avorable organi-
may provide opportunities to re ne coping skills to draw energy zational climate and relationship with nonphysician staf . The PWS
rom other restorative aspects o daily li e. showed that support rom colleagues has a signi cant protective
ef ect against stress. To highlight the importance o coworker sup-
■ WORK ROLE AND AUTONOMY port, solo practitioners were ound to be particularly susceptible
to stress. Furthermore, Hof and colleagues show that burnout in
The demand-control model o occupational stress of ers a rame-
hospitalists is more closely associated with less avorable social rela-
work in which control o the work environment mitigates stress
tions involving colleagues and coworkers than negative experiences
created by ongoing work demands. According to the 1999 national
related to the economically induced pressures o the job, such as
survey o hospitalists, 97% o hospitalists reported being “highly
reduced autonomy and the use o nancial incentives. Shana elt and
autonomous” or “autonomous” in their clinical role. Since that time,
colleagues ound that physicians and scientists’ negative relation-
the role o some hospitalists may have evolved in ways that diminish
ship with their leaders is a strong predictor o burnout.
control over their work. This is re ected in the raction o hospitalists
in 2009 who indicated having “much” or “very much” in uence over
the ollowing aspects o their work: quality o work, 83%; order o ■ ORGANIZATIONAL CHARACTERISTICS
tasks to per orm, 69%; pace o work, 50%; schedule, 33%; amount Hospitalist practice models are as diverse as the number o practices
o work, 13%. that exist in the United States. Although, hospitalist groups are uni-
In addition to workload described above, the perceived decline in ed in the goals o containing cost while improving quality o care,
autonomy may be related to growing administrative oversight over the organizational culture, incentive structures, hierarchies, and
hospitalists in the orm o per ormance assessments, volume-based operating rules that govern job demands vary widely. All o these
reimbursements, privacy rules, patient sa ety evaluations, resource actors can in uence individual hospitalists’ t with their organiza-
stewardship, and malpractice litigations. While requirements or tion. Additional negative impact on burnout can sur ace when
billing or service documentation are aimed at eliminating excessive organizations violate basic expectations o airness and equity.
275
Various studies demonstrate how economic pressures adversely HOW CAN BURNOUT IN HOSPITALISTS BE ADDRESSED?
af ect attitudes related to physician career quality and longevity. For
■ FOCUS ON THE JOB AND THE INDIVIDUAL
instance, job satis action is negatively impacted by more restrictive
orms o reimbursement such as capitation, less time available to Considerable evidence supports the act that organizational and
P
job actors play a greater role in burnout than individual worker
A
spend with patients, and the use o nancial incentives related to
characteristics and may be more remediable. It can be said that the
R
productivity. These ndings help to anticipate the potential impact
T
o bundled payment models on hospitalist work-li e. organization acts upon the physician, whose personality, coping
I
The lack o t between a hospitalist and his or her job is another style, early experiences, skills, and competencies lter or exacerbate
determinant o burnout. Hospitalists have been shown to vote with its ef ects. Solutions to the burnout problem, there ore, should be
their eet to nd jobs that align with their personal goals. A study addressed primarily at the level o the job or organization. The goal
by Shana elt and colleagues ound that physicians in an academic should be to balance the demands o the job with the control,
T
reward, and support needs o the worker. This may not be an easy
h
department o internal medicine were less likely to show burnout
e
when they spent more time engaged in the activities—patient care, task; most burnout interventions in the past have ocused on indi-
S
viduals rather than jobs or organizations because it is perceived to
p
research, education, or administration—they ound most meaning-
e
ul. The practice o Hospital Medicine of ers ample opportunities or be easier and less expensive to change people. Nevertheless, given
c
i
the evolving nature o hospitalist programs and their relationships
a
physicians to pursue customized careers in unconventional areas o
l
t
with organizations, job design should be a primary ocus. An expert
y
medicine like management and advocacy. While some hospitalists
o
nd the variety o activities to be satis ying, others risk mismatch in generalist physician work-li e, Mark Linzer, advocates or a process
f
H
when they commit to jobs to which they are not well suited. improvement approach to addressing physician burnout through
o
his “10 Bold Steps to Prevent Burnout in General Internal Medicine”
s
Fairness re ers to the extent to which work decisions are per-
p
(Table 41-2). This requires systematically measuring the known
i
ceived to be just. When organizational policies are air, emotional
t
a
predictors and outcomes o burnout and using them or continu-
l
exhaustion correlates with job per ormance but when they are
M
consistently un air, poor job per ormance is prevalent even where ous eedback. Four o his 10 steps ocus on work conditions: ensure
e
adequate resources, plan rather than react to reductions in FTE
d
employees show no signs o burnout. Hospitalists may perceive
i
c
injustice through dif erential treatment o physician groups by hos- rom predictable li e events, resist thoughtless ef orts to standardize
i
n
processes by allowing physician exibility, and control workload.
e
pitals. Disproportionate distribution o responsibilities or rewards
Additionally, he endorses cultivating “career t” to allow physicians
a
among members o a hospitalist group can breed resentment that
n
to engage in meaning ul career development. These, in addition,
d
can af ect quality o care and burnout status. The earnings disparity
to sel -care through resilience building are advocated to nurture a
S
that disadvantages women in hospital medicine as they do in the
y
s
rest o society is another example o organizational injustice that renewed pro essionalism. Many o these recommendations have
t
e
been incorporated into publicly available online resources of ered
m
persists.
by the STEPS Forward program o the American Medical Association
s
o
■ PERSONAL CHARACTERISTICS (www.steps orward.org). Resources or monitoring burnout and
f
C
Hospitalist jobs usually do not require special training beyond the other work-li e parameters are also available rom the AMA and the
a
Society o Hospital Medicine through its Hospital Medicine Engage-
r
typical hospital-based residency experience. As a result, it is an
e
attractive career opportunity or young physicians who can eel ment Benchmarking Program.
pro cient immediately out o residency training. However, studies
consistently show that younger individuals are more susceptible ■ OPTIMIZE WORK LOAD, FLOW, AND FIT
to burnout compared with more experienced counterparts. The The strongest evidence to reduce burnout supports controlling job
reason or this has not been studied thoroughly but one possible demands: workload and time pressure. But as stated above, simply
explanation is that young physicians are closer to their training
experiences that reward per ectionism and denial o personal vul-
nerability, while being surrounded by peers accustomed to delayed
TABLE 41-2 Ten Steps to Prevent Physician Burnout for
grati cation. Young workers with high expectations o a work envi- Hospital Medicine Practices
ronment and a desire to express their skills and abilities can become
burned out when expectations are unmet. Institutional Metrics
Unmarried workers, especially men, seem to be more prone to 1. Make clinician satis action and wellbeing quality indicators
burnout; and single workers seem to experience higher burnout
2. Incorporate mind ulness and teamwork into practice
levels than those who are divorced. Family can be protective o
stress and burnout, presumably because social support can buf er 3. Decrease stress rom electronic health records
the pathologic in uences o stress ul events. Social support, espe- Work Conditions
cially rom in ormal contacts, is positively related to good health and 4. Allocate needed resources to reduce health care disparities
e cacy, irrespective o the presence or absence o work stressors. 5. Hire physician loats to cover predictable li e events
Individuals with certain psychological dispositions are more 6. Promote physician control o the work environment
prone to burnout. Traits like compulsiveness, guilt, and sel -denial
7. Maintain manageable workload
may con er short-term advantages to physicians, but o ten uel
eelings o inadequacy over the long term making individuals Career Development
with these traits less resilient to external pressures. Those with an 8. Preserve physician “career it”
internal locus o control attribute daily events to personal ef ort 9. Promote lexibility o work schedule
while those with an external locus o control attribute them to Self-Care
outside orces or chance. Not surprisingly, burnout is higher among 10. Make sel -care a part o medical pro essionalism
individuals with an external locus o control. Persons with low sel -
esteem and low levels o resilience or hardiness are associated with Adapted rom Linzer M, Levine R, Meltzer D, Poplau S, Warde C, West CP.
higher burnout due to not being open to change and eelings o 10 Bold steps to prevent burnout in general internal medicine. J Gen Intern
not being in control. Med. 2013;29(1):18-20. With permission o Springer.
276
reducing work hours or patient encounters does not alleviate the asking physicians to better tolerate stress neglects root causes and
consequence o burnout and high demands are inherent to the MBSR should not be considered a su cient work-li e intervention
C
medical pro ession. The Healthy Work Place Study was a cluster on its own.
H
randomized trial o an intervention that was primarily designed to
A
According to sel -determination theory, people are growth-
P
improve work ow in the outpatient setting. Some speci c modi- oriented and predisposed to using their capacities ully to connect
T
cations to existing work ow included of -loading administrative with others and integrate their experiences in a relative unity. The
E
tasks to nonphysician assistants, removing bottlenecks o care at the theory proposes that intrinsic motivations autonomously acilitate
R
bedside, extending visit times, and scheduling standing meetings this growth orientation while extrinsic motivations compel behav-
4
dedicated to addressing management issues. Sites with improved iors or its instrumental value. Individuals are more vulnerable to
1
work ow exhibited odds o improving their physician burnout rate burnout under conditions dominated by extrinsically motivated
by 5.9 (95% con dence interval: 1.4, 24.6) compared to controls. behaviors. Consequently, cultivating the sources o hospitalists’
F
o
Adapting these interventions to the inpatient setting may involve intrinsic motivations may potentially protect them rom burnout.
r
employing administrators to handle nonclinical tasks, making the Pro essional coaching is a strategy used by athletes and business
I
n
d
workload more predictable on a day by day or a week by week i not leaders and recently being adopted by physicians to ampli y internal
i
v
an hourly basis, adjusting patient census, using electronic resources locus o control, sel -e cacy, and sel -determination. By challeng-
i
d
ergonomically, and addressing system aws that restrict exibility ing xed thoughts and circumstances, coaches empower clients to
u
a
or individual hospitalists. reclaim control and innate resilience. Hospitalists can learn coping
l
s
Addressing job t in mismatched areas may be an alternative skills or stress in a number o other ways. Evidence suggests that
a
n
strategy. For example, although early-career hospitalists are known skills to improve relaxation, time management, assertiveness, and
d
to switch jobs to success ully align personal and organizational social skills can be learned. Recruiting resilient individuals with inter-
P
r
a
goals, hospitalists established in a practice have been shown to tai- personal skills and avorable personality pro les may be a strategy
c
t
lor their jobs to their strengths. As a result, some hospitalists diversi y or organizations. Although medical schools and training programs
i
c
e
their responsibilities across clinical, administrative, and academic routinely select based on characteristics like personality traits, an
s
:
domains while others choose to work in settings that allow them exclusionary approach to burnout prevention may sacri ce diver-
C
a
to handle larger than average work load or commensurate control sity in the hospitalist work orce without addressing the underlying
r
e
and rewards. Hospitalists, in general, are likely to tolerate greater systems problems. The most durable and easible interventions may
e
r
workload i they have control over the work they do, value the involve tailoring practice models to individual needs, but evidence
S
u
work they are doing, and eel they are doing something important, to guide such interventions is still sparse. The t o the job to the
s
t
or i they eel well rewarded or their ef orts. Creative scheduling worker is key to job satis action and avoiding burnout.
a
i
n
may decrease perceived workload by introducing necessary time
a
or recovery. This is particularly important given evidence linking
b
CONCLUSION
i
l
burnout and health problems with nonstandard work shi ts (nights
i
t
The work o typical hospitalists is characterized by high work
y
and weekends) in workers o almost every industry. Providing
a
demands, decreased autonomy, and potentially di cult interper-
n
hospitalists with skills to care or nonmedical patients and to orm
d
better relationships with consultants may relieve the demands o sonal interactions that puts them at risk or burnout. The conse-
A
quences o burned-out hospitalists include poor job per ormance,
v
comanagement.
o
dissatis action, turnover, and various health-related problems like
i
d
depression and substance abuse. Evidence to guide ef orts to
i
n
■ CULTIVATE A SUPPORTIVE CULTURE prevent or minimize the impact o burnout is emerging. More
g
B
Hospitalists are also social creatures, thus ocusing on relationships studies examining the work-li e impact o organizational processes
u
are important. Fostering a mutually supportive organizational cul- and material incentives are necessary to guide ongoing health
r
n
ture ocused on common goals, mutual respect, and implementing
o
care re orm. Organizational ef orts to protect work-li e balance
u
systems to acilitate rewarding interpersonal interactions are likely and support vulnerable individuals rom destructive behavior are
t
to minimize burnout. Hospitalists who eel they are working as part worthwhile.
o a cohesive team, supported and appreciated by those around
them (other hospitalists, nurses, consultants, primary care providers,
practice leaders, etc), and who eel positively recognized by patients
SUGGESTED READINGS
and amilies will have less burn out. The most important relation- Bakker AB, Schau eli WB, Sixma HJ, Bosveld W, Van Dierendonck
ships that impact hospitalist work-li e are with patients. Satis action D. Patient demands, lack o reciprocity, and burnout: a ve-year
with one’s ability to deliver the highest quality o care is consistently longitudinal study among general practitioners. J Organ Behav.
the strongest driver o job satis action or hospitalists and other 2000;21:425-441.
specialists, alike. Strengthening the patient-doctor relationship is
Elliott DJ, Young RS, Brice J, Aguiar R, Kolm P. Ef ect o hospitalist
an ef ective strategy to enhance physicians’ personal accomplish-
workload on the quality and e ciency o care. JAMA Intern Med.
ments, as proven by palliative care specialists and pediatricians
2014;174(5):786-793.
whose examples should in orm the practices o adult hospitalists.
These and other interventions may be particularly important in sup- Hinami K, Whelan CT, Miller JA, Wolosin RJ, Wetterneck TB. Job char-
porting younger hospitalists who are more vulnerable to emotional acteristics, satis action, and burnout across hospitalist practice
exhaustion. models. J Hosp Med. 2012;7(5):402-410.
Hof T, Whitcomb WF, Nelson JR. Thriving and surviving in a new
■ RECLAIM RESILIENCE medical career: the case o hospitalist physicians. J Health Soc
At the individual level, numerous studies have ound the bene ts Behav. 2002;43:72-91.
o mind ulness training and other stress reduction interventions Lindenauer PK, Pantilat SZ, Katz PP, Wachter RM. Hospitalists and
to reduce burnout rates. Courses in Mind Body Stress Reduction the practice o inpatient medicine: results o a survey o the
(MBSR) are proli erating as a result. But where burnout is an a ter- National Association o Inpatient Physicians. Ann Intern Med.
math o chronic stress rom modi able work conditions, simply 1999;30:343-349.
277
Linzer M, Levine R, Meltzer D, Poplau S, Warde C, West CP. 10 bold Shana elt TD, Boone S, Tan L, et al. Burnout and satis action with
steps to prevent burnout in general internal medicine. J Gen Intern work-li e balance among US physicians relative to the general US
Med. 2013;29(1):18-20. population. Arch Intern Med. 2012;172(18):1377-1385.
P
Linzer M, Poplau S, Grossman E, et al. A cluster randomized trial o West CP, Huschka MH, Novotny PJ, et al. Association o per-
A
interventions to improve work conditions and clinician burnout in ceived medical errors with resident distress and empathy: a
R
primary care: results rom the Healthy Work Place (HWP) study. J prospective longitudinal study. J Am Med Assoc. 2006;296(9):
T
Gen Intern Med. 2015;30(8):1105-1111. 1071-1078.
I
Maslach C, Leiter MP. Early predictors o job burnout and engage-
ment. J Appl Psychol. 2008;93:498-512.
T
h
e
S
p
e
c
i
a
l
t
y
o
f
H
o
s
p
i
t
a
l
M
e
d
i
c
i
n
e
a
n
d
S
y
s
t
e
m
s
o
f
C
a
r
e
278
PART II
Medical Consultation
SECTION 1 Surgery Prevention, Assessment, and

SECTION 4 Management of Common
42 Physiologic Response to Surgery . . . . . . . . . . . . . . . . . . . . . 283 Complications in Noncardiac Surgery
43 Perioperative Hemostasis . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286
55 Antimicrobial Prophylaxis in Surgery ................. 361
44 Postoperative Complications ........................ 292
56 Venous Thromboembolism (VTE) Prophylaxis
45 Surgical Tubes and Drains . . . . . . . . . . . . . . . . . . . . . . . . . . . 297 for Nonorthopedic Surgery . . . . . . . . . . . . . . . . . . . . . . . . . . 368
46 Surgical Critical Care ................................ 300 57 Postoperative Blood Transfusion . . . . . . . . . . . . . . . . . . . . . 373
58 Nutrition and Metabolic Support . . . . . . . . . . . . . . . . . . . . . 377
SECTION 2 Anesthesia 59 Cardiac Complications after Noncardiac Surgery. . . . . . . 385
47 Anesthesia: Choices and Complications . . . . . . . . . . . . . . . 309 60 Management of Postoperative Pulmonary
Complications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 391
48 Perioperative Pain Management ..................... 313
61 Assessment and Management of Patients with
Renal Disease . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 397
Perioperative Risk Assessment and
SECTION 3 Management 62 Postoperative Neurologic and Psychiatric
Complications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 412
49 Role of the Medical Consultant . . . . . . . . . . . . . . . . . . . . . . . 325
50 Preoperative Cardiac Risk Assessment and Specialty Consultation—What the
Perioperative Management . . . . . . . . . . . . . . . . . . . . . . . . . . 329 SECTION 5 Consulting Hospitalist Needs to Know
51 Perioperative Pulmonary Risk Assessment
and Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 336 63 Surgical Management of Obesity . . . . . . . . . . . . . . . . . . . . . 421
52 Perioperative Risk Assessment and Management 64 Common Postoperative Complications
of the Diabetic Patient . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 342 in Neurosurgery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 430
53 Preoperative Evaluation of Liver Disease . . . . . . . . . . . . . . 348 65 Management of Common Perioperative
Complications in Orthopedic Surgery. . . . . . . . . . . . . . . . . 437
54 Preoperative Assessment of Patients with
Hematologic Disorders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 352 66 Transplant Surgery. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 447
67 Urology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 453
279
SECTION 1
Surgery
281
CHAP TER
42 INTRODUCTION
A Scottish physician and metabolism researcher named Dr David
Cuthbertson rst reported on the metabolic consequences o
surgery in 1932 when he coined the term “ebbs and ows” when
studying the ef ects o lower limb injury. Since that time scienti c
study has added a great deal more to what we know about the
physiological response to surgery, and understanding this response
lends itsel to more ef ective management, including mitigating
risks o the surgical patient.
Physiologic Response This chapter outlines the primary physiologic responses to sur-
gery, and their impact on the management o the postoperative
to Surgery patient, including the typical surgical stress response, ef ects on

uids and electrolytes, and common organ-speci c ef ects.
SURGICAL STRESS RESPONSE

Gil Freitas, MD The surgical stress response has three key physiologic components:
Vihas Patel, MD • Sympathetic nervous system activation
Allan B. Peetz, MD • Endocrinologic activation
• Immunologic activation with the production o cytokines and
acute phase reactants, and resultant neutrophil release and
demargination, and lymphocyte proli eration
■ SYMPATHETIC NERVOUS SYSTEM ACTIVATION

Surgical stress stimulates the hypothalamic-pituitary-adrenal (HPA)
axis and leads to sympathetic nerve activation, which triggers the
release o catecholamines. In addition to their cardiovascular af ects
(tachycardia, hypertension), these hormones also af ect the liver,
kidneys, and pancreas.
■ ENDOCRINOLOGIC ACTIVATION
Surgical stress results in a variety o changes in serum levels o endo-
crinologic hormones (Table 42-1).
These endocrine responses are normal a ter surgery, but can
sometimes result in clinically signi cant complications. For example,
increased cortisol can lead to hyperglycemia, which is associated
with surgical complications and poorer outcomes. Cortisol also
stimulates protein catabolism, whereby both skeletal and visceral
muscle are broken down to release amino acids or energy or to
be used by the liver to orm new protein including the acute phase
reactants. This process can result in weight loss, muscle wasting,
and impaired healing. Arginine vasopressin can result in ree water
retention, which can result in hypervolemia and hyponatremia.
■ IMMUNOLOGIC ACTIVATION
Surgical stress also results in a variety o immunological changes.
Such activation is essential or recovery and wound healing, but can
also have untoward physiologic ef ects (such as ever). White blood
cells and endothelial cells produce interleukins and inter erons,
which contribute to an in ammatory cascade. Interleukin-1 (IL-1),
tumor necrosis actor-α (TNF-α), and IL-6 are the primary cytokines
released a ter surgery. Increased IL-6 production a ter surgery acti-
vates acute phase proteins such as C-reactive proteins (CRP), brino-
gen, and α-2 macroglobulin, which act as in ammatory mediators
and scavengers in tissue repair. CRP levels rise approximately 4 to
12 hours a ter surgery, peak at 24 to 72 hours, and are elevated
or approximately 2 weeks. D-dimer protein, a brin degradation
283
TABLE 42-1 Principal Hormonal Responses to Surgery TABLE 42-3 Intravenous Crystalloid Fluids Commonly Used
in the Perioperative Period
Endocrine Gland Hormone Change in Secretion
P
Type of Fluid Indications Composition per Liter
A
Anterior pituitary ACTH Increases
R
Growth hormone Increases Lactated Typically used 130 mEq Na,
T
Ringer’s (LR) or the irst 48 h 109 mEq Cl, 4 mEq K,
TSH May increase or
postoperatively; 3 mEq Ca,
I
decrease
I
replacement luids 28 mEq lactate,
FSH and LH May increase or pH 6.4, 273 mOsm
decrease
D5 1/ 2NS+20K Hypotonic 50 g dextrose;
Posterior pituitary AVP Increases maintenance luids 77 mEq Na,
M
Adrenal cortex Cortisol Increases used a ter initial 77 mEq Cl,
e
d
Aldosterone Increases luid resuscitation 20 mEq K; pH 5.7,
i
c
completed 452 mOsm
a
Pancreas Insulin O ten decreases
l
Normal saline Alternative to 154 mEq Na,
C
Glucagon Usually small
o
(NS) LR; watch or 154 mEq Cl, pH 5.7,
n
increases hyperchloremic 308 mOsm
s
u
Thyroid Thyroxine Decrease metabolic acidosis
l
t
a
½ NS Hypotonic 77 mEq Na,
t
i
o
maintenance solution 77 mEq Cl, pH 5.7
n
3% NS Used with 513 mEq Na, 513 mEq
product, also increases in the postoperative period, and may remain neurosurgical or Cl, pH 5.7,
elevated or several weeks. This normal in ammatory state response neurology teams to 1027 mOsm
to surgery is thought to be the cause or the mild evers and leuko- treat cerebral swelling
cytosis commonly seen in the rst 48 hours a ter surgery. D5Water Free water; no role in 50 g dextrose,
resuscitation 278 mOsm
FLUIDS AND ELECTROLYTES Human 140 mEq Na, 103 mEq Cl, 4 mEq K, 5 mEq Ca,
plasma 2 mEq Mg, 25 mEq HCO3, pH 7.4, osmolality
■ FLUIDS 290 mOsm
Fluid balance is a key element or perioperative management.
Many surgical patients ast or 12+ hours prior to surgery. This
preoperative ast, in combination with intraoperative blood loss PRACTICE POINT
and evaporative losses, can lead to signi cant uid de cits. Postop- • Perioperative uid replacement:
erative redistribution o uids, or third spacing, may also contribute
to intravascular hypovolemia. Patients in the perioperative period • Intraoperative uid losses are typically isotonic.
should be continuously monitored or volume status, as under- or Lactated Ringer’s (LR) solution is typically the intravenous
overcorrection are both problematic. Signs and symptoms o hyper- crystalloid o choice or uid resuscitation in the rst 48 hours
and hypovolemia are outlined in Table 42-2. a ter surgery.
I the patient is hypovolemic and needs uid resuscitation, crys-
talloids are pre erable to colloids; most surgeons pre er Lactated • Ongoing uid losses are typically hypotonic.
Ringer’s solution or resuscitation. Maintenance luids should D51/ 2NS + 20 mEq/L o K+ is typically the intravenous crystalloid o
be D51/ 2NS + 20 mEq/L K+ at an appropriate rate based on patient choice or uid maintenance, a ter the 48 hours a ter surgery.
weight. The purpose o the dextrose component is it maintains
tonicity, and prevents catabolism, ketosis, and hypoglycemia (a liter
o 5% dextrose solution provides about 170 kcal). Typical IVF com-
position rates are outlined in Table 42-3. ■ ELECTROLYTES
Electrolyte and acid-base imbalances are common in the periop-
erative period, and are exacerbated by being NPO, and by uid-
electrolyte losses rom tubes, drains, or stulas.
TABLE 42-2 Signs and Symptoms of Volume Disturbances
Sodium
System Volume Deficit Volume Excess High-output nasogastric tubes, emesis, or enteric stulas may result
Generalized Weight loss Weight gain in a hyponatremic, hyperchloremic metabolic alkalosis. Leakage
Decreased skin turgor Peripheral edema rom large wounds may also deplete extracellular water and sodium.
Cardiac Tachycardia Increased cardiac ADH secretion rom surgical stress and resuscitation with large
output volumes o LR may also contribute to postoperative hyponatremia.
Orthostasis/ Increased central Potassium
hypotension venous pressure
Hypokalemia is common in the surgical patient, and can be exac-
Collapsed neck veins Distended neck veins
erbated by diarrhea, stula, or vomiting losses. Most patients are
Renal Oliguria Polyuria asymptomatic until plasma K+ alls below 3.5 mEq/L. Symptoms o
Azotemia hypokalemia are primarily related to ailure o normal contractility
GI Ileus Bowel edema o muscle bers, and present as ileus, constipation, atigue, weak-
Pulmonary – Pulmonary edema ness, or cardiac arrest. Potassium replacement therapy is most sa ely
administered orally; intravenous replacement should be reserved or
284
patients unable to tolerate oral administration, and rate o intrave-
nous in usion should not exceed 20 mEq/h. TABLE 42-4 Surgical and Medical Conditions Predisposing to
C
Perioperative Acute Renal Failure
H
Magnesium
A
Trauma Vascular disease
Magnesium levels should also be monitored perioperatively and
P
Cardiopulmonary bypass Hypotension
T
supplemented i low. Hypermagnesemia perioperatively is unusual
E
and is usually spurious. Symptoms o hypomagnesemia include Underlying renal disease Hypovolemia
R
hyperactive re exes, muscle tremors, and tetany. Magnesium should Renal transplantation Liver ailure
4
be aggressively replaced to restore potassium or calcium homeostasis Urologic surgery Sepsis
2
when hypokalemia or hypocalcemia are concomitant with hypomag- Pigment nephropathy Contrast agents
nesemia. Hypomagnesemia can be repleted orally with 50 to 100 mEq,
Drugs (eg, aminoglycosides)
P
but i it results in diarrhea, intravenous replacement is pre erable.
h
y
s
i
Calcium
o
l
o
Calcium is another important electrolyte that requently becomes
g
the symptoms or duration o an ileus and can have side ef ects, so
i
out o balance in the perioperative period and needs to be closely
c
should not be routinely prescribed.
R
monitored. Hypocalcemia is de ned as a serum calcium level
e
s
below 8.5 mEq/L, or an ionized calcium level below 4.2 mg/dL
p
■ RENAL
o
(which is more physiologically important). Symptoms include tet-
n
any, Chvostek’s sign, Trousseau’s sign, laryngospasm (stridor), and Postoperative renal ailure is uncommon but signi cantly increases
s
e
con usion. Transient hypocalcemia commonly occurs a ter removal perioperative morbidity and mortality. Risk actors or postoperative
t
o
o a parathyroid adenoma, due to atrophy o the remaining glands renal ailure are outlined in Table 42-4. Perioperative acute renal
S
and avid bone remineralization, and may require high doses o ailure is o ten initiated by perioperative hemodilution ( rom aggres-
u
r
sive uid resuscitation) and hypovolemia ( rom hemorrhage or inad-
g
supplementation. Symptomatic hypocalcemia is best corrected
e
with intravenous calcium gluconate or calcium chloride. equate volume replacement); this reduces the viscosity o capillary
r
y
blood ow, and redistributes blood rom the renal medulla to the
COMMON ORGAN-SPECIFIC RESPONSES TO SURGERY renal cortex, subjecting the medulla to ischemia. Ischemic kidney
injury can also be incited by the constriction o the af erent and
■ NEUROLOGIC AND PSYCHIATRIC
ef erent arterioles by circulating catecholamines in response to sur-
See Chapter 62 (Neurologic and Psychiatric Complications) and gical stress. Renal tubular injury, or acute tubular necrosis, is another
Chapter 81 (Delirium). common cause o perioperative acute renal ailure. Once tubular
damage occurs, a number o actors contribute to ongoing tubular
■ CARDIOVASCULAR damage, including intraluminal obstruction rom cell swelling and
Preoperatively patients should be assessed or cardiac risk and sloughing, persistent vasoconstriction, back-leakage o luminal uid
monitored or cardiac complications. See Chapter 50 (Preopera- across damaged tubular epithelium, and decreased glomerular cap-
tive Cardiac Risk Assessment and Risk Reduction) and Chapter 59 illary membrane permeability.
(Cardiac Complications). Perioperative management should include close monitoring
o urine output and electrolytes, daily weights, avoidance o all
■ PULMONARY nephrotoxic medications, and appropriate adjustment o all renally
Surgical patients are susceptible to a myriad o postoperative pul- cleared medications. For patients that become oliguric, manage-
monary complications. See Chapter 51 (Perioperative Pulmonary ment includes the placement o a bladder catheter, and an isotonic
Risk Assessment and Management). uid challenge (500 mL o normal saline or Ringer’s lactate). A uri-
nalysis should be evaluated or speci c gravity, casts, or evidence o
■ GASTROINTESTINAL in ection, a ractional excretion o sodium should be calculated, and
Paralytic ileus, the temporary disruption o normal peristalsis, is com- nephrology should be consulted.
mon in surgical patients. The symptoms o ileus may include bloat-
ing, abdominal distention, inability to pass atus and intolerance to CONCLUSION
an oral diet. Although the underlying mechanism or postoperative Surgical stress is associated with predictable metabolic and hor-
ileus is not well understood, catecholamine surges, electrolyte monal changes and af ects nearly every organ system. Understand-
changes, and uid shi ts are likely contributors. Ileus is particularly ing these physiologic changes that result rom surgery are critical or
common a ter gastrointestinal surgery and is potentially related to appropriately managing perioperative patients.
bowel manipulation and inhibiting spinal re ex arcs via adrenore-
ceptors. Other contributors include metabolic abnormalities such as SUGGESTED READINGS
hypokalemia or hypomagnesemia.
Although postoperative ileus is usually sel -limited, it can impede Jamieson RA, Ledigham I, Kay AW, MacKay C. Jamieson and Kay’s
enteric eeding and there ore signi cantly slow normal recovery Textbook of Surgical Physiology, 4th ed. Philadelphia, PA: Churchill
rom surgery. Postoperative ileus is also associated with impaired Livingstone; 1988.
wound healing, prolonged hospital stays, and overall postoperative
Kanani M, Elliott M. Applied Surgical Physiology Vivas. London:
morbidity. Treatment is supportive and includes adequate hydra-
Greenwich-Medical; 2004.
tion, correction o electrolyte imbalances (especially potassium and
magnesium), use o epidurals during and a ter surgery, and limiting Morgan GE, Mikhail MS, Murray MJ. Clinical Anesthesiology (Lange
perioperative opiate use (by using multimodal pain regimens). Early Series), 4th ed. New York: McGraw-Hill; 2002.
ambulation is highly encouraged and can signi cantly improve Townsend CM, Beauchamp D, Evers MB, Mattox KJ, eds. Sabiston
both the symptoms and the duration o the ileus. Promotility medi- Textbook of Surgery: The Biologic Basis of Modern Surgical Practice,
cations such as metoclopramide and erythromycin do not improve 18th ed. Philadelphia, PA: Saunders; 2007.
285
43
Perioperative bleeding is a dreaded surgical complication. More
o ten than not you will hear, “It was dry when we closed…” as
your surgical colleagues struggle to identi y the source. Hospi-
talists are o ten involved in managing patients be ore and a ter
surgery, hence they need a good working knowledge o how to
predict, evaluate, and manage perioperative bleeding. This chap-
ter will ocus on the preoperative evaluation o bleeding risk,
intraoperative risk actors or bleeding, typical presentations o
Perioperative postoperative bleeding, and how a hospitalist needs to evaluate
and manage the bleeding.
Hemostasis
PREOPERATIVE EVALUATION OF BLEEDING RISK
To identi y patients at increased risk o perioperative bleeding com-
Brian K. Yorkgitis, DO plications, inquire about any history o bleeding problems, such as
a known bleeding diathesis, excessive bleeding rom minor trauma,
Allan B. Peetz, MD menorrhagia, gingival bleeding, hemarthoses, excessive bruising,
petechiae, liver, or renal disease. Review medications that can a ect
normal coagulation, such as antiplatelet agents and anticoagulants,
and review the risks and bene ts o stopping these agents with the
surgeon. All patients identi ed with risk actors or bleeding should
have preoperative laboratory evaluations including a complete
blood count, liver unction tests, chemistry panel, prothrombin time
(PT), activated partial prothrombin time (aPTT), and international
normalized ratio (INR). The laboratory test results should be inter-
preted based on the in ormation in Table 43-1.
INTRAOPERATIVE RISK FACTORS FOR BLEEDING

Ideally, hemostasis is achieved be ore the patient leaves the
operating room. In cases o minimally invasive surgery, it may
be di icult or the surgeon to appreciate injury to vascular
structures rom trocar insertion, such as retroperitoneal vascular
structures or super icial vessels such as the epigastric vessels. In
pelvic and retroperitoneal surgery, it can be di icult or the sur-
geon to obtain hemostasis due to loss o anatomical planes and
rich blood supplies.
Patients who receive anticoagulation during surgery, experience
large-volume blood loss (>500 cc) during surgery, or require >10
units o blood have a signi cant risk or postoperative bleeding.
Certain procedures are also associated with higher risk o postop-
erative bleeding, including vascular surgery, cardiac surgery, any
surgery with cardio-pulmonary bypass (CPB), orthopedic surgeries
with muscle and bone bleeding, liver surgeries, oncologic tumor
resections, prostate surgery (due to the location and release o
urokinase), and obstetrical cases (due to the rich vascular supply
to the pelvic organs). Any bleeding associated with surgeries o
the neck, oropharynx and upper respiratory tract can compromise
the patient’s airway and needs immediate evaluation. Emergency
surgery is associated with higher blood loss compared to elective
operations, due to requent coagulopathy and hypothermia that
o ten accompany emergency surgery.
286
T
B
A
R
M
L
d
D
V
t
d
i
D
V
C
D
H
D
n
A
i
i
i
o h
i
b
b
b
e
i
I
l
e
i
o
n
e
I
e
i
e
v
t
I h
r
e
y
n
e
C
r
B
,
n
r
r
r
a
e m
e
p
t
e
e
e
i
i
i
i
i
d
r
X
i
i i
m
b
c
a
n
n
n
c
L
r
c
c
a
c
p
a
l d
a
,
o
b
i
l
t
i
i
i
o
o
o
E
t
p
r e
l
t
e
e
i
I
p
i
a
a
i
p i
n
I
t
o
n
l
g
l
n
y
n
n
a
n
n y
y
h
4
i
t
a
a
(
r
l
,
r g
e
i
,
e
c
e
p
s
t
u
c
c
K
o
r
c
l
r
s
3 a
i
u
D
i
y
n
i
l
o
d
y
y
c
r
t
r
l
s n
e
D
r
-
d
i
,
o
e
o
r
,
m
a
g 1
t
u
e
o
o
a
t
U
e
e
i
t
D r
b
a
h
(
s
a
x
g
h
r r
b
+
F
X
o
t
i
I
a
i
r E
g
g
e
r
i
C
c
r
i
(
H
o
n
/
a
a
t o
r
n
o
s
x
e
n
i
a
–
)
i
d m
c
e
l
h
v
m
p
b
n
i
i
t
a
t a
t
n
g
e
n
e
r
i
v
s n
o
a
t
e
b
b
n
a
b
h
r
h c
i
s
n
(
t
d
d
r
c
n i
o
d
y
i
t
i
t
t
s
a n
b
,
t
P
i
)
,
i
c
, s
m
w
o
e
o
V
)
A
i
A
y
w
o a
t
,
n
d
r
I t
a
o
i
r
c
)
I
V
r
a n
a
o
I
g
d
r
,
t
r
o
L
,
l r
p
s
o
a
I
e
a
K
o a
X
,
a
t
e
r
r r
r
b
r
,
r
i
s n
a
X
n
o
i
o
o
n c
i
a
,
b
a
r r
t
a
, o
a
X
n r
t
I s
o r y
+
w
s
h
s
h
M
N
C R
c
c
e
e
B
/
o
i
A h
h
e
t
–
l
l
C
r m
m
h
i
i
H
s
m
s
s M
u
t
t
A
e
e o
o
a
l
A
t
t
t l
c
c
w
s H
c
c
y
y
e
e i
b
t
t
A
t e
e
l
l
h
l
l
y s
s
s
s ,
,
T yp e
N
N
M
N
N
a
q
a
L
d
t
s
a
d
L
(
N
P
Q
U
U o
e
o
o
b
b
n
l
u
i
e
s
o
o
o
o
o
o
u
s
a
a q
F
w
w
d
u
n
n
a
s
e
r
r
r
r
r
a
f
y
t
i
H u
p
m
m
m
m
m
c
l
a
o
o
e
a
l
i
B
T i
b
e
i
i
t
)
l
l
s
r
r
t
l e
n
l
e
P
a
a
a
a
a
a
e
l
m
m
a
y
s
e
e e
n
n
t
l
l
l
l
l
O
t
t
t
l i
h
r
e
i
i
d
c
a
a
l
v
i v
v
C
a
o
c
d
y
i
l
l
e
u
e
e g
t
w
o i
l
i
e
r
l
o
y
h n
y
u
n
, g
n t
D i
N
N
P
P
P
N
N
P
o
a
a
p
u
r
d
c
o
t
e
V
N
P
P
s a
a
o
o
n
n
o
r
r
r
r
r
I
n
T
u
r
o
o
o
o
o c
I
t
o
o
o
o
o
o
n
p
d
t
e r
*
c
i
p
e
p
r
t
r
r
r
r
o
i
l
l
l
l
l t
d
m
m
m
m
m
o
o
c
co
s
o
r
o
o
o
o
r
r e
s
t
o
u
e
e
t
e
r
a
n
n
n
n
n o
)
i
a
a
a
a
a
s
c
n
d
d
m g
r
a
o
g
g
g
g
g
l
l
l
l
l o
s
i
g
o
u
n
e
e
e
c
e
e p
a
u
m
i
l
t
d
d
d
d
d n
a
t
g
a
l
a i
t
a
o
i
u
b (
d
c
i
n
n
l
o n
l
a
l
t
u
a
a
e t
n
o
c
e n
n
k
r t
t P
N
N
N
N
P
P
(
p
a
a
p
u
r
P
d
c
o
N
p
s
P
a
l
a
a e
o
n
n
a
P
r
r
r
r
r
n
u
r
r
r
o
o
o
o
o
c
t
o
o
o
o
o o
o
o
v
c
n
d
t
T
c
i
p
e
t
r
r
r
r
r o
i
e
l
l
l
l
l
k
c
t
l
m
m
m
m
m
o
o
c
s
T
o
o
o
o
o o
r
e
r
s
t
o
u
e o
r
a
o
n
n
n
n
n
e
o
n
)
i
a
a
a
a
a
s a
n
d
m
g
a
g
g
g
g
g
g
l
l
l
l
l g
s
i
g
u
e
e
e
c
e
e
i
e p
u
u
i
l
t
d
d
d
d
d
d n
a
t
l
m
a
l a
i
t
a
o
b i
i
n
n
i
o n
l
l
t
e
d t
n , N
P
N
N
P
P
T
T
r
r
r o
o
o
o
o
o r
r
r
l
l
l
m
m
m o
o
o n
n
n
a
a
a g
g
g
l
l
l e
e
e d
d
d N
R
e
o p
r
m t i
l a
a l
s e a
i
i
L
L
F
n
n
i
i
i o
o
b
b
b
b
i
b
c
c w
w
n
r
r
r
r
r
r
i
i
i e
e
n
n
n
o
i
n d
d
a
a
o
o
o
r
m o
s
s
u
u
g
l
l
e
e
y
y g
e
e
e
a
s
s
d
d e
i
i
n
l
t
t
s
s n
o
o
,
, *
D
* I
C P
C
T
V
P
P i
r
r
r
a
l m
o
o
o
o
r
i l
l
l
t
a
o
o
o
e t
b i
n
n
n
n l
g
g
g
e g
e
e
e d
d
d a
R
C
D
a
o
a
F
O
c
p
p
s
o
p
a
d
p
S
B
T
i
p P
a
t
d
n
F
k
r
b
a
l
t
e
e
r
r
a
o
u
o
e a
h
o
o
i
t
P
e
T
r
t
m
s
t
r
p
t
n
m e
z
e
h
n
e
i
i
h
i
e
e
i
l
t
c
T
c
p
n t
r
o
i
h
d
l
e
s
n
i
c
w
t r
h
o
e
t
i
y
m
n
h
o
c
n
i
e
r o
r
t
i
h
e
g
s
,
d
e
l
n
u o
e
i
a
a
o
g
o
n
g
s
e
a
i
n m
r
g
p
o
m
n
a
y
r
m
s
e
t
n
c
e
u
c n
s
t
r
r
e
p
o
i
n d
o
,
g
q
o
a
a
g
m
e
t
b
c ,
a
s
h
a
o
o
t n
m
e
u
r
s
s
o
e
u
i
c
u
n
i
r o
r
i
n
r
o
d
d
i
s
h
c
t
r r
o
r
s
t
l
n
d
r
o
s
g
n e
s
/
c
i
i
a
e
d
e
I
p
b o
i
u
r
r
N
h
d
e
m
g
t e
c
c
s
a
o
i
o
i
e
r
t
r
y
d
o
R e
t
o
i
r
y
a
i
s
n i
i
t
m
i
l
,
t
c
o
m ,
t
h
d
V
n
s
w
b
i
s b
o
i
i
c
t
n
y
o
o
r
e
i d
(
r
r
m
t
s
l
&
i
o
n
C
o a
e
t
n
n
s
t K
h
m
a
d
e
on
u
a
e
o
c
,
r n
a
s
d
r
m
d
y
a
n
b
o u
d
t
n
i
i
n o
i
m
a
m
n
n d
s
g
u e
i
a
s e
, d )
3 4 R E T P A H C
s
i s a t s o m e H e
v i t a
r e p o i r e
P
287
n o i t a t l u s n o C l a c i d e M I I T R A P
288
T B
V
L
I
X
I
a
d
n
n
A
a u
m
W
l
e
I
c
h
h
,
e B
p t
o
y
e
i
i
i
D o
b
b
c
u
L
r
l d
o
i
r
i
i
s
E
e X
t
t
s
i
i o
o
n
n
d
a
I
4
V
I r
r
n
g
c
o
s
s
3 y
t
o
s D
-
i
o
o
i
co
r 1
s
i X
s
a
P a
o
,
a
E
s T
g
p
c
s
r x
,
d o
u
t
r
p o
i
e
c
i
l
m b
a
e
r
i
r a
s
n
r
c
a i
t
t
t
V
n r
e
e y
I
o
d I
d
I g
, I
e
L
a
X n
a
c , t
,
b o
o r r
X a t o r y C R B
e C
s u l t s b y T yp e
N
N
N
N
e
s
w
V
l
a
w
P o
o s
x
W
l
o
o
o
o
o i
i
w
a
s
m
c
t
t r
r
r
r
o
f
h
h
t
D
e m
m
m
m
e
i
e
c
B
p t
S
l
i a
a
a
a
e
y
l
a
t
i
p
L
e n
l
l
l
l
p
t
t E
i
,
a
e
i
n e
o
o
C
t
d i
r
n
o
e
2 i n
n
u
V t
g n
s t
D i
N
P
N
N
P
s o
r
T o
o
o
o r
*
r
r
r
l d
m
m
m
o e
n a
a
a
r
g l
l
l (
e
C d o n t i
n u e P
P
P
P
i
n
F
a
a
d i
c
a
b
P
r
r
r
r
o )
l
o
o
o
o
t
c o
T
r
r
i
i
l
l
l
l m
t
v
w
T
n
o
o
o
o o
i o
n
n
n
n
t
a
r y
g
g
g
g
g
l V
>
i
e
e
e
e
e
I I
n
d
d
d
d
4
I 0
; % N
T T
o r
m a l R e p t i
l a s e I
<
F i
3
b i
0
b r
%
i r
n i
n o o
g g e
e nn * * C
T i
l
m o t
e t i
n g C o m m e n t
a
t
*
* P
M
*
P F
A
T
o
i ,
T
b r
r
, e
r
e
i
a n
c
s
c o
o
e
t i
m
g
n
v s
e
a b
i
n
t
t e
i
i
n v
h
d
e a
a n
p
s
t t
o
a t
r
t o
d
t
i s
i
a
e s
a
l
u i
l
t
c
l
e h i
<
-
e r
t
1
o y
n 0
p
m
c 0
i
e
e b
m
s
p o l
o
p g
a s
l
/
a m
c
d
s
o L
t
i i
m
n
n b
o m
e
t
g i
m o
e
o r
n
n
e
e ;
a
p P
D
c
a
T t
I
t
C i
h
v
p ,
a
w r
d
o t a
o
i
l
s o
y
r
s
; n
e
T
a g
m
P
c s
O
t
i
, n
o
e
, a
r
v
t
s t
h
e ,
e r
n
d
o i
b m
l
i o
r
n
i n
w
b
t r
o
o
a
e g
p
v r e
a
o
b s
n
i
e
e
cu t
t
o h
i
l
n r
a
a
e ;
r
n T
a
c
T a
P
o
, P
T
a
t T
T
h
g
T r
u
p
.
o l
r o
m
o p
l
b
o a
o
n t h
p
g y
l
s
a ;
. s
d
t i
i
r n
e ct
t i
m a e
ct ; U
o F
r H X ,
a u i
n
n h
r i
a
b c
i
t t
i o
o r
n
s ,
a r
t
i e
v d
a r
h
o e
x p
a b
a r
a
i n
n ,
p a
; p
V i
W
x a
D b
, a
v n
o , n
e d W
o i
x
l l
a
e b
b a
r n
a n
; M
d ’
A
s H
d i
A
s e , m
a s
i
e cr
. o a n g i
o p a t h i
c h e m o l
y t i
c a n e m i
a ; P T
, p r
o t h r
o m b i
n t i
m e ;
PRACTICE POINT No matter the potential source, the operating surgeon should be
noti ed immediately about any change in the patient’s condition or
C
Hospitalists should know a patient’s intraoperative actors that
possibility o postoperative bleeding.
H
increase the risk o postoperative bleeding, including:
A
• EVALUATION AND MANAGEMENT OF PERIOPERATIVE
P
Type o surgery
BLEEDING
T
•
E
Use o anticoagulants ■ EVALUATION
R
• Blood loss >500 cc In evaluating a patient with suspected perioperative bleeding, the
4
3
• >10 units o blood administration hospitalist should quickly review the history and physical exam,
including the operative note. Key points to evaluate in the operative
Any bleeding with surgeries o the neck, oropharynx, or upper note should include type o procedure, use intraoperative antico-
P
respiratory tract can compromise the airway and needs immediate
e
agulants, volume o blood loss and/or blood products administered,
r
i
evaluation.
o
and i the patient had any hypotension or reduced urine output dur-
p
ing the surgery (Table 43-3). Figure 43-1 provides management
e
r
a
assistance when postoperative bleeding is encountered.
t
PRESENTATION OF PERIOPERATIVE BLEEDING
i
The physical exam should ocus on vital signs (hypotension,
v
e
Nonsurgical bleeding—commonly called “medical bleeding”— tachycardia, hypoxia), any external signs o bleeding (including
H
usually results rom dys unction in coagulation and clot ormation.
e
rom any surgical drains), or any internal signs o bleeding (such
m
The presentation is typically slow oozing, particularly capillary bleed- as abdominal ullness or tenderness). The evaluation o the patient
o
ing rom traumatized sur aces. This o ten presents a ter many hours
s
should be rapid, and resuscitation should begin simultaneous to
t
a
or days postoperatively, although more rapid blood loss can occur the evaluation.
s
i
s
(depending on the type and degree o the dys unction).
Surgical bleeding usually results rom a blood vessel that the sur-
geon was unable to adequately identi y and/or control. This o ten PRACTICE POINT
presents within hours postoperatively and is more brisk and robust Postoperative bleeding should be suspected in any postoperative
than medical bleeding. Physical exam may reveal signs o hypovole- patient with hypotension, low urine output, tachycardia, or other
mic shock due to loss o circulating volume (Table 43-2). signs o shock.
Patients with drains in the surgical bed may have bloody output.
I the bleeding is intra-abdominal, the patient may have increasing • Surgical bleeding is usually rom a blood vessel that the
abdominal girth or a tense/tender abdomen. A tense abdomen with surgeon was unable to adequately identi y and/or control and
impaired renal unction and ventilation may be signs o abdominal commonly occurs in the early postoperative period.
compartment syndrome. Patients with a retroperitoneal hematoma • Medical bleeding is usually more insidious is associated
may have weakness o hip exion and knee extension due to a with coagulation dys unction due to medications, medical
emoral neuropathy. conditions, or large volume resuscitations with resultant
Extremity bleeding may be external (ie, on the dressing, oor, coagulation actor depletion.
or in the bed) or may present with tense compartments or symp- • Resuscitation should occur simultaneously with the
toms o compartment syndrome. Cardiac and thoracic patients evaluation o suspected bleeding to prevent hemodynamic
may present with bleeding into a chest tube, or signs-symptoms collapse.
o a hemothorax (shortness o breath, cough, hypoxia, diminished
breath sounds). Retroperitoneal bleeding poses a special problem
because its presentation is o ten nonspeci c—sometimes the only ■ MEDICAL MANAGEMENT
sign is a decreasing hemoglobin/hematocrit without any obvi- Resuscitation o the patient should begin simultaneous with the
ous source. Because retroperitoneal bleeding can be an insidious evaluation. The patient needs secure IV access (with two large bore
source o signi cant hemorrhage, a high index o suspicion should IVs or central venous access) and circulatory volume should be
be maintained especially when the patient has recently undergone maintained with isotonic IV uid or blood products. To accurately
aortic, kidney, or prostate surgery. Likewise, obese patients’ bleed- monitor uid resuscitation, a Foley catheter should be inserted.
ing may be subtle, since there is more subcutaneous tissue that can The patient should be made nil per os (NPO) in the event there is
allow blood to accumulate, and the abdominal examination can be a need to return to the operating room or undergo other invasive
compromised by their body habitus. Di use oozing and petechiae procedures, and the patient should be considered or the need or
may be signs o disseminated intravascular coagulopathy (DIC). ICU monitoring.
TABLE 43-2 Classifications of Hemorrhagic Shock and Associated Findings
Class I II III IV
Blood loss (mL) <750 750-150 1500-2000 >2000
Blood loss (% body volume) <15 15-30 30-40 >40
Pulse rate (per min) <100 100-120 120-140 >140
Blood pressure Normal Normal Decreased Decreased
Pulse pressure Normal or increased Decreased Decreased Decreased
Respiratory rate (per min) 14-20 20-30 30-40 >40
Urine output (mL/h) >30 15-30 5-15 <5
Mental status Slightly anxious Mildly anxious Anxious, con used Con used, lethargic
289
The degree o resuscitation in the bleeding patient depends on the
TABLE 43-3 Procedure-Related Risk Factors for Perioperative suspected pace o the bleeding and degree o end organ e ect. Ther-
Bleeding apy should be guided by physiologic endpoints such as heart rate,
P
blood pressure, urine output, oxygen saturation, end-organ per u-
Intraoperative blood loss >500 cc
A
sion, and electrocardiography. More advanced monitoring including
R
Blood products administered >10 units echocardiography and invasive hemodynamic measurements may
T
Procedure type be needed i the patient is not responding or is clinically unstable.
I
I
• Trauma surgery Hypothermia can exacerbate and complicate surgical bleeding due
• Emergency surgery to the enzymatic reactions o coagulation inhibition by low tempera-
• Vascular, cardiac surgery, cardiopulmonary bypass tures. The patient should be made normothermic through the use o
warm IV uids, warming blankets, or warming devices. Additionally,
• Obstetrical
M
any electrolytes should be corrected, particularly calcium and acidosis,
e
• Orthopedic surgery
d
to maintain an optimal milieu or coagulation reactions to occur.
i
c
• Liver surgery Additional interventions should be undertaken based on the
a
l
• Prostate results o coagulation studies (Table 43-1), and blood products
C
o
• Tonsillectomy should be administered in conjunction with clinical assessment and
n
s
based on the American Society o Anesthesia (ASA) guidelines or
u
Perioperative medications
l
perioperative bleeding (Table 43-4).
t
a
• Antiplatelet agents
Patients with qualitative platelet dys unction (who have been
t
i
o
• Anticoagulants taking antiplatelet agents or have uremia) may require desmopres-
n
Hypothermia (T < 35°C) sin, which can partially reverse the platelet dys unction. Tranexamic
acid (TXA) is an agent with some utility in ameliorating perioperative
Ge ne ral Me as ure s
Monitor urine output
Review ope ra tive re port
Ma inta in te mpe ra ture >35*C
Review me dica tions (pa rticula rly
a ntipla te le t/a nticoa gula tion)
Pos t-ope ra tive ble e ding Obta in La bora tory Te s ts :
Type a nd cros s -ma tch
He moglobin/he ma tocrit/pla te le ts
Ele ctrolyte s
Re na l/he pa tic/coa gula tion function
Fibrinoge n (if DIC s us pe cte d)
He modyna mic S ta tus A, B, C’s, Vita l S igns
EKG a nd bioma rke rs
His tory from prima ry nurs e
(if ca rdia c is che mia s us pe cte d)
Uns table S table

In a ddition to “STABLE” inte rve ntions, Ens ure IV a cce s s
a ls o do the following: Atte mpt to loca te s ource of ble e ding
Reve rs e coa gulopa thy (if pre s e nt) Be ing volume re s us cita tion
S ta rt va s opre s s ors for hypote ns ion Tra ns fus e blood products (if ne e de d)
re fra ctory to volume re s us cita tion Hold a ge nts a ffe cting coa gula tion
Pre pa re for pos s ible re turn to OR or (ris k/be ne fit mus t be eva lua te d)
e mboliza tion (if fe a s ible ) Notify s urge on
Vis ible ble e ding Co ag ulo pathic ble e ding No n-vis ible ble e ding
Control loca l ble e ding (be ds ide or OR) Re pla ce de ficie nt fa ctors (if ide ntifie d) S e ria l exa mina tions
Apply pre s s ure S pe cific reve rs a l a ge nts : Eva lua te for compa rtme nt syndrome
Topica l he mos ta tic a ge nt/ca ute ry DDAVP for ure mia /pla te le t dis orde rs Abdomina l
S uture liga tion Prota mine for He pa rin Extre mity
S pe cific reve rs a l a ge nt or P CC for Ima ging s tudie s
NOAC Che s t ra diogra ph
Vita min K/FFP /P CC for VKA’s CT s ca n
Cryopre cipita te for DIC Angiogra phy
Wo rs e ning e xam o r c o nc e rning imag ing finding s :

Ope ra tive explora tion or e mboliza tion (if fe a s ible )
Figure 43 1 Approach to managing postoperative bleeding.
290
pressure, or with bedside cautery. Gauze applied to the bleeding
TABLE 43-4 American Society of Anesthesiologists (ASA) wound with constant, continuous pressure maintained or 10 minutes
C
Guidelines for Perioperative Transfusion in is usually e ective. A ter 10 minutes, pressure can be released and
H
Excessive Bleeding
A
the wound reassessed—i bleeding has not stopped then pressure
P
Laboratory Value Laboratory Value should be reapplied and the surgeon should be noti ed.
T
Indicates need for Exceeds Indication for Deep surgical bleeding requires prompt evaluation by a surgeon
E
Therapy Transfusion Transfusion so that exploration and de nitive control o the culprit blood vessel
R
Packed red Hemoglobin Hemoglobin ≥10 g/dL can be established. Invasive interventions are likely required in cases
4
o postoperative hemodynamic instability, hemothorax, hemoperi-
3
blood cells ≤6 g/dL
Platelets ≤50,000 cells/mm 3* ≥100,000 cells/mm 3 toneum and compartment syndromes.
P
Fresh rozen INR ≥ 2.0* Normal values
CONCLUSION
e
plasma
r
i
o
Cryoprecipitate Fibrinogen Fibrinogen ≥150 mg/dL Postoperative bleeding is a serious surgical complication. Preop-
p
erative evaluation can help determine which patients present the
e
≤80-100 mg/dL
r
a
highest risk o su ering this complication. Bleeding should be
t
i
v
Central nervous system surgeries: platelets <100,000 and INR >1.5.
* considered in any postoperative patients with tachycardia, hypo-
e
tension, oliguria, or evidence o shock. Postoperative bleeding may
H
e
be medical, surgical, or both. Physical exam and laboratory tests
m
bleeding; however, it is associated with the risk o thrombosis, and should guide a ocused management o the problem, and resusci-
o
s
its use should be decided upon in conjunction with the surgeon. tation should be initiated immediately. The surgical team should be
t
a
For patients who have received intraoperative un ractionated involved in the work-up and decision making as soon as postopera-
s
i
s
heparin (UFH) administration within 6 to 12 hours o the bleeding, tive bleeding is suspected.
it should be reversed with the use o protamine, and the activated
clotting time (ACT) can guide the appropriate dosage. Protamine SUGGESTED READINGS
can also reverse the e ects o low-molecular weight heparins, but it
is less e ective than it is or UFH. Bougle A, Harrois A, Duranteay J. Resuscitative strategies in trau-
Coagulopathy rom war arin should be reversed with resh rozen matic hemorrhagic shock. Ann Intensive Care. 2013;3(1):1.
plasma and/or vitamin K. Prothrombin concentrate complex (PCC) Hammond KL, Margolin DA. Surgical hemorrhage, damage control,
contains vitamin K-dependent clotting actors and can be used or and the abdominal compartment syndrome. Clin Colon Rectal
urgent correction o bleeding associated with acquired coagulation Surg. 2006;19:188-194.
actor de ciency. Recombinant actor VII can also be e ective in
ameliorating bleeding associated with acquired coagulation actor Hunt BJ. Bleeding and coagulopathies in critical care. N Engl J Med.
de ciency; however, it is very expensive and is associated with a risk 2014;370:847-859.
o thrombosis. Its use should be decided upon in conjunction with Ker K, Prieto-Merino D, Roberts I. Systematic review, meta-analysis
the surgeon. and meta-regression on the e ect o tranexamic acid on surgical
The newer oral direct actor inhibitors include actor Xa inhibitors blood loss. Br J Surg. 2013;100:1271-1279.
and direct thrombin inhibitors. They pose challenges due to lack o Levy JH, Faraoni D, Spring JL. Managing new oral anticoagulants in
routine laboratory testing to monitor their e ect on coagulation, the perioperative and intensive care unit setting. Anesthesiology.
and lack o speci c reversal agents; as such, these drugs should be 2013;118:1466-1474.
stopped 3 to 5 days prior to major surgery. I bleeding does occur
Practice Guidelines or Perioperative Blood Management. An
on these agents, traditional measures to control bleeding locally
updated report by the American Society o Anesthesiologists
should be employed such as direct pressure, topical hemostatic
Task Force on perioperative blood management. Anesthesiology.
agents, cuatery, or embolization. Due to their short hal -lives, stop-
2015;122:241-300.
ping the agent and providing expectant support is usually suf cient
in minor bleeding. In cases o more severe hemorrhage; speci c or Phillips LE, Zatta AJ, Schembri NL, Noone AK, Isbister J. Uncontrolled
non-speci c reversal agent should be employed to abate bleeding. bleeding in surgical patients: the role o recombinant activated
See Chapter 78 “Bleeding and Coagulopathy” and Chapter 57, actor VIIa. Curr Drug Targets. 2009;10(8):744-770.
“Postoperative Blood Trans usion.” Shah A, Stanworth SJ, McKenchnie S. Evidence and triggers or
the trans usion o blood and blood products. Anaesthesia.
■ SURGICAL MANAGEMENT 2015;70(Suppl 1):10-19.
Visible bleeding rom a surgical wound itsel may represent a small Spinella PC, Holcomb JB. Resuscitation and trans usion principles
“skin bleeder” that can be managed with a stitch, by holding direct or traumatic hemorrhagic shock. Blood Rev. 2009;23(6):231-240.
291
44
Postoperative complications are common and costly. Recent stud-
ies suggest that, on average, each avoidable surgical complication
costs payers >$10,000. Their incidence, risk actors, and impact on
patient outcomes are as varied as the eld o surgery itsel . Surgical
site in ections (SSIs) alone a ect >500,000 patients annually and are
associated a 2 to 11 times increase in the risk o postoperative mor-
tality. Good communication among all providers caring or surgical
patients is undamental to both the prevention and management
Postoperative o surgical complications. Hospitalists caring or surgical patients
should, accordingly, understand what surgical procedure was per-
Complications ormed, the indication or the procedure, and any perioperative

concerns rom the operating surgeon, based on the circumstances
o that particular patient or procedure. This chapter will review some
o the more common postoperative complications that a hospital-
Peter Najjar, MD ist needs to recognize and manage. Each section will review the
risk actors o the complication, how to mitigate those risks in the
Allan B. Peetz, MD perioperative period, and how the hospitalist should identi y and
manage such complications.
Elective procedures provide more o an opportunity to identi y
and mitigate risk actors be ore surgery, although an attempt
should take place be ore any surgery (even urgent or emergent).
A thorough history and physical examination should aim to iden-
ti y risks outlined in this chapter, and optimize the risk-bene t
pro le o the procedure. For example, when deciding whether to
hold antiplatelet or anticoagulation medications in patients with
cardiac indications, the risk o cardiac complications is weighed
against the risk o bleeding. This decision making should occur in
concert with the surgeon, and with the patient, to ensure all par-
ties have a common understanding o the risks and the bene ts
o the surgery.
PRACTICE POINT
• Prevention o postoperative complications begins in the
preoperative period.
• Surgical complications are common and costly; interdisciplinary
teams must work together on their prevention and
management.
COMMON COMPLICATIONS WITHIN THE POST

ANESTHESIA CARE UNIT (PACU)
A ter surgery, patients usually stay in a postanesthesia care unit
(PACU) or close monitoring while they recover rom the e ects o
anesthesia. Common problems managed in the PACU include post-
operative pain, hyper- and hypotension, respiratory insu ciency,
and nausea and vomiting.
■ POSTOPERATIVE PAIN
Patients in this phase o care are o ten unable to verbalize pain due
to e ects o anesthesia; accordingly, pain assessments are o ten
based on other objective assessments such as blood pressure, heart
rate, respiratory rate and signs o agitation. See Chapter 48 (Periop-
erative Pain Management).
292
■ POSTOPERATIVE HYPER- AND HYPOTENSION PRACTICE POINT
C
Pain and elevated catecholamines can contribute to hypertension Low urine output in the immediate postoperative setting
H
and tachycardia. β-blockers should be continued in the periopera-
•
A
tive setting or patients who took them preoperatively. Intravascular volume depletion may occur concurrently with
P
Hypertension in the PACU is most commonly caused by pain pulmonary edema due to increased vascular permeability
T
and/or a history o hypertension. Certain procedures, such as carotid associated with perioperative inf ammation; administration o
E
diuretics or postoperative pulmonary edema can exacerbate
R
endarterectomy, require immediate and aggressive control o sys-
tolic blood pressure regardless o etiology to avoid catastrophic intravascular depletion, hypotension, and inadequate end-
4
organ per usion.
4
vascular, cardiac, or neurologic complications. Invasive monitoring
with an arterial line may be necessary or such patients. Clear com- • Postoperative oliguria (less than the equivalent o 0.5 cc/kg/h)
munication with the surgical team regarding target blood pressure requires urgent evaluation.
P
o
is essential and hemodynamic agents including vasodilators and
s
t
negative chrono- and inotropes are requently used as rst-line
o
p
agents in these settings. Absent an indication or strict hemody- ■ POSTOPERATIVE NAUSEA AND VOMITING
e
r
namic control, the patient should be assessed or pain and treated
a
Postoperative nausea and vomiting (PONV) is also common in the
t
i
appropriately be ore treating blood pressure directly. For patients
v
PACU, the causes o which are multi actorial. Prior history is the
e
with pre-existing hypertension requiring medication, it is generally
C
most signi cant risk actor; other risk actors include longer dura-
most appropriate to gradually reintroduce the preoperative antihy-
o
tion procedures, use o volatile anesthetics (such as isof urane), and
m
pertensive regimen with the exception o diuretics in the immediate
procedures involving the inner ear, eye, and abdominal viscera.
p
postoperative period.
l
i
Patients at moderate to high risk o PONVbene t rom prophylactic
c
Hypotension in the PACU is usually due to hypovolemia, narcotic
a
antiemetics, motility agents, or a scopolamine patch be ore emerg-
t
i
and benzodiazepine administration, or epidural anesthesia; postop-
o
ing rom anesthesia.
n
erative bleeding must also be considered. Markers or hypovolemia
s
include low urine output, signs o shock, and altered mental status,
which can be masked by the residual e ects o anesthesia. Invasive POSTOPERATIVE FEVER
monitoring with a urinary catheter, central line, or arterial line should Low-grade evers in the rst 48 hours a ter surgery are a normal
be utilized i a patient remains hypotensive despite initial resuscita- sequelae o inf ammation, atelectasis, or hematoma absorption
tion with crystalloid. Epidural anesthesia can cause hypotension ollowing surgery, and usually not rom an in ectious process. In
by blunting sympathetic tone and decreasing vascular resistance. the absence o any localizing signs or symptoms, sel -limited ever
In the absence o hypovolemia, however, epidurals do not usually within the rst 48 hours postoperatively usually does not need
cause hypotension. Treatment o epidural-related hypotension in ectious work-up. A ter 48 hours, temperatures greater than 38.5°C
should include administration o a f uid bolus; temporarily hold- should prompt a complete ever workup. In the postoperative
ing the anesthetic in usion can also be help ul until euvolemia is patient, the surgical wound and site o venous access are potential
obtained. sources o in ection and need to be care ully examined. See Chapter
206 (Undiagnosed Fever in Hospitalized Patients).
■ POSTOPERATIVE RESPIRATORY INSUFFICIENCY
Although most patients will require some supplemental oxygen SURGICAL SITE INFECTIONS
immediately a ter surgery, dyspnea, tachypnea, wheezing, and Surgical site in ections (SSIs) account or approximately 30% o nos-
signs o respiratory distress are not normal postoperative signs ocomial in ections and are the most common in ections a ter sur-
and symptoms, and need to be addressed in the PACU. The gery. They are associated with a 7-day increased length o stay and
causes and degree o risk o postoperative respiratory insu i- cost $400 or a super cial SSI and $30,000 or organ space SSIs.
ciency are complex and patient speci ic, but all patients recover- SSIs are classi ed as super cial, deep, or organ space in ections.
ing rom anesthesia require close monitoring o their respiratory Super cial in ections are wound in ections involving the skin and
status, with personnel and equipment or reintubation readily subcutaneous tissues. Deep in ections involve the ascia or muscle
available. The primary actors that contribute to postoperative below. Organ space in ections involve organs below the cutaneous
respiratory insu iciency include use o general anesthesia, upper and muscular layers (Figure 44-1).
abdominal and thoracic surgeries, longer duration surgeries, use
o endotracheal intubation, and use o narcotics. Signi icant pain
also puts patients at increased risk, as pain impairs respiratory PRACTICE POINT
unction by limiting vital capacity and can result in hypoxia and Wound infections
dyspnea.
Pulmonary edema is a common etiology or respiratory distress • Despite the most rigorous aseptic technique, all wounds are
in the postoperative period, o ten secondary to the e ects o f uid contaminated to some degree and have some risk o in ection.
shi ts and sometimes overload rom intraoperative resuscitation. Even “clean”wounds have a 1.5% risk o in ection (Table 44-1).
Initial evaluation o pulmonary edema should include physical • Wound in ections commonly occur between 5 and 10 days
exam and chest x-ray; some patients may need urther work-up or a ter an operation.
evaluation by a cardiologist. Pneumothorax should be considered • Antibiotics are not necessary or simple wounds that have been
in patients with a central line recently placed, and should also be drained.
evaluated with a Chest x-ray. For patients with a postoperative chest • Deep space in ections usually require drainage; antibiotics
tube (usually cardiac or thoracic surgery patients) a poorly unction- alone are insu cient.
ing chest tube or residual pneumothorax can also cause or exacer-
bate respiratory distress. This can be detected by physical exam and
con rmed with a chest x-ray ollowed by prompt contact with the Risk actors or wound in ection are patient and operation
operating surgeon to relay the concern. dependent. Patient related risk actors include large body habitus,
293
Deep space in ections occur in enclosed spaces with some
degree o isolation rom blood supply, making them relatively
impervious to antibiotics. Such in ections usually require drainage
P
either percutaneously or in the operating room. Anastomotic leaks
A
S kin S upe rficia l typically occur between postoperative days 5 and 7 and should be
R
incis iona l suspected in surgical patients with tachycardia, abdominal pain,
T
SSI ever, and elevated white count. This constellation o symptoms
I
S ubcuta ne ous
I
tis s ue a ter an enteric anastomosis should prompt a CT scan. These leaks
can o ten be managed with percutaneous drainage, but inability to
control the in ection may require operative drainage.
De e p s oft tis s ue De e p incis iona l
M
(fa s cia & mus cle ) SSI
e
WOUND DEHISCENCE, HEMATOMAS, SEROMAS
d
i
c
Wounds typically heal to a maximum o 80% o the tensile strength
a
l
Orga n/S pa ce
within 6 weeks among healthy, well-nourished patients. Because
C
Orga n/S pa ce SSI
o
o this, most surgeons restrict postoperative activities to avoid stress
n
s
on the wound or 4 to 6 weeks. Wounds that have been closed
u
l
primarily should be kept clean, dry, and covered or a minimum o
t
a
t
Figure 44 1 Categories of surgical site infections. 48 hours a ter surgery. Dry, sterile operative dressings may be kept
i
o
in place until postoperative day 2; therea ter, patients can usually
n
shower without submerging the wound.
diabetes, disability, immunosuppression, malnutrition, and smoking. Wound dehiscence is disruption o any layer o the surgical wound.
Certain operations, such as those involving the colon or small bowel, This rare complication results rom increased pressure on the wound
are higher risk than others. Other operative risk actors include and can arise or a variety o reasons. Suspected dehiscence should
operating room conditions, surgical technique (eg, laparoscopic be promptly evaluated by a surgeon and may require a return to
or open), administration o antibiotic prophylaxis, and hypoxia or the operating room. Poor wound healing o ten leads to dehiscence.
hypotension during the procedure. Malnutrition, liver disease, diabetes, immunosuppression, and chronic
Prophylactic antibiotics are very e ective at reducing the risk steroid use inhibit normal wound healing and are risk actors.
o SSIs; they should be administered within 1 hour o incision and The most common layers involved are the skin and ascia. Sud-
continued or no more than 24 hours a ter surgery. In the event o den output o serosanguinous f uid rom the wound is usually the
signi cant contamination in the OR, wounds may be le t open and rst sign o dehiscence. There ore, daily evaluation by multiple team
managed with delayed primary closure or wet to dry dressings. members (including surgeons, hospitalists, and nurses) may help
The hallmarks o a wound in ection are ever, pain, tenderness, or identi y subtle changes that may be harbingers o dehiscence.
purulent drainage. The typical presentation is between 5 and 10 days The management o postoperative wound dehiscence depends
postoperatively. Some in ections can present earlier; clostridial nec- on the size and location o the wound as well as the patient’s con-
rotizing wound in ection should be suspected when a patient has a dition. Fascial dehiscence—separation o the deepest layer o the
very high ever in the immediate postoperative period; these require abdominal wall—typically requires urgent closure in the operating
immediate surgical evaluation and drainage. room. In the most severe o cases, dehiscence leads to extrusion o
intra-abdominal contents (eg, evisceration). Evisceration is a surgical
emergency that requires immediate return to operating room.
TABLE 44-1 Surgical Wound Classification Hematomas are more common, and can be caused either by
Infection inadequate hemostasis during surgery or disruption o hemostasis
Wound Classification Risk Procedure Type postoperatively; risk actors or hematomas include bleeding disor-
ders and anticoagulant use. A hematoma can result in wound eleva-
Clean 1.5% Vascular surgery
tion, pressure, pain, dehiscence, and in ection. The surgeon should
No in ection or in lammation always be alerted i there is a suspected hematoma; depending on
Respiratory, gastrointestinal, biliary, the size and location, treatment can vary rom watch ul waiting to
and urinary tracts not entered re-exploration in the OR. Hematomas ollowing neck exploration
Clean contaminated 10% Appendectomy may rapidly compromise the airway in the postoperative period. Pre-
Entry into respiratory, biliary, cipitating actors include abrupt increases in intrathoracic pressure
gastrointestinal, urinary tracts with rom coughing, emesis, or Valsalva maneuvers. Treatment involves
minimal spillage No evidence o emergent evacuation o the hematoma prior to reintubation.
in ection or major break in aseptic Seromas are collections o serous f uid that orm a ter procedures
technique involving disrupted lymphatic f ow and raised skin f aps. They are
Contaminated 20% Foreign body in generally the result o a normal physiologic response to anatomic
In lammation, gross spillage rom a wound dead space, and their incidence is dependent on the anatomic loca-
GI tract, break in technique tion o the wound. Examples o procedures commonly associated
Fresh traumatic wound with seromas include inguinal hernia repair, groin exploration, and
Dirty or infected 40% Abscess, mastectomy. Suction drains may be le t in place at the end o the
per orated procedure to increase tissue apposition and remove f uid. Compres-
Purulent drainage, ecal
contamination, per orated viscous sion dressings can also reduce the risk o seroma ormation. Seromas
viscous, delayed or contaminated may increase the risk or wound disruption and in ection but are
traumatic wound, presence o usually nothing more than a nuisance. Management may be expect-
devitalized tissue ant or include serial aspirations. Rarely, return to the OR is indicated
to ligate contributing lymphatics.
294
PULMONARY AND CARDIAC COMPLICATIONS general anesthesia, which results in prolonged immobility, hyperco-
agulability, and endothelial damage. Patients with known hyperco-
C
See Chapters 60 (Pulmonary Complications) and 59 (Cardiac
aguable states, prior VTE, and malignancy are at especially high risk.
H
Complications).
High-risk surgical procedures include orthopedic surgery, trauma,
A
P
and neurosurgical treatment o head injury and brain tumors. Pro-
GASTROINTESTINAL COMPLICATIONS
T
phylaxis starts be ore surgery with the application o pneumatic
E
Postoperative ileus is expected a ter gastrointestinal surgery but is compression devices and subcutaneous heparin 2 hours prior to
R
also common a ter other procedures, usually due to narcotic use anesthetic induction. In the absence o procedure- or patient-
4
and/or immobility. Ileus usually presents with abdominal disten- speci c concerns, pharmaceutical prophylaxis should not be held
4
tion, nausea, belching, and inability to pass f atus. Resolution o the preoperatively. Unless there are clear contraindications, such as
symptoms usually occurs within 5 days o surgery but can be longer, increased bleeding risk, patients should receive pharmacologic
P
particularly in debilitated patients or those with signi cant narcotic prophylaxis and pneumatic boots throughout and perioperative
o
s
use. Classically, an ileus on plain abdominal x-ray is associated with period. See Chapter 56 (VTE Prophylaxis or Patients Requiring Non-
t
o
uni orm distribution o air throughout the bowel, but this nding
p
Orthopedic Surgery).
e
is not speci c. Abdominal CT scan with enteric contrast has close Pulmonary embolus (PE) still causes considerable mortality in
r
a
to 100% sensitivity and speci city in distinguishing postoperative
t
hospitalized patients. PE should be suspected in all surgical patients
i
v
obstruction rom ileus.
e
presenting with symptoms o dyspnea, tachycardia, and hypoxemia.
Colonic pseudo-obstruction is a rare postoperative complication,
C
The decision to start anticoagulation should be made with the
o
and most likely occurs in older patients with prolonged immobility,
m
operating surgeon, while pending urther diagnostic testing. See
preoperative institutionalization and unctional limitation. It more
p
Chapter 253 (Diagnosis and Treatment o VTE).
l
i
commonly occurs a ter orthopedic surgeries than other types o
c
a
surgeries. Patients present with signs o an ileus, and radiographs
t
URINARY TRACT INFECTIONS (UTI)
i
o
demonstrate a decompressed small bowel and uni ormly dilated
n
Urinary tract in ections (UTI) are most common a ter vaginal or uro-
s
colon. Treatment involves hydration, bowel rest, and decompres-
sion (distally with rectal tube placement and proximally with naso- logic surgery and any surgery with the use o indwelling catheters.
gastric tube placement i there is evidence o gastric distension). Women and obese patients are at highest risk. The most common
In re ractory cases, neostigmine is an option but can be associ- pathogens are Escherichia coli, Staphylococcus saprophyticus, and
ated with bradycardia. I the cecum is dilated >11 cm, additional Proteus mirabilis. However, hospitalized and immunosuppressed
measures should be undertaken to avoid per oration, such as a patients are also susceptible to Klebsiella, Proteus vulgaris, Candida
cecostomy tube or a cecal resection. General surgeons should be albicans, and Pseudomonas. The standard or prevention is the
engaged early in the management o suspected postoperative removal o indwelling catheters within 48 hours o insertion. The
colonic pseudo-obstruction. need or continued urinary catheterization should be assessed at
least daily to prevent needless prolongation o catheter placement
DELIRIUM and increased risk o catheter-associated UTI. Criteria or diagnosis
and treatment are the same as or UTIs in the nonpostoperative
Postoperative delirium is most common in patients with advanced
period. See Chapter 197 (Urinary Tract In ection and Pyelonephritis).
age, a history o sensitivity to narcotics or anesthetics, heavy alcohol
use, dementia or prior delirium. Delirium can occur a ter any surgical
POSTOPERATIVE URINARY RETENTION
procedure, but is most common a ter cardiovascular and thoracic
surgery. The causes are multi actorial, including a myriad o medi- Postoperative is common but rarely prolonged. Common risk ac-
cations, loss o environmental cues, insomnia, and recovery rom tors or postoperative urinary retention include male sex, prostatic
cardiopulmonary bypass. enlargement, epidural/spinal/prolonged anesthesia, use o antihis-
Delirium presents with waxing and waning severity with periods o tamines or narcotics, and pelvic/perineal procedures. An overdis-
lucidity, recurrent con usion, agitation, hallucination, and con abula- tended bladder (>500 mL) and disruption o the neural pathways
tions. These patients may present a risk to themselves and others by that control voiding impairs urinary contraction and micturition.
trying to get out o bed without assistance, or trying to remove tubes Prophylactic catheterization in the operating room is recommended
and drains. Prevention is the best strategy and entails maintaining a or any procedure lasting more than 3 hours, or when interruption
regular sleep-wake cycle, restoring environmental cues (such as day- o the sacral plexus is anticipated (eg, abdominoperineal resection).
light and clocks), and removing extraneous environmental stressors I a catheter is not present, patients should be encouraged to void
(minimizing bright lights and loud noises). Some patients will respond soon a ter the procedure. I the patient has not voided or more
well to reassurance and emotional support. I at risk or sel -harm, low- than 6 hours, it is appropriate to evaluate retention with a bedside
dose antipsychotics are rst line agents. It is best to start at low doses ultrasound; alternatively, an in-out catheter may be used to deter-
and titrate up as needed, as these medications can be sedating, may mine the extent o retention. The treatment or bladder distention is
worsen delirium at higher doses, and may inter ere with pulmonary intermittent catheterization along with mitigation o any contribut-
unction and physical therapy. These pharmacologic treatments may ing actors. Some patients may have prolonged urinary retention in
provide symptom control but do not treat the underlying causes the postoperative period (>48 hours). In this scenario, appropriate
o delirium and at times can exacerbate delirium. When managing pharmacologic treatment should be initiated and an indwelling
patients with postoperative delirium it is also important to rule out Foley catheter should be placed. Some may require subsequent
other organic disease triggers such as sepsis, stroke, or metabolic outpatient urologic ollow-up or a void trial a ter discharge. See
derangements. See Chapter 81 (Delirium). Chapter 67 (Urology).
DEEP VENOUS THROMBOSIS AND PULMONARY CONCLUSION

EMBOLUS Postoperative complications are associated with signi cant morbid-
Venous thromboembolism (VTE) is a leading cause o preventable ity, prolonged length o stay, and hospital costs. Hospitalists should
death in the postoperative setting. Surgical patients are at high risk know how to identi y and manage the most common postop-
or VTE due to the surgical procedure itsel as well as induction o erative complications. Comprehensive assessment preoperatively,
295
attention to surgical technique and anesthetic management, and Güldner A, Pelosi P, De abreu MG. Nonventilatory strategies to pre-
a multidisciplinary approach to medication management, in ec- vent postoperative pulmonary complications. Curr Opin Anaesthe-
tion control, and patient mobility can reduce the risk o surgical siol. 2013;26(2):141-151.
P
complications. Mattei P, Rombeau JL. Review o the pathophysiology and manage-
A
ment o postoperative ileus. World J Surg. 2006;30(8):1382-1391.
R
SUGGESTED READINGS
T
Najjar PA, Smink DS. Prophylactic antibiotics and prevention o surgi-
I
cal site in ections. Surg Clin North Am. 2015;95(2):269-283.
I
Arozullah AM, Khuri SF, Henderson WG, et al. Development and
validation o a Multi actorial risk index or predicting postopera-
tive pneumonia a ter major noncardiac surgery. Ann Intern Med.
2001;135:847-857.
M
e
d
i
c
a
l
C
o
n
s
u
l
t
a
t
i
o
n
296
CHAP TER
45 INTRODUCTION
Surgical drains are used to monitor or postoperative leaks or
abscesses, collect normal physiologic uid, or to minimize dead space.
Table 45-1 lists various types o drains, with their indications or use.
This chapter will review the most common types o surgical drains,
and the basic care o these drains rom a hospitalist perspective.
PRACTICE POINT
Surgical Tubes and • Hospitalists should know the location and purpose o all
surgical drains in their patients, but should not manipulate
Drains these drains without input rom the surgeon who placed them.
CHEST TUBES
Alexandra Columbus, MD Chest tubes are placed in the pleural space to evacuate air or uid.
Joaquim M. Havens, MD They can be as thin as 20 French or as thick as 40 French ( or adults).
Chest tubes are typically placed between the ourth and th inter-
Allan B. Peetz, MD costal spaces in the anterior axillary or mid-axillary line, however,
location may vary according to the indication or placement. The
tubes can be straight or angled.
The tubes are connected to a collecting system with a three-way
chamber. The water chamber holds a column o water which pre-
vents air rom being sucked into the pleural space with inhalation.
The suction chamber can be attached to continuous wall suction to
remove air or uid, or it can be placed on “water seal” with no active
suction mechanism. The third chamber is the collection chamber
or uid drainage.
Indications or a chest tube include pneumothorax, hemothorax,
or a persistent or large pleural e usion. Pneumothorax and hemo-
thorax usually require immediate chest tube placement. Chest tubes
are also commonly placed at the end o thoracic surgeries, to allow
or appropriate re-expansion o the lung tissue.
A chest x-ray should be obtained a ter any chest tube insertion to
ensure appropriate location. Chest tubes are equipped with a radi-
opaque line along the longitudinal axis, which should be visible on
x-ray. Respiratory variation in the uid in the collecting tube, called
“tidling,” should also be seen in a correctly placed chest tube, and
should be monitored at the bedside to reassure continued appropri-
ate location. The interventional radiologist or surgeon who placed
the tube should determine the subsequent requency o serial chest
x-rays required to monitor the location o the chest tube.
I the patient has a pneumothorax, air bubbles will be visible in the
water chamber (called an “air leak”) which is o ten more apparent when
the patient coughs. The chest tube should initially be set to continu-
ous suction at –20 mm Hg to evacuate the air. Once the “air leak” has
stopped, the chest tube should be placed on water seal to con rm the
pneumothorax is resolved (water seal mimics normal physiology). I the
pneumothorax is not resolved, the chest tube should be placed back
on continuous suction. A chest x-ray should be obtained anytime the
chest tube is changed rom suction to water seal or vice versa.
I the patient experiences ongoing or worsening pain, ever, or
inadequate drainage, a chest computed tomographic (CT) scan may
be warranted to identi y inappropriate positioning or other compli-
cations, such as occlusion or e usion, o the tube. Chest tubes may
become clogged by blood or other debris; the surgical team may
be able to evacuate the tube with suction tubing at the bedside.
I unsuccess ul, the tube may need to be removed and reinserted.
297
TABLE 45-1 Surgical Tubes and Drains
Type Location Clinical Indication Clinical Scenario

P
A
Chest tube Pleural space Pneumothorax Trauma, cardiac surgery, thoracic surgery,
R
Mediastinal space Hemothorax malignant e usion, empyema
T
Pleural e usion
I
I
Nasogastric tube Stomach Intestinal decompression, Small bowel obstruction, ileus, temporary
gastric eeding dysphagia
Gastric tube Stomach Prolonged enteral access, Prolonged mechanical ventilation, gastric outlet
(gastrostomy) gastric decompression obstruction
M
Jejunal tube Jejunum Prolonged postgastric eeding, Prolonged mechanical ventilation, malignant
e
d
(jejunostomy) gastric outlet obstruction, high gastric outlet obstruction, recurrent aspiration
i
c
aspiration risk pneumonia
a
l
C
Duodenal tube Duodenum Postgastric eeding, gastric outlet Mechanical ventilation, dysphagia, acute
o
obstruction, high aspiration risk aspiration risk
n
s
Penrose drain Peritoneal space, Used to maintain surgical tract or
u
l
t
small surgical adequate drainage
a
t
space
i
o
n
Closed suction drain Surgical space Evacuate serous luid or blood, Mastectomy, ventral hernia repair, plastic
(eg, Jackson Pratt, prevent seroma ormation, tissue surgery laps, gastrointestinal anastomoses,
Hemovac) apposition to improve wound healing, orthopedic surgery
drain GI secretions
The team that placed the tube should help the hospitalist deter- • Plastic surgery to prevent seroma ormation and to promote
mine the timing o the chest tube removal. I the patient has a tissue apposition
pleural e usion, the chest tube can usually be removed when the • Cholecystectomy to drain bile
output is less than 100 to 200 mL per day, and the lung is expanded. • Inadvertent postoperative leakage ollowing a dif cult rectal
The tube should usually be taken o suction and placed on water anastomosis
seal (to rule out pneumothorax) prior to tube removal. • Post pancreatic surgery
Typically, closed suction drains will be le t in place until the drain-
age is less than 20 mL per day. These drains can be le t in or weeks
PRACTICE POINT
i necessary and will o ten be removed upon the patient’s sched-
Chest tubes uled surgical ollow-up. Rare complications include erosion into
surrounding tissues and inadvertent suturing o the drain in place
• Contact the team who placed the chest tube i it needs
such that re-exploration is required to remove it. I a closed suction
adjustment or removal.
drain becomes occluded, contact the team who placed the drain or
• Never advance a chest tube into the pleural space, due to the
urther recommendations on adjustment, replacement, or removal.
risk o introducing in ection.
• Do not clamp a tube when pneumothorax is suspected, due to
the risk o precipitating a tension pneumothorax. PRACTICE POINT
Penrose and closed suction drains
• Always check with the surgeon who placed the drain be ore
PENROSE DRAINS readjustment or removal.
Penrose drains are o ten used to drain uid or to keep a space open • Noti y the surgeon i a patient has bloody drainage and/or a
or drainage. Surgeons may anchor penrose drains to skin with alling hematocrit.
sutures. Common indications include:
• Ventral hernia repair
• Debridement o in ected pancreatitis NASOGASTRIC AND DUODENAL TUBES
• Drainage o super cial abscess cavities
Nasogastric tubes (NGTs) are o ten used in the nonoperative
Penrose drains are simple, exible tubes open at both ends; in management o small bowel obstruction or ileus. NGTs should be
contrast to closed drains, they permit ingress as well as egress, acili- placed in the most dependent portion o the gastric lumen, and
tating colonization. con rmed by chest or abdominal x-ray. NGTs are sump pumps and
have a double lumen, which includes an air port to assure ow. The
air port should be patent or optimal unctioning. The tube may be
CLOSED SUCTION DRAINS
connected to continuous wall suction or intermittent suction, set to
Closed suction drains with a plastic bulb attachment (ie, Jackson- low (<60 mm Hg) to avoid mucosal avulsion.
Pratt, Blake, Hemovac) are used to collect uid rom a postoperative NGT output should decrease during the resolution o obstruc-
cavity. Common indications include: tion or ileus, and symptoms o nausea, vomiting, and abdominal
• Postmastectomy to drain subcutaneous uid distention should concomitantly improve. Persistently high output
• Abdominal surgery in a patient with other indicators o bowel unction (eg, atus) may
298
suggest postpyloric placement (and placement should be checked Jejunostomy tubes are used exclusively or eeding and are usu-
by an x-ray). A gastric uid loss can cause electrolyte disturbances, a ally placed 10 to 20 cm distal to the ligament o Treitz. These tubes
C
daily electrolyte panel should be checked and repleted as needed. are indicated in patients who require distal eedings, due to gastric
H
The timing o NGT removal depends on resumption o bowel unc-
A
dys unction or ollowing a surgery in which a proximal anastamosis
P
tion. NGTs may also be help ul in gastric lavage and in diagnosing requires time to heal. These tubes are more apt to clog and can be
T
the source o gastrointestinal bleeding. Bloody output indicates an more dif cult to manage because the lumen o the small bowel is
E
upper bleed, proximal to the ligament o Treitz, whereas clear or smaller than the stomach. Some pre er not to put pills down the
R
bilious output suggests a lower gastrointestinal bleed. tube to mitigate this risk. Routine ushes (30 mL every 4-6 hours)
4
Duodenal tubes are small-bore tubes used when postpyloric with water or saline are also help ul in mitigating the risk o clog-
5
eeding is desired. Like NGT’s, small-bore duodenal tubes are placed ging. In the event that they do get clogged, similar to gastrostomy
through the nares. They are very narrow caliber and require a long tubes, carbonated liquids, meat tenderizer, or enzymes can help
S
u
wire or insertion. The wire should be removed as soon as placement dissolve the obstruction.
r
g
is con rmed by x-ray. They are very so t and exible, but the wire Percutaneous tube sites should be examined requently or
i
c
used or placement is very sti , increasing the risk o inadvertent signs o in ection. Though gastrostomy and jejunostomy tubes
a
l
insertion into the airway. There ore, placement by an experienced are typically well secured intra-abdominally, it is possible or them
T
u
operator is particularly important. In patients who are intubated or to become dislodged. I a gastrostomy or jejunostomy tube has
b
e
who have undergone tracheostomy placement, nasoenteric eed- been in place or more than 2 weeks, it can easily be replaced at
s
ing tubes should be placed under bronchoscopic or uoroscopic the bedside with a tube o comparable caliber by a member o the
a
n
guidance to ensure that the tube is properly positioned. surgical team or by an experienced hospitalist. I the tube has been
d
in place less than 2 weeks, it requires replacement with radiographic
D
r
guidance, as the risk o creating a alse lumen is high. Over time,
a
PRACTICE POINT
i
n
tubes can become loose and all out. I they need replacement, the
s
Nasogastric tube removal in the management of abdominal preceding guidelines apply.
ileus or obstruction
• The NGT can likely be removed i the patient is having atus
and/or bowel movements. PRACTICE POINT
• Place the NGT to gravity and check the aspirate a ter 4 hours; Gastrostomy and jejunostomy tubes
NGT output less than 100 cc indicates passage o gastric
contents through the GI tract. • Noti y the surgeon i a gastrostomy or jejunostomy tube
• Recurrence o nausea, abdominal distention, or pain prior to becomes dislodged.
the 4 hours indicates ongoing ileus or obstruction, and NGT • Correct positioning o these tubes should be con rmed
suction should continue. radiographically be ore use.
GASTROSTOMY AND JEJUNOSTOMY TUBES

CONCLUSION
Gastrostomy tubes are most commonly used or eeding but may
also be used or decompression o unctional or anatomic gastric There are several types o tubes and drains used or a variety o indi-
outlet obstruction. They are indicated when patients need pro- cations in surgical patients. Hospitalists should know the location
longed enteral access (such as prolonged mechanical ventilation or and purpose o all surgical drains, but should not manipulate these
head and neck pathology that prohibits oral eeding). They are also drains without input rom the team who placed them. Drain output
rarely used or gastropexy, to tack an atonic or patulous stomach to should be closely monitored; i unexpected changes in the quantity
the abdominal wall or to prevent recurrence o paraesophageal her- or quality o the output are observed, the team who placed the tube
nias. These tubes can be placed percutaneously by interventional or drain should be noti ed immediately.
radiologists, endoscopically by surgeons and gastroenterologists,
or via laparoscopy or laparotomy by surgeons. This last option is
o ten reserved or patients with dif cult anatomy or who are having SUGGESTED READINGS
laparotomy or another reason.
Brunicandi FC, Andersen DK, Billiar TR, et al. Schwartz’s Principles of
Because o the stomach’s generous lumen, gastrostomy tubes
Surgery, 10th ed. New York: McGraw-Hill; 2014.
rarely clog. In the event that they do get clogged, carbonated liq-
uids, meat tenderizer, or enzymes may help dissolve the obstruction. Sugarbaker D, Bueno R, Krasna M, et al. Adult Chest Surgery. New York:
I a gastrostomy tube is le t to drainage, this can result in signi cant McGraw Hill; 2009.
uid and electrolyte losses; a daily electrolyte panel should be Wiley W, Souba MP, Fink GJ, et al. ACS Surgery: Principles and Practice.
checked and repleted as needed. WebMD; 2006.
299
CHAP TER
46 INTRODUCTION
Caring or critically ill patients began through recognizing the
unique needs o the acutely injured and postoperative patient. In
the 1850s during the Crimean War, Florence Nightingale placed
the most seriously ill patients in beds near the nursing station. This
stressed the importance o a separate geographic location or criti-
cally ill and injured patients. Dr Walter E. Dandy, in 1923, at the Johns
Hopkins Hospital created a three bed postoperative unit or neuro-
surgical patients and sta ed the unit with specially trained nurses
Surgical Critical Care to manage and monitor these patients. The Second World War
brought about the creation o specialized shock units to provide
resuscitation or the large number o critically injured soldiers. The
1950s experienced the widespread development o shock units and
Evert A. Eriksson, MD, FACS, FCCP postoperative recovery units. In addition, respiratory units were cre-
ated due to the large number o polio patients requiring mechani-
Alicia Privette, MD
cal ventilation. In 1986, the American Board o Medical Specialties
Brent Jewett, MD approved certi cation in Critical Care or the our primary boards:
Samir M. Fakhry, MD, FACS anesthesiology, internal medicine, pediatrics, and surgery.
Surgical Critical Care is a core competency o surgical training
and relates to the care o patients with acute, li e-threatening or
potentially li e-threatening surgical conditions. Surgical Critical Care
brings together the art o critical care management o severely ill
patients with the science o surgical procedures targeted at improv-
ing their altered physiology. These surgeons are well versed in the
pre and postoperative management o patients a ter undergoing
surgical procedures rom any surgical discipline and o any age
group. While much o the knowledge base is shared with other criti-
cal care specialists, ellowship training provides the surgical critical
care specialist with speci c expertise relating to the interactions
between the patients disease process and the pathophysiologic
response to in ections, in ammation, ischemia, trauma, burns, and
operations. Given the rising rate o Hospitalist comanagement o
surgical patients, this chapter will cover the most common surgical
critical care patient types, and the management o the most com-
mon surgical conditions encountered among these patients or the
Hospitalist.
SURGICAL CRITICAL CARE ADMIT TYPES

■ SURGICAL ICU ADMITS
Patients may be admitted to the ICU rom the Emergency Depart-
ment, preoperatively, immediately postoperatively, or postopera-
tively a ter initial admission to the postanesthesia care unit or the
ward. Preoperative admission may be required or resuscitation in
the event o preoperative respiratory ailure, shock, or sepsis. ICU
admission may also be required or patients who need invasive
monitoring or hemodynamic optimization prior to undergoing
surgical procedures.
Postoperatively, patients may be admitted to the ICU or respi-
ratory ailure, hemodynamic instability, or close monitoring or
complications such as bleeding or other physiologic derange-
ments. Patients may also require admission to the ICU due to
exacerbation o underlying comorbidities or a ter procedures with
signi cant blood loss or massive uid shi ts. Some patients need
prolonged mechanical ventilation due to the e ects o general
anesthesia, airway edema, dys unctional pulmonary mechanics,
acute lung injury, traumatic injury to the respiratory tract, cardiovas-
cular disease, or volume overload. Patients with preoperative renal,
hepatic, or pulmonary insuf ciency may also bene t rom elective
300
postoperative ICU admission, as they may be more susceptible to history o impaired mentation, and/ or con nement in a burning
the adverse e ects o narcotics and less likely to tolerate hemody- environment, explosion with burns to head and torso, and carboxy-
C
namic uctuations. hemoglobin level greater than 10%. Stridor or circum erential burns
H
A
to the neck are absolute indications or endotracheal intubation.
P
■ TRAUMA ICU ADMITS Breathing and oxygenation issues arise rom several possible
T
injuries a ter burns. Direct thermal injury results in upper airway
In the United States, injury is the most common cause o death
E
edema and obstruction. Inhalation o combustion products and
R
up to the age o 44 years and the ourth most common cause o
toxic umes can cause a chemical tracheobronchitis, edema, and
4
death overall based on data rom the Centers or Disease Control
predispose patients to develop pneumonia. Carbon monoxide
6
and Prevention. Three million people were hospitalized or injury
(CO) poisoning should be considered in any burn patient rom an
in 2013 and many o these patients required admission to a critical
enclosed area. Patients’ with CO levels o less than 20% usually do
care unit. While most severely injured trauma patients are managed
S
u
not exhibit symptoms. CO levels o 20% to 30% result in headache
at Trauma Centers with comprehensive resources, some will receive
r
g
and nausea, 30% to 40% con usion, 40% to 60% coma, and more
at least part o their care at hospitals that are not trauma centers.
i
c
than 60% death. The classically described cherry-red skin is rare.
a
In most trauma systems, severely injured patients are identi ed in
l
CO displaces the oxyhemoglobin dissociation curve to the le t and
C
the eld and taken to a Level 1 trauma center because there is an
r
as a result hemoglobin has an increased af nity or hemoglobin.
i
t
approximately 25% improvement in survival at such centers. Once
i
c
The hal -li e or dissociation is approximately 4 hours at room air
a
a trauma patient arrives at a trauma center, they undergo a highly
l
but decreases to 40 minutes or patients breathing 100% oxygen.
C
standardized resuscitation process using the Advanced Trauma Li e
High- ow oxygen should be initiated immediately i CO poisoning
a
Support (ATLS) program.
r
e
is suspected. Inhalational injuries o ten require bronchoscopy i
The modern day care o trauma patients requently involves
intubation is needed. An adequately sized endotracheal tube should
critical care admission. Trauma patients may require ICU admission
be used to allow subsequent bronchoscopy. I intubation is delayed
or many reasons such as observation and management or severe
until respiratory distress begins intubation is o ten not possible and
and/or multiple trauma, management o acute respiratory ailure,
a surgical airway must be established.
resuscitation o hemorrhagic shock, medical management o trau-
Blood pressure may be dif cult to measure in patients with burn
matic brain injury, and pain control. Patients with multiple injuries
injuries and endovascular volume is o ten dif cult to assess. O ten
requently require ICU care or both management o their injuries
patients are hypovolemic and require large volumes o crystalloid to
and control o pain, agitation, and delirium. Management o speci c
normalize hemodynamic parameters. For patients with more than
injuries requires experience and training in trauma care. When a
10% BSA burns, uid requirements or initial resuscitation can be
trauma surgeon is not available, a general surgeon can provide
estimated using the Parkland ormula. The Parkland ormula is cal-
needed expertise as the principles o trauma care are part o general
culated by multiplying our times the patient’s weight in kilograms
surgery training.
times the percent total body sur ace area involved in second and
The management o pain and agitation is important or all ICU
third degree burns. This volume o uid should be administered
patients but especially injured patients. Consensus guidelines or
over the rst 24 hours a ter the burn. Hal o the total volume should
the management o pain, agitation, and delirium are available and
be administered in the rst 8 hours a ter the burn. The remaining
apply to trauma patients with little i any modi cation (see sug-
hal should be administered over the ollowing 16 hours. The goal
gested readings). Because they have high requirements or pain
urinary output is 1 mL/kg/h or adult patients. In the experience o
medication with its attendant risks o respiratory depression and
the authors, this ormula e ectively approximates the uid needs or
altered sensorium, trauma patients requiring moderate to large
patients with burns between 10 and 40% o total body sur ace area.
doses o analgesia and/or sedation are best managed in an ICU set-
Patients with burns larger than 40% o ten need additional uid to
ting to optimize outcomes and minimize complications.
maintain adequate urine output and hourly urine production should
Since traumatic brain injury (TBI) accounts or one third o injury
be care ully monitored.
deaths, TBI is a requent indication or ICU admission. In addition,
In evaluating the degree o injury, the amount o body sur ace
nearly all trauma patients undergoing acute neurosurgical interven-
injured must be evaluated. First degree burns are erythematous skin
tion will require ICU admission or postoperative care. In general,
that is moderately pain ul to touch without blistering. Partial thick-
TBI patients need admission to the ICU i they require intubation
ness burns, or second degree burns consist o blistered or broken
and mechanical ventilation, i they need requent (more than every
skin that has a red appearance, is very pain ul and has wet weeping
4 hours) neurologic examination and i they have severe TBI requir-
sur aces. Full thickness burns, or third degree burns are o ten white
ing modern-day care. The principles o severe TBI management are
or pale and leathery in consistency without pain to pinprick evalu-
best outlined in the consensus guidelines developed by the Brain
ation and are dry on the sur ace. As a rule o thumb, the area o the
Trauma Foundation. In addition to emphasizing adequate oxygen-
palm o the hand with the ngers extended represents 1% body
ation and brain per usion, the guidelines describe the role o many
sur ace area or each person. This can be used as a gauge to deter-
ICU interventions in the care o patients with severe TBI.
mine the total burn area. For calculating initial resuscitation uids,
only the sur ace o second and third degree burns are used or the
■ BURN ICU ADMITS calculation. Some clinicians nd Figure 46-1 use ul in estimating
As with any critically injured patient assessment o burn patients total body sur ace area burned help ul in evaluating these patients.
begins with the ABCs (airway, breathing, and circulation). Airway In adults, a reasonable system or calculating the percentage o
compromise must be considered in any burn patient. Signs o body sur ace burned is the “rule o nines”: Each arm equals 9%, the
airway compromise may not be immediately obvious but may head equals 9%, the anterior and posterior trunk each equal 18%,
progress rapidly. Supportive measures or potential airway injury and each leg equals 18%; the sum o these percentages is 99%. 1%
should be initiated and early intubation should be considered i is made up o the perineum.
signs o airway compromise are present. Clinical signs o potential Patients should receive adequate supplementary oxygen and
inhalational injury include: ace and/or neck burns, singeing o the endovascular volume replacement. Pain control is extremely impor-
eyebrows and nasal vibrissae, carbon deposits, and acute in amma- tant in these patients as well. Second degree burns are extremely
tory changes in the oropharynx, carbonaceous sputum, hoarseness, sensitive to air- ow so patients should be covered with clean
301
A A
1
P
1 A A
A
R
13 13 1
T
2 2 1
2 2
I
I
1 1 /2 1 1 /2 2 13 2 2 2
1 1 13
1 /2 1 /2
2 1 /2 2 1 /2 1 1 1 1
1 1 1
1 /4
B B 1 1 /4
M
1 /4 2 2
1 1 /4 1
1 /4 1 1 /4
1
1 1 /4 1 1 /4
e
B B B B
d
B B
i
c
a
l
C C C C
C
C C
o
C C
n
1 1 13/4
s
13/4
u
l
t
a
t
i
1 3 /4 1 3 /4 13/4 13/4
o
n
Re la tive pe rce nta ge s of a re a s a ffe cte d by growth Re la tive pe rce nta ge s of a re a s a ffe cte d by growth
Ag e Ag e
Are a 10 15 Adult Are a 0 1 5
A = ha lf of he a d 5 1 /2 4 1 /2 3 1 /2 A = ha lf of he a d 9 1 /2 8 1 /2 6 1 /2
B = ha lf of one thigh 4 1 /4 4 1 /2 4 3 /4 B = ha lf of one thigh 2 3 /4 3 1 /4 4
C = ha lf of one le g 3 3 1 /4 3 1 /2 C = ha lf of one le g 2 1 /2 2 1 /2 2 3 /4
Figure 46 1 Estimating extent of burns. (From Demling RH. Burns & other thermal injuries. In Doherty GM, ed. Current Diagnosis &Treatment:
Surgery, 14th ed. New York: McGraw-Hill; 2015:227-240.)
linens or nonadherent dressings to decrease the ow o air across CLINICAL MANAGEMENT OF SELECTED CONDITIONS
the wounds. Blisters should not be purpose ully ruptured and cold IN THE SURGICAL ICU
packs should not be applied to burned skin. Systemic antibiotics are ■ RIB FRACTURES
not indicated. Tetanus status should be evaluated and vaccination
Rib ractures are a commonly encountered traumatic injury across
given i needed. Criteria or trans er are presented in Table 46-1.
many medical disciplines. The overall mortality o rib ractures is
approximately 10%, with mortality increasing with each additional
rib racture and worsening prognosis among patients older than
TABLE 46-1 Criteria for Transfer of Burn Patients
55 years old. Patients with limited pulmonary reserve at baseline or
Burn injuries that should be re erred to a burn center include: due to lung contusion are at higher risk or adverse outcome. Flail
1. Partial thickness burns greater than 10% total body sur ace chest de ormity, commonly de ned as three or more consecutive
area (TBSA). ribs ractured at two or more locations results in impaired pulmo-
2. Burns that involve the ace, hands, eet, genitalia, perineum, nary physiology. This impairment can also be produced when ribs
or major joints. are ractured on each side o the sternum. These racture patterns
result in paradoxical chest wall movement which limits the ability
3. Third degree burns in any age group.
o patients to com ortably breathe. Morbidity is increased in this
4. Electrical burns, including lightning injury. patient population rom short and long term disability and dis-
5. Chemical burns. ease related complications, with up to 60% o patients remaining
6. Inhalation injury. disabled. Common complications include prolonged mechanical
7. Burn injury in patients with preexisting medical disorders ventilation, pneumonia and acute respiratory distress syndrome.
that could complicate management, prolong recovery, or Epidural anesthesia has been shown to decrease need or mechani-
a ect mortality. cal ventilation and retain baseline pulmonary status. Despite aggres-
8. Any patient with burns and concomitant trauma (such as sive analgesia and excellent pulmonary hygiene, up to 60% o
ractures) in which the burn injury poses the greatest risk o patients inevitably require prolonged mechanical ventilation. The
morbidity or mortality. In such cases, i the trauma poses the duration o this ventilation on average is 13 days. Surgical rib xation
greater immediate risk, the patient may be initially stabilized has recently grown in popularity due to recent data showing that
in a trauma center be ore being trans erred to a burn unit. this procedure may decrease the length o mechanical ventilation.
Physician judgment will be necessary in such situations and Surgical rib xation has also been described as a salvage therapy to
should be in concert with the regional medical control plan decrease the duration o mechanical ventilation a ter ailing medical
and triage protocols.
management or ail chest.
9. Burned children in hospitals without quali ied personnel or
equipment or the care o children.
10. Burn injury in patients who will require special social,
■ POSTOPERATIVE BLEEDING
emotional, or rehabilitative intervention. Early recognition o postoperative bleeding and initiation o appro-
priate therapy is imperative as patients can progress quickly to
302
li e-threatening hemorrhagic shock. In the immediate postoperative negative consequences or patients. As a result, hemoglobin o 7 g/dL
period (<24 hours), the presence o hypotension and tachycardia in nonbleeding patients without evidence o cardiovascular disease
C
should constitute a presumptive diagnosis o bleeding until another is used routinely as a trans usion trigger. (See Chapter 57: Blood
H
cause is clearly identi ed. The patient should be examined immedi-
A
Products in the Postoperative Period.)
P
ately or signs o hemorrhage rom surgical incisions, drains, tubes, Massive trans usion is most commonly de ned as the administra-
T
or intravenous lines. I signi cant (>100-200 mL) or ongoing drain- tion o 10 or more units o packed red blood cells within a 24-hour
E
age is noted, the surgical team should be noti ed immediately. The period. The need or massive trans usion is most commonly associ-
R
patient should also be examined or evidence o ecchymosis, so t ated with gastrointestinal bleeding and severe traumatic injuries.
4
tissue swelling, and abdominal distension. In the case o trauma These patients o ten receive large volumes o crystalloid in addition
6
patients, missed external (ie, lacerations) or internal (solid organ) to packed red blood cells which can result in a dilutional coagulopa-
injuries may also be present and need to be rapidly identi ed. A thy. In addition, as many as 25% o trauma patients arrive at the hos-
S
u
large volume o blood can be sequestered within the chest, abdo- pital already coagulopathic and are at an increased risk o mortality.
r
g
men, retroperitoneal, and thigh compartments. Hemorrhage into The presence o coagulopathy needs to be addressed early among
i
c
these areas may be dif cult to identi y and o ten require advanced patients requiring massive trans usion with the addition o FFP and
a
l
imaging such as x-ray, bedside ultrasound, or CT scan. Imaging platelets in a balanced ratio. Over the last 10 years, data rom the
C
r
studies that can be per ormed in the ICU should be used pre eren- military experience in Iraq and A ghanistan has suggested improved
i
t
i
c
tially and transporting the patient should be avoided unless they outcomes with the early use o a 1:1:1 ratio o plasma to red blood
a
l
have been appropriately stabilized. I a patient is stable enough to cells to platelets in patients at high risk or needing massive trans u-
C
undergo a CT scan, it should be per ormed with IV contrast as this sion. This ratio was selected in order to more closely mimic the com-
a
r
e
will demonstrate both old hematoma and the presence o active position o the whole blood lost by the patient. Subsequent analysis
bleeding. in civilian trauma centers has demonstrated a 3- to 4- old decrease
Postoperative bleeding can be the result o bleeding rom the in early mortality with the administration o a 1:1:1 ratio within the
surgical site or surrounding tissues, di use coagulopathy, or a com- irst 6 hours although overall mortality was not a ected. Thus,
bination o these actors. Coagulopathic, or “nonsurgical,” bleeding the early use o a balanced blood product trans usion ratio should
is o ten propagated by actors such as hypothermia and acidosis. be considered in massively bleeding patients.
Hypothermia causes platelet dys unction via decreased platelet In order to e ectively administer blood products in emergent
adhesion and aggregation and acidosis decreases thrombin gen- situations, many institutions have developed a protocol or massive
eration, actors which combine to signi cantly impair the clotting trans usion in order to streamline rapid delivery. The protocol is
cascade. Bleeding also propagates coagulopathy through the ongo- activated ollowing clinical evaluation, requently without waiting
ing loss and consumption o clotting o actors as well as dilutional or laboratory con rmation, as this can lead to unacceptable delays.
coagulopathy rom resuscitation. Hypothermia should be addressed As part o the protocol, the blood bank has blood products available
immediately by administering warmed intravenous uids, applying that can be delivered rapidly in large quantities and in prede ned
warmed blankets or orced-air rewarming units, and increasing the ratios. The ideal ratio and the quantity o blood products are deter-
ambient room temperature. Acidosis can be addressed with volume mined by individual institutions. A typical protocol may consist o
resuscitation to improve tissue per usion, mechanical ventilatory an initial container containing 10 units o PRBCS, 10 units o plasma,
adjustments, and correction o the underlying source (bleeding, and two to our packs o platelets. Subsequent containers may
in ection, etc). contain smaller quantities o these components. Blood products
As noted above, postoperative coagulopathy is o ten multi ac- continue to be delivered until the protocol is discontinued by the
torial. The patient’s history should be reviewed or the presence ICU team, ideally once hemostasis has been achieved. The use o a
o known congenital bleeding disorders, such as Von Willebrand massive trans usion protocol has demonstrated improved patient
disease or hemophilia A or B, and these should be addressed as survival and reduced rates o organ ailure. The development o
indicated. In the setting o suspected ongoing bleeding, serial labo- a massive trans usion protocol requires institutional support and
ratory tests (CBC, PT, PT, INR, brinogen) should be drawn every 4 to cooperation among multiple services (physicians, nurses, trans u-
6 hours. Platelets should be administered in order to maintain a level sion services, and laboratory) in order to be success ul.
more than 50,000. Cryoprecipitate should be given in the setting
o consumptive coagulopathy to maintain a brinogen level more
than 100 g/L. An INR o ≤1.5 should be achieved using resh rozen ■ DAMAGE CONTROL SURGERY AND THE
plasma (FFP), which contains all o the clotting actors, or prothrom- OPEN ABDOMEN
bin complex concentrate (PCC). PCC contains the vitamin K-depen- Critically ill surgical patients and trauma patients o ten require emer-
dent coagulation actors (II, VII, IX, and X) and stored as a lyophilized gent li e-saving surgery. These patients may have extensive injuries
powder ollowing extraction rom large donor-pooled plasma. PCC that necessitate time-consuming and complex repairs. Historically,
can be used to rapidly reverse war arin-induced coagulopathy with a de nitive repair o these injuries was attempted during the initial
a signi cantly lower volume o administration as compared to FFP. operation despite prolonged operative times. Un ortunately, many
In addition, there is some data to suggest it is also e ective or hem- o these patients deteriorated intraoperatively or in the acute post-
orrhage or trauma induced coagulopathy. operative period due to the detrimental physiological e ects o
long operative time combined with extensive comorbidities and
traumatic injuries. There ore, surgical management o these patients
■ BLOOD TRANSFUSION—MASSIVE TRANSFUSION shi ted rom prolonged de nitive repair to damage control tech-
The use o blood trans usion in the ICU is extremely common, with niques to avoid urther physiological derangement.
as many as hal o all ICU patients receiving at least one trans usion In 1993, Rotondo et al are credited with modernizing the tech-
during their stay. In nonbleeding postsurgical patients, the criteria niques o damage control surgery or trauma patients. Damage
or blood trans usion may vary widely among individual surgeons. control surgery is de ned as the use o surgical techniques to rapidly
Historically, many patients were managed using a liberal trans usion control hemorrhage and contamination and de er de nitive repair
strategy in order to achieve hemoglobin o 10 g/dL. However, recent in an e ort to temporize the patient and leave the operating room
data suggests that the use o liberal trans usion criteria actually has expeditiously to initiate aggressive resuscitation in the intensive care
303
unit postoperatively. De nitive repair is de erred to avoid the lethal
triad, which consists o hypothermia, acidosis, and coagulopathy.
Prolonged time in the operating room inevitably leads to progres-
P
sive hypothermia which leads to dys unctional coagulation. As the
A
coagulation cascade begins to ail, the patient will lose more blood
R
which leads to urther hypothermia and increased acidosis second-
T
ary to inadequate per usion. Acidosis urther uncouples the coagula-
I
I
tion cascade leading to more hemorrhage. By employing damage
control surgery, bleeding and contamination is rapidly controlled to
avoid the inevitable death spiral o the lethal triad.
Preoperative identi cation o patients who would bene t rom
M
damage control surgery is paramount to ully mobilize the essential
e
d
personnel and equipment to prepare or these critically ill patients.
i
c
There ore, current recommendations or a trauma patient in which
a
l
damage control surgery should be considered include patients
C
o
arriving rom the emergency department with a revised trauma
n
s
score 5 or greater, patients requiring greater than 2 L o crystalloid
u
l
or 2 units o blood or resuscitation, or have a pH less than 7.2. For
t
a
t
nontrauma patients, indications or damage control surgery include
i
o
uncontrolled hemorrhage or elective general surgery, complica-
n
tions during complex duodenal ulcer operations, generalized peri-
tonitis, and other orms o severe intra-abdominal sepsis.
Intraoperative indications to abort a traditional de nitive opera-
tive procedure and transition to damage control surgery include
patients who require more than 10 units o blood or more than 12 L Figure 46 2 An open abdomen covered with a vacuum assisted
o resuscitation, continued acidosis o less than 7.2 and hypothermia closure device.
o less than 34, major inaccessible venous bleeding, re ractory ooz-
ing rom coagulopathy, need to reassess intra-abdominal contents ■ PREVENTION AND MANAGEMENT OF STRESS
postoperatively, and the concern or likely abdominal compartment GASTRITIS
syndrome i the abdomen was closed. Patients admitted to critical care units are at risk or developing
Once the decision has been made to proceed with damage stress-related mucosal damage (SRMD) leading to stress gastropa-
control surgery, the intraoperative goals are appropriately nar- thy. Approximately 1.5% o critically ill patients develop SRMD as a
rowed to expeditiously restore hemostasis and leave the operating result o severe physiologic stress. The cause o SRMD is splanchnic
room as quickly as possible. There ore, a ter opening the patient’s hypoper usion and mesenteric ischemia making SRMD a orm o
abdomen, the rst step is to control hemorrhage. This is initially organ- ailure. With hypoper usion, gastric mucosal cells cannot neu-
accomplished with packing to tamponade bleeding ollowed by tralize acid, perpetuating cellular toxicity. Stress gastropathy may
various rapid maneuvers such as splenectomy, blood vessel isola- occur as a result leading to gastrointestinal bleeding which may
tion, liver packing, etc. Once bleeding has been temporarily con- be severe leading to hemodynamic instability and, in severe cases,
trolled, this provides a crucial window or the anesthesia team to death. Stress gastropathy should there ore be distinguished rom
“catch up” and try to replace lost intravascular volume and blood peptic ulcer disease where increased acid production is the norm.
products. Following hemorrhage control, the next step in dam- Most patients with stress gastropathy are actually achlorhydric.
age control surgery is to stop enteric contamination. O ten, this
means removing portions o damaged or necrotic bowel, stapling
o the ends, and leaving the intestine in discontinuity. Taking
additional time to reconnect the remaining intestine would only
add unnecessary time to the operation and have a high chance o
anastomotic ailure.
With bleeding and enteric contamination controlled, the abdomi-
nal ascia is le t open with the plan o returning to the operating
room in 24 to 48 hours or de nitive repair once the patient can
be adequately resuscitated in the intensive care. In general, most
surgeons temporarily cover the visceral contents o the abdomen
with a vacuum assisted dressing (Figure 46-2). This aids in keep-
ing abdominal domain and preventing desiccation o the intra-
abdominal contents. I the intestine is edematous secondary to the
massive resuscitation required in these critically ill patients, several
return trips to the operating room are o ten necessary or abdomi-
nal washout and replacement o temporizing dressings. Ideally,
the abdomen should be closed within 7 days o the initial surgery
because complications o ascial closure rise dramatically a ter this
time rame. I this is not achieved, the abdomen sometimes has to
remain without ascial closure and be allowed to granulate over Figure 46 3 An open abdomen due to severe abdominal edema;
time. The bowel must not become dry and wound care is extremely granulation tissue and fibrous exudate can be seen over this abdomen.
important. These patients are at high risk or gastrointestinal atmo- Over time the wound will contract and re-epithelize if fistulas do not
spheric stula ormation (Figure 46-3). form.
304
The currently accepted risk actors or SRMD in ICU patients are surgical critical care patients are postoperative, trauma, and burn
mechanical ventilation or at least 48 hours and primary coagulopa- patients. E ective comanagement requires the hospitalist to know
C
thies. Other variables that have been associated with increased risk how to recognize and manage common conditions and complica-
H
include the use o high dose glucocorticoids, severe head trauma, tions in these patient populations.
A
P
extensive thermal burn injury, organ transplantation and severe
T
liver dys unction. Although now uncommon, gastrointestinal bleed- SUGGESTED READINGS
E
ing secondary to stress gastropathy is associated with a our old
R
increase in ICU. Advanced trauma li e support (ATLS(R)): the ninth edition. J Trauma
4
Stress gastritis prophylaxis reduces SRMD and is indicated in ICU
6
Acute Care Surg. 2013;74:1363-1366.
patients who are on mechanical ventilation or in those with coagu-
Alexander RH, Proctor HJ, American College o Surgeons. Committee
lopathy or one o the other risk actors described above. The most
S
on Trauma. Advanced Trauma Life Support Program for Physicians:
u
commonly employed prophylaxis is pharmacologic, and both hista-
ATLS, 5th ed. Chicago, IL: American College o Surgeons; 1993.
r
g
mine type 2 receptor antagonists and proton pump inhibitors are
i
c
utilized. Other options or prophylaxis include antacids administered American College o Surgeons. Committee on Trauma. Resources or
a
l
every 4 hours and titrated to an alkaline gastric pH, and sucral ate, optimal care o the injured patient. Chicago, IL: American College
C
o Surgeons, Committee on Trauma; 2006.
r
an orally administered cytoprotective agent that coats the gastric
i
t
i
c
mucosa providing protection against damage. There are data to Barr J, Fraser GL, Puntillo K, et al. Clinical practice guidelines or the
a
l
suggest that once a patient is tolerating enteral eedings, they likely management o pain, agitation, and delirium in adult patients in
C
do not need stress gastritis prophylaxis. It is important to emphasize
a
the intensive care unit. Crit Care Med. 2013;41:263-306.
r
e
that stress gastritis prophylaxis is not indicated in patients who are Guidelines or the management o severe traumatic brain injury. J
NPO or have an NG tube, unless they are on mechanical ventilation Neurotrauma. 2007;24(Suppl 1):S1-S106.
or have one o the other recognized risk actors. As a result, stress
MacKenzie EJ, Rivara FP, Jurkovich GJ, et al. A national evaluation
gastritis prophylaxis should very rarely be used outside the ICU.
o the e ect o trauma-center care on mortality. N Engl J Med.
2006;354:366-378.
CONCLUSION
Vincent JL. Critical care—where have we been and where are we
With the rising rate o hospitalist comanagement o surgical patients,
going? Crit Care. 2013;17(Suppl 1):S2.
many hospitalists will need to be able to manage patients in and
out o the surgical ICU. The most commonly encountered types o
305
SECTION 2
Anesthesia
307
CHAP TER
47 INTRODUCTION
Hospitalists are o ten involved in perioperative patient care and
should be amiliar with techniques and complications o anesthesia,
as well as preoperative and postoperative considerations. Current
modalities include general anesthetics, neuraxial techniques (spinal
and epidural), regional anesthetics (nerve blocks), and monitored
anesthetic care (MAC), or so-called conscious sedation. Each mode
o anesthesia has bene ts and risks that must be weighed in view o the
operative procedure and the condition and comorbidities o each
Anesthesia: Choices patient. The administration o regional or local anesthetics does
not preclude the necessity or general anesthesia in the event o
and Complications un oreseen events or complications. There ore, patients undergoing

all but the most minor procedures should be assessed as potential
candidates or general anesthesia.
Aeron A. D. Doyle, MD, CM, FRCPC PRACTICE POINT

Cardiovascular and psychiatric medications in the
perioperative period
• Beta-blockers, calcium channel blockers, and amiodarone
should be continued in the perioperative period. Patients who
receive perioperative angiotensin-converting enzyme inhibitors
and angiotensin receptor blockers may be at greater risk o
intraoperative hypotension. Some authorities recommend
holding these drugs on the day o surgery, particularly or
operations with signi cant uid shi ts or using techniques
associated with systemic in ammatory responses, such as
cardiopulmonary bypass. It is traditionally recommended to
stop monoamine oxidase inhibitors (MAOIs) 2 weeks prior to
surgery. Patients who take MAOIs perioperatively are at risk
o serotonergic toxicity and hypertension, especially with
vasopressor use, as well as excessive sedation rom inhibition
o opioid metabolism by MAOIs. Some anesthesiologists
continue MAOIs perioperatively, avoiding indirect-acting
sympathomimetics such as ephedrine, and using narcotics
such as morphine with lesser degrees o interaction with
MAOIs, instead o meperidine. Tricyclic antidepressants (TCAs)
and selective serotonin reuptake inhibitors (SSRIs) may be
continued perioperatively. TCAs have rarely been associated
with intraoperative hypotension, requiring norepinephrine
or reversal. SSRIs are occasionally implicated in perioperative
serotonin syndrome, particularly when given with serotonin
5-HT3 receptor antagonists such as ondansetron, and
phenylpiperidine opioids such as entanyl.
GENERAL ANESTHESIA
General anesthesia is usually induced with a short-acting intrave-
nous agent such as propo ol and maintained with inhaled haloge-
nated ethers or intravenous propo ol. The mechanism o action o
inhalational anesthetics remains unclear and may be a membrane
e ect, a receptor e ect, or both. These agents may be used in con-
junction with narcotics and muscle relaxants to achieve balanced
anesthesia and may also be supplemented with inhaled nitrous
oxide. Airway protection may be obtained by endotracheal intuba-
tion; airway patency, but not protection, may be ensured with a
laryngeal mask airway, or oropharyngeal airway with mask.
309
Complications o general anesthesia include postoperative nau- re ux disease (GERD), obesity, intra-abdominal obstruction or other
sea and vomiting (PONV); aspiration; complications o intubation, pathology, pregnancy, and trauma. American Society o Anesthesi-
such as dental, mucosal, or laryngeal trauma; atelectasis and com- ologists guidelines or nil per os (NPO) status preoperatively recom-
P
plications o positive pressure ventilation, such as barotrauma; com- mend 2 hours or clear uids, 6 hours or a light meal (essentially
A
plications o positioning during surgery; and allergic or idiosyncratic toast and clear uids), and 8 hours or ull meals. These guidelines
R
reactions to anesthetic agents. Additionally, ischemic or thrombo- are or healthy elective patients with no GERD or other risks. Routine
T
embolic events may occur perioperatively because o physiologic antire ux prophylaxis is not recommended, but in patients with
I
I
stresses rom surgery or anesthesia. GERD, histamine H2-receptor antagonists, proton pump inhibitors,
physical antacids, or promotility agents such as metoclopramide
may be indicated. Patients on these medications should have them
PRACTICE POINT ordered preoperatively.
M
Side ef ects o anesthesia induction
e
d
■ COMPLICATIONS OF INTUBATION AND AIRWAY
•
i
c
Propo ol is the most commonly used agent or anesthesia
a
MAINTENANCE
l
induction. A common side e ect is hypotension rom
C
vasodilation and decreased cardiac output, which occurs In closed claim studies, the most common awards or anesthetic
o
n
in up to 16% o patients. Barbiturates such as methohexital complications are those or dental trauma (approximately 1 in 5000).
s
u
and thiopental sodium are occasionally used or anesthesia Laryngeal injury may have an incidence as high as 6% in general
l
t
a
induction but also have cardiovascular depressant e ects. anesthesia but is usually minor and sel -limiting, such as sore throat
t
i
or vocal cord hematoma. Hoarseness lasting longer than 7 days
o
Etomidate and ketamine are sometimes used or anesthesia
n
induction in patients at higher risk o hypotension rom should be evaluated by an otolaryngologist. Mucosal lacerations
propo ol, such as the elderly. Etomidate rarely has signi cant have an incidence o 1 in 1000, but again are usually sel -limiting.
cardiovascular side e ects, but it does inhibit 11-β-hydroxylase,
an enzyme involved in steroid synthesis, thus attenuating ■ ATELECTASIS
the adrenal stress response and potentially leading to Atelectasis, alveolar collapse generally in dependent areas o the
postoperative hypotension. This e ect may persist or up to lung, is common with general anesthesia, particularly in surgery
24 hours in elderly patients a ter a single dose or anesthesia involving the upper abdomen. Postoperative ever and hypoxia
induction. Ketamine does not depress respiratory drive, unlike due to physiologic intrapulmonary shunt are typical mani estations.
most anesthetic agents, and it actually has bronchodilator Mobilization is the most e ective treatment, with chest physio-
e ects that make its use attractive in patients with reactive therapy and incentive spirometry generally having disappointing
airway disease. However, ketamine has side e ects that make results. Continuous or bilevel positive airway pressure (CPAP or
its use in elderly patients problematic, including increases in BiPAP) devices may be bene cial i tolerated by the patient.
heart rate and blood pressure, myocardial depression that is
masked by its sympathomimetic e ects, postoperative delirium ■ POSITIONING COMPLICATIONS
and hallucinations, and neurodegenerative apoptosis in animal Nerve injuries may occur as a complication o various positions
models. under anesthesia. Ulnar neuropathies are the most common, with an
incidence varying rom 4 to 50/10,000 patients. Risk actors include
male gender, extremes o body weight, and prolonged hospitaliza-
■ POSTOPERATIVE NAUSEA AND VOMITING tion. The common peroneal nerve may be injured by pressure at
PONV occurs a ter approximately 10% o surgeries. Risk actors the bular head in the lithotomy position ( eet in stirrups). Brachial
include younger age; emale gender; intra-abdominal, ophthal- plexus injuries may occur rom sternotomy and retraction o the
mic, or ear, nose, and throat (ENT) surgery; past history o PONV or chest wall during coronary bypass surgery. Compartment syndrome
motion sickness; and being a nonsmoker. Strategies to lower the in the extremities may occur i tissue swelling causes obstruction o
risk o PONV include avoidance o general anesthesia in avor o venous and lymphatic drainage, with ailure o capillary per usion
regional anesthesia, use o propo ol, avoidance o nitrous oxide and and cell death. This may result rom improper positioning, or more
volatile anesthetics, minimization o opioids, and adequate hydra- o ten rom extensive and prolonged surgery. It presents as intracta-
tion. Intraoperative prophylaxis and postoperative treatment may ble pain, with examination revealing an absence o pulses. Surgical
include central dopaminergic antagonists such as prochlorperazine, asciotomies are required to decrease tissue pressure.
peripheral dopaminergic antagonists such as metoclopramide,
serotonin 5-HT3 receptor antagonists such as ondansetron, and cor- ■ EYE INJURY
ticosteroids such as dexamethasone. These agents may also be used
Eye injury is reported in up to 0.05% o patients postoperatively. The
as rescue agents a ter emesis to prevent urther symptoms. PONV
majority o cases are corneal abrasions. These resolve within days
usually abates in 24 to 48 hours. The examination o a patient with
with topical antibiotics and rarely cause permanent visual impair-
presumed PONV should assess or bowel sounds and the presence
ment. Rarely, patients develop ischemic optic neuropathy or central
or absence o abdominal distention to avoid missing a diagnosis o
retinal artery occlusion. Risk actors include prone or lateral position,
postoperative ileus.
prolonged surgery, hypotension, spine surgery, and anemia. Un or-
tunately, the prognosis or recovery is poor in these instances.
■ ASPIRATION
Aspiration is the entry o gastric contents into the trachea and lower ■ ALLERGIC AND OTHER REACTIONS
airways. It may occur prior to or during induction o anesthesia, Anaphylactic reactions under anesthesia are rare but potentially li e
intraoperatively i the airway is unprotected, or during emergence threatening. As many medications are given in the perioperative
rom anesthesia and postoperatively. A chemical pneumonitis usu- period, it may be dif cult to determine the causal agent, which may
ally results with severity increasing with lower pH or particulates. include antibiotics, muscle relaxants, or latex. Opioids cause hista-
Risk actors include a ull stomach, preexisting gastroesophageal mine release, resulting in redness and itching, but rarely anaphylaxis.
310
Malignant hyperthermia (MH) is a rare reaction triggered by NEURAXIAL ANESTHESIA
inhalational anesthetics (except nitrous oxide) or succinylcholine
C
Neuraxial anesthesia techniques include the introduction o medi-
(a depolarizing muscle relaxant). It generally presents with increased
H
cations, usually local anesthetics or opioids, into the subarachnoid
metabolic rate, acidosis, and nally rhabdomyolysis. Treatment
A
space (spinal uid) or epidural space. In the epidural space, a
P
consists o discontinuing triggering agents, the administration o
catheter may be inserted or injections or in usions. Local anesthet-
T
dantrolene, and supportive care in an intensive care setting. MH is
ics then act on the spinal cord and nerve roots to inhibit sodium
E
an autosomal dominant disorder, so patients with relatives with sus-
R
channel conduction and block nerve impulses. Opioids act directly
pected or con rmed MH should receive nontriggering anesthetics.
4
on spinal cord receptors, as well as having varying degrees o ros-
Abnormalities o the enzyme pseudocholinesterase may result
7
tral and systemic spread. The bene ts o neuraxial anesthesia may
in the prolongation o action o the muscle relaxant succinylcho-
include preemptive analgesia (the prevention o establishment o
line, leading to postoperative weakness or paralysis. Heterozygous
pain pathways), decreased sympathetic activation, hypercoagulabil-
A
individuals may have mild prolongation, lasting 20 to 30 minutes,
n
ity and in ammation caused by the stress o surgery, and potential
e
whereas homozygous individuals may require sedation and ventila-
s
avoidance o airway manipulation. Although there are short-term
t
h
tion support or hours, depending on the dosage initially given. Suc-
bene ts o postoperative analgesia through an indwelling epidural
e
cinylcholine should be avoided in these patients i possible.
s
catheter inserted just be ore surgery, there are con icting mortality
i
a
:
and morbidity bene ts. Contraindications to neuraxial anesthesia
C
■ ACUTE RENAL FAILURE
h
include untreated sepsis, bacteremia, in ection at the injection site,
o
Acute renal ailure, de ned as a all in creatinine clearance to 50 mL/min bleeding diatheses, increased intracranial pressure, and patient
i
c
e
or less, occurs within the rst week a ter major noncardiac surgery re usal.
s
in approximately 0.8% o patients with previously normal renal unc-
a
Relative contraindications include preexisting neurological de -
n
tion, with 0.1% o patients requiring renal replacement therapy. The cit, hypovolemia, le t ventricular out ow obstruction such as aortic
d
development o postoperative acute kidney injury (AKI) is an inde-
C
stenosis, and lack o cooperation or communication. Complications
o
pendent predictor or hospital mortality. Preoperative predictors o o neuraxial anesthesia include pain at insertion site, dural puncture
m
AKI in various studies have included age, emergent surgery, liver
p
headache, hypotension, high level o block, epidural or spinal hema-
l
i
disease, diabetes mellitus, an elevated body mass index, high-risk toma or abscess, and pruritis rom opioids.
c
a
surgery, congestive heart ailure, ischemic heart disease, peripheral
t
i
o
vascular occlusive disease, and chronic obstructive pulmonary
n
■ DURAL PUNCTURE HEADACHE
s
disease necessitating chronic bronchodilator therapy. Intraopera-
tive strategies to minimize the risk o AKI include maintenance o Dural puncture headache results rom persistent leakage o cere-
euvolemia, avoidance o nephrotoxins, and maintenance o optimal brospinal uid (CSF) ollowing spinal anesthesia or inadvertent dural
blood pressures. puncture during epidural anesthesia. It is typically positional and
severe. Rates ollowing spinal anesthesia are about 1 in 400. The
■ PERIOPERATIVE MYOCARDIAL INFARCTION risk is reduced by use o a smaller-gauge needle with a pencil-point
tip, rather than a cutting tip. The risk also decreases with patient
Myocardial in arction occurs in two distinct clinical settings in the age. Dural puncture a ter epidural is operator dependent; a rate o
perioperative period. Patients may develop an acute coronary 1 in 200 is usually quoted. Treatment includes bed rest, hydration,
syndrome related to plaque rupture. Major contributing actors ca eine, and epidural blood patch i headache persists despite
are physiologic stress and high levels o catecholamines leading to conservative therapy. Epidural blood patch involves injecting the
tachycardia, hypertension, and coronary thrombosis, coupled with patient’s own blood into the epidural space to occlude and clot the
increased coagulability because o tissue trauma and cessation o CSF leak, with about 70% lasting success the rst time per ormed.
antiplatelet agents. The other scenario is the patient with severe The procedure may be repeated i initial results are not satis actory.
but stable coronary disease who develops subendocardial ischemia
because o an imbalance between myocardial oxygen supply and
demand. Causative actors may include tachycardia, hypotension ■ HYPOTENSION AND HIGH BLOCK
or hypertension, anemia, and hypoxemia. In this latter setting, anes- Hypotension occurs secondary to loss o vascular tone. It may be
thesia may be a contributing actor, leading to hypotension and mitigated by preprocedural uid loading and treated with vasopres-
decreased cardiac output. sors and uid a ter it develops. In epidural anesthesia, incremental
dosing may also limit hypotension, particularly in patients with a
■ STROKE xed cardiac output. High block occurs when nerve conduction
Perioperative stroke has many contributing actors, including hyper- is lost at a higher level than intended. It may result in perceived
coagulability rom tissue trauma, emboli rom vascular manipula- respiratory distress due to loss o chest wall proprioception, loss o
tion, temporary cessation o antiplatelet and anticoagulant drugs, diaphragm unction at midcervical levels, and unconsciousness i
and hypotension rom bleeding or anesthetic agents. The risk o the brain stem level is reached. Shock may ensue rom pro ound
stroke is highest in vascular surgery: 1.4% to 3.8% in patients under- vasodilation accompanied by block o cardioaccelerator sympa-
going coronary artery bypass gra ting (CABG), 7.4% in combined thetic bers. Treatment includes physiologic support and sedation
CABG and valve replacement, and 9.7% in multiple valve replace- until the block recedes.
ment. The risk o stroke in patients undergoing general surgery
ranges rom 0.08% to 0.7%. Atrial brillation, valvular disease, renal ■ HEMATOMA OR ABSCESS
disease, and prior stroke are the most robust predictors o periopera- Spinal hematoma or abscess may cause nerve root or spinal cord
tive stroke in general surgery patients. Most perioperative strokes are compression and constitutes a medical emergency, as decompres-
embolic in origin, with only 9% being related to hypoper usion. The sion must occur within 8 hours or reasonable chance o recovery.
risk o perioperative stroke may be increased when high doses o The overall rate o hematoma is 1:220,000 or spinals and 1:150,000
beta-blockers are begun just prior to surgery in patients who have or epidurals but is higher in certain groups o patients and proce-
not taken β-blockers previously. Chronic β-blocker use does not dures. Hal o all cases o epidural hematoma result in devastating,
appear to increase the risk o perioperative stroke. preventable, and permanent neurologic injury, even a ter prompt
311
and may involve a single injection o local anesthetic or the place-
TABLE 47-1 Timing of Cessation of Antiplatelet Drugs and ment o a catheter to provide or multiple injections or in usions
Thromboprophylaxis in Patients Receiving Spinal or postoperative analgesia. Traditionally, injections were guided by
or Epidural Anesthesia
P
anatomical landmarks with or without electrical nerve stimulation
A
Drug Perioperative Management to con rm proximity to the nerve. With the advent o ultrasound-
R
guided techniques, the success rate o regional nerve blocks can be
T
Aspirin Stop 6 d prior to procedure in patients
taking or primary prophylaxis; platelet improved, and the risk o nerve injury lessened. Contraindications
I
I
rebound phenomenon may occur in include in ection at the injection site or patient re usal or inability
patients taking aspirin or secondary to cooperate. Preexisting nerve injury or progressive neuropathies
prophylaxis, requiring individualized may also be contraindications. Certain block techniques may have
assessment and risk strati ication speci c contraindications, such as interscalene block and respira-
M
Nonsteroidal anti- Should be stopped 5 hal -lives prior to tory ailure, as hemidiaphragm paralysis results. Complications o
e
d
in lammatory drugs procedure regional blocks include direct injury to nerve bers by injection
i
c
needles or local anesthetics, systemic anesthetic toxicity, and hema-
a
Clopidogrel Stop 7 d be ore neuraxial block
l
toma or abscess at the injection site. Anticoagulant guidelines are
C
Prasugrel Stop 7-10 d be ore neuraxial block
o
similar to those or neuraxial techniques.
n
Glycoprotein (GP) Stop 8-24 h (epti ibatide, tiro iban) to
s
u
IIb/IIIa inhibitors 2-5 d (abciximab) be ore neuraxial block
l
t
a
War arin Stop 4-5 d be ore procedure; document MONITORED ANESTHETIC CARE
t
i
normal INR be ore initiation o neuraxial
o
MAC, also re erred to as conscious sedation or local-combined anes-
n
block
thesia, re ers to the use o periprocedural sedation, with or without
Subcutaneous Stop 8-10 h be ore procedure local, regional, or neuraxial anesthesia. Medications are the same as
heparin
those used or induction o general anesthesia, including narcotics,
Intravenous heparin Stop 2-4 h prior to procedure benzodiazepines, and propo ol. In some cases, only the dosage o
Low-molecular- Stop 24 h prior to procedure sedative agents or the absence o airway maintenance distinguishes
weight heparin this rom general anesthesia. There ore, ull monitoring, airway
(LMWH) equipment, and resuscitative medications should be available. MAC
Fondaparinux Stop 4 d be ore procedure does not preclude the need to sometimes progress to general anes-
Dabigatran Stop 4-5 d be ore procedure (6 d i thesia, and preoperatively should be treated similarly with regard to
impaired renal unction) NPO status, medications, or resuscitation.
Rivaroxaban Stop 3 d be ore procedure
Apixaban Stop 3-5 d be ore procedure SUGGESTED READINGS
Data rom Narouze S, Benzon HT, Provenzano DA, et al. Reg Anesth Pain Bateman BT, Schumacher HC, Wang S, Shae S, Berman MF. Peri-
Med. 2015;40:182-212.
operative acute ischemic stroke in noncardiac and nonvascular
surgery: incidence, risk actors, and outcomes. Anesthesiology.
surgical intervention. As mentioned, bleeding diatheses are a con- 2009;110:231-238.
traindication to neuraxial anesthetics, but i time permits, a bleeding Blessberger H, Kammler J, Domanovits H, et al. Perioperative beta-
tendency may be corrected preoperatively with plasma or speci c blockers or preventing surgery-related mortality and morbidity.
coagulation actors. The use o anticoagulant drugs signi cantly Cochrane Database Syst Rev. 2014;18;9:CD004476.
increases the risk o spinal hematoma. Un ractionated heparin may Green L, Machin SJ. Managing anticoagulated patients during neur-
be reversed with protamine, but low-molecular-weight heparin axial anaesthesia. Br J Anaesth. 2010;149:195-208.
must be held or 24 hours preprocedure. Platelet inhibitors should
Kheterpal S, Tremper KK, Englesbe MJ, et al. Predictors o postopera-
be held preoperatively, though low-dose aspirin is not contraindi-
tive acute renal ailure a ter noncardiac surgery in patients with pre-
cated i no other anticoagulant is present. Speci c perioperative
viously normal renal unction. Anesthesiology. 2007;107:892-902.
anticoagulant therapy guidelines are published by the American
Society o Regional Anesthesia and Pain Medicine (Table 47-1). Landesberg G, Beattie WS, Mosseri M, et al. Perioperative myocardial
in arction. Circulation. 2009;119:2936-2944.
REGIONAL ANESTHESIA Narouze S, Benzon HT, Provenzano DA, et al. Interventional spine
In regional anesthesia, speci c nerves or a nerve plexus is blocked and pain procedures in patients on antiplatelet and anticoagulant
to create a discrete area o anesthesia. These techniques may be medications. Reg Anesth Pain Med. 2015;40:182-212.
utilized alone or in combination with sedation or general anesthesia Selim M. Perioperative stroke. N Engl J Med. 2007;356:706-713.
312
48
Pain is the most common presenting symptom o disease. It is
de ned as an unpleasant sensory and emotional experience, asso-
ciated with actual or potential tissue damage. There are sound
medical and legal reasons to treat pain aggressively in hospitalized
patients. The Joint Commission, which certi es health care institu-
tions in the United States, mandates that all patients have the right
to adequate pain assessment and management (Table 48-1).
In the inpatient setting, patients may be more concerned about
Perioperative Pain pain relie than the outcome o their underlying illness. Poor pain
control has adverse physiologic consequences that lead to worse
Management outcomes (Table 48-2).

In postoperative patients, better analgesia improves cardiovas-
cular, respiratory, endocrine, immunologic, gastrointestinal, and
hematologic status. Following many common surgeries, acute pain
Darin J. Correll, MD that is not satis actorily treated may become persistent.
PATHOPHYSIOLOGY: NOCICEPTIVE AND

ANTI-NOCICEPTIVE PATHWAYS
Nociception, the perception o noxious stimuli, is a preconscious
neural activity that is normally necessary, but not su cient, or pain.
It is more accurate to re er to nociceptive pathways, rather than
pain pathways. The peripheral nerve bers acting as nociceptors are
lightly myelinated A-delta and unmyelinated C bers, which are trig-
gered or sensitized (peripheral sensitization) by several substances,
including adenosine triphosphate (ATP), prostanoids, bradykinin,
serotonin, histamine, and hydrogen ions. Heat, pressure, or nerve
damage also results in activation.
The primary nociceptors synapse in the dorsal horn o the spinal
cord (Figure 48-1), where the excitatory amino acids glutamate and
aspartate and peptides such as substance P serve as neurotransmit-
ters. Noxious impulses ascend in the lateral spinothalamic tract to
the medial and lateral thalamus and spread to sensory regions o
the cerebral cortex. Parts o the limbic system are also activated;
most likely, this is where nociception is associated with emotion and
arousal and is then perceived as pain.
The nociceptive system has built-in positive and negative eed-
back loops. Prolonged ring o nociceptors enhances synaptic trans-
mission to dorsal horn neurons. This process o central sensitization
involves glutamine and a host o other mediators. Central sensitiza-
tion is an adaptive response that prevents urther injury during a
vulnerable period o tissue healing. This heightened sensitivity gen-
erally returns to baseline over time. However, i central sensitization
is prolonged beyond the healing phase, chronic pain may result.
Substance P–mediated nociception is antagonized by local pro-
duction o endogenous opiates, such as enkephalins and endorphins,
in the dorsal horn and the brain stem. Binding o opioids to opiate
receptors in these locations may account or the analgesic e ects o
these drugs. As well, power ul top-down, endogenous mechanisms
o pain modulation originate in the cortex and travel through brain
stem and midbrain structures en route to the spinal cord. These
descending pain control pathways are mediated by nor-adrenergic
and serotonergic transmission, as well as endogenous opiates.
CHARACTERIZING PAIN INTENSITY

The most commonly used measures o pain intensity in the acute
setting are single-dimension scales (Tab le 48-3). A numerical
rating scale has the numbers 0 to 10 spaced evenly across a page,
313
TABLE 48-1 Joint Commission Pain Assessment and A B
Management Standards for Hospitals F
C
P
1. The hospital respects the patient’s right to pain management.
A
R
2. The hospital educates all licensed independent practitioners ss
Tha la mus
T
on assessing and managing pain.
Hypotha la mus
I
3. The hospital assesses and manages the patient’s pain. The
I
hospital conducts a comprehensive pain assessment that is
consistent with its scope o care, treatment, and services and
the patient’s condition. Midbra in
M
4. The hospital uses methods to assess pain that are consistent
S pinotha la mic
e
with the patient’s age, condition, and ability to understand.
d
tra ct
i
c
5. The hospital assesses and reassesses its patients. The hospital Me dula
a
de ines, in writing, criteria that identi y when additional,
l
Injury
C
specialized, or more in-depth assessments are per ormed
o
n
or pain.
s
u
6. Based on the patient’s condition and assessed needs, the
l
t
a
education and training provided to the patient by the
t
i
hospital include any o the ollowing: discussion o pain, the
o
S pina l
n
risk or pain, the importance o e ective pain management,
the pain assessment process, and methods or pain cord
management.
Figure 48 1 Nociception transmission and pain modulatory path-
ways. (A) Transmission system for nociceptive messages. Noxious stimuli
activate the sensitive peripheral ending of the primary afferent nocicep-
where 0 is “no pain at all” and 10 is “the worst pain imaginable.” tor by the process of transduction. The message is then transmitted over
Patients are instructed to circle the number that represents the the peripheral nerve to the spinal cord, where it synapses with cells of
amount o pain they are currently experiencing. A common varia- origin of the major ascending pain pathway, the spinothalamic tract.
tion is the verbal numeric scale, where patients are asked to verbally The message is relayed in the thalamus to the anterior cingulate (C),
state a number between 0 and 10 to correspond to their present frontal insular (F), and somatosensory cortex (SS). (B) Pain-modulation
pain intensity. Some people pre er to use words to describe the network. Inputs from frontal cortex and hypothalamus activate cells in
intensity o their pain; these are termed verbal descriptor scales. the midbrain that control spinal pain-transmission cells via cells in the
Another variant that may be use ul in the elderly or cognitively medulla. (Reproduced with permission rom Fauci AS, Braunwald E,
impaired are scales with drawings o aces, ranging rom a con- Kasper DL, et al. Harrison’s Principles of Internal Medicine, 17th ed.
tented smiling ace to a distressed-looking ace. New York: McGraw-Hill; 2008, Fig. 12-4.)
Single-dimensional scales are quick and simple to use, an impor-
tant bene t in the acute setting when repeated measures are
needed over a brie period o time. One disadvantage is that they Questionnaire, Brie Pain Inventory). These multidimensional scales
attempt to assign a single value to a complex, multidimensional attempt to take into account the complex nature o pain. However,
experience. Another is that patients can never know i the present in the inpatient setting, they are too time consuming or rapid or
experience is the “worst.” I a value o “10” is chosen and the pain repeated use. One compromise is to address a limited number
worsens, the patient has no o cial means to express this; in practice o the dimensions o pain, using a ew single-dimensional scales
many patients state a number over 10. to address issues that are important to hospitalized patients— or
Several multidimensional scales exist that attempt to assess example pain, anxiety, depression, anger, ear, and inter erence with
various aspects o the patient’s pain experience (eg, McGill Pain physical activity.
DIAGNOSIS
TABLE 48-2 Physiologic Consequences of Uncontrolled Pain ■ HISTORY
Tachycardia, hypertension, increased The history o the patient in pain includes the pain’s location and
Cardiovascular cardiac workload the presence, or absence, o radiation rom the primary site. Inten-
Pulmonary Hypoxia, hypercarbia, atelectasis, decreased sity should be determined using appropriate scales, as already
cough described. The patient should describe the pain’s character (eg,
Gastrointestinal Decreased gastric emptying, nausea/ aching, burning, dull, electric-like, sharp, shooting, stabbing, ten-
vomiting, ileus der, and throbbing). This may provide clues to help diagnose the
generator o the pain which aides in deciding the correct therapy
Renal Urinary retention
(Table 48-4). Does the pain have a pattern (constant, intermittent, or
Endocrine Increased adrenergic activity, catabolic state, better or worse at certain times o day) and aggravating and alleviat-
sodium/water retention
ing factors? Does the pain have an impact on functional status? Are
Immunologic Impairment, slowed wound healing the patient’s activities o daily living a ected as an outpatient, or is
Musculoskeletal Splinting, contractures, decreased mobility it hampering their ability to cough, get out o bed, and ambulate
(deep vein thrombosis) while in the hospital? The patient’s prior analgesic history, in particu-
Hematological Increased coagulability lar, what therapies have either worked or not worked in the past,
Neurological Anxiety, ear, anger, atigue, delirium helps to decide what agents may be e ective now. Exact doses o
ongoing analgesics should also be determined.
314
C
TABLE 48-3 Single-Dimension Pain Scales
H
A
Numerical Rating Scale
P
0 1 2 3 4 5 6 7 8 9 10
T
E
Verbal Descriptor Scales
R
No pain None Mild Moderate Severe Worst pain imaginable
4
8
P
e
r
i
o
p
0 2 4 6 8 10
e
r
a
t
i
The Faces Pain Scale—Revised. From Pain 2001;93:173-183. Used with permission rom IASP.
v
e
P
a
i
n
■ PAST MEDICAL HISTORY activation is a common and nonspeci c nding in hospitalized
M
The patient’s medical history should be obtained. Medical con- patients, it o ers little help in the diagnosis and treatment o pain
a
n
ditions that cause pain include cancer, diabetes, osteoarthritis, in an awake, competent patient. These measures may be used as
a
surrogates in patients who cannot express their pain experience.
g
rheumatoid arthritis, herpes zoster (shingles), and spinal cord injury.
e
m
Psychological conditions may adversely impact a patient’s pain
e
experience and need appropriate diagnosis and therapy. These PRACTICE POINT
n
t
include anxiety (especially in acute pain states), depression (most
prevalent in persistent pain states), ear, catastrophizing (assuming
• Patients may not exhibit any alterations in vital signs despite
signi cant levels o pain, especially patients who have
the worst-case scenario), and personality disorders. A amily history
persistent pain.
that is positive or substance abuse in the patient’s relatives is a risk
actor or addiction in the patient. A social history positive or alco-
hol, tobacco, or other drugs may indicate a need to prescribe agents ■ DIAGNOSTIC TESTING
to prevent withdrawal (ie, benzodiazepines or nicotine patches). Diagnostic tests to determine the etiology o pain may be use ul in
Even patients with a history o addiction still need to be appropri- some situations (eg, radiographs to assess or racture, magnetic
ately treated or pain in the acute setting. In this setting, it may be resonance imaging [MRI] to diagnose nerve impingement in the
use ul to enlist the help o a psychiatrist or psychologist trained in spinal cord, or electromyography [EMG] to diagnose a neuropathy).
addiction management. However, normal test results should not be used to discount a
patient’s report o pain.
■ PHYSICAL EXAMINATION
A directed physical examination o the pain ul site, and a generalized CARDINAL PRINCIPLES OF PAIN MANAGEMENT
physical exam o the patient as appropriate, should be per ormed. Pain is a subjective phenomenon, resulting rom the processing,
Pain (especially acute) may be associated with tachycardia, hyper- ltering and modulating o nociceptive input through the a ective
tension, diaphoresis, and tachypnea. However, since sympathetic (limbic system) and cognitive processes unique to each individual.
TABLE 48-4 Determining the Mechanism and Treatment of Pain
Pain Mechanism Character Examples Treatment Options

Somatic Usually well localized and • Laceration • Heat/cold
constant • Fracture • Acetaminophen
Aching, sharp, stabbing • Burn • NSAIDs
• Abrasion • Opioids
• Localized in ection or in lammation • Local anesthetics (topical or in iltration)
Visceral Not well localized— • Muscles/spasm • NSAIDs
constant or intermittent • Colic or obstruction (gastrointestinal or • Opioids
Generalized ache, pressure renal) • Muscle relaxants
or cramping, can be sharp • Sickle cell • Local anesthetics (nerve blocks)
• Internal organ in ection or in lammation
Neuropathic Can be localized (ie, • Trigeminal • Anticonvulsants
dermatomal) or radiating, • Postherpetic • Tricyclic antidepressants
can also be generalized • Postamputation • Muscle relaxants
and not well localized • Peripheral neuropathy • NMDA antagonists
Burning, tingling, electric • Nerve in iltration • Neural/neuraxial blockade
shock, lancinating
NMDA, N-methyl-d -aspartate; NSAIDs, nonsteroidal anti-in lammatory drugs.
315
The patient’s report o pain must be respected and believed. As pain chronic pain, particularly pain ul diabetic neuropathy. Relaxation
is an a ective and cognitive experience, the placebo response to and guided imagery have shown little bene t in the acute setting.
analgesics is real and may be help ul. However, using the placebo Attention techniques can be complicated in that one needs to
P
response does not mean misleading patients, or administering an determine which approach is better or a particular patient. Some
A
inactive substance to determine whether they are lying or to punish patients do better when instructed to shi t attention away rom the
R
them. Rather, the placebo e ect in contemporary medicine is that pain, whereas others do better i instructed to attend to a particular
T
patient belie in a particular therapy makes it more likely to work. portion o the pain (eg, the sensory component, as opposed to the
I
I
Physician attempts to truth ully “talk up” genuine attempts at anal- emotional component). Acupuncture and electroacupuncture have
gesia are thus likely to enhance the e ects. The reverse is also true. been shown to be bene cial in the acute setting, reducing both
I a patient states that a particular therapy “never works or them,” it pain and common side e ects rom opioid analgesics. However,
is less likely to be e ective. these are labor intensive, and speci c training is required. The use o
M
The patient’s pain level and degree o pain relie should be virtual reality has been shown to reduce levels o pain and unpleas-
e
d
assessed appropriately and regularly. Pain should be treated quickly. antness or burn care procedures, common pain ul cancer proce-
i
c
Therapy should not be withheld while the diagnosis is unclear; dures and treatments, and routine medical procedures. The major
a
l
pain treatment does not impede the ability to diagnose disease. A hindrance to its use is the cost and availability o the equipment.
C
o
comprehensive plan should be used that addresses the multidimen-
n
s
sional aspects o pain. This may require an interdisciplinary team
u
■ NONOPIOID ANALGESICS
l
approach (eg, hospitalist, pain specialist, anesthesiologist, surgeon,
t
a
In the absence o a contraindication, all patients in pain should
t
psychiatrist or psychologist, and physical therapist), especially or
i
o
patients with persistent pain. be prescribed a nonopioid analgesic. These agents have analgesic
n
The analgesic plan should be discussed with the patient and, e ects and are opioid sparing, leading to decreased side e ects.
when appropriate, the patient’s amily. The patient’s expectations They are the primary analgesics or low-intensity pain associated
or pain management should be understood, and patients should with headache or musculoskeletal disorders and are use ul adjuncts
be o ered reasonable goals or the outcomes o therapy. in moderate to severe pain. These agents have a plateau e ect, such
that doses beyond the recommended range increase the incidence
o side e ects but do not improve analgesia.
PRACTICE POINT Acetaminophen does not inhibit peripheral prostaglandin syn-
thesis. This explains its lack o side e ects on gastric mucosa and
• A multimodal approach or managing pain, employing both
platelets, but it also means that it is not active at peripheral sites
pharmacologic and nonpharmacologic measures, is better
o inf ammation. In diseases where inf ammation plays a major role
than using just one modality. This approach allows or optimal
in generating pain (eg, rheumatoid arthritis), acetaminophen is o
analgesia with the lowest incidence o side e ects. In the
minimal bene t. The analgesic mechanism o acetaminophen is
absence o a contraindication, all patients in pain should be
not well characterized but may involve acilitation o central antino-
prescribed a nonopioid analgesic. Clinicians should be amiliar
ciceptive pathways via serotonin, increasing levels o endogenous
with several agents within each class o analgesics, including
cannabinoids, or inhibition o nitric oxide synthesis in the spinal
possible side e ects, because individual responses vary greatly.
cord, which may inter ere with substance P–related nociception.
The nonacetylated salicylates (eg, choline magnesium trisalicy-
I pain is present most o the time or expected to last or an late) have a relatively low incidence o gastrointestinal bleeding,
extended period o time (eg, more than a ew weeks), long-acting perhaps related to their lack o inhibition o platelet aggregation.
agents or round-the-clock dosing o short-acting agents should be The nonselective nonsteroidal anti-inf ammatory drugs (NSAIDs) are
used. When long-acting drugs are used, immediate-release agents potent anti-inf ammatory analgesics with signi cant risk or gastro-
will also be needed or breakthrough pain. When pain is intermittent intestinal bleeding and renal insu ciency. No single NSAID appears
or expected to be o brie duration (eg, less than a ew weeks), then to be more e ective as an analgesic than any other, but as there
as-needed dosing o immediate release agents can be used alone. is great interpatient variability in response, thus changing agents
may be o bene t i one does not seem to be e ective. The COX-2
selective NSAIDs have a reduced risk o peptic ulceration compared
TREATMENT to nonselective NSAIDs, but an equivalent chance o renal toxicity.
Pharmacologic and nonpharmacologic treatment measures are Celecoxib is currently the only COX-2 selective NSAID available in
o ten used together. Pain medication alls into three categories: the United States. It should not be considered a rst-line agent
nonopioid analgesics, opioids, and adjuvant analgesics. Therapy given it’s cost, and should not be used long term at high doses, as
should be individualized in a multimodal, stepwise approach, add- it increases the risk o major cardiovascular events. Table 48-6 lists
ing or changing agents when pain control is inadequate, and with- the dosing regimens and adverse e ects o selected nonopioid
drawing agents as pain resolves (Table 48-5). analgesics.
■ NONPHARMACOLOGIC MEASURES ■ OPIOIDS: TERMINOLOGY

Although scienti c data on nonpharmacologic measures are lim- Tolerance is the diminished response to a drug over time, such
ited, most have little risk. At a minimum, they may have placebo that, in order to maintain the same e ect, the drug dose needs to
bene t, due to the cognitive and a ective inf uence on pain. Appli- be increased. Dependence is a state o physiologic adaptation that
cation o cold (to reduce inf ammation) or heat (to reduce spasms) develops with continued use o a drug, presenting as a withdrawal
to muscles or joints are commonly used, but the evidence or an syndrome i the drug is abruptly stopped, the dose is dramatically
analgesic bene t is mixed. Hypnosis has been shown to reduce reduced, or an antagonist is given. Addiction is a primary, chronic,
pain associated with procedures. However, it requires speci c train- neurobiologic disease with many actors inf uencing its develop-
ing and time to administer. In the acute setting, the results with ment. It mani ests as drug-seeking behaviors, impaired control
transcutaneous electrical nerve stimulation (TENS) are conf icting, over the drug, and continued use despite negative e ects. Tol-
with somewhat better evidence o e ectiveness in the setting o erance to and dependence on opioids do not equal addiction!
316
TABLE 48-5 Suggested Pain Management Schemes
C
H
The World Health Organization devised the analgesic ladder or the treatment o cancer pain. The concepts behind its use are help ul in
A
the management o all types o pain, both persistent and acute.
P
S trong Opioid + Non-opioid
T
E
± Adjuva nts
R
4
Ina de qua te Ana lge s ia
8
We a k Opioid + Non-opioid
P
e
± Adjuva nts
r
i
o
p
e
r
Ina de qua te Ana lge s ia
a
t
i
v
e
Non-opioid Ana lge s ic
P
a
± Adjuva nts
i
n
M
The World Federation o Societies o Anesthesiologists devised another analgesic ladder to use or the treatment o acute/postoperative
a
pain.
n
a
g
S trong P a re nte ra l Opioid
e
m
Loca l a ne s the tic
e
n
t
P a in De cre a s e s
Ente ra l Opioid
P a in De cre a s e s
Non-opioid Ana lge s ic
Data rom World Health Organization. Cancer Pain Relief. Geneva, Switzerland: World Health Organization; 1986 and Data rom Charlton JE. WFSAUpdate in
Anesthesia. 1997;7:2-17.
TABLE 48-6 Select Nonopioid Analgesics
Agent Adult Dosing Maximum Daily Dose Comments

Acetaminophen 650-1000 mg every 6 h 4000 mg Single doses above 1000 mg do not improve analgesia
Choline magnesium 1000-1500 mg twice a day 3000 mg Caution in liver disease, avoid in severe liver disease
trisalicylate
Diclo enac 50 mg twice a day- our 200 mg Low GI e ect incidence, but possible increased renal
times a day e ects, recent data suggest increased negative CVe ects
Etodalac 200-400 mg every 6-8 h 1000 mg Low GI and renal e ect incidence, sa est NSAID in liver
disease
Ibupro en 400-600 mg every 4-6 h 3000 mg <1500 mg daily has low risk o GI e ects, possible
increased renal e ects, inhibits CVbene its o aspirin when
given concomitantly
Ketorolac 30 mg every 6 h 120 mg High risk o renal and GI complications; use or no more
than 5 d; 15 mg every 6 h in renal impairment, age >65,
weight <50 kg
Nabumetone 750-1500 mg daily or 1500 mg Low GI e ect incidence
twice a day
Naproxen 250-500 mg every 6-12 h 1500 mg Possible increased liver and renal e ects, probably least
negative CVe ects
Celecoxib 100-200 mg daily 200 mg Use 100 mg dose i possible; long-term use has increased
negative CVe ects
CV, cardiovascular; GI, gastrointestinal.
317
accumulation o the metabolite, morphine-6-glucuronide, which
TABLE 48-7 Opioid Classification can lead to sedation and respiratory depression.
Not Recommended Weak Strong
P
Meperidine (Demerol) is not recommended for pain manage-
A
Meperidine Codeine Fentanyl ment. Its active metabolite, normeperidine, can accumulate in 24
R
Hydrocodone Hydromorphone to 48 hours to levels that produce nervous system excitation (trem-
T
Tramadol Methadone ors, muscle twitching, convulsions). Meperidine causes a strong
I
I
Morphine euphoric eeling, especially when given by intravenous push. It is
a weak agonist that is usually ine ective or more than mild pain,
Oxycodone
and it causes more nausea than other agents. Hydrocodone should
Oxymorphone be used with caution. In the United States, it is o ten co ormulated
M
with acetaminophen, aspirin, or ibupro en, and adverse events
e
d
and toxicity may result rom these other agents, in addition to the
i
c
hydrocodone itsel . Hydrocodone has also become a avored drug
a
l
Pseudoaddiction denotes iatrogenically induced patient behaviors o abuse in the United States.
C
o
that mimic drug seeking, due solely to the under treatment o pain.
n
Opioid administration
s
When pain is adequately managed, the behaviors resolve.
u
l
t
Whenever possible, the enteral route o administration is best, as it is
a
t
the easiest route with the most stable pharmacokinetics. I a patient
i
o
Opioid therapy basics
n
cannot take anything by mouth or adequate analgesia cannot be
When treating moderate to severe pain, pure agonists should be obtained in a timely manner, then intravenous (IV) administration
used, as opposed to agonist/antagonists. The commonly used ago- should be used. Intramuscular administration should be avoided or
nists are shown in Table 48-7. several reasons: It is pain ul; there are wide f uctuations in absorp-
The optimal analgesic dose varies widely among patients, even tion; it takes a long time to reach peak e ect; there is a rapid all-o
the opioid naïve. Side e ects rom opioids also vary widely between o action therea ter. When a patient is competent, the use o an IV
patients. It is, there ore, help ul to be amiliar with the characteristics patient-controlled analgesia (PCA) o ers the best overall pain man-
o several di erent agonists (Table 48-8). agement option (see later discussion).
Opioid dosing
PRACTICE POINT Recommended starting doses or moderate to severe pain in the opi-
Opioids oid-naïve are listed in Table 48-9. I a patient is not receiving enough
• Patients should be asked which opioids have worked or not pain relie at a given dose, subsequent doses should be increased by
worked in the past, or have given them intolerable side e ects. 25% to 50%. I a patient is having pain be ore the next dose is due, the
dosing interval should be reduced, or the dose increased.
• Whenever possible, the enteral route o administration is best, A switch to another opioid may be necessary in several circum-
as it is the easiest route with the most stable pharmacokinetics.
stances. First, patients on therapeutic opioid doses who are not
I a patient cannot take anything by mouth or adequate
receiving any pain relie may not have a receptor population at
analgesia cannot be obtained in a timely manner, then
which that particular opioid is e ective. A di erent opioid may
intravenous (IV) administration should be used. Intramuscular
provide better analgesia. Second, i a patient is having intolerable
administration should be avoided.
side e ects, rotation to a di erent opioid may provide relie . In this
• I pain is present most o the time or expected to last or an case, the patient’s receptor population may bind a particular opioid
extended period o time (ie, more than a ew weeks), long-
in regions that cause side e ects. Third, i a particular opioid cannot
acting agents or round-the-clock dosing o short-acting agents
be given by the route o administration required, then changing to
should be used. When long-acting drugs are used, immediate-
another opioid will be necessary. Finally, i a patient has been on an
release agents will also be needed or breakthrough pain.
opioid or a long time and has developed tolerance, rotation to a
• Patients should be monitored closely or e ectiveness and di erent opioid may provide better analgesia, usually at less than
adverse events whenever there is a change o agent or route o the expected equianalgesic dose. A similar e ect may also be seen
administration. in patients on long-term opioid therapy or chronic pain who have
• When pain is intermittent or expected to be o brie duration an episode o acute pain; better analgesia may also be experienced
(eg, less than a ew weeks), then as-needed dosing o in this setting with a switch to a di erent opioid.
immediate release agents can be used alone. Equianalgesic-dosing charts (see Table 48-8) are based on the
• When a patient is competent, the use o an IVpatient- relative potency o opioid agonists, as determined by single-dose
controlled analgesia (PCA) o ers the best overall pain clinical studies and experience. These calculations are estimates
management option or postoperative hospitalized patients. only, and clinical judgment is always required or use. Incomplete
cross-tolerance exists between the various opioids. This means
patients will not be as tolerant to a new opioid agonist as they are
Selecting an opioid to the one they were on previously. Thus, when converting between
Codeine is not a good rst choice due to the act that 10% to 20% opioids, the calculated equianalgesic dose o the new agent must
o the population lacks an active orm o the enzyme (cytochrome be reduced by 25% to 75% to prevent over sedation and respiratory
P450 2D6) necessary to convert codeine into an active drug in the depression. Table 48-10 shows an example o opioid conversion.
body (ie, morphine). All opioids should be used with caution in
patients with renal or hepatic insu ciency; lower doses or longer Sustained release or long acting opioids
dosing intervals are wise in this setting. Morphine is relatively con- Episodic pain or pain expected to be o a brie duration should
traindicated in patients with severe renal insu ciency due to the be treated with immediate-release agents alone. Sustained-release
318
TABLE 48-8 Opioid Characteristics
C
H
Agonist Route Equianalgesic Dose (mg) Onset (min) Peak Effect (min) Duration of Effect (h)
A
Morphine IV 10 5-10 10-30 3-5
P
T
Oral 30 15-60 60-120 4-6
E
Oral CR – 30-120 180-240 8-12
R
Oral SR – 30-120 480-600 8-24
4
8
Codeine IM 120 10-30 90-120 4-6
Oral 200 30-45 60 3-4
P
Hydromorphone IV 1.5 5-20 15-30 3-4
e
r
Oral 7.5 15-30 90-120 4-6
i
o
p
Oxycodone Oral 20 15-30 30-60 4-6
e
r
Oral CR – 30-60 90-180 8-12
a
t
i
v
Methadone IV 10* 10-20 60-120 4-6
e
Oral 20* 30-60 90-120 4-12
P
a
i
Fentanyl IV 0.1 <1 5-7 0.75-2+
n
M
TD (see Table 48-11) 720-1080 1440-4320 48-72
a
Oxymorphone IV 1 5-10 30-60 3-6
n
a
Oral 10 “meaning ul relie ” = 60 4-6
g
e
m
CR, controlled release; IM, intramuscular; IV, intravenous; SR, sustained release; TD, transdermal.
e
n
*
These doses are based on single administrations and should only be used to convert between oral and IVmethadone. I converting a patient who has been
t
on a di erent opioid, use the ollowing table that takes into account the dose-dependent potency changes seen with methadone:
Equianalgesic Conversion to Methadone §

Oral Morphine Equivalent Oral Methadone Oral Morphine
That Patient Is Taking (mg) (mg)
< 100 mg/d 1 4
101-300 mg/d 1 8
301-600 mg/d 1 10
601-800 mg/d 1 12
801-1000 mg/d 1 15
> 1000 mg/d 1 20
To determine the starting dose o oral methadone:

• Convert the patient’s daily opioid dose into oral morphine equivalents.
• Convert the daily oral morphine equivalents to a daily oral methadone dose using the table.
• Reduce the calculated daily oral methadone dose by 33%-50%.
• Divide the resulting reduced daily dose by 3.
• Prescribe this dose o oral methadone (in mg) every 8 h.
§
This table is not meant to be used to convert rom methadone to other opioids. There is limited data on the conversion rom methadone to other agents,
and inadequate analgesia o ten results. Thus, i it is necessary to convert rom methadone to another opioid agonist, it is best per ormed in stages, with close
monitoring o the patient or e ectiveness (eg, introduce the new agent over a 3-d period as the methadone dose is tapered by one-third each day).
ormulations should be initiated in the acute setting i pain is present

TABLE 48-9 Recommended Starting Doses of Opioids for most o the time, and pain is expected to last or an extended period
Adults Over 50 kg o time (2-3 weeks or more). When using a sustained-release opioid,
an immediate-release opioid equivalent to 10% to 15% o the 24-hour
Agonist Oral IV
total every ew hours on an as-needed basis should also be prescribed.
Codeine 15-60 mg every 3-4 h n/a I more than our to ve rescue doses o immediate-release opioid are
Hydrocodone 5-10 mg every 3-6 h* n/a needed in 24 hours, the dose o sustained-release agent should be
Tramadol 50-100 mg every 4-6 h? n/a increased by 50% to 100% o the total 24-hour breakthrough dose used.
Oxycodone 5-10 mg every 3-4 h n/a Transdermal entanyl is not appropriate to treat acute pain, espe-
Morphine 10-30 mg every 3-4 h 5-10 mg every 2-4 h cially in the opioid-naïve. Its use in the acute setting may lead to
Hydromorphone 2-6 mg every 3-4 h 1-1.5 mg every 3-4 h severe respiratory depression rom the delayed peak e ect o the
drug. It should only be used in patients already tolerant to opioids
Oxymorphone 10-20 mg every 4-6 h 1 mg every 3-4 h
o comparable potency. Table 48-11 gives recommendations or
conversion rom other opioids to transdermal entanyl.
*Daily dose limited by acetaminophen component in available preparations.
?
Maximum recommended 24-h dose: 400 mg in adults < 75 y old; 300 mg Methadone is also not appropriate as a rst-line agent in the
in adults > 75 y old. acute setting, especially in the opioid-naïve. Its use requires an
319
to achieve com ort, because the incremental dosing o the PCA will
TABLE 48-10 Example of Opioid Conversion not be e ective in a reasonable period o time. The use o a PCA
helps overcome the wide interpatient variation in opioid require-
1. Patient used 15 mg o IVhydromorphone in the past 24 h.
P
ments by allowing the patient to control the dosing regimen. Mor-
A
2. According to the equianalgesic table: phine is the most common rst-line agent. It is not the best choice in
R
1.5 mg o IVhydromorphone = 20 mg o oral oxycodone patients with renal insu ciency, due to accumulation o the active
T
1.5mg of IVhydromorphone 1.5mg of IVhydromorphone metabolite. Fentanyl has a quicker onset and shorter duration o
I
I
= action than morphine. This decreases the likelihood o oversedation,
20mg of oraloxycodone X
but the patient must activate the PCA more o ten, making it di cult
X = 200 mg o oral oxycodone/d or some patients to sleep at night. Hydromorphone is generally
more e ective in opioid-tolerant patients and given its pharmacoki-
M
3. Taking into account incomplete cross-tolerance, decrease the
netics is an excellent choice or use in PCAs. Recommended starting
e
total daily opioid dose by 25%-75%:
d
doses in the opioid-naïve patient are listed in Table 48-12, along
i
200 – (0.25 × 200) = 150 mg o oral oxycodone/d
c
with suggestions or dose titration.
a
l
200 – (0.75 × 200) = 50 mg o oral oxycodone/d The lockout interval, or minimum time between doses, is typically
C
o
4. Dose initially every 4 h: set at 5 to 10 minutes. Even though the time to peak e ect may be
n
s
150/6 25 mg oxycodone every 4 h longer than this, in practice no major di erences are seen with lon-
u
l
ger lockouts. There have also been no good studies to suggest that
t
50/6 8 mg oxycodone every 4 h
a
t
a particular lockout interval is better than any other.
i
o
There ore, order: oxycodone 10-25 mg every 4 h as needed or
A basal rate on the PCA may be needed in opioid-tolerant
n
pain.
patients or in patients receiving entanyl, given its short hal -li e.
Basal rates are not recommended in the opioid-naïve, elderly, or
patients with obstructive sleep apnea or morbid obesity. Some
understanding o the unique pharmacology o the drug, especially
people advocate use o basal rates at night however this is not
its extended duration o action and its dose-dependent potency.
recommended as there is no data that basal rates help patients get
Also, as it takes several days to reach a stable plasma concentra-
more sleep at night, pain scores are not improved and the risk o
tion, patients need to be monitored closely or e cacy and side
respiratory depression is increased. Basal rates should be decreased
e ects. As methadone is a racemic mixture o a mu agonist and
or discontinued i a patient is not activating the PCA, or i the patient
an N-methyl-d -aspartate (NMDA) antagonist (see later discussion),
is becoming excessively sedated.
patients develop less analgesic tolerance. Patients must be made
aware o the long duration o action o methadone, be warned not
■ COMPLICATIONS/OPIOID-INDUCED SIDE EFFECTS
to take extra doses or mix it with other medications, and be amiliar
with signs o overdose. Nausea, vomiting, pruritus, constipation, sedation, and respiratory
depression are common opioid-related side e ects. They occur
■ PATIENT-CONTROLLED ANALGESIA BASICS more o ten in opioid-naïve patients, as tolerance eventually devel-
ops to all these e ects, except constipation. Adverse e ects can be
PCA is intended as maintenance therapy. I the patient is in moder-
ameliorated by changing the drug dose or schedule, switching to a
ate to severe pain when it is begun, IVloading doses must be given
di erent agent (side e ects o di erent opioid agonists vary among
patients), using speci c therapy to counteract the side e ect, or
adding another analgesic or adjuvant to allow a lower opioid dose.
TABLE 48-11 Dose Conversion Guidelines from Another Constipation should always be expected with opioids. Prophylac-
Opioid to Transdermal Fentanyl tic use o stool so teners, such as docusate, and stimulant laxatives,
such as senna preparations, is recommended. Nausea and vomiting
24-h Oral Morphine Equivalent Transdermal Fentanyl
can be treated with any o the available agents (eg, prochlorpera-
Dose (mg/d)* Initial Dose (mcg/h)
zine, ondansetron, metoclopramide, and promethazine), as none
60-134 25 has been shown to be more or less e ective. Metoclopramide is a
135-224 50 promotility agent with limited antinausea e ects and is most e ec-
225-314 75 tive i there is vomiting. Promethazine or possibly a scopolamine
315-404 100 patch (especially i used be ore the symptoms onset) may be
405-494 125 e ective i the patient has a history o motion sickness, or i nausea
is provoked by movement, as opioids sensitize the inner ear laby-
495-584 150
rinthine system. Extreme caution must be used with promethazine
585-674 175 because o its possibility to cause severe tissue damage i extravasa-
675-764 200 tion occurs. Pruritus is thought to be a central mu opioid receptor-
765-854 225 related phenomenon. Diphenhydramine is only e ective i the
855-944 250
945-1034 275
1035-1124 300 TABLE 48-12 Suggested Starting Patient-Controlled
Analgesia Dose and Dose Changes
*Convert other opioid to oral morphine equivalents using an equianalgesic Morphine Hydromorphone Fentanyl
dose table.
This table should not be used to convert rom transdermal entanyl to an- Staring 1.0-1.5 mg 0.2 mg 20-25 mcg
other opioid because the conversion to transdermal entanyl in this table PCA dose
is conservative. There ore, use o this table to convert rom transdermal PCA dose 0.5 mg 0.1 mg 5-10 mcg
entanyl to another opioid can overestimate the amount o the new agent, change
resulting in overdosage and respiratory depression.
320
than the antidepressant e ects (weeks). Skeletal muscle relaxants
TABLE 48-13 Treatment of Suspected Opioid-Induced are use ul or muscle injury or spasms. The antispasmodic baclo en
C
Respiratory Depression
is use ul or the treatment o pain with a spastic component or in
H
A
Suggested definition of respiratory depression certain neuropathic pain states. Antagonism o the NMDA receptor
P
has no primary analgesic e ect, but it has opioid-sparing, opioid
• Oxygen saturation below 90% or decrease o more than 5%
T
rom baseline in patients with baseline oxygen saturation o tolerance-reversing, and antihyperalgesic e ects. Ketamine, in addi-
E
<90%. tion to being an NMDA antagonist, interacts with opioid and other
R
receptors, and thus it has true analgesic properties in addition to
4
AND
the NMDA class e ects. Ketamine use improves pain scores and has
8
• Respiratory rate less than 8 breaths per minute.
an opioid-sparing e ect o up to 50% although there are equivocal
Primary, nonpharmacologic treatments of respiratory bene ts in reduction o opioid side e ects. The alpha-2 agonist,
P
depression clonidine, has analgesic and opioid-sparing e ects. Dexmedetomi-
e
r
• I patient is taking e ective breaths but at a rate o <8 per
i
dine has documented opioid-sparing e ects when given by IVin u-
o
p
minute: sion. It may have analgesic e ects as well, although this e ect may
e
r
Tactile and verbal stimulation, naloxone administration may only occur at sedating doses, restricting its use to sedated intensive
a
t
not be essential.
i
care patients. Glucocorticoids are used in cancer pain management
v
e
• I patient is taking ine ective breaths and/or with a respiratory to reduce inf ammation rom tumor invasion o nerves.
P
rate <4 per minute: Benzodiazepines may reduce the insomnia and anxiety that
a
i
n
May require ventilatory assist with bag-valve mask and o ten accompany acute pain. However, these agents do not have
M
supplemental oxygen. This should be instituted while analgesic properties. They must be used with extreme caution in
a
n
diluting and administering naloxone. acute pain, especially when high doses o opioids are required, as
a
signi cant sedation and respiratory depression can occur in the
g
Naloxone should only be considered in the following
e
situations benzodiazepine-naïve patient. In the anxious patient with pain,
m
adequate titration with analgesics should occur be ore the addition
e
• Patient is unarousable or minimally arousable to tactile/verbal
n
stimulation. o a benzodiazepine.
t
• Patient is requiring ventilatory assistance.
Proper naloxone dilution and dosing ■ ACUTE PAIN IN THE OPIOID TOLERANT
• 1 ampule (0.4 mg) o naloxone must be diluted with 9 mL When opioid-tolerant patients experience an event resulting in pain
saline to yield 0.04 mg/mL. escalation, opioid use is expected to be higher than mere replace-
• Administer to patient in 1-2 mL increments (0.04-0.08 mg) at ment o what the patient was receiving be ore. The additional
2-3 min intervals until response. doses o opioids required may be much higher than in opioid-naïve
• I no change in respiratory depression a ter 0.4 mg naloxone patients. More complaints o pain and high pain scores should be
has been titrated, consider another etiology other than opioid expected. Discussion o reasonable goals and expectations o anal-
induced. gesic therapy with the patient is crucial. Multimodal therapy in this
• I there is some, but not enough, improvement a ter 0.4 mg o patient population is help ul to achieve the best pain control, and
naloxone has been titrated, continue titration. have the least escalation o home opioid dose as possible.
• Naloxone’s hal -li e is less than most o the opioid agonists,
so be aware that rebound respiratory depression may recur.
There ore, be prepared or the need to readminister naloxone CONSULTATION
boluses or consider use o a naloxone in usion. Involvement o a pain specialist may be appropriate in the patient
with severe pain that remains uncontrolled a ter several escalations
o drug doses and use o multiple classes o agents. Concomitant
psychiatric illness may warrant input rom a psychiatrist or psycholo-
etiology is de nitely due to histamine release, which is usually only gist. Certain diagnostic tests, such as diagnostic epidural injections,
the case or large doses o morphine given quickly, or a true aller- require an interventional pain physician. Physical therapy, surgery,
gic reaction. Nalbuphine 5 mg IV every 4 hours as needed is more complementary therapies, or invasive treatment modalities, such as
e ective in that it treats the cause, by antagonism o the central mu epidural injections, intrathecal pumps, and spinal cord stimulators,
receptors. Sedation may be a troublesome side e ect, particularly will require re erral to the appropriate provider.
when using opiates to alleviate persistent pain in terminal illness.
The proper treatment o respiratory depression rom opioid agonists DISCHARGE CONSIDERATIONS
is described in Table 48-13.
Communication with the patient’s primary care provider about the
discharge analgesic plan is essential, especially i the prior analgesic
■ ADJUVANT ANALGESICS regimen has been changed. Follow-up should be arranged to ensure
Adjunctive agents are use ul or additional analgesia especially in e ectiveness o the analgesic regimen, monitor or side e ects a ter
opioid-tolerant patients, and have a particular role in the treatment discharge, and taper the patient o analgesics, or reduce doses to
o neuropathic symptoms and chronic pain. Examples o adjuvant baseline i the patient was on chronic analgesics be ore.
analgesics with dosing guidelines and common side e ects are
listed in Table 48-14. ■ WEANING OPIOIDS
The most commonly used antiepileptic agents are gabapentin I the cause o pain is gone, patients need discontinuation o
and pregabalin. They are e ective or neuropathic pain and may opioids in a manner that prevents the occurrence o withdrawal
have bene ts in the acute setting as well. Antidepressants are also symptoms, such as abdominal pain, diarrhea, tachycardia, vomiting,
e ective in neuropathic pain. Analgesic doses are lower than those diaphoresis, runny nose, muscle cramps, piloerection, anxiety, and
or depression treatment, and the onset o analgesia is aster (days) irritability. When weaning a patient rom long-acting agents, the
321
TABLE 48-14 Select Adjuvant Analgesics
Class Agent Adult Dosing Side Effects/Comments

P
A
Antiepileptics Gabapentin Start with 300 mg orally every 8 h, increase Dizziness and somnolence; do not stop abruptly
R
by 300 mg daily a ter a ew days to a max o
T
3600 mg/d in divided doses
I
Pregabalin Start with 50 mg orally every 8 h or 75 mg
I
orally every 12 h; in 1 wk increase to max o
300 mg/d in divided doses
Tricyclic Amitriptyline 25 mg orally every night at bed time; Anticholinergic symptoms (eg, dry mouth,
M
antidepressants Nortriptyline increase to max o 150 mg/d in a single or con usion, sedation, and hypotension)
e
divided doses
d
i
Local Lidocaine 2.5% and 2-2.5 g per 10-25 cm 2 skin or 1-2 h be ore Localized skin reactions; rare cardiovascular
c
a
anesthetics prilocaine 2.5% procedure and/or CNS toxicity; prilocaine may contribute
l
C
cream to methemoglobinemia in patients treated with
o
other agents known to cause this
n
s
u
Lidocaine patch 5% Up to three patches or up to 12 h within a Localized skin reactions; rare cardiovascular
l
t
24-h period and/or CNS toxicity; only FDA indication is or
a
t
treatment o postherpetic neuralgia
i
o
n
Glucocorticoids Dexamethasone 4-8 mg orally every 8-12 h Typical steroid-induced side e ects rom
long-term use (>2-3 mo) usually outweigh
bene its; concomitant use with NSAIDs not
recommended
Skeletal muscle Cyclobenzaprine 5-10 mg orally every 8 h Long-term use can lead to the development o
relaxants dependence
Tizanidine 4-8 mg orally every 6-24 h
Orphenadrine 100 mg orally every 12 h
60 mg IVevery 12 h
Antispasmodic Baclo en 10 mg orally every 8 h, titrate slowly to max Drowsiness; may impair renal unction; abrupt
o 80 mg/d in divided doses discontinuation may cause seizures
NMDA Ketamine 0.1-0.2 mg/kg/h IV Sedation, dreams, and hallucinations possible
antagonists but in requent at analgesic (low) dose, treat
with the addition o benzodiazepine or dose-
reduction
Dextromethorphan Start with 30-90 mg orally every 8 h, increase Best dose and regimen not well de ined
to max o 360 mg/d in divided doses
Alpha-2 agonist Clonidine 0.2 mg/d via a transdermal patch, le t on or Hypotension and sedation; monitor or rebound
1 wk hypertension on discontinuation i used or >1 wk
CNS, central nervous system; FDA, US Food and Drug Administration; NSAID, nonsteroidal anti-in lammatory drug.
dose o the long-acting agent should be decreased by 25% to 50% SUGGESTED READINGS
every 2 days. Once the patient is o the sustained-release orm, the
immediate-release agent can also be weaned. In weaning a patient American Pain Society. Principles of Analgesic Use in the Treatment of
rom immediate-release agents, the opioid dose should be reduced Acute Pain and Cancer Pain, 6th ed. Glenview, IL: American Pain
by 50% or 2 days, and then reduced by 25% every 2 days therea ter Society; 2008.
until the total dose in oral morphine equivalents is 30 mg/d. The Australian National Health and Medical Research Council’s Acute pain
drug may be discontinued a ter 2 days on the 30 mg/d dose. management: scienti c evidence website. https://www.nhmrc.gov
.au/guidelines-publications/cp104. Accessed May 26, 2015.
QUALITY IMPROVEMENT Gordon DB, Dahl JL, Miaskowski C, et al. American Pain Society
Data on pain management quality should be collected periodically recommendations or improving the quality o acute and cancer
to assess the quality o care, to establish baseline data, and to iden- pain management. Arch Intern Med. 2005;165:1574-1580.
ti y areas in which care can be improved. The American Pain Society Gruener D, Lande SD, eds. Pain Control in the Primary Care Setting.
has proposed the ollowing six quality indicators or hospital-based Glenview, IL: American Pain Society; 2006.
pain management: (1) pain intensity is documented with a numeric
Morrison RS, Meier DE, Fischberg D, et al. Improving the management
or descriptive rating scale; (2) pain intensity is documented re-
o pain in hospitalized adults. Arch Intern Med. 2006;166:1033-1039.
quently; (3) pain is treated by a route other than intramuscular; (4)
pain is treated with regularly administered analgesics, and whenever United States Food and Drug Administration/Center or Drug
possible, a multimodal approach is used; (5) pain is prevented and Evaluation and Research website. http://www. da.gov/Drugs/.
controlled to a degree that acilitates unction and quality o li e; and Accessed May 26, 2015.
(6) patients are involved in the treatment plan and are in ormed and Whelan CT, Jin L, Meltzer D. Pain and satis action with pain control in
knowledgeable about pain management. hospitalized medical patients. Arch Intern Med. 2004;164:175-180.
322
SECTION 3
Perioperative Risk Assessment
and Management
323
CHAP TER
49 INTRODUCTION
Medical consultation has become an important component o
Hospital Medicine. These consultations include preoperative evalu-
ation, perioperative management, and medical care o patients on
various nonmedical services. Previous surveys ound that many
primary care physicians and hospitalists elt inadequately trained in
perioperative medicine, and as a result, this area received additional
emphasis as part o the core competencies or Hospital Medicine.
With the growth o the hospitalist movement, the role o the con-
Role of the Medical sultant has evolved rom providing evaluation and advice to include
comanagement o the patient in certain settings. The goal o this
Consultant chapter is to review the role and responsibilities o the medical con-
sultant, ocusing on the principles o consultation and techniques to
improve e ectiveness.
Steven L. Cohn, MD, FACP, SFHM GENERAL PRINCIPLES OF CONSULTATION

More than 25 years ago, Goldman and colleagues described the
concepts or per orming medical consultations. His “Ten Command-
ments” or e ective consultation included the ollowing:
1. Determine the question.
2. Establish urgency.
3. Look or yoursel .
4. Be as brie as appropriate.
5. Be specif c and concise.
6. Provide contingency plans.
7. Honor thy tur .
8. Teach with tact.
9. Talk is cheap and e ective.
10. Follow-up.
These concepts, which incorporated many o the ethical prin-
ciples described by the American Medical Association (AMA), are
important and remain valid or the traditional consultation. How-
ever, some modif cations are necessary to cover the new role o
hospitalists as comanagers.
■ TYPES OF CONSULTATION
The traditional or standard medical consultation consisted o a
ormal request rom the requesting physician to evaluate a patient
and answer a specif c question (Table 49-1). The consultant was
expected to address the question and to provide advice and
recommendations, but not to write orders or bring in other con-
sultants; the requesting physician remained in control and respon-
sible or the patient’s overall care and treatment. The consultant
also ocused on the specif c problem rather than looking or and
addressing other issues. Consultations were requested only when
necessary and not or routine management. The ollow-up period
was usually brie and did not involve daily visits or the duration o
hospitalization.
This traditional role o the consultant has been changing over the
past 5 to 10 years. A survey by Salerno and colleagues revealed that
many surgeons wanted the medical consultant to assume more
o a comanagement role. Specif cally, they wanted the consultant
to address all medical issues as necessary as well as to write orders
and continue to ollow the patient. Comanagement arrangements
have most o ten been with orthopedic surgeons and more recently
with neurosurgeons. Comanagement has potential advantages o
325
TABLE 49-1 Roles and Responsibilities of Different Types on Consultations
Traditional Comanagement Curbside

P
A
MD in charge overall Requesting physician Shared responsibility Requesting physician
R
Primary care of medical Requesting physician Medical consultant Requesting physician
T
problems Surgical—requesting physician
I
I
Question addressed Speci ic Broader issues—other medical Should not address either but o er to do
problems ormal consult or give only general advice
Order writing No Yes No
Follow-up Limited-as needed Daily until discharge No—no ormal relationship
M
e
d
i
c
a
l
C
o
decreasing length o stay and reducing complications. Surgeons that ails to answer the question being asked by the requesting
n
physician.
s
and nurses o ten pre er comanagement; however, one possible dis-
u
l
t
advantage is that the comanaging consultant may eel subservient
a
t
to the surgeon and may be asked to assume responsibilities outside
i
o
PRACTICE POINT
n
his area o training.
Yet another type o consultation is the so-called curbside or In view o the multiple types o consultations, it is imperative that
in ormal consult in which the consultant is asked to provide an the requesting physician speci y:
opinion or advice without personally seeing the patient. Although
• The expected role o the consultant
these should be avoided rom a medicolegal standpoint, they occur
requently. Ideally the consultant should o er to per orm a ormal • The question to be answered by the consultant
consult but i any advice is given, it should be generic and simple. I there is any uncertainty, the consultant should clari y this
The requesting physician should not re er to the consultant in the question by communicating directly with the requesting
medical record i he has not seen the patient, and i he has had any physician. The consultant should avoid making recommendations
contact with the patient, the consultant should write a note in the about the type o anesthesia and other areas outside his or her
chart. area o expertise.
PRACTICE POINT ■ ANSWERING THE QUESTION

I the consultant is asked to provide an opinion or advice without Traditionally, the consultant restricted his or her advice to the spe-
personally seeing the patient (the “curbside consult”), the cif c problem or question. However, more requently the consultant
consultant should: is addressing other issues noted during the evaluations. Assuming
these other f ndings and recommendations are relevant and impor-
• O er to per orm a ormal consult. tant, most surgeons are in avor o this approach. What the request-
• Provide only generic and simple advice. ing physician does not want is a laundry list o things to do or minor
problems or issues that do not need to be addressed during the
• Document any patient encounter in the chart. current hospitalization.
The requesting physician should not re er to the consultant in the I the consultation is or preoperative evaluation, the consultant
medical record i the consultant has not seen the patient. needs to:
1. Assess the severity and degree o control o the patient’s medi-
cal problems.
■ DETERMINING THE QUESTION 2. Estimate surgical risk.
3. Determine i the patient is in his or her optimal medical condi-
In view o the multiple types o consultations, it is imperative that
tion or surgery.
the requesting physician speci y exactly what is being requested,
4. Decide whether urther tests or interventions are indicated.
and i there is any uncertainty, the consultant should clari y this
5. Make recommendations regarding the patient’s medications
question by communicating directly with the requesting physi-
and any necessary prophylaxis.
cian. In addition to speci ying the role o the consultant, the
requesting physician should be speci ic as to the question being The consultant should avoid making recommendations about
asked o the consultant. For example, a request or preoperative the type o anesthesia and other areas outside his or her area o
consultation may be or surgical risk assessment, a “green light” expertise. Also, the consultant should re rain rom using the term
to proceed with anesthesia and surgery, a diagnostic or man- “cleared or surgery,” even i consulted or that reason, as this implies
agement issue, reassurance, or documentation or medicolegal a guarantee that the patient will not have a complication.
purposes. As obvious as this may be, disagreement regarding the
primary purpose or the consult still occurs between the request- ■ OPTIMIZING EFFECTIVENESS
ing physician and the consultant. Several studies noted that the
consult requests were vague and nonspeci ic (eg, clearance or Factors influencing or improving compliance
evaluation), or did not even ask a question. Without clari ying the Various studies ound a number o actors that have been associated
reason or the consult, the consultant may respond in a manner with improved compliance with the consultant’s recommendations
326
4. Follow-up : Appropriate ollow-up visits will reassess the
TABLE 49-2 Factors that Influence or Improve Compliance patient’s condition and ensure that recommendations were
C
with Consultant Recommendations ollowed. The consultant should clearly document his ind-
H
A
Prompt response (within 24 h) ings and update recommendations in the medical record.
P
There is no standard regarding how o ten the consultant
Limit number o recommendations (≤ 5)
T
needs to see the patient, but this should be determined
E
Identi y crucial or critical recommendations (vs routine) by the patient’s medical condition, type o surgery, and
R
Focus on central issues whether the requesting physician wants comanagement
4
Make speci ic relevant recommendations or not. When the patient is medically stable and there is no
9
Use de initive language longer a need or the medical consultant, he should sign
o and document this in the chart. Recommendations and
Speci y drug dosage, route, requency, duration
R
arrangements or long-term ollow-up can also be noted at
o
l
Frequent ollow-up including progress notes
e
this time.
o
Direct verbal contact
t
Therapeutic (vs diagnostic) recommendations
h
e
Severity o illness PRACTICE POINT
M
e
The consultant should document:
d
From Cohn SL, Macpherson DS. Overview o the principles o medical con-
i
c
sultation. In: Basow DS, ed. UpToDate. Waltham, MA: UpToDate; 2009; with • Specif c and precise recommendations listed in order o
a
l
permission. importance
C
o
•
n
Name, initial dose, requency, route o administration, titration,
s
u
and duration o recommended therapy
l
t
(Table 49-2). In general, ollowing Goldman’s Ten Commandments
a
n
or Salerno’s modif cation (see Salerno SM, Hurst FP, Halvorson S, The consultant should provide:
t
Mercado DL. Principles o e ective consultation: an update or the • Prompt service
21st-century consultant. Arch Intern Med. 2007;167(3):271-275) will
result in e ective consultation. • Direct verbal communication with the requesting physician
Determine and clari y the question: As noted, the reason or the upon completion o the initial consult
consultation needs to be clearly def ned by the requesting physician • Updates and ollow-up as appropriate depending on
and understood and addressed by the consultant. requested role
Punctual response: The consultant should be available to respond
in a timely ashion, depending on the urgency o the consultation.
Truly “stat” consults should be answered in less than 30 minutes, and
in general, elective consults should be answered within 24 hours, CONCLUSION
pre erably the same day they were requested.
Recommendations: The ideal medical consultant will “render a report that in orms
without patronizing, educates without lecturing, directs without
1. Prioritize and limit: The consultant should make specif c, ordering, and solves the problem without making the re erring
precise recommendations that should be listed in order o physician appear to be stupid.” It is hoped that by ollowing these
importance. Crucial or critical recommendations are more principles, the medical consultant will be e ective in providing
likely to be ollowed, as are those at the top o the list. For use ul in ormation and recommendations to the requesting phy-
this reason, it was previously elt that the number o recom- sician who will then implement them in an attempt to improve
mendations should be limited to no more than f ve, but more patient outcome.
recently the eeling is to leave as many recommendations as
needed to answer the consult and o er to help with writing
and implementing them (comanagement). Therapeutic rec- SUGGESTED READINGS
ommendations are more likely to be ollowed than diagnostic
ones. Choi JJ. An anesthesiologist’s philosophy on “medical clearance” or
2. Language: The consultant should use def nitive language, be surgical patients. Arch Intern Med. 1987;147(12):2090-2092.
specif c with his recommendations, and provide contingency Cohn SL, Macpherson D. Overview o the principles o medical
plans. For example, recommendations or medications should consultation. In: Basow D, ed. UpToDate. Waltham, MA: UpToDate;
speci y the drug name, dose, requency, route o adminis- 2015.
tration, and duration o therapy. The requesting physician Devor M, Renvall M, Ramsdell J. Practice patterns and the adequacy
should be told what response to expect, how long it will take, o residency training in consultation medicine. J Gen Intern Med.
as well as how and when to adjust the medication dose i 1993;8(10):554-560.
necessary.
Goldman L, Lee T, Rudd P. Ten commandments or e ective consul-
3. Communication: Direct verbal communication with the
tations. Arch Intern Med. 1983;143(9):1753-1755.
requesting physician is crucial and pre erable to just leaving
a note in the chart. A quick call to the requesting physician Kleinman B, Czinn E, Shah K, Sobotka PA, Rao TK. The value to the
will let him know that the consult has been answered, what anesthesia-surgical care team o the preoperative cardiac consul-
the recommendations are, and what needs to be done so tation. J Cardiothorac Anesth. 1989;3(6):682-687.
the orders can be written and the process expedited. It is also Kuo D, Gi ord DR, Stein MD. Curbside consultation practices and
important to communicate with other members o the health attitudes among primary care physicians and medical subspecial-
care team to coordinate care. ists. JAMA. 1998;280(10):905-909.
327
Lee T, Pappius EM, Goldman L. Impact o inter-physician communi- Rudd P, Siegler M, Byyny RL. Perioperative diabetic consultation: a
cation on the e ectiveness o medical consultations. Am J Med. plead or improved training. J Med Educ. 1978;53(7):590-596.
1983;74(1):106-112. Salerno SM, Hurst FP, Halvorson S, Mercado DL. Principles o e ec-
P
Plauth WH, Pantilat SZ, Wachter RM, Fenton CL. Hospitalists’ percep- tive consultation: an update or the 21st-century consultant. Arch
A
tions o their residency training needs: results o a national survey. Intern Med. 2007;167(3):271-275.
R
Am J Med. 2001;111(3):247-254.
T
I
I
M
e
d
i
c
a
l
C
o
n
s
u
l
t
a
t
i
o
n
328
CHAP TER
50 INTRODUCTION
Hospitalists and internists are requently called upon to per orm
preoperative medical consultations, and cardiac risk assessment is
what is most o ten requested. Preoperative evaluation is now part
o the core curriculum or Hospital Medicine, but when surveyed
a number o years ago, many hospitalists elt inadequately trained
to do this.
Preoperative cardiac risk assessment has evolved over the past
40 years rom a simple global assessment o a patient’s physical
Preoperative Cardiac status (the ASA classi cation) to multivariate risk analyses (Goldman,
Detsky) to a simpli ed scoring system (Lee RCRI) to guidelines rom
Risk Assessment the American College o Cardiology/American Heart Association,

American College o Physicians (ACC/AHA, ACP). The most current
and Perioperative o these is the 2014 ACC/AHA guidelines or perioperative cardiac

evaluation and management, which was originally published in
1996 and incorporated the RCRI actors in the 2007 update. Using
Management these guidelines and selective cardiac testing (pharmacologic stress
tests), physicians are now better able to provide a more accurate
assessment o perioperative risk, and the ocus has turned to risk
reduction strategies. These include revascularization by coronary
Steven L. Cohn, MD, FACP, SFHM artery bypass gra ting (CABG) or percutaneous coronary interven-
tion (PCI), medical therapy (β-blockers, α-agonists, statins), and other
intraoperative measures (normothermia, anesthetic technique).
Although surgical and anesthetic techniques have improved and
perioperative cardiac events have decreased, operative mortality
and cardiac morbidity remain signi cant, especially among high-risk
patients or high-risk procedures.
This chapter reviews the current state o the art or periopera-
tive risk assessment and risk reduction in patients with cardiac
disease.
PREOPERATIVE RISK INDICES

Goldman and colleagues published the rst large prospective
multivariate analysis o preoperative cardiac risk. They identi ed
nine independent predictors o death or major postoperative
cardiac complications. These risk actors were assigned points
based on their relative importance, and the event rates were
correlated with the point total in this risk index. Detsky and col-
leagues modi ed this risk index by expanding the list o risk actors
and combining this with the pretest probability o complications
based on the risk o the surgery itsel . Eagle and colleagues identi-
ed ve actors—age, diabetes mellitus (DM), angina, myocardial
in arction (MI), and heart ailure—associated with perioperative
cardiac events and used these to strati y risk and decide when to
do urther cardiac testing.
Lee and colleagues identi ed and validated six actors associated
with increased risk o perioperative complications. These actors
were high-risk surgery, coronary artery disease (CAD), heart ailure,
cerebrovascular disease (stroke or transient ischemic attack), DM
requiring insulin, and renal insu ciency (creatinine >2.0 mg/dL).
These studies, using simple clinical evaluation (history, physical
examination, and basic laboratory studies) ound many similar ac-
tors predicting increased risk o perioperative cardiac complications
and helped re ne preoperative risk strati cation.
The newest risk calculators were derived rom the National
Surgical Quality Improvement Program (NSQIP) database. The Gupta
calculator (http://www.surgicalriskcalculator.com/miorcardiacarrest)
was derived rom over 200,000 patients and validated on another
329
250,000 patients in the database. It estimates risk o MI or cardiac FUNCTIONAL CAPACITY
arrest based on ve independent predictors: type o surgery,
Goldman and colleagues noted that patients with good exercise
dependent unctional status, abnormal creatinine, American Society
capacity, even with mild, stable angina, tend to do well. This ollows
P
o Anesthesiologists (ASA) class, and increasing age. The American
the concept o the ischemic threshold in which a patient develop-
A
College o Surgeons (ACS) surgical risk calculator (http://riskcalcula-
ing ischemia on a stress test at a lower exercise level and with a
R
tor. acs.org) was derived rom over 1.4 million patients in the NSQIP
T
lower rate-pressure product is at higher risk than someone who can
database. It is more comprehensive and is based on the CPT code
I
per orm 8 to 10 metabolic equivalents (METS) be ore developing
I
or the surgical procedure and an additional 20 variables. It predicts
symptoms. Reilly and colleagues ound that a patient’s sel -reported
mortality, serious complication, cardiac complications, and several
exercise capacity correlated with the risk o postoperative compli-
others and states whether these risks are average, above average,
cations, and the ACC guidelines use this in their risk assessment
or below average.
algorithm.
M
e
d
i
CLINICAL RISK FACTORS PRACTICE POINT
c
a
l
A detailed history and ocused physical examination are key in •
C
Because a patient’s sel -reported exercise capacity correlates
o
clinical risk assessment, and a ew basic diagnostic tests may also with the risk o postoperative cardiac complications, clinicians
n
be help ul. Current risk assessment is usually based on the ACC/
s
should actor unctional capacity into their risk assessment.
u
AHA guidelines which now include the Lee RCRI and the NSQIP risk
l
t
a
calculators.
t
i
o
The 2014 ACC/AHA guideline algorithm is now speci cally or
n
evaluation o patients with CAD and there ore removed heart ail- LABORATORY TESTS
ure, arrhythmias, and valvular disease, recommending that those Many preoperative screening blood tests are per ormed unneces-
conditions be evaluated and managed based on current guideline sarily, but a ew may be help ul in assessing cardiac risk. These
directed medical therapy (GDMT). O the previous “active cardiac include measures o renal unction, blood urea nitrogen (BUN) and
conditions,” only the presence o an acute coronary syndrome (ACS) creatinine, and glucose (as a screen or diabetes). Unless serum
remains in the algorithm. potassium is signi cantly abnormal (<3.0 or >5.5 mEq), it is unlikely
to increase risk or alter management. Anemia has been noted as a
risk actor in some studies, but there is no evidence that treating it
PROCEDURAL RISK
with trans usions alters risk. An electrocardiogram (ECG) looking or
Independent o the patient’s clinical risk actors, the type o surgery evidence o CAD or conduction de ects may be o value in at-risk
has its own inherent risk that needs to be taken into account. This patients. Other ndings can either be identi ed by clinical exam
concept was used in Detsky’s modi ed cardiac risk index and is also (arrhythmias) or do not change management (le t ventricular hyper-
considered in the ACC algorithm. For example, a patient undergoing trophy [LVH], nonspeci c ST-T changes).
cataract surgery, a low-risk operation, is unlikely to have a complica-
tion even i the patient’s clinical risk is high. Conversely a patient
2014 ACC/AHA ALGORITHM
with no clinical risk actors undergoing high-risk surgery, such as a
Whipple procedure, is more likely to have a postoperative compli- The new ACC guidelines use a stepwise approach to preoperative
cation than would have been predicted based on clinical pretest cardiac risk evaluation or patients with CAD or risk actors or it
probability alone. There ore, the risk o the surgery itsel may alter using the in ormation obtained rom the history, physical examina-
management and in uence the decision to do urther testing. The tion, and laboratory tests (Figure 50-1). The underlying theme is to
2014 ACC guidelines now de ne surgical risk as either low risk (<1% minimize testing and not to order a test i the result will not change
MACE) or elevated risk (>1% MACE) based on a combined procedural management. The approach is as ollows:
and clinical risk. The 2007 ACC guidelines de ned three groups 1. Is the surgery emergent (or urgent, meaning within 24 hours)?
or surgical risk—vascular, intermediate, and low. The previous I it is, time does not permit diagnostic testing or revascularization
designation o high risk was changed to vascular surgery to re ect and the patient will proceed to surgery. In the short time period
that the preponderance o evidence or cardiac testing was done available, the physician can try to medically optimize the patient’s
or patients undergoing aortic and major vascular surgery, and the problems.
approach to these patients may be somewhat dif erent than that 2. Assuming surgery is not emergent, has the patient had an
or nonvascular surgery. The intermediate risk category includes acute coronary syndrome (ACS) (as opposed to any o the
most intrathoracic, intra-abdominal, head and neck, orthopedic, and active cardiac conditions in the 2007 guidelines)?
urologic procedures as well as some lower-risk vascular procedures
I so, elective surgery should be delayed or urther diagnos-
such as carotid endarterectomy and endovascular abdominal aortic
tic workup and therapy. Most patients do not have these
aneurysm repair. Low-risk surgery includes procedures not invading
conditions.
the chest or abdomen such as endoscopic or super cial procedures,
3. Is the estimated perioperative risk o MACE based on combined
eye surgery, and breast surgery. The risk indices recommended by
clinical/surgical risk low (<1%)?
the new ACC/AHA guidelines incorporate the type o surgery and
those using the NSQIP database include even more speci c data I it is, the patient should proceed to surgery without any urther
based on procedural risk. testing or intervention because we cannot urther reduce risk that
is already low (<1%).
PRACTICE POINT 4. I elevated risk, does the patient have moderate or greater
unctional capacity (≥4 METS)?
• The ACC/AHA guidelines now classi y perioperative risk as low
(<1% complication rate) or elevated (>1%) based on combined I so, or the most part, these patients will do well and do not need
clinical and procedural risk. They suggest using either the RCRI to undergo urther cardiac testing.
or NSQIP calculators to determine risk. 5. I the patient has poor or unknown exercise capacity, the next
question is whether or not urther testing will impact decision
330
Pa tie nt s che dule d for s urge ry with
C
known or ris k fa ctors for CAD*
H
(S te p 1)
A
P
T
E
R
Ye s Clinica l ris k s tra tifica tion
5
Eme rge ncy
a nd proce e d to s urge ry
0
No
P
r
e
o
p
e
ACS † Ye s Eva lua te a nd tre a t
r
a
(S te p 2) a ccording to GDMT†
t
i
v
e
C
No
a
r
d
i
a
Es tima te d pe riope ra tive ris k of MACE
c
ba s e d on combine d clinica l/s urgica l ris k
R
No furthe r
i
s
(S te p 3) te s ting
k
Exce lle nt (Clas s lIa)
A
(>10 METs )
s
s
e
s
Mode ra te
s
m
of gre a te r Proce e d to
Low ris k (<1%) Eleva te d ris k
e
(≥4 METs )
n
(S te p 4) (S te p 5) s urge ry
t
functiona l
a
ca pa city
n
Mode ra te /Good
d
(≥4–10 METs )
P
No furthe r
e
No furthe r
r
No or te s ting
i
o
te s ting unknown (Clas s lIb)
p
(Clas s lII:NB)
e
r
a
t
i
v
e
Poor OR
M
Proce e d to unknown
a
functiona l ca pa city
n
s urge ry Pharmac o lo g ic
a
(<4 METs ): Ye s
g
s tre s s te s ting
e
Will furthe r te s ting impa ct
(Clas s IIa)
m
de cis ion ma king OR
e
pe riope ra tive ca re ?
n
If If
t
(S te p 6)
norma l a bnorma l
No Co ro nary
revas c ularizatio n
Proce e d to s urge ry ac c o rding to exis ting
a ccording to GDMT OR CPGs (Clas s I)
a lte rna te s tra te gie s
(noninva s ive tre a tme nt,
pa llia tion) (S te p 7)
Figure 50-1 ACC/AHA Stepwise approach to perioperative cardiac assessment for CAD. (Reproduced, with permission, rom Fleisher LA,
Fleischmann KE, Auerbach AD, et al. American College o Cardiology; American Heart Association. 2014 ACC/AHA guideline on perioperative
cardiovascular evaluation and management o patients undergoing noncardiac surgery: a report o the American College o Cardiology/
American Heart Association Task Force on practice guidelines. J Am Coll Cardiol. 2014;64(22):e77-e137.)
making or perioperative care. This decision should be based procedure, a nonsurgical treatment, or palliation. I the test is
on a discussion with the physician, patient, and perioperative normal, proceed to surgery according to GDMT.
team. 6. I urther testing will not impact decision making or manage-
I the answer is yes, then pharmacologic stress testing is indicated. ment, then proceed to surgery according to GDMT or consider
I the stress test is abnormal, coronary angiography and revascu- alternative as previously mentioned.
larization may be indicated based on the degree o abnormality, These last two steps allow or individualizing management, but
and a ter that, the patient can proceed to surgery ollowing are somewhat vague and may lead to signi cant dif erences in
GDMT. Other alternatives include changing to a lesser surgical opinion and approach.
331
DIAGNOSTIC TESTS high-risk patients, and stress test results only changed the prob-
ability o complications (con rmed by outcomes) or intermediate-
PRACTICE POINT risk patients. Boersma and colleagues also demonstrated that the
P
clinical risk score correlated with outcomes and was rarely changed
•
A
The underlying theme o preoperative testing is to minimize by stress testing. However, a study by Poldermans and colleagues
R
unnecessary testing and not to order a test i the result will randomized intermediate risk patients, all o whom were treated
T
not change management. A preoperative stress test may be with bisoprolol (titrated to a target heart rate o 60 to 65 beats per
I
considered or patients with a low unctional capacity
I
minute), to preoperative stress testing or no testing. They ound no
(<4 METs) undergoing an intermediate-to-high risk procedure, dif erence in perioperative MI or cardiac death regardless o whether
but only i it will change management. they had a preoperative stress test assuming they were adequately
β-blocked. The results o this study beg the question o whether
M
anyone needs NIT i they can be optimally treated medically. How-
e
Stress tests were designed to diagnose CAD and myocardial isch-
d
ever, questions have been raised about the scienti c integrity and
i
emia and not to predict short-term events. Because the patho-
c
results o this and other β-blocker studies (DECREASE trials) by the
a
physiology o perioperative MIs includes unstable plaque rupture
l
Poldermans group.
C
or coronary thrombosis as well as myocardial ischemia related to
o
n
coronary stenosis and an imbalance o oxygen supply and demand,
■ WHEN IS PREOPERATIVE CORONARY ANGIOGRAPHY
s
u
stress tests are poor at predicting which patients are at increased
l
INDICATED?
t
a
risk or surgery. Although they are good at identi ying a patient
t
The ultimate goal o the preoperative evaluation process is to iden-
i
o
with CAD, they have a low positive predictive value (PPV), usually
n
between 15% and 20% or perioperative MI, and most patients, ti y the high-risk patient and intervene in some manner to reduce
even with an abnormal test, will not have a postoperative cardiac risk o postoperative complications. Assuming stress testing identi-
complication. On the other hand, a normal or negative stress test is es a patient at high risk or postoperative complications, the next
usually associated with a high negative predictive value (NPV), rang- step should be to urther de ne risk using cardiac catheterization
ing rom 95% to 99%, indicating that there is a low likelihood that with a goal o possible revascularization. Otherwise, i the patient is
these patients will have a perioperative event. to be treated medically, the stress test was probably unnecessary as
Because many patients with cardiac disease undergoing surgery it did not change management. Potential candidates or coronary
have suboptimal exercise capacity and would be unable to achieve angiography are those high-risk patients with demonstrated isch-
85% o their target heart rate on an exercise test, pharmacologic emia on stress testing or those with unstable coronary syndromes
stress testing is usually used. Furthermore, i these patients had (recent MI, severe angina, or unstable angina) whose clinical risk is
adequate exercise capacity, they probably would not be candidates high enough to bypass stress testing and who have independent
or stress testing in the rst place. The tests most commonly used criteria or coronary angiography independent o their need or non-
are dobutamine stress echocardiography (DSE) and dipyridamole, cardiac surgery. I coronary angiography demonstrates signi cant
adenosine, or regadenoson nuclear imaging (either thallium or anatomic lesions, a decision must be made regarding revasculariza-
technetium). These tests are ef ective in identi ying CAD, but as tion options—PCI or CABG.
noted earlier, are poor at identi ying patients who will develop
postoperative cardiac events and should be used selectively in RISK REDUCTION STRATEGIES
conjunction with the patient’s pretest probability or the results to Once a patient has been identi ed as being at high risk, either by
be meaning ul. The test characteristics are in uenced by patient clinical evaluation or a ter stress testing or coronary angiography,
selection and pretest probability. In general, results are similar with the next step is to take measures to reduce that risk. The two main
DSE and dipyridamole thallium (DPT), and test selection should be options include revascularization or medical therapy or both. The
based on the local expertise available; however, DSE tends to have question is whether these therapies are ef ective.
ewer alse-positives except in the case o le t bundle branch block
(LBBB) where DPT is pre erred. On the other hand, DSE is pre erred ■ CORONARY REVASCULARIZATION
or patients with asthma or chronic obstructive pulmonary disease Results rom the Coronary Artery Surgery Study (CASS) showed that
(COPD) because DPT can cause bronchospasm. The role o cardiac patients who underwent CABG, were symptom ree, and then went
CT or MRI angiography in the perioperative setting is unclear. Rest- on to have noncardiac surgery had a lower risk o postoperative
ing two-dimensional (2D) echocardiography is not recommended mortality and non atal MI than similar study patients treated medi-
to predict perioperative ischemic complications and should only be cally. This bene t was only or high-risk surgery, and the protective
used to evaluate valvular heart disease or heart ailure. ef ect o CABG appeared to last or 4 to 6 years. However, the mor-
bidity and mortality a ter CABG were not taken into account, and
■ WHO NEEDS STRESS TESTING BEFORE SURGERY? these patients did not have prophylactic CABG in preparation or
The goal o NIT is to urther re ne clinical judgment and identi y noncardiac surgery.
patients at signi cantly increased risk. Bayes theorem can be These results dif er rom those o the Coronary Artery Revascular-
applied to preoperative evaluation to decide which patients might ization Prophylaxis (CARP) trial in which patients with stable cardiac
bene t rom stress testing. A patient with a low pretest probability symptoms scheduled to undergo elective major vascular surgery
will usually remain at low risk or perioperative complications even were evaluated by NIT, and those with abnormal tests went on to
i the NIT is positive. Similarly, a patient with high pretest probability coronary angiography. Patients with suitable anatomy or revascu-
will remain at relatively high risk even i the stress test is negative larization were then randomized to medical therapy with or without
(may represent a alse-negative). There ore, NIT should probably revascularization (CABG or PCI). Exclusion criteria included >50%
be restricted to intermediate-risk patients where a positive test can stenosis o the le t main coronary artery, signi cant aortic stenosis,
move a patient into a higher-risk category and a negative test can or le t ventricular ejection raction (LVEF) less than 20%. Prophylactic
reclassi y a patient as being at lower risk. The utility o this theoretical revascularization was associated with a mortality o 1.7%, periopera-
approach was supported by L’Italien and colleagues who showed tive MI o 5.8%, and reoperation rate o 2.5%. In this particular study,
that clinical assessment correlated with outcomes or low- and there were no perioperative strokes, but typically this occurs in up
332
to 2%. A subgroup analysis o this study, as well as a recent meta- Early studies with prophylactic β-blockers demonstrated ben-
analysis showed that CABG appeared to be more protective than e cial ef ects. Mangano and colleagues randomized 200 patients
C
PCI, possibly because o a more complete revascularization. undergoing various operations to placebo or atenolol, started
H
Because o the associated morbidity and mortality, prophylactic
A
immediately preoperatively, titrated to a heart rate o 55 to 65 beats
P
revascularization would only be expected to bene t patients at per minute and continued or less than 7 days postoperatively. There
T
high risk undergoing high-risk surgery. However, DECREASE V, a was less ischemia in the atenolol-treated group but no reduction in
E
small study o these very high-risk patients (three or more risk actors short-term outcomes o death or non atal MI. However, surviving
R
and extensive stress-induced ischemia on DSE undergoing major patients in the β-blocker group went on to have ewer cardiovascu-
5
vascular surgery) in whom a previous study showed no bene t rom lar events by 2 years. Poldermans and colleagues randomized 112
0
perioperative β-blockers, ailed to demonstrate improved short- or patients with abnormal DSEs undergoing major vascular surgery to
long-term outcomes with revascularization in addition to optimal placebo or bisoprolol, started at least 7 days preoperatively, titrated
P
medical therapy. Criticisms o these studies raised concerns that the to a heart rate between 55 and 65 beats per minute, and continued
r
e
o
CARP trial patients did not have severe enough CAD, whereas the or at least 30 days postoperatively. The trial was stopped early
p
DECREASE Vpatients were too sick. because the bisoprolol-treated group had a signi cant reduction in
e
r
a
Another trial ound a bene t to routine versus selective cardiac postoperative MI and cardiac death (3% vs 34%). Despite the small
t
i
v
catheterization or screening patients be ore elective aortic surgery. numbers o patients in these trials and methodologic criticisms,
e
Intention-to-treat analysis showed decreased cardiac mortality and various agencies and society guidelines began recommending
C
a
a similar trend in major adverse cardiac events (MACE) in the routine prophylactic β-blockers.
r
d
cardiac catheterization group at 30 days and at ollow-up 4 years later. Three subsequent studies involving approximately 1500 patients
i
a
It was elt that prophylactic PCI, with its lower risk or adverse (POBBLE, DIPOM, MaVS) using metoprolol, started at most 1 day
c
R
events related to the procedure than CABG, might be better, but be ore surgery and not titrated to a speci c heart rate, showed no
i
s
k
there are no studies to con rm this. On the other hand, numerous bene t in various cardiovascular outcomes. Lindenauer and col-
A
studies and a meta-analysis reported an increased risk associated leagues using an administrative database o over 600,000 patients,
s
s
with noncardiac surgery soon a ter PCI. This is related to stent throm- reported that being on a β-blocker within 2 days o surgery was
e
s
bosis in patients who have prematurely discontinued the recom- associated with decreased in-hospital mortality in high- but not low-
s
m
mended course o dual antiplatelet therapy (DAPT) and, to a lesser risk patients (strati ed by RCRI score).
e
n
degree, bleeding in patients who were taking aspirin and clopido- The POISE trial was expected to resolve this controversy but
t
grel. The ACC/AHA guidelines recommend delaying elective surgery instead raised more questions. In this study over 8000 patients with
a
n
or at least 2 weeks a ter balloon angioplasty, 4 to 6 weeks a ter atherosclerotic heart disease (ASHD) or risk actors or it who were
d
placement o a bare metal stent (BMS), and 12 months a ter a drug- scheduled or various surgical procedures were randomized to
P
e
eluting stent (DES) in order to complete the course o DAPT. Should metoprolol controlled release (CR) or placebo. Patients received the
r
i
o
surgery be required be ore these time intervals in a patient who had rst dose (metoprolol CR 100 mg or placebo) 2 to 6 hours be ore
p
e
PCI, the recommended options, in priority order, are to try to per- surgery ollowed by a second dose (100 mg) within 6 hours a ter
r
a
orm the surgery on dual antiplatelet therapy i possible, to continue the end o surgery, and then a maintenance dose o 200 mg daily
t
i
v
aspirin but discontinue the second antiplatelet agent (discontinue started 12 hours a ter the postoperative dose. The drug was with-
e
clopidogrel 5 to 7 days be ore surgery, prasugrel 7 days be ore, and held or heart rate less than 45 beats per minute or systolic blood
M
a
ticagrelor 5 days be ore), or to discontinue both antiplatelet drugs pressure (BP) less than 100 mm Hg and then restarted at hal the
n
a
5 to 7 days be ore surgery. However, the 2014 ACC/AHA guidelines dose 12 hours later i BP and pulse improved. Primary outcome, a
g
now state that elective noncardiac surgery may be considered with composite o cardiac death, non atal MI, and cardiac arrest, was sig-
e
m
DES a ter 180 days o DAPT i the risk o surgical delay is greater than ni cantly better in the metoprolol-treated group (see Figure 50-1)
e
n
the expected risks o ischemia and stent thrombosis (IIb). The anti- but was primarily due to a reduction in non atal MIs. However, this
t
platelet therapy should be resumed as soon as possible a ter surgery, bene t came at the expense o a signi cant increase in stroke and
assuming adequate hemostasis has been assured. total mortality (two o the secondary outcomes) in the treatment
group, in part due to more episodes o hypotension and bradycar-
dia. Mortality was increased in patients with sepsis, and stroke risk
PRACTICE POINT appeared to be increased in patients with prior strokes and intraop-
The ACC/AHA guidelines recommend delaying elective surgery erative hypotension. This study generated signi cant commentary
or patients with recent stent placement at least: and criticism, mainly related to the high dose o metoprolol that was
started shortly be ore surgery in β-blocker-naïve patients, many o
• Two weeks a ter balloon angioplasty whom underwent emergency surgery or had sepsis. This tempered
• Four to six weeks a ter placement o a bare metal stent (BMS) the enthusiasm or prophylactic β-blockers.
The DECREASE IV trial using bisoprolol, started well in advance
• Twelve months a ter a drug-eluting stent (DES)* in order to (median 34 days) o surgery and titrated to a heart rate between 50
complete the course o dual antiplatelet therapy and 70 beats per minute reduced cardiac death and non atal MI in
*Except i a ter 6 months, the risk o delaying surgery exceeds intermediate risk patients. I β-blockers are to be bene cial based
expected risks o ischemia and stent thrombosis, surgery may be on the positive although controversial trials, it appears that they
considered. need to be started more than 1 and probably at least 7 days be ore
surgery and potentially titrated to a heart rate somewhere in the
range o 55 to 70 beats per minute to minimize the risk o signi cant
hypotension or bradycardia (although most patients remained on
■ MEDICAL THERAPY
their initial β-blocker dose). Higher-risk patients or those undergo-
Beta-blockers ing vascular or higher-risk surgical procedures would be most likely
Note: The scienti c integrity o the DECREASE trials rom the Polder- to bene t.
mans group has been questioned but they will be mentioned in the Multiple meta-analyses have come to dif erent conclusions based
text that ollows to illustrate the controversy. on the studies included or excluded in the analysis. Although the
333
nal answer is not in, the ACC/AHA published a systematic review event in determining whether or not to continue aspirin in patients
o randomized controlled trials with perioperative β-blockers ana- already on it ( or secondary prevention).
lyzing the results with and without inclusion o the DECREASE and
P
POISE trials. Their recommendations were to continue β-blockers in Statins
A
patients taking them chronically, that it may be reasonable to begin In addition to lowering cholesterol, statins have a number o so-
R
them be ore surgery in patients with ischemia on stress testing or called pleotropic ef ects. These include reduced platelet aggrega-
T
with three or more RCRI actors, and that initiating β-blockers in tion, improved endothelial unction, and reduced in ammation. It is
I
I
patients with other indications but no RCRI actors was o uncertain thought that this latter ef ect in particular may help stabilize plaques
bene t. I initiating β-blockers preoperatively, they should be started and prevent plaque rupture, which might lead to an MI.
ar enough in advance to assess sa ety and tolerability. This should Most observational studies report that perioperative statin use
be done more than 1 day be ore surgery and not started on the is bene cial in reducing postoperative cardiac complications and
M
day o surgery, which was ound to be harm ul. Although not spe- death in both cardiac and noncardiac surgery. A meta-analysis by
e
d
ci cally recommended in the ACC/AHA guidelines, a cardioselective Kapoor and colleagues con rmed this bene t; however, there are
i
c
β-blocker (bisoprolol or atenolol) was recommended over metopro-
a
ew randomized controlled trials. The rst o these, a small study
l
lol by the ESC guidelines.
C
o 100 patients using atorvastatin 20 mg started 30 days be ore
o
surgery, demonstrated a bene cial ef ect on a composite outcome
n
s
including some weaker end points.
u
PRACTICE POINT
l
t
The DECREASE III trial randomized 497 patients to uvastatin 80
a
t
The 2014 ACC/AHA recommendations for administration of mg extended release or placebo started approximately 1 month
i
o
n
β-blockers be ore vascular surgery and showed that the statin-treated group
had less ischemia and a statistically signi cant reduction in the
Class I recommendations:
composite end point o cardiac death and non atal MI. It also dem-
• Continue β-blockers or patients already on them. onstrated that statins reduced LDL and total cholesterol, multiple
in ammatory markers including C-reactive protein (CRP) and inter-
Class IIa recommendations:
leukin (IL)-6, but also were sa e in that there were no cases o rhab-
• Reasonable or management o β-blockers postoperatively to domyolysis or signi cant hepatic injury. Although no sa ety issues
be guided by clinical circumstances, regardless o when they were ound in a review o statin use in vascular surgery patients, the
were started preoperatively. drug manu acturers still recommend discontinuing statins be ore
major surgery due to these potential sa ety concerns. However,
Class IIb recommendations:
the ACC/AHA recommends continuing them perioperatively as
• Reasonable to start β-blockers be ore surgery in patients with the potential bene t outweighs the theoretical risk. Based on the
ischemia on stress testing or with three or more RCRI actors. DECREASE III trial it appears that patients who are not on a statin
but are scheduled or vascular surgery would bene t rom starting a
• Use ulness uncertain or patients with other long-term
statin preoperatively. However, the uvastatin arm o the DECREASE
indications or β-blockers but no RCRI actors.
IV trial in intermediate risk patients only showed a trend toward
• Reasonable to initiate β-blockers ar enough in advance o decreased MI and death that was not statistically signi cant. The
surgery to assess sa ety and tolerability—pre erably more than ACC/AHA also states that preoperative initiation o statin therapy
1 day be ore surgery. is reasonable in patients undergoing vascular surgery and may be
considered in patients with clinical indications (CAD, DM, peripheral
Class III recommendations (Harm):
arterial disease (PAD), hyperlipidemia) undergoing elevated risk sur-
• Do not start β-blockers on the day o surgery. gery. Unanswered questions regarding perioperative statin use are
whether this is a class ef ect, what dose should be used, how long
in advance to start it prophylactically or it to be ef ective, and which
Alpha-2 agonists and aspirin
patients are most likely to bene t rom them.
Although some earlier studies suggested that clonidine might
be bene cial in reducing postoperative cardiac complications,
OTHER CARDIOVASCULAR CONDITIONS
the POISE-2 trial, which randomized 10,010 patients to clonidine,
aspirin, both, or neither be ore noncardiac surgery, demonstrated ■ HYPERTENSION
that prophylactic clonidine did not reduce MI or cardiac death Hypertension is at best a minor risk actor. Although various recom-
but was associated with an increase in non atal cardiac arrest and mendations mention blood pressures (diastolic BP > 110 mm Hg
hypotension. The ACC/AHA guidelines state that clonidine is not or systolic BP > 180 mm Hg) when cancellation o elective surgery
recommended to prevent perioperative cardiac events. However, in should be considered or which might be associated with increased
patients already taking clonidine it should be continued to prevent risk, there is no hard evidence to support them. Hypertensive
rebound hypertension. patients are more likely to have more labile blood pressure peri-
The two aspirin groups—patients already on it and those who operatively and intraoperative hypotension. The etiology o the
were aspirin naïve—also showed no bene t in terms o reducing hypertension, and presence o end organ damage are more likely
MI or death but did have an increased risk o bleeding. The earlier to be associated with any cardiac morbidity than the preoperative
the aspirin was started postoperatively, the higher the bleeding risk. blood pressure itsel .
Patients with cardiac stents who had not completed their course o Most antihypertensive medications should be continued, includ-
antiplatelet therapy, recent strokes, and patients undergoing carotid ing on the morning o surgery, with the possible exceptions o
endarterectomy were excluded. Some controversy exists because diuretics and angiotensin-converting enzyme (ACE) inhibitors or
there was no subgroup analysis o patients on aspirin or primary angiotensin receptor blockers (ARBs). O note, the 2014 ACC/
versus secondary prevention or or higher versus lower bleeding risk AHA guidelines state that continuation o ACEIs or ARBs is reason-
procedures. The ACC/AHA guidelines state that the risk o a cardio- able perioperatively although many anesthesiologists recommend
vascular event needs to be weighed against the risk o a bleeding stopping them at least 10 hours be ore surgery due to concerns
334
or potential hypotension with induction o anesthesia. I they are AHA did not recommend routine postoperative screening in patients
stopped, they should be restarted as soon as sa e postoperatively. without signs or symptoms o myocardial ischemia.
C
H
■ ARRHYTHMIAS
A
CONCLUSION
P
Hemodynamically signi cant arrhythmias including tachyarrhyth- Using the ACC guidelines and risk calculators, the patient can be
T
mias (rapid atrial brillation (AF), supraventricular tachycardia (SVT), classi ed as low versus elevated risk based on combined clinical and
E
R
ventricular tachycardia (VT)) as well as bradyarrhythmias (symp- surgical risk. The physician can then decide not only whether urther
tomatic sinus bradycardia, high-degree atrioventricular (AV) block) testing is indicated but also whether it is likely to change manage-
5
0
should be evaluated and treated as per GDMT be ore elective ment. Prophylactic revascularization is rarely necessary just to get
surgery. a patient through surgery, and the majority o the patients will be
managed medically. It is important or uture studies to determine
P
r
■ HEART FAILURE optimal use o β-blockers, statins, and other therapies in order to
e
o
have patients in their optimal medical condition prior to elective
p
Heart ailure is a predictor o postoperative MACE. Systolic and
e
symptomatic heart ailure carry a greater risk than diastolic or noncardiac surgery to reduce postoperative complications.
r
a
t
asymptomatic dys unction. Decompensated heart ailure requires
i
v
SUGGESTED READINGS
e
urther evaluation and treatment as per GDMT be ore proceeding
C
with elective surgery. Beta-blockers should not be started preop-
a
r
eratively in these patients, and in a subgroup analysis o the CIBIS II Bilimoria KY, Liu Y, Paruch JL, et al. Development and evaluation o
d
i
the universal ACS NSQIP surgical risk calculator: a decision aid and
a
study, patients with heart ailure had little bene t rom β-blockers.
c
in ormed consent tool or patients and surgeons. J Am Coll Surg.
R
i
■ VALVULAR HEART DISEASE 2013;217(5):833-842.
s
k
A
Patients with valvular heart disease should be evaluated and treated Fleisher LA, Fleischmann KE, Auerbach AD, et al. American College o
s
s
as per GDMT. The lesion most likely to be associated with periopera- Cardiology; American Heart Association. 2014 ACC/AHA guideline
e
s
tive cardiac complications is symptomatic, severe aortic stenosis (AS). on perioperative cardiovascular evaluation and management o
s
m
Patients with a systolic murmur suggestive o AS, particularly i they patients undergoing noncardiac surgery: a report o the American
e
have chest pain, dyspnea, or syncope, should have a 2D echocar- College o Cardiology/American Heart Association Task Force on
n
t
diogram per ormed. I they have symptomatic severe AS and the practice guidelines. J Am Coll Cardiol. 2014;64(22):e77-e137.
a
n
planned noncardiac surgery is elective, the recommendation is or Grines CL, Bonow RO, Casey DE Jr, et al. Prevention o premature
d
valve replacement. Should the patient re use or i the surgery is more
P
discontinuation o dual antiplatelet therapy in patients with
e
urgent, it has been possible to get most o these patients through coronary artery stents: a science advisory rom the American
r
i
o
surgery success ully using medical therapy and intraoperative moni- Heart Association, American College o Cardiology, Society or
p
toring. Patients with asymptomatic aortic stenosis tend to tolerate
e
Cardiovascular Angiography and Interventions, American College
r
a
surgery much better than those with symptoms and comparable o Surgeons, and American Dental Association, with representa-
t
i
v
severity o stenosis. Mitral stenosis, when associated with atrial bril- tion rom the American College o Physicians. J Am Coll Cardiol.
e
lation and heart ailure, may also increase risk, but most o the other
M
2007;49(6):734-739.
valvular lesions do not require surgical intervention be ore noncardiac
a
Gupta PK, Gupta H, Sundaram A, et al. Development and validation
n
surgery. The ACC/AHA guidelines state that it is reasonable to per orm
a
o a risk calculator or prediction o cardiac risk a ter surgery.
g
elevated risk elective noncardiac surgery in patients with asymptom-
e
Circulation. 2011;124(4):381-387.
m
atic valvular disease (even i severe) with appropriate intraoperative
e
and postoperative monitoring. Pulmonary hypertension is now being Kristensen SD, Knuuti J, Saraste A, et al. 2014 ESC/ESA Guidelines
n
t
recognized as a risk actor as well, but studies are limited. on non-cardiac surgery: cardiovascular assessment and manage-
Endocarditis prophylaxis is only indicated or patients undergo- ment: the Joint Task Force on non-cardiac surgery: cardiovascular
ing dental and upper respiratory procedures who have a prosthetic assessment and management o the European Society o Cardiol-
valve, previous endocarditis, complex congenital heart disease that ogy (ESC) and the European Society o Anaesthesiology (ESA).
has not been repaired (or was repaired in the past 6 months), or Eur Heart J. 2014;35(35):2383-2431.
valvular disease in a transplanted heart. L’Italien GJ, Paul SD, Hendel RC, et al. Development and validation
o a Bayesian model or perioperative cardiac risk assessment in
BIOMARKERS a cohort o 1,081 vascular surgical candidates. J Am Coll Cardiol.
Elevated natriuretic peptides (BNP, NT-proBNP) preoperatively and 1996;27(4):779-786.
postoperatively have been associated with MACE. These tests have Nuttall GA, Brown MJ, Stombaugh JW, et al. Time and cardiac risk o
been used in conjunction with the RCRI and may provide incremen- surgery a ter bare-metal stent percutaneous coronary interven-
tal value in risk assessment. However, which test to use, when to tion. Anesthesiology. 2008;109(4):588-595.
obtain it, what cutof is best, and what to do with the results are not Rabbitts JA, Nuttall GA, Brown MJ, et al. Cardiac risk o noncardiac
known, and at this time their role, i any, in preoperative risk assess- surgery a ter percutaneous coronary intervention with drug-eluting
ment is unclear. stents. Anesthesiology. 2008;109(4):596-604.
Most postoperative myocardial in arctions are not associated with
Wilson W, Taubert KA, Gewitz M, et al. Prevention o in ective
chest pain, which raises the question o whether patients should be
endocarditis: guidelines rom the American Heart Association:
screened postoperatively with troponins or EKGs. An entity termed
a guideline rom the American Heart Association Rheumatic
MINS—myocardial injury a ter noncardiac surgery—de ned as an ele-
Fever, Endocarditis, and Kawasaki Disease Committee, Council on
vated troponin presumed to be o ischemic etiology but not requiring
Cardiovascular Disease in the Young, and the Council on Clinical
diagnostic criteria or an MI, is associated with adverse outcomes.
Cardiology, Council on Cardiovascular Surgery and Anesthesia,
However, it is unclear what to do in these situations and whether or
and the Quality o Care and Outcomes Research Interdisciplinary
not any treatment is bene cial or harm ul. For these reasons the ACC/
Working Group. Circulation. 2007;116(15):1736-1754.
335
CHAP TER
51 INTRODUCTION
A comprehensive preoperative evaluation must include assessment
o the risk o postoperative pulmonary complications. While ew
would argue this point, pulmonary risk is o ten underappreciated as
clinicians typically ocus the majority o their energy on the preop-
erative cardiac evaluation and preventing venous thromboembolic
complications. Highlighting the risk o this approach, postoperative
respiratory problems occur with similar requency and greater mor-
bidity than cardiovascular complications.
Perioperative Pulmonary complications ollowing anesthesia and surgery result
rom central nervous system suppression and altered respiratory
Pulmonary Risk dynamics. Administration o sedating agents and neuromuscular

blockade exposes the patient to the risk o aspiration. Furthermore,
Assessment and regardless o the type o anesthetic technique utilized, patients will
experience a reduction in lung volumes perioperatively. Reduction
in lung volumes is the primary mechanism that may lead to atel-
Management ectasis and predispose a patient to the additional complications o
pneumonia and respiratory ailure. This reduction in lung volumes
is greatest or patients undergoing thoracic and upper abdominal
surgery. Table 51-1 lists speci c postoperative pulmonary compli-
Kurt Pfeifer, MD, FACP, FHM cations and diagnostic considerations or each.
Gerald W. Smetana, MD, FACP
RISK STRATIFICATION
Clinicians intuitively recognize several risk actors or pulmonary
complications, but some predictors o postoperative respiratory
problems are not obvious. Additionally, clinicians may struggle with
the appropriate utilization o preoperative pulmonary diagnostic
testing. Recently published risk indices and practice guidelines pro-
Key Clinical Questions vide valuable assistance in the identi cation o risk actors and the
per ormance o evidence-based preoperative evaluation.
1 What risk actors predict postoperative pulmonary
complications? ■ PATIENT-SPECIFIC RISK FACTORS
2 What role does diagnostic testing play in the evaluation Several di erent patient characteristics increase postoperative pul-
o perioperative pulmonary risk? monary risk (Table 51-2). While most o these patient-speci c ac-
3 What objective tools are available or per orming a tors are nonmodi able, their identi cation is important or providing
preoperative pulmonary risk assessment? patients, surgeons and anesthesiologists with an accurate assess-
4 What strategies are e ective to reduce and manage ment o perioperative risk and to identi y patients or whom one
postoperative pulmonary risk? should employ risk reduction strategies.
Chronic lung disease

Multiple studies have identi ed chronic obstructive pulmonary
disease (COPD) as a signi cant risk actor or postoperative respira-
tory problems. However, the risk attributable to COPD is lower than
several other patient-related risk actors, including total unctional
dependence, obstructive sleep apnea and age 70 years or greater.
The impact o other orms o chronic lung disease on perioperative
pulmonary risk has been much less studied, but at least one retro-
spective review o interstitial lung disease patients ound them to
have rates o postoperative pneumonia higher than the general
population. Additionally, pulmonary hypertension has recently been
shown to increase the risk o postoperative pulmonary complica-
tions as well as perioperative mortality and cardiac complications.
Advanced age
Advanced age is a major independent predictor o pulmonary com-
plications. Contrary to previous belie , this risk is not a result o the
cumulative morbidities associated with aging. Elevated risk begins
336
less than that rom most other patient-speci c actors including
TABLE 51-1 Common Postoperative Pulmonary COPD. The adjusted odds ratio (OR) or pulmonary complications
C
Complications in smokers is 1.3, and the risk is greatest among active smokers
H
A
Condition Diagnostic Considerations and recent (<4 weeks) quitters. Patients who have abstained rom
P
smoking or at least 2 months have similar rates o postoperative
Atelectasis • Potential cause o mild hypoxia
T
pulmonary complications as nonsmokers. Clinicians should consider
E
• Generally not a cause o postoperative ever or a recent history o cigarette smoking to be a moderate predictor o
R
moderate to severe hypoxia pulmonary complications.
5
Pneumonia • Diagnostic criteria vary—utilize same as
1
nosocomial pneumonia
Medical comorbidities
• O ten polymicrobial—common pathogens
As the number o comorbid illnesses increases, so does the risk o
P
include Pseudomonas, Staphylococcus aureus,
e
perioperative complications. The American Society o Anesthesiolo-
r
Streptococcus pneumoniae and enteric Gram-
i
o
negative bacilli gists (ASA) classi cation system is easy-to-use and predicts the risk
p
e
• Though aspiration o secretions is a likely o postoperative respiratory problems, as well as overall morbidity
r
a
and mortality. Compared to ASA class I patients (no comorbidities),
t
contributor to development, anaerobic
i
v
bacteria rarely cause postoperative pneumonia patients with an ASA class o II or greater have a nearly ve- old
e
increase in the rate o postoperative pulmonary complications. The
P
Respiratory • Inability to wean o ventilator support within
u
risk increases with each higher ASA class.
l
ailure 48 h o surgery or unplanned reintubation
m
o
• Typically, a combination o hypoxic and
n
hypercapnic respiratory ailure Functional status
a
r
y
COPD • Diagnostic criteria and assessment same as in Large multivariate studies have also identi ed dependence on oth-
R
exacerbation nonoperative population ers or assistance with activities o daily living as a signi cant risk
i
s
k
actor or postoperative pulmonary complications. The contribution
A
COPD, chronic obstructive pulmonary disease. o unctional dependence to perioperative pulmonary risk is greater
s
s
e
than cigarette smoking and chronic lung disease. For patients with
s
s
a completely dependent unctional status, the adjusted OR or
m
at age 50 and is especially high in patients aged 70 years and older
respiratory complications is greater than 4, and or partial unctional
e
(adjusted odds ratio o 3). Recognition that age, even in the absence
n
dependence the adjusted OR is approximately 2.
t
o other comorbidities, is an important source o pulmonary risk is
a
n
critical as the population ages and more elderly patients undergo
d
Congestive heart failure
surgical procedures. The consequence o these observations is that
M
even healthy older patients are at risk or postoperative pulmonary Congestive heart ailure (CHF) is not only an obvious risk actor or
a
n
complications. perioperative cardiac complications but also a predictor o postop-
a
g
erative respiratory ailure. Although the cause o respiratory ailure
e
m
Smoking may be CHF-related pulmonary edema, a history o heart ailure is
e
nonetheless an independent risk actor or pulmonary complica-
Cigarette smoking is associated with increased postoperative pul-
n
t
tions that con ers a relative risk o approximately 6 in the geriatric
monary complications, but the degree o risk elevation is actually
population.
Obstructive sleep apnea

TABLE 51-2 Patient-Specific Risk Factors for Postoperative
Pulmonary Complications For years, clinicians have presumed that obstructive sleep apnea
(OSA) increases the risk o postoperative complications. Only
Chronic obstructive pulmonary disease recently have high-quality studies established this to be true. In
Advanced age (age >50) several studies OSA patients have increased rates o postoperative
Functional dependence pulmonary complications as well as statistically signi cant increases
Comorbid disease (ASA class ≥2) in cardiac complications and unplanned ICU admissions. This risk
extends not only to airway compromise in the immediate post-
Smoking
operative period, but also traditional postoperative pulmonary
Obstructive sleep apnea
complications including pneumonia and respiratory ailure. Other
Pulmonary hypertension recent studies also suggest that patients who screen positive or
Congestive heart ailure OSA but do not yet carry the diagnosis have an increased risk o
Impaired sensorium postoperative complications. For example, patients with at least
Respiratory in ection within past month two clinical eatures o OSA (snoring, crowded oropharynx, daytime
Preoperative anemia and/or trans usion somnolence, or witnessed apneas) who have abnormal overnight
oximetry are more likely to develop postoperative hypoxemia or
Preoperative sepsis
airway management problems than patients without these eatures.
Preoperative hypoxemia Whether ormally diagnosed or suspected based on clinical eatures,
Consumption o >2 alcoholic drinks per day within 2 wk OSA should be considered at least a moderate risk actor or periop-
Preoperative trans usion o >4 units o packed red blood cells erative respiratory problems.
Chronic corticosteroid use
Weight loss o >10% within 6 mo Other risk factors
History o stroke In addition to the previously described risk actors, several other
patient characteristics have been less consistently linked to post-
ASA, American Society o Anesthesiologists. operative pulmonary complications. For instance, one large study
337
ound consumption o more than two alcoholic drinks per day
within 2 weeks o surgery, recent weight loss exceeding 10% o TABLE 51-3 Procedure-Specific Risk Factors for Postoperative
total body weight, chronic corticosteroid use, preoperative trans u- Pulmonary Complications
P
sion o more than 4 units o blood, acute alteration o mental status,
Surgery type Head and neck
A
and a history o stroke to be modest risk actors or postoperative
R
pneumonia. More recent risk models have not identi ed these as Thoracic, including esophageal
T
independent predictors but demonstrated that respiratory in ec- Abdominal
I
I
tion within a month prior to surgery as well as preoperative sepsis, Abdominal aortic aneurysm
hypoxemia and anemia signi cantly contribute to the risk o post- repair
operative pulmonary complications. Neurosurgical
Vascular
M
Factors not associated with increased pulmonary risk General anesthesia
e
d
Notably absent rom the described risk actors are our conditions
i
Nasogastric intubation
c
a
o ten mistaken as predictors o postoperative pulmonary com- Emergency surgery
l
C
plications: well-controlled asthma, obesity, diabetes mellitus and
o
Prolonged surgery (>2-3 h)
immunosuppression. Although a potential relationship between
n
s
perioperative pulmonary risk and morbid obesity and/or asthma
u
l
t
has been assumed by some clinicians, multivariate investigations
a
t
have consistently ound no such correlation. The current literature
i
o
the surgical approach should be made by the surgeon based on
n
provides little data on the contributions o immunosuppression and
other actors.
diabetes mellitus to pulmonary surgical complications.
■ PROCEDURE-SPECIFIC RISK FACTORS PRACTICE POINT

Patient characteristics impact perioperative pulmonary risk sig-
Strongest Predictors of Postoperative Pulmonary
ni cantly, but procedure-speci c actors are even greater predictors
o postoperative pneumonia and respiratory ailure (Table 51-3). Complications
Unlike the majority o patient-speci c risk actors, some o these The strongest predictors o postoperative pulmonary
surgery-related actors may be modi able. complications are:
• Surgery type: A location anywhere rom the mouth to upper
Surgery type abdomen disrupts the airway and/or normal respiratory
The type o surgery is the most important determinant o postsurgi- dynamics.
cal pulmonary risk. As expected, intrathoracic procedures and sur- • Advanced age: Beginning at age 50, perioperative pulmonary
geries in close proximity to the diaphragm carry the highest risk o risk begins to increase and becomes particularly high in
postoperative pulmonary complications. Head and neck operations patients over the age o 70 years.
are high-risk procedures due in part to impairment o the upper • Obstructive sleep apnea: Known or presumed OSA, based on
airway. Restriction o diaphragmatic motion by pain rom abdomi- a positive screen, is an important risk actor.
nal procedures leads to reduced lung volumes, the precursor o • Functional dependence: Reliance upon others or assistance
pulmonary complications. Particularly high risks are esophageal and with activities o daily living increases the risk o multiple
aortic surgeries. complications, including postoperative respiratory problems.
Anesthetic technique
Several studies have identi ed general endotracheal anesthesia
(GETA) as an independent predictor o postoperative pulmonary ■ DIAGNOSTIC STUDIES
complications. In some analyses, neuraxial blockade (spinal and A thorough history and physical examination are the cornerstones
epidural anesthesia), either alone or in combination with GETA, has o the preoperative evaluation. In most instances, diagnostic testing
been associated with lower rates o postoperative morbidity and adds little to the risk assessment as established by clinical evalua-
mortality compared to GETA alone. However, randomized con- tion. Testing has a role when the risk is uncertain a ter history and
trolled trials (RCTs) have not consistently identi ed GETA as a risk physical examination.
actor. While these data suggest general anesthesia may contribute
to pulmonary risk, selection o anesthetic technique is the primary
responsibility o the anesthesiologist and is dependent on multiple Chest radiography
other actors outside the purview o the hospitalist. The medical Chest radiography is part o many clinicians’ routine preoperative
consultant should share his or her risk assessment with the anes- evaluation despite evidence that chest x-ray results rarely change
thesiologist so they can incorporate this into anesthesia planning, perioperative management. Abnormalities on preoperative chest
but in general should de er the selection o anesthetic type to the radiography are relatively common, occurring on 10% to 20% o
anesthesiologist. x-rays. However, a minority o these ndings are unexpected based
on the clinician’s physical assessment and history, and an even smaller
Other risk factors number in uence perioperative care. Clinicians should reserve chest
Regardless o the surgical site or anesthetic technique, procedures imaging or the same indications as in the nonoperative setting: eval-
per ormed on an emergent basis carry a higher risk o respiratory uation o new or changed cardiopulmonary symptoms. As part o
complications. Surgeries with duration greater than 2 to 3 hours the American Board o Internal Medicine’s Choosing Wisely initiative,
are also associated with an increased risk or postoperative pneu- three di erent pro essional societies have speci cally recommended
monia and respiratory ailure. There is probably no di erence in risk not obtaining preoperative chest radiographs in patients without
between laparoscopic versus open surgeries; decisions regarding clinical ndings suggestive o pulmonary abnormalities.
338
Arterial blood gas analysis
TABLE 51-4 ARISCAT Index for Predicting Postoperative
C
Early studies o preoperative arterial blood gas (ABG) analysis sug- Pulmonary Complications
H
gested that hypercarbia and hypoxemia predicted postoperative
A
pulmonary complications. Subsequent reviews concluded that clini- Risk Factor Risk Score
P
cal assessment was equally accurate in predicting postsurgical respi- Age (y) 51-80 3
T
ratory problems. Taking into account that no arterial CO2 or O2 value
E
>80 16
R
is an absolute contraindication to surgery, clinicians should per orm
Preoperative SpO2 (%) 91-95 8
ABG analysis only i the results will signi cantly in uence periop-
5
<91 24
1
erative management (eg, excluding hypercarbia and respiratory
acidosis in a patient with COPD and increased shortness o breath). Respiratory in ection in past month 17
Location o surgery Upper 15
P
e
Pulmonary function testing abdominal
r
i
o
Formal pulmonary unction testing has an established role in the Thoracic 24
p
e
preoperative evaluation o patients be ore lung resection, but its Duration o surgery (h) >2-3 16
r
a
role in nonthoracic surgery is questionable. Analogous to ABG analy-
t
>3 23
i
v
sis, initial studies o preoperative spirometry suggested it was use ul
e
Emergency surgery 8
P
or predicting postoperative morbidity, but later data comparing
Preoperative hemoglobin ≤10 g/dL 11
u
spirometry to history and physical examination showed no clear
l
m
incremental bene t rom pulmonary unction testing. Further, no Postoperative
o
Pulmonary
n
spirometric values are a prohibitory threshold or noncardiothoracic
a
Risk Class Risk Score Complications (%)
r
surgery. Indications or pulmonary unction testing are the same as
y
Low <26 1.6-3.4
R
in the general setting: evaluation o unexplained respiratory symp-
i
s
toms and characterization o lung disease when it is unclear i a Intermediate 26-44 13-13.3
k
A
patient’s air ow obstruction is optimally reduced. High >44 38-42.1
s
s
e
s
Used with permission rom Canet J, Gallart L, Gomar C, et al. Anesthesiology.
s
m
PRACTICE POINT 2010;113(6):1338-1350.
e
n
Diagnostic Testing for Perioperative Pulmonary Risk
t
a
clinician can determine which o three di erent risk classes a
•
n
Diagnostic testing rarely adds value to perioperative pulmonary
d
patient its into. The ARISCAT index was originally derived and vali-
risk assessment in noncardiothoracic surgery. A thorough
M
dated rom a geographically localized population but has subse-
history and physical examination is suf cient in almost all cases.
a
quently been validated in a sample o patients rom across Europe.
n
a
Although the above risk indices are help ul or estimating risk in
g
e
the general population, they do not account or risks associated
m
■ RISK ASSESSMENT TOOLS with obstructive sleep apnea. The ASA recommends screening all
e
n
In recent years, the scienti c community has paid more attention surgical patients or OSA. Several tools or this purpose are avail-
t
to the importance o postoperative pulmonary complications. The able, including the STOP-BANG, Berlin and Flemons surveys. The
result has been the generation o practical risk assessment tools. STOP-BANG questionnaire accurately identi es patients at risk or
In 2006, the American College o Physicians developed clini- moderate to severe OSA and postoperative complications (Table 51-5).
cal guidelines or perioperative pulmonary risk assessment and A score o 5 or higher (o eight possible points) increases the likeli-
reduction or patients undergoing noncardiothoracic surgery. This hood o moderate to severe OSA, as well as rates o perioperative
systematic review provides a valuable resource or per orming complications.
an evidence-based preoperative evaluation. Several perioperative
pulmonary risk prediction models have also been developed to
assist clinical risk strati cation. Most o these are restricted to either TABLE 51-5 STOP-BANG Screening for Obstructive Sleep
speci c types o surgery or unique respiratory complications (eg, Apnea
acute respiratory distress syndrome), but three recent models show
promise or being airly generalizable. S Snoring: Do you snore loudly (loud enough to be
heard through closed doors or your bed-partner
Using data rom the American College o Surgeons National
elbows you or snoring at night)?
Surgical Quality Improvement Program, Gupta and colleagues
developed separate models or prediction o postoperative respira- T Tired: Do you o ten eel tired, fatigued or sleepy
during the daytime (such as alling asleep during
tory ailure and pneumonia. Predictive risk actors or postoperative
driving)?
respiratory ailure were: site o surgery, emergency surgery, ASA clas-
O Observed: Has anyone observed you stop
si cation, sepsis and unctional status (ie, ability to per orm activities
breathing or choking/gasping during your sleep?
o daily living). The pneumonia risk model used the same variables
except that age, COPD and smoking were included and emergent P Pressure: Do you have or are being treated or high
blood pressure?
surgery was not. Both models were incorporated into risk calculators
that are reely available online (www.surgicalriskcalculator.com). The B BMI > 35 kg/m 2?
Gupta postoperative pulmonary risk calculators per ormed very well A Age older than 50 y?
in their initial study but have not been externally validated. N Neck size large? For male, shirt collar ≥17 in/43 cm;
In 2010, Canet and colleagues developed the ARISCAT index or or emale, shirt collar ≥16 in/41 cm?
predicting postsurgical pulmonary complications. This model uses G Gender = male
seven predictive actors, each with a weighted score based on its
risk contribution (Table 51-4). Tallying the score in this model the Used with permission rom University Health Network.
339
RISK REDUCTION STRATEGIES respiratory problems with the use o thoracic epidural analgesia.
However, evidence o bene t rom other regional analgesia is scant.
Many pulmonary risk actors are nonmodi able, but or patients
Administration o intravenous opioids through PCA reduces the risk
with elevated perioperative risk several evidence-based risk reduc-
P
o pulmonary complications when compared to intermittent intra-
tion methods exist. Although some o these strategies are outside
A
venous administration on a PRN basis.
the scope o hospitalists, medical consultants should be aware
R
T
o all potential risk reduction interventions. For all patients with
■ LUNG EXPANSION TECHNIQUES
I
increased pulmonary risk, special consideration should be given to
I
the patient’s postoperative triage (home vs general hospital admis- The largest body o evidence or a perioperative pulmonary risk
sion vs intensive care unit placement), and the hospitalist should reduction strategy is or lung expansion techniques. Incentive spi-
incorporate the patient’s preoperative pulmonary risk assessment rometry, intermittent positive pressure breathing, deep breathing
into postoperative care decision making. For instance, more aggres- exercises and continuous positive airway pressure each reduce the
M
risk o postoperative pulmonary complications. None o these strate-
e
sive work-up o postoperative ever or hypoxia may be initiated in
d
the patient who was preoperatively assessed to be at increased risk gies has been shown to be superior to another. Given the minimal
i
c
a
or respiratory ailure or pneumonia. risk associated with these techniques, the American College o Phy-
l
C
sicians recommends that incentive spirometry or deep breathing
o
exercises be used or all patients at risk or pulmonary complications.
n
■ PREOPERATIVE SMOKING CESSATION
s
Furthermore, a number o studies have demonstrated the value o
u
l
The role o smoking cessation in perioperative risk reduction
t
lung expansion techniques or cardiopulmonary conditioning ther-
a
remains an area o debate. Some studies have shown that pre-
t
apy started 1 to 2 weeks prior to surgery. This strategy reduces rates
i
o
operative smoking cessation decreases the risk o postsurgical
n
o pneumonia and all pulmonary complications by approximately
respiratory complications, while others have demonstrated similar one-hal . Preoperative initiation o lung expansion or cardiopulmo-
pulmonary morbidity between smokers and nonsmokers. An older nary physiotherapy should be considered or patients with elevated
study actually suggested an increase in postoperative pulmonary pulmonary risk.
complications i smoking cessation occurred shortly be ore surgery.
Recent meta-analyses o preoperative smoking cessation have ■ SELECTIVE NASOGASTRIC DECOMPRESSION
identi ed no statistically signi cant increase in risk or patients who
The American College o Physicians also recommends use o naso-
quit smoking less than 4 to 8 weeks be ore surgery. Patients who
gastric tubes only as needed or postoperative nausea or vomiting,
stop smoking more than 4 weeks prior to surgery have a signi cant
oral intake intolerance or symptomatic abdominal distention. This
decrease in postoperative pulmonary complications. Randomized
recommendation results rom observations that routine, rather than
controlled trials have con rmed that smoking cessation interven-
selective, nasogastric decompression ollowing abdominal surgery
tions started at least 1 month be ore surgery lower the incidence
led to increased rates o pneumonia and atelectasis.
o overall complications. The greatest bene t occurs with durations
o cessation o at least 8 weeks be ore surgery. Preoperative smok-
■ LUNG-PROTECTIVE VENTILATION
ing cessation interventions (including the use o tobacco cessation
pharmacotherapy) cause no harm, and increase rates o long-term A growing body o evidence also supports the use o lung-pro-
abstinence. For any hospitalized surgical patient, clinicians should tective ventilation (lower tidal volume, higher levels o positive
advise smoking cessation. end-expiratory pressure [PEEP] and lung recruitment maneuvers)
in patients who require perioperative mechanical ventilation. This
approach has long been utilized in patients with acute respiratory
■ ANESTHETIC TECHNIQUES
distress syndrome to reduce ventilator-associated lung injury. More
Anesthesia considerations are the responsibility o the anesthesi- recently, studies and meta-analyses have also shown reduced post-
ologist, but it is important or hospital medicine practitioners to operative pulmonary complications in patients ventilated with tidal
understand potential risk reduction strategies related to anesthetic volumes o 6 to 8 cc/kg and PEEP o 5 to 10 cm o water pressure
techniques. Previous studies have shown that avoidance o long- (cwp). Data rom a 2014 randomized controlled trial indicates that
acting neuromuscular blockade medications (pancuronium) can use o lower tidal volume (8 cc/kg) with PEEP levels more than 10 cwp
reduce the risk o postoperative pulmonary complications. This results in an increased risk o hypotension. There ore, multiple ac-
results rom less residual neuromuscular blockade and resultant tors in uence the selection o an optimal postoperative ventilation
hypoventilation a ter surgery. Data supporting the use o neuraxial strategy (eg, risk o hypotension).
(epidural or spinal) anesthesia as a risk reduction modality has been
mixed as well. However, a 2012 Cochrane review determined that ■ OBSTRUCTIVE SLEEP APNEA MANAGEMENT
neuraxial (spinal or epidural) anesthesia either alone or when added
As the impact o OSA on perioperative risk is now well established,
to general anesthesia signi cantly reduced the risk o postoperative
experts have developed a number o strategies to reduce the risk
pneumonia.
o sleep apnea-related postoperative complications. Most o these
interventions were not derived rom RCTs, but prospective studies
■ ANALGESIC TECHNIQUES o patients screened as high risk or OSA provide some evidence o
Sedating analgesics (including opioids and some adjunctive agents) their ef cacy. High-risk patients in whom such OSA-targeted inter-
increase the risk o hypoventilation which can lead to atelectasis and ventions were employed have complication rates similar to patients
other respiratory complications. Thus, limitation o systemic opioids at low risk or OSA. The 2014 ASA practice advisory or perioperative
through epidural or regional analgesia or patient-controlled analge- OSA management suggests several strategies or mitigating sleep
sia (PCA) o ers the potential or pulmonary risk reduction. Neuraxial apnea-related risks (Table 51-6). For patients likely to have severe
analgesia o ers more risk reduction than PCA. In addition to less OSA, empiric use o positive airway pressure (PAP) ventilation is
potential or hypoventilation, these strategies reduce postopera- recommended i signs o severe hypoxia or airway obstruction are
tive pain, and increase the ability to take deep breaths and increase evident. At least one recent RCT demonstrated no reduction in
lung volumes a ter surgery. Meta-analyses o abdominal aortic and postoperative pulmonary complications rom empiric PAP initiation
cardiac surgery patients have demonstrated reduced postsurgical postoperatively (ie, without signs or symptoms o hypoxemia or
340
TABLE 51-6 ASA Recommendations for Postoperative However, identi cation o patients at increased pulmonary risk
C
Management of Patients with Known or remains critical or the ollowing reasons:
H
Suspected Sleep Apnea • Communication o increased pulmonary risk to the
A
anesthesiologist and surgical team prompts consideration o
P
Elevate head o bed (unless contraindicated by surgical risk reduction strategies under their control.
T
requirements)
• Modi cation o the postoperative care unit team’s triage
E
R
Continuous pulse oximetry (with centralized monitoring) decisions. For example, a ter a surgery that would usually be
5
Use adjunctive analgesics (eg, NSAIDs) to reduce systemic opioid done as an outpatient, a high-risk patient may be admitted or
1
needs extended observation.
Minimize use o nonopioid sedatives • Heightened postoperative vigilance or respiratory problems.
P
PAP therapy Previously on home PAP: have patient Speci c examples o e ects on management include:
e
r
use home PAP device whenever Lower threshold or initiating workup or pneumonia in
i
o
sleeping
p
setting o postoperative ever (which might otherwise be
e
No previous PAP: initiate PAP or attributed to benign causes)
r
a
requent or severe airway obstruction
t
Admission to unit specializing in respiratory care or
i
v
or hypoxemia
e
mobilization o respiratory therapy resources beyond
P
traditional scope
u
ASA, American Society o Anesthesiologists; NSAIDs, nonsteroidal anti-
l
m
in lammatory drugs; PAP, positive airway pressure. Lower threshold or treatment o postoperative pneumonia
o
when chest radiographs are equivocal
n
a
r
airway obstruction). When possible, surgery should be delayed or
y
R
preoperative evaluation and management o OSA or patients likely
i
SUGGESTED READINGS
s
k
to have severe sleep apnea and are undergoing major surgery. Ret-
A
rospective studies suggest that patients who are ormally diagnosed
s
American Society o Anesthesiologists Task Force on Perioperative
s
and started on appropriate PAP therapy preoperatively have lower
e
Management o patients with obstructive sleep apnea. Practice
s
rates o postoperative complications compared to those who were
s
m
guidelines or the perioperative management o patients with
diagnosed and treated a ter surgery.
e
obstructive sleep apnea: an updated report by the American
n
t
Society o Anesthesiologists Task Force on Perioperative Manage-
CONCLUSION
a
ment o patients with obstructive sleep apnea. Anesthesiology.
n
d
Postoperative pulmonary complications are an under-recognized 2014;120:268.
M
source o signi cant surgical morbidity and mortality. Estimation o
Chung F, et al. High STOP-Bang score indicates a high probability o
a
pulmonary risk is an essential part o the preoperative evaluation
n
obstructive sleep apnea. Br J Anaesth. 2012;108:768.
a
and includes assessment o both patient- and procedure-related
g
e
risk actors. Most risk actors are not modi able, but the preopera- Gupta H, et al. Development and validation o a risk calculator
m
tive pulmonary evaluation provides an accurate estimation o risk predicting postoperative respiratory ailure. Chest. 2011;140:1207.
e
n
or patients and physicians that improves in ormed decision making. Gupta H, et al. Development and validation o a risk calcula-
t
Moreover, identi cation o high-risk patients can heighten postop- tor or predicting postoperative pneumonia. Mayo Clin Proc.
erative vigilance or respiratory ailure and pneumonia and lead to 2013;88:1241.
the use o those strategies proven to reduce risk. Mazo V, et al. Prospective external validation o a predictive score
or postoperative pulmonary complications. Anesthesiology.
PRACTICE POINT 2014;121:219.
Minai OA, et al. Perioperative risk and management in patients with
Why Preoperative Pulmonary Evaluation Matters
pulmonary hypertension. Chest. 2013;144:329.
Pulmonary evaluation has been traditionally undervalued or a Qaseem A, et al. Risk assessment or and strategies to reduce
number o reasons: perioperative pulmonary complications or patients undergoing
• Most o the risk actors or postoperative pulmonary noncardiothoracic surgery: a guideline rom the American College
complications are nonmodi able. o Physicians. Ann Intern Med. 2006;144:581.
• Many o the available risk reduction strategies are either Wong J, et al. Short-term preoperative smoking cessation and
beyond the responsibility o the primary care physician/ postoperative complications: a systematic review and meta-
hospitalist (choice o anesthesia and surgical approach) or analysis. Can J Anaesth. 2012;59:268.
routinely utilized (incentive spirometry and deep breathing
exercises).
341
52
About 21 million adults have diabetes in the United States, and
5.5 million hospital discharges include diabetes as one o the listed
diagnoses. Annual inpatient health care costs directly related to
diabetes were estimated at $76 billion in 2012. People with diabetes
have a higher li etime probability o requiring surgery compared to
the general population due to increased cardiovascular, ophthalmic,
renal, and neuropathic complications rom their disease. Diabetes
mellitus itsel may complicate surgical wound healing and recovery.
Perioperative Risk Surgical site in ections may be twice as high in patients with diabetes,
potentially rom small vessel disease impairing oxygen and nutrient
Assessment and delivery to tissues, impaired leukocyte and monocyte unction due
to hyperglycemia, and decreased release o neuropeptides in those
Management of the with peripheral neuropathy. Patients with diabetes also have higher
postoperative in-hospital mortality compared to those without
diabetes. Patients with higher postoperative blood sugars a ter
Diabetic Patient coronary bypass surgery have an increased rate o complications.
Surgery and anesthesia provoke release o counter-regulatory
hormones, which cause hyperglycemia and increased catabolism.
Care or the surgical patient with diabetes should ocus on avoid-
Katherine Lewis, MD ance o marked hyperglycemia, which can alter wound healing
and disrupt uid balance, while also avoiding hypoglycemia, which
Kathie L. Hermayer, MD can cause cardiac stress. Perioperative evaluation or patients with
diabetes allows or the development o an individualized plan to
reduce perioperative complications and to determine a sa e and
e ective diabetes discharge plan.
PREOPERATIVE EVALUATION
A thorough history remains a key component in the preoperative
evaluation o the patient with diabetes. A history o comorbidities
and complications associated with diabetes may provide additional
insight into surgical risk, and there ore perioperative management.
For example, elective surgery should be delayed in patients with
diabetes a ter a recent cardiac event (see Chapter 50: Preoperative
Cardiac Risk Assessment and Perioperative Management). Patients
with concomitant renal disease need to be monitored care ully,
with precautions to avoid contrast materials and other nephrotoxic
agents. Diabetic autonomic neuropathy may predispose patients
to perioperative hypotension. Gastroparesis with impaired gastric
emptying may predispose patients to aspiration during intubation
and extubation.
Clarif cation o the patient’s home diabetes regimen and assess-
ment o their home glycemic control and adherence will help
guide perioperative management. Ideally, patients will have good
glycemic control prior to undergoing elective surgery, but this will
not always be possible. Some experts advocate delaying elective
surgery i the patients HbA1C is >8.5%, but there is no good litera-
ture to support or negate this practice. In addition, the management
plan should take into consideration the patients’ typical diet and
activity, which may change drastically in the perioperative period.
The type and duration o surgery is also important to consider, as
longer surgeries and recovery times (including ICU admission or
prolonged NPO status) will a ect the perioperative glycemic man-
agement plan.
Specif c physical exam components or patients with diabetes
should include inspection o injection sites or insulin pump in u-
sion sites (i applicable) or evidence o lipohypertrophy, which
may a ect insulin absorption. Evaluation or signs and symptoms o
342
peripheral neuropathy will help plan or all prevention tactics dur- his or her normal routine, or i they endorse asting hypoglycemia
ing the hospital stay. Inspection or associated wounds will help plan on their home regimen. In one study, use o 80% o the basal dose
C
or wound prevention and care during the stay. In addition, obtain- was sa e and e ective in patients with both Type 1 and Type 2 dia-
H
ing orthostatic blood pressures may identi y patients who may have
A
betes undergoing noncardiac surgery. Table 52-2 lists commonly
P
a component o autonomic neuropathy, which should be taken into used insulin ormulations, approach to preoperative dosing, and
T
account or blood pressure management during and a ter surgery. recommendations or when and how to resume insulin regimens
E
An inpatient HbA1C measurement is recommended i not postoperatively and at the time o discharge.
R
obtained within the last 3 months. The HbA1C correlates with the
5
risk or complications, perioperative hyperglycemia, and increased
2
PREOPERATIVE RECOMMENDATIONS FOR
length o stay. More importantly, it may guide both perioperative THE CARE TEAM
and discharge diabetes management. Clinicians must be aware o Once a preoperative diabetes plan has been devised and commu-
P
the limitations o HbA1C accuracy, such as in the settings o recent
e
nicated with the patient, the plan should be relayed to the rest o
r
i
trans usion, anemia, and hemoglobin variants.
o
the patient care team, including the surgical team and anesthesi-
p
Immediate preoperative glucose values predict postoperative
e
ologists. Type o diabetes, duration o diabetes, and complications
r
glucose and may lead to an alteration o management plan in cases
a
related to diabetes should also be relayed to the surgical care team,
t
i
o extreme hyperglycemia or hypoglycemia. A comprehensive
v
and well documented in the medical record.
e
chemistry should be checked preoperatively, and will provide in or- Given the limitations o oral diabetic agents in the periopera-
R
mation on renal unction, liver unction, electrolytes, and bicarbon-
i
s
tive setting, insulin management is pre erred. Patients with Type 1
k
ate status; impaired renal unction reduces insulin clearance and diabetes must continue with some orm o basal insulin to prevent
A
increases a patient’s risk or hypoglycemia, and impaired liver unc-
s
diabetic ketoacidosis. Basal insulin may be in the orm o long act-
s
e
tion increases the risk o hypoglycemia. Correcting hypokalemia is ing insulins such as once a day glargine, detemir, or twice daily NPH.
s
s
important or patients receiving insulin, as insulin treatment lowers
m
Rapid acting insulin analogs (lispro, aspart, and glulisine) act over
plasma potassium levels. Patients in diabetic ketoacidosis will have
e
several hours and are used or prandial insulin and or supplemental
n
an increased anion gap with decreased bicarbonate and hypergly-
t
(correction) insulin. Regular insulin given subcutaneously can also
a
cemia; this should be corrected prior to any surgical procedure.
n
be used or prandial and supplemental insulin.
d
Basal bolus regimens have been recommended or the majority
M
PREOPERATIVE RECOMMENDATIONS FOR PATIENTS o noncritically ill-diabetic patients. This consists o basal insulin +
a
n
In addition to receiving recommendations rom their surgical team, prandial insulin + supplemental insulin. Basal bolus regimens are
a
g
patients need guidance as to what to do with their home diabetes superior to sliding scale insulin alone (which is not recommended).
e
m
regimen prior to surgery. In general, oral diabetic agents and other Basal bolus regimens con er better glycemic control and reduced
e
noninsulin diabetic medications should be held prior to surgery complications such as wound in ection, renal ailure, pneumonia,
n
t
(Table 52-1). However, some controversy in the literature exists or respiratory ailure, and bacteremia. In patients with Type 2 diabetes
o
certain medications. on home oral agents and/or relatively low insulin, basal insulin +
t
h
• Metformin: It is generally recommended that met ormin be
supplemental insulin only is comparable to basal bolus insulin, and
e
held up to 48 hours be ore surgery and restarted 48 hours is a reasonable regimen or these patients.
D
i
a
a ter surgery, due to risk o lactic acidosis; however, in diabetic
b
cardiac surgery patients, one study showed better outcomes MONITORING AND GLYCEMIC TARGETS
e
t
FOR NONCRITICALLY ILL
i
in patients who inadvertently took met ormin. These patients
c
P
had a lower incidence o prolonged intubation, lower risk o During surgery, glucose should be monitored hourly or procedures
a
t
in ections, and lower overall morbidity compared to matched lasting >60 minutes. Glycemic targets or noncritically ill patients on
i
e
patients not taking met ormin. While these f ndings are inter- insulin should be <140 mg/dL be ore a meal and <180 mg/dL ran-
n
t
esting, routine use o preoperative met ormin is not recom- domly. Dose reductions should be considered or any blood glucose
mended, but this in ormation is reassuring should met ormin levels o <100 mg/dLto prevent hypoglycemia, and dose reductions
inadvertently be taken preoperatively. should be mandatory or any blood glucose levels o <70 mg/dL
• Glucagon-like receptor 1 (GLP-1) agonists: These agents (unless the event is easily explained by an identif ed preventable
are generally held prior to surgery due to the gastric empty- actors). In-patient point-o -care monitoring should be per ormed
ing delays and nausea seen with these agents; however, some our or more times a day, including be ore meals and at bedtime. For
studies have shown better glucose control with (compared patients who are not eating, monitoring should occur every 4 to
to without) these agents in cardiac and noncardiac surgery 6 hours (around the clock). A plan or preventing hypoglycemia, and a
patients. A current study may help clari y the potential role or protocol or treating hypoglycemia, should be in place; any episode o
these agents in perioperative management in patients under- hypoglycemia should be documented in the medical record.
going noncardiac surgery. Several ormulations o GLP-1 ago-
nists are dosed once weekly; there ore, it may not be possible ■ INSULIN PUMP THERAPY
to e ectively hold these medications given their long hal -lives. Continuous subcutaneous insulin in usion (CSII, eg, insulin pump
• Dipeptidyl peptidase-4 (DPP- IV) inhibitors: These agents therapy) has become much more commonly used among patients
are o ten held perioperatively, though one study o sitagliptan with diabetes. It provides exibility, convenience, and ease o
showed it was sa e and e ective in selective medical-surgical adjustments in insulin-treated patients. O ten it is associated with
hospitalized patients (without signif cant renal disease, liver less glycemic variability and improved glycemic control. It is o ten
disease or high degree o insulin resistance). recommended that CSII be discontinued during inpatient hospital
For patients on insulin, basal insulin in many cases can be given stays i the patient is not able to sel -manage their pump; however,
at the regular dose or slightly reduced doses or short procedures. use during surgery has been less clear. For patients admitted or
In patients with Type 1 diabetes, basal insulin should not be held same day surgery or procedures o <120 minutes, CSII appears to be
due to the risk o diabetic ketoacidosis. A reduction in basal insulin sa e when used with a protocol that allows or the addition o IVinsulin
may be needed i the patient gives a history o requent snacking in by anesthesiology or glucose >250 mg/dL or any extension o surgery
343
TABLE 52-1 Recommend ations for Oral Diab etic Agents and Other Noninsulin Diab etic Therap ies in the
Periop erative Setting
P
Perioperative
A
Medication Mechanism of Action Recommendation Explanation When to Resume
R
T
Sulfonylureas Increases insulin secretion Hold medication day Risk o hypoglycemia At time o discharge
o surgery and during when oral intake is
I
Glipizide (Hal -li e 2-5 h)
I
hospitalization back to normal
Glyburide (Hal -li e 10 h)
Glimeperide May need to hold Possible adverse
(Hal -li e 9.2 h) chlorpropamide the day cardiovascular e ects
M
Tolazamide (Hal -li e 7 h) be ore surgery as well
e
due to long hal -li e
d
Tolbutamide
i
c
(Hal -li e 4.5-6.5 h)
a
l
C
Chlorpropamide
o
(Hal -li e 36 h)
n
s
u
Meglitinides Increases insulin secretion Hold medication day Risk o hypoglycemia At time o discharge
l
t
(given at time o a meal) o surgery and during with asting when oral intake is
a
Repaglinide (Hal -li e 1 h)
t
hospitalization normal
i
o
Nateglinide
n
(Hal -li e 1.5 h)
Biguanides Sensitizes tissues to Hold medication or surgery Risk o lactic acidosis, At time o discharge
Metformin insulin and prior to contrast particularly in the i luid status and
administration (up to setting o renal renal unction are
48 h in advance) and during insu iciency reasonable (hold or
hospitalization Cr >1.4 in emale and
>1.5 in male)
Alpha-glucosidase Reduces oligo- and Hold medication day No role in asting At time o discharge
inhibitors disaccaridoses in o surgery and during patient when oral intake is
Acarbose intestinal brush border hospitalization normal
Miglitol Caution in renal
insu iciency
Thiazolidinediones Sensitizes tissues to Hold medication day May cause additional At time o discharge
Rosiglitazone insulin o surgery and during luid retention i no heart ailure
hospitalization Concern or heart concern
Pioglitazone
ailure and MI Caution in hepatic
increased risk impairment
Dipeptidyl peptidase-4 Work in glucose Hold medication day Limited role in asting With resumption o
inhibitors dependent ashion to o surgery and during patient oral intake
Sitagliptan increase insulin hospitalization There is one small
Saxagliptan study o sa ety and
e icacy in selective
Linagliptan hospitalized patients
Alogliptan when used with basal
insulin
GLP-1 agonists Increases insulin Hold medication day o Delay o gastric At time o discharge
Exenatide secretion, decreases surgery or ormulations given emptying when oral intake is
glucagon secretion, daily or twice daily Recommendations normal
Exenatide QW delays gastric emptying
(Hal -li e 2 wk) Hold once weekly medications may change in uture:
in the days leading up to One study o bene it
Liraglutide surgery using IV orm
Albiglutide Hold medications during
(Hal -li e 5-8 d) Ongoing study
hospitalization looking at role in
Dulaglutide (May not be possible to post-surgical patients
(Hal -li e 4-7 d) e ectively hold once weekly
ormulations given long hal -
li e o these ormulations)
Sodium glucose Increase glucose Hold medication up to several Reported cases o At time o discharge
cotransporter-2 excretion in the urine days in advance euglycemic DKA when oral intake is
inhibitors in post-operative normal
Canagli lozin patients
Empagli ozin Fluid loss rom
increased urine
Dapagli lozin output
(Continued)
344
TABLE 52-1 Recommend ations for Oral Diab etic Agents and Other Noninsulin Diab etic Therap ies in the
C
Periop erative Setting (Continued )
H
A
Perioperative
P
Medication Mechanism of Action Recommendation Explanation When to Resume
T
Bromocriptine-QR Short-acting dopamine Hold medication day Short hal -li e o just At time o discharge
E
R
(quick release) agonist which lowers o surgery and during a ew hours with low
postprandial blood hospitalization risk o hypoglycemia
5
2
sugars; given in morning Causes modest blood
to reset circadian pressure lowering
rhythms o hypothalamic
P
dopamine and serotonin May cause nausea
e
r
resulting in reduced
i
o
p
insulin resistance;
e
reduced cardiovascular
r
a
events in Type 2 diabetes
t
i
v
compared to placebo
e
R
i
s
k
A
s
s
TABLE 52-2 Recommendations for Insulin Regimens in the Perioperative Setting
e
s
s
m
Onset and Duration Preoperative Adjustment Perioperative Management
e
Insulins of Action of Home Dose in the Hospital Discharge Dosing
n
t
Long-acting
a
n
basalinsulins
d
Glargine
M
Onset 1 h, duration Consider reduction by Basal regimen pre erred Consider dosing based on
a
24 h 20% o dose prior to over sliding scale alone preoperative A1C:
n
a
Detimir Onset 1 h, duration surgery Dosing 0.15-0.3 units/kg/d A1C <7%: home insulin regimen
g
e
6-24 h Do not hold basal insulin Little known about and home oral agents
m
Degludec Hal -li e o 24 h, or patients with Type 1 degludec in perioperative A1C 7-9%: home oral agents with
e
n
duration o >42 h diabetes setting addition o 50% o inpatient basal
t
o
insulin
t
A1C >9%: home oral agents plus
h
e
80% inpatient basal insulin plus
D
80% inpatient prandial insulin
i
a
Intermediate-
b
e
acting basal
t
i
c
insulins
P
NPH Onset 1-2 h, peak Consider ull evening May be used as basal insulin Consider similar dosing to
a
t
4-14 h, duration dose (or slight reduction) dosed twice daily long-acting basal noted above
i
e
n
10-24 h and hal morning dose Twice daily dosing allows
t
on day o surgery or more requent dose
adjustments
Mixed insulins:
Combined NPH
with rapid-acting
analogs or
regular insulin
70/30 See separate Give ull evening meal Recommend substituting Consider inpatient insulin needs,
components: dose the day be ore separate basal and prandial home regimen, oral intake, degree
given with meals, surgery insulin or increased o post-operative stress and A1C
typically twice Consider giving hal lexibility o dosing Consider using 50%-80% o
daily morning dose on day o total daily inpatient insulin dose
75/25 NPH is acting as surgery converting to twice daily regimen
basal insulin Consider providing NPH o mixed insulin regimen with
50/50 Other components only on day o surgery home oral agents
are prandial insulin Consider discharge instead on
basal bolus regimen
Patient should be eating meals to
resume mixed insulin
(Continued)
345
TABLE 52-2 Recommendations for Insulin Regimens in the Perioperative Setting (Continued )
Onset and Duration Preoperative Adjustment Perioperative Management

P
Insulins of Action of Home Dose in the Hospital Discharge Dosing
A
R
Concentrated
T
insulins
I
Glargine Onset 6 h, duration Consider 20% reduction Consider using alternate Concentrated glargine: consider
I
(300 units/mL) 24-36 h o concentrated glargine basal insulin dosing strategy similar to standard
U-500 Onset 30 min, peak the dose prior to surgery Consider perioperative basal therapy
1-3 h, duration 8 h For U-500 insulin, insulin in usion because o U-500: Consider inpatient needs
M
consider converting to degree o baseline insulin and oral intake and dosing at 80%
e
a U-100 basal since this resistance o inpatient needs; take special
d
insulin has a long hal -li e precautions to clari y dosing as
i
I using U-500 in the
c
a
and acts as both a basal hospital because o extreme dosing errors are common
l
C
and prandial insulin insulin resistance, take
o
n
special precautions to avoid
s
dosing errors
u
l
t
Rapid-acting
a
t
i
insulin analogs
o
n
Aspart Onset 0.25 h, peak Hold prandial dosing Used or prandial and Consider home regimen, A1C,
1-3 h, duration while npo or surgery supplemental insulin in inpatient needs and oral intake
3-5 h addition to basal insulin I A1C >9%, consider adding to
Lispro Onset <0.5 h, peak Prandial insulin dosing: home regimen at 80% o prandial
0.5-1.5 h, duration 0.1 unit/kg/meal (adjust or inpatient dosing
<6 h renal insu iciency and poor
Glulisine Onset <0.5 h, peak oral intake) OR 0.15-
0.5-1.5 h, duration 0.25 units/kg/d divided over
<6 h 3 meals
Regular insulin Hold prandial dosing Used in IVinsulin Consider home regimen, A1C,
while npo or surgery in usions or perioperative inpatient needs and oral intake
management and dose similar to rapid-acting
May use as prandial insulin analogs
instead o rapid acting Consider using rapid-acting
analogs but has slower analogs as alternative to regular or
onset and longer duration prandial insulin
beyond 2 hours. Patients with insulin pumps should be instructed ■ IV INSULIN INFUSION THERAPY
to place a resh in usion set in an area not a ected by surgery. A In critically ill patients, glycemic management with an IV insulin
retrospective review o CSII patients ound similar control between in usion is recommended or the majority o patients; it should
perioperative CSII versus conversion to perioperative IVinsulin drip. be initiated at a threshold o 180, with target blood sugars
between 140 and 180. The use o insulin protocols are highly
MONITORING AND GLYCEMIC TARGETS recommended, to reduce variability and increase the amount o
FOR CRITICALLY ILL time within target.
The American College o Physicians endorses a blood glucose target
o 140 to 200 mg/dL in surgical and medical intensive care unit (ICU)
DISCHARGE PLANNING
patients. The American Diabetes Association and American Associa-
tion o Clinical Endocrinologists endorse a blood glucose target o Discharge planning should begin shortly a ter admission, so that
140 to 180 mg/dL or most ICU patients. A target o <110 mg/dL is very clear instructions are provided to the patient at time o dis-
never recommended, due to increased hypoglycemia and mortality charge. Preoperative HbA1C can be a valuable tool in devising a dis-
seen with the NICE-Sugar trial. See Table 52-3 or urther recom- charge diabetes regimen. Umpierrez et al demonstrated improved
mendations or glycemic targets or critically ill. HbA1C at 4 and 12 weeks postdischarge with the use o a discharge
In a review o patients undergoing cardiac surgery treated with protocol based on initial HbA1C in surgical and medical patients
IV insulin in usion postoperatively with a lower glycemic target with Type 2 diabetes. Follow-up included nursing phone calls every
(80-110 mg/dL) compared with those treated using a higher target 2 weeks or the f rst 2 months and visits at 1 and 3 months:
(110-140 mg/dL), patients in the higher target group experienced • Patients with HbA1C <7% were discharged on their home regi-
similar mean glucose values, no signif cant di erences in 30-day men (oral agents or insulin).
morbidity and mortality (with the possible exception o increased • Patients with HbA1C between 7% and 9% were discharged on
reintubation), and less hypoglycemia than the lower target group. their home oral agents as well as basal insulin at 50% o the
Similarly, in a study o cardiac patients treated with IV insulin inpatient hospital regimen.
in usion to a target o 110 to 140 versus 140 to 180, there were • Patients with HbA1C >9% were discharged on their home oral
no signif cant di erences in morbidity or mortality between the agents as well as 80% o inpatient basal insulin or basal bolus
groups. There ore lower targets have been essentially abandoned therapy (using 80% basal insulin and 80% prandial insulin used
or ICU patients. in the hospital).
346
TABLE 52-3 Summary of the Most Current Professional Society Guidelines for Glycemic Control
C
H
Organization Year Patient Population Rx Threshold (mg/dL) Target Glucose (mg/dL)
A
American Diabetes Association 2015 ICU patients 180 140-180; 110-140, select pts
P
T
Surviving Sepsis Campaign 2009 ICU patients 180 <180
E
Inst.Healthcare Improvement 2009 ICU patients 180 <180
R
American Heart Association 2009 ICU patients with ACS 180 90-140
5
2
European Society o Cardiology 2009 Patients a ter major 180 140-180
noncardiac surgery
American College Physicians 2014 ICU patients 180 140-200
P
e
Societyo Thoracic Surgeons 2009 ICU patients a ter 180 <180
r
i
o
Adult cardiac surgery <150 i requires
p
e
> 3 days ICU care or
r
a
t
i
ventilatory, inotrophic
v
e
Ormechanical support,
R
i
s
Renalin su iciency,
k
A
on antiarrhythmics
s
s
e
s
s
m
e
CONCLUSION Fleisher LA, Fleischmann KE, Auerbach AD, et al. 2014 ACC/AHA
n
guideline on perioperative cardiovascular evaluation and man-
t
High-quality perioperative care o diabetic patients can reduce
a
agement o patients undergoing noncardiac surgery: a report
n
morbidity, mortality, and cost. A thorough preoperative diabetes
d
evaluation along with clear communication with all members o o the American College o Cardiology/American Heart Associa-
M
the surgical care team remains necessary or the development and tion Task Force on Practice Guidelines. Circulation 2014;130(24):
a
n
implementation o an individualized perioperative care plan. Hospi- e278-e333.
a
g
talists have the opportunity to collaborate and coordinate care or Handelsman Y, Bloomgarden ZT, Grunberger G, et al. American
e
m
these patients throughout the perioperative process. Hospitalists association o clinical endocrinologists and american college
e
can play a key role in identi ying areas o need and collaborating o endocrinology—clinical practice guidelines or developing
n
t
with surgical teams and diabetologists in improving perioperative a diabetes mellitus comprehensive care plan. Endoc Prac. 2015;
o
and inpatient care. 21(Suppl 1):1-87.
t
h
Qaseem A, Chou R, Humphrey LL, Shekelle P. Inpatient glycemic
e
control: best practice advice rom the Clinical Guidelines Com-
D
SUGGESTED READINGS
i
mittee o the American College o Physicians. Am J Med Qual.
a
b
2014;29(2):95-98.
e
Barker P, Creasey PE, Dhatariya K, et al. Peri-operative management
t
i
c
o the surgical patient with diabetes 2015: Association o Anaes- Udovcic M, Castro JC, Apsey HA, Stearns JD, Schlinkert RT, Cook CB.
P
thetists o Great Britain and Ireland. Anaesthesia. 2015. online early Guidelines to improve perioperative management o diabetes
a
t
release. doi: 10.1111/anae.13233. mellitus: assessment o the impact o change across time. Endoc
i
e
n
Bock M, Johansson T, Fritsch G, et al. The impact o preoperative Prac. 2015;21(9):1026-1034.
t
testing or blood glucose concentration and haemoglobin A1c Underwood P, Seiden J, Carbone K, et al. Early Identif cation o
on mortality, changes in management and complications in Individuals with Poorly Controlled Diabetes Undergoing Elective
noncardiac elective surgery: a systematic review. Eur J Anaesth. Surgery: improving A1C testing in the preoperative period. Endoc
2015;32(3):152-159. Prac. 2015;21(3):231-236.
Evans AS, Hosseinian L, Mechanick JI. Emerging paradigms on Umpierrez GE, Reyes D, Smiley D, et al. Hospital discharge algorithm
glucose managemnt in the intensive care unit. Minerva Endocrinol. based on admission HbA1c or the management o patients with
2014;39:261-273. type 2 diabetes. Diabetes Care. 2014;37(11):2934-2939.
347
53
Many hospitalists will be asked to assess the operative risk o
patients who have acute or chronic liver disease. The ollowing
chapter outlines an assessment plan and a basis or predicting
operative morbidity and mortality. Evaluation o patients with liver
disease prior to surgery is crucial to estimate perioperative morbid-
ity and mortality. The operative risk o liver disease can be related to
the rapid changes in liver unction that can occur in acute hepatitis,
or can be related to chronic complications o portal hypertension
Preoperative and parenchymal liver disease in patients with cirrhosis. There ore,
establishment o a risk pro le should be based on the etiology o
Evaluation of the underlying liver disease, the degree o hepatic decompensation

associated with the presence o cirrhosis and portal hypertension,
Liver Disease and the type o surgery the patient is undergoing.
FACTORS THAT AFFECT PERIOPERATIVE

OUTCOMES IN LIVER DISEASE PATIENTS
Amir A. Qamar, MD ■ CHANGES IN HEPATIC BLOOD FLOW
Norman D. Grace, MD The liver receives dual blood supply rom the portal vein and the
hepatic artery. Unlike most other organs, the majority o hepatic
oxygen supply in normal individuals is venous, via the portal vein.
Administration o anesthesia and surgery in uences both portal and
hepatic blood ow; usually, when ow through the portal vein is
reduced, the hepatic artery vasodilates to increase oxygen supply
to the liver. This compensatory vasodilatation is reduced in patients
with altered hepatic architecture (such as brosis and nodular or-
mation associated with cirrhosis). Due to intraoperative decreases
in blood pressure and cardiac output, blood ow in patients with
cirrhosis is decreased in the portal vein, splanchnic vessels, and
hepatic artery; anesthetics also reduce the hepatic artery’s ability to
vasodilate in response to these changes in portal blood ow.
These changes in hepatic blood ow may lead to hepatic isch-
emia and necrosis. The release o in ammatory mediators may result
in multiorgan system ailure. In a study o 733 cirrhosis patients
undergoing surgery, Ziser et al ound an 11.6% mortality rate.
Intraoperative hypotension correlated strongly with perioperative
complications and decreased survival.
■ TYPE OF SURGERY
Postoperative morbidity and mortality in patients with cirrhosis is
also in uenced by the type o surgery.
Abdominal surgery
During abdominal surgery, direct trauma due to surgical retrac-
tion can lead to hepatic injury. Manipulation o the splachnic and
portal vasculature may also reduce portal or hepatic ow leading
to ischemic injury. In particular, patients with Child-Pugh class C
cirrhosis who undergo abdominal surgery have a 75% perioperative
mortality.
Cardiovascular surgery
Cardiovascular surgery, due to e ects on portal and hepatic artery
blood ow, is also associated with high perioperative morbidity and
mortality. The need or perioperative pressor support and prolonged
cardiopulmonary bypass are actors that strongly correlate with
hepatic injury.
348
Emergency Surgery
TABLE 53-2 Operative Mortality Rates in Patients Undergoing
C
Many patients who require emergency surgery may be hemody- Abdominal Surgery Based on Child-Pugh Class
H
namically unstable rom systemic vasodilation (eg, sepsis) or hypo-
A
tensive due to hemorrhage (eg, trauma, abdominal surgery), and Child-Pugh Class Risk of Operative Mortality
P
their outcome is o ten poor regardless o underlying conditions. In A 10%
T
a study by Demetriades and colleagues o 46 patients with cirrhosis
E
B 30%
R
who underwent emergency laparotomy, the postoperative mortal-
C 80%
ity rate was 45%, which was signi cantly greater than noncirrhotic
5
3
control patients.
Adapted rom Garrison RN, Cryer HM, Howard DA, et al. Clari ication o risk
Orthopedic surgery actors or abdominal operations in patients with hepatic cirrhosis. Ann Surg.
P
1984;199(6):648-655 and Mansour A, Watson W,; Shayani V, et al. Abdominal
r
There is little in ormation in the literature regarding speci c opera-
e
operations in patients with cirrhosis: still a major surgical challenge. Surgery.
o
tive risks or patients with cirrhosis who undergo orthopedic surgery. 1997;122(4):730-735.
p
e
Hsieh et al reviewed 38 patients who underwent hip arthroplasty
r
a
over a 20-year period and ound the 30-day complication rate was
t
i
with cirrhosis, compared to 31 age- and sex-matched control
v
27%. Advanced cirrhosis, age, elevated serum creatinine, low serum
e
patients, the Child-Pugh score accurately predicted morbidity a ter
E
albumin, low platelet count, ascites, hepatic encephalopathy, and
v
cholecystectomy. In another study o 44 patients with cirrhosis who
a
high operative blood loss correlated with the high complication rate.
l
underwent cardiac surgery, a preoperative Child-Pugh score ≥8 was
u
a
predictive o postoperative mortality. Table 53-2 summarizes the
t
■ TYPE OF ANESTHETIC
i
o
operative mortality risk o patients with di erent Child-Pugh scores
n
Anesthetic agents may reduce hepatic blood ow by reducing car- who undergo abdominal surgery.
o
diac output. Even spinal and epidural anesthesia, by reducing the
L
mean arterial pressure, may a ect hepatic blood ow. In patients
i
v
MELD score
e
with liver disease, e ects on hepatic metabolism may lead to pro-
r
The Model or End Stage Liver Disease (MELD) score predicts short
D
longed action o anesthetic agents or production o toxic radicals
term mortality in patients with cirrhosis. The scoring system uses
i
s
resulting in increased morbidity and mortality.
e
serum bilirubin, creatinine, and prothrombin time (international nor-
a
s
malized ratio) to assess hepatic unction. Recent data also suggests
e
■ CAUSE AND SEVERITY OF LIVER DISEASE
the presence o hyponatremia urther increases the risk o mortality.
Perioperative morbidity and mortality is highly in uenced by the Teh et al assessed 772 patients with cirrhosis who underwent
etiology and severity o the patient’s liver disease. The presence o major digestive, cardiac or orthopedic surgery. They ound that the
cirrhosis or acute hepatitis at the time o surgery adversely in u- 30-day mortality ranged rom 5.7% or MELD score <8 to 50% or a
ences surgical outcomes. Generally, patients with chronic hepatitis MELD score >20. Table 53-3 summarizes the mortality risk or surgi-
rom any etiology, without eatures o hepatic decompensation, do cal patients with di erent preoperative MELD scores.
very well with surgery and speci c precautions are not necessary.
However, in patients with acute liver disease or cirrhosis, it is critical PREOPERATIVE MANAGEMENT
to assess their perioperative risk as part o in ormed consent. Acute
hepatitis, especially alcoholic hepatitis, or decompensated cirrhosis The preoperative management o patients with cirrhosis should
are relative contraindications to elective surgery. A number o scor- involve aggressive treatment o portal hypertension and hepatic
ing and staging systems are use ul in assessing the perioperative insuf ciency to reduce operative morbidity and mortality. Depend-
risk, the most common o which are the Child Pugh Score and the ing on the speci c etiology o the liver disease, certain preoperative
MELD score. Measurement o the hepatic venous pressure gradient changes in management may be considered (more below).
(HVPG) has excellent prognostic value but its use is con ned to a
limited number o medical centers. A recent study showed that ■ DIAGNOSTIC STUDIES
patients with compensated cirrhosis with clinically signi cant portal Table 53-4 lists laboratory and radiology studies that should be
hypertension, de ned by a HVPG ≥10 mm Hg, were at signi cant risk considered in the preoperative assessment o a patient with liver
o developing clinical decompensation de ned as the occurrence o disease to ascertain their perioperative risk.
jaundice, ascites, variceal bleeding, or hepatic encephalopathy.
■ COAGULOPATHY
Child Pugh Score Hepatic insuf ciency leads to inadequate production o actors II,
The Child-Pugh Score combines a subjective and objective assess- V, VII, and IX resulting in the development o coagulopathy with
ment o liver unction (Table 53-1). In a recent study o 33 patients an increased risk o perioperative bleeding. Thrombocytopenia
TABLE 53-1 Child Pugh Scoring System
1 2 3
Bilirubin <2.0 mg/dL 2.0-3.0 mg/dL >3.0 mg/dL
Albumin >3.5 g/dL 3.5-2.8 g/dL <2.8 g/dL
PT <4 s greater than control 4-6 s greater than control >6 s greater than control
Ascites Absent Mild-Moderate Moderate-Severe
Encephalopathy Absent Mild (Grade I-II) Severe (Grade III-IV)
Score 5-6: Child Class A; 7-9: Child Class B; ≥10: Child Class C.
349
TABLE 53-3 Preoperative MELD Scores and 7-day and 30-day TABLE 53-5 Risks Factors for Morbidity and Mortality in
Mortality Rates Patients with Biliary Obstruction
P
MELD Score 7d 30 d • Hct < 30
A
R
0-7 (n = 351) 1.9 (314) 5.7 (301) • Bilirubin > 11 mg/dL
T
8-11 (n = 257) 3.3 (236) 10.3 (219) • BUN > 90
I
I
12-15 (n = 106) 7.7 (94) 25.4 (78) • Creatinine > 1.4 mg/dL
16-20 (n = 35) 14.6 (29) 44.0 (19) • Albumin < 3.0 g/dL
21-25 (n = 13) 23.0 (7) 53.8 (4) • Age > 65 y
≥26 (n = 10) 30.0 (6) 90.0 (1) • AST > 90
M
e
• Malignancy
d
MELD, Model or End-stage Liver Disease.
i
c
a
With permission rom Teh SH, Nagorney DM, Stevens SR, et al. Risk actors
l
or mortality a ter surgery in patients with cirrhosis. Gastroenterology.
C
o
2007;132(4):1261-1269. With permission rom Elsevier.
n
and spironolactone) and paracentesis. Transjugular intrahepatic
s
u
portosystemic shunting (TIPS) in the management o ascites prior to
l
t
surgery is not recommended except or patients who may be under-
a
in liver disease is multi actorial, including portal hypertension-
t
i
going liver transplantation. For patients with hypoalbuminemia,
o
induced sequestration o platelets, which urther increases the risk
n
ascites or edema, perioperative uid management should include
o bleeding. Administration o platelets, resh rozen plasma and
the use o colloids such as albumin. In patients without third space
cryoprecipitate should be administered prior to surgery to reduce
accumulation o uids, crystalloids (ie, saline) are appropriate.
the risk o bleeding during and a ter surgery. There has been inter-
est in the use o recombinant actor VIIa in situations o emergency
uncontrollable bleeding in cirrhotic patients. However, studies have ■ TRANSJUGULAR INTRAHEPATIC PORTOSYSTEMIC
SHUNT PRIOR TO SURGERY
not shown any improved outcomes with the use o this actor in
patients undergoing hepatic resection, liver transplantation, and Transjugular intrahepatic portosystemic shunts are e ective in
upper gastrointestinal bleeding. This agent should be used with reducing the complications o portal hypertension (such as variceal
caution due to an increased risk o arteriothromboembolic events hemorrhage and ascites). Preoperative placement o a TIPS may
including myocardial ischemia, myocardial in arction, cerebral reduce the risk o portal hypertensive complications in patients with
ischemia, and cerebral in arction. cirrhosis, but data regarding its e ectiveness is limited.
■ VARICES ■ BILIARY OBSTRUCTION

The risk o bleeding in patients with cirrhosis who have esophageal Biliary obstruction preoperatively is associated with increased mor-
varices is approximately 8% per year. Once variceal bleeding occurs, bidity and mortality and should be treated with decompression
there is a 20% mortality associated with the bleeding event. Preven- prior to surgery. Speci c actors that increase operative risk are
tive therapies include the use o nonselective beta blockers and noted in Table 53-5.
endoscopic variceal ligation. There are no speci c treatments to
reduce the risk o perioperative variceal bleeding. Preventing “over ■ HEPATIC ENCEPHALOPATHY
trans usion” with a target hematocrit o no more than 25 is recom- Hepatic encephalopathy is provoked by a number o actors listed in
mended. In patients who do experience postoperative gastrointes- Table 53-3. The risk o hepatic encephalopathy a ter surgery can be
tinal bleeding, the use o antibiotics is recommended to reduce the reduced by avoiding speci c precipitants (Table 53-6).
risk o mortality rom spontaneous bacterial peritonitis (SBP).
■ ADRENAL INSUFFICIENCY
■ ASCITES AND SPONTANEOUS BACTERIAL Adrenal insuf ciency can be ound in 33% o patients with ulmi-
PERITONITIS
nant hepatic ailure and 66% o patients with cirrhosis. In patients
Ascites is the most requent and generally the rst event to occur with decompensated cirrhosis or acute liver ailure who develop
in patients with cirrhosis and hepatic decompensation. Treatment
includes salt restriction, the use o diuretics (including urosemide
TABLE 53-6 Precipitants of Hepatic Encephalopathy

TABLE 53-4 Laboratory and Radiologic Testing to Be
Considered for Preoperative Risk Assessment in • Gastrointestinal bleeding
the Patient with Liver Disease • Hypovolemia
Liver chemistries • Renal ailure
Complete blood count • Use o sedating agents
Prothrombin time and international normalized ratio • Hypokalemia
Electrolytes and renal unction • Alkalosis
Abdominal ultrasound and Doppler study o hepatic vasculature • Trauma
CT scan or MRI o liver • In ection
Hepatic venography or hepatic venous pressure gradient • Constipation
(HVPG) • Colon surgery
350
liver enzyme abnormalities (without cirrhosis) and most compen-
TABLE 53-7 Etiology-Specific Perioperative Management sated Child-Pugh Class A patients can sa ely undergo surgery. For
C
all other patients, a care ul assessment o the bene ts o surgical
H
Etiology of Liver
A
Disease Perioperative Management intervention must be weighed against the risk o hepatic decom-
P
pensation and mortality. These risks should be enumerated as part
Alcoholic liver Should be abstinent ≥3 mo preoperative
T
o the in ormed consent process.
disease (i possible)
E
R
Should be assessed or alcohol
SUGGESTED READINGS
5
dependence preoperative
3
I evidence o alcoholic hepatitis, surgery
should be delayed Angeli P, Gines P, Wong F, et al. Diagnosis and management o
acute kidney injury in patients with cirrhosis: revised consensus
P
Hepatitis B and C Treatment with nucleoside/nucleotide
r
recommendations o the International Club o Ascites. J Hepatol.
e
analogues or direct antiviral agents
o
2015;62:968-974.
p
should continue perioperative
e
Garcia-Tsao G. Current management o the complications o cirrhosis
r
All new perioperative medications
a
t
should be reviewed or interactions with and portal hypertension: variceal hemorrhage, ascites, and sponta-
i
v
e
Hepatitis B and C drugs neous bacterial peritonitis. Gastroenterology. 2001;120(3):726-748.
E
Inter eron should be held perioperative Garcia-Tsao G, Sanyal AJ, Grace ND, et al. Prevention and manage-
v
a
l
Autoimmune Treatment with prednisone or ment o gastroesophageal varices and variceal hemorrhage in
u
a
hepatitis azathioprine should continue cirrhosis. Hepatology. 2007;46(3):922-938.
t
i
o
perioperative (stress dose steroids i Gustot T, Fernandez J, Garcia E, et al. Clinical course o acute-on-
n
warranted) chronic liver ailure syndrome and e ects on prognosis. Hepatology.
o
α-1 antitrypsin Should be evaluated or underlying lung 2015;62(1):243-252.
L
i
deficiency disease
v
Mansour A, Watson W, Shayani V, et al. Abdominal operations in
e
Hemochromatosis Should be evaluated or underlying
r
patients with cirrhosis: still a major surgical challenge. Surgery.
D
cardiomyopathy
i
1997;122(4):730-735.
s
e
Wilson’s disease Treatment with penicillamine should
a
continue, but at reduced dose Qamar AA, Grace ND. Abnormal hematological indices in cirrhosis.
s
e
perioperative Can J Gastroenterol. 2009;23(6):441-445.
Treatment with zinc or trientene should Surman A, Barnes DS, Zein NN, et al. Predicting out a ter cardiac
continue perioperative surgery in patients with cirrhosis: a comparison o Child-Pugh and
MELD score. Clin Gastroetnerol Hepatol. 2004;(8):719-723.
Teh SH, Nagorney DM, Stevens SR, et al. Risk actors or mortality
a ter surgery in patients with cirrhosis. Gastroenterology.
signs o adrenal insuf ciency (including hypotension and hypogly- 2007;132(4):1261-1269.
cemia a ter surgery), treatment with stress dose steroids should be Ziser A, Plevak DJ, Wiesner RH, et al. Morbidity and mortality in cir-
considered. rhotic patients undergoing anesthesia and surgery. Anesthesiology.
1999;90(1):42-53.
■ ETIOLOGY-SPECIFIC MANAGEMENT
Perioperative management by liver disease etiology is outlined in
Table 53-7.
CONCLUSION
It is important to identi y the operative risk o patients with liver
disease based on the acuity, etiology and severity o the liver disease
and the urgency and type o surgery. Generally, patients with mild
351
54
Hematologic disorders are diseases o circulating blood cells and
plasma proteins that play a role in oxygen delivery, in ammation,
in ection control, hemostasis, and thrombosis. Given that surgery
can result in bleeding and induce hemostatic changes that pro-
mote thrombosis, it is not surprising that patients with hematologic
disorders present serious preoperative management challenges.
The ability o hospitalists to manage diseases associated with blood
disorders, and hemostatic and thrombotic risks associated with sur-
Preoperative gery is vital to the wel are o these patients. This chapter will discuss
preoperative assessment o patients with hematologic disorders,
Assessment of and review the risk o perioperative hematologic complications and

speci c management strategies to reduce perioperative risks in this
Patients with vulnerable patient population.
Hematologic PREOPERATIVE ASSESSMENT OF PATIENTS

WITH HEMATOLOGIC DISORDERS
Disorders ■ HISTORY AND PHYSICAL EXAMINATION

The most important aspect in the preoperative assessment o
patients with a hematologic disorder is a thorough history. This is
especially true or patients with hemostatic disorders (Table 54-1).
Ming Y. Lim, MB BChir Though the patient may have a known hemostatic diagnosis, the
clinical phenotype o these patients may vary considerably. Occa-
Charles S. Greenberg, MD sionally, a patient may report an unexpected personal and/or amily
history o bleeding or thrombosis, or a hematologic diagnosis as a
child with no subsequent ollow-up. Depending on the severity o
the presumed diagnosis, these subjective accounts may need to
be con rmed objectively. The physical examination is rarely help ul
in such situations. However, the presence o petechiae, purpura,
ecchymoses, jaundice, ascites, and splenomegaly may alert one to
the presence o potential hematologic disorder.
Similarly, given the high perioperative complications in patients
with sickle cell disease (SCD), a care ul history is critical to ensure
optimal surgical outcomes. Speci c questions include recent cough,
wheezing, dyspnea on exertion, ever, ankle edema, right upper
quadrant pain, change in stool or urine color, hematuria, or dysuria.
In addition, details on past SCD-related complications should be
documented, including acute chest syndrome, requency o pain
episodes, strokes, hyperhemolytic crisis, pulmonary hypertension,
aplastic crisis, trans usion reactions, and alloimmunization.
In addition to the speci c questions above, all patients with
hematologic disorders should have the ollowing documented:
alcohol use, smoking history, past anesthesia history, current medi-
cations (including vitamins, supplements, and herbal prepara-
tions), and allergies. Herbal remedies are o ten sel -administered by
patients and can modi y hemostasis. These medications should be
stopped prior to surgery.
■ DIAGNOSTIC STUDIES
It has been established by multiple prospective studies that the
use o routine preoperative coagulation testing is not help ul in the
absence o a bleeding history. In contrast, a preoperative complete
blood count and coagulation testing is essential in patients with
known hematologic disorders. Table 54-2 list laboratory and radiol-
ogy studies that should be considered in the preoperative assess-
ment o these patients.
352
PERIOPERATIVE RISK ASSESSMENT AND MANAGEMENT
TABLE 54-1 Recommended Questions to Ask in Patients with FOR HEMATOLOGIC DISORDERS
C
Known or Suspected Hemostatic Disorders
H
Patients with hematologic disorders are at increased risk or compli-
A
• Have you or anyone in your amily ever been labeled a cations related to changes in blood cells and plasma actors initiated
P
“bleeder”? Has someone in the amily ever experienced by surgery. The magnitude o the risk may vary considerably with
T
abnormal bleeding? the underlying hematologic disorder, nature o the surgery, age o
E
R
• Have you ever bled with surgery or ollowing childbirth? What the patient and other comorbidities. These disorders may require
surgical procedures have you had, including major surgery, speci c treatment and consultation with hematologists to ensure
5
4
minor surgery, biopsies, and dental extractions? an accurate assessment o risk o perioperative complications and
• Did a surgeon or dentist ever have to re-explore the wound to improve outcomes. This involves understanding the nature and
site or did you ever have to return to the operative suite or severity o the hematologic disorders with the goal to minimize
P
r
hemorrhagic control? the risk o intraoperative bleeding and postoperative complications
e
o
Have you ever had excessive menstrual periods? How long including bleeding, in ection, thrombosis, and abnormal wound
p
•
e
do your periods last? How many pads or tampons are needed healing. This section discusses several common hematologic disor-
r
a
each day? Have you ever required iron supplementation or ders that hospitalists may encounter, the nature o the perioperative
t
i
v
anemia due to a menstrual blood loss? risk and speci c perioperative management.
e
A
• Do you bruise excessively? Are these bruises multiple? Are
s
they con ined only to the outer thighs or other areas that are ■ DISORDERS OF RED BLOOD CELLS
s
e
subject to trauma? Are any o these bruises palpable (ie, are
s
Anemia is a very common abnormality in patients undergoing sur-
s
they true hematomas) or are they level with the sur ace o the
m
skin? gery and is even more common in patients with active hematologic
e
n
disorders due to treatment, marrow involvement by the disease,
t
• Do you have nosebleeds now or was there ever a time in your
bleeding, renal ailure, in ammation, or advanced age. Adequate
o
li e when you did have spontaneous nosebleeds?
red cell mass is needed to promote tissue oxygen delivery and is
P
• Have you ever required a blood or plasma trans usion and, i
a
a vital actor to aide wound healing and prevent myocardial and
t
so, why?
i
e
central nervous system injury in surgical patients. Red cells can also
n
• Have you ever bruised or experienced hemorrhage ollowing
t
promote hemostasis and play an important role in postoperative
s
trauma, car accidents, alls, organized or unorganized sports,
w
recovery.
altercations, or any acts o violence?
i
t
Given the importance o red cells, the presence o anemia
h
needs to be recognized preoperatively and corrected i surgery is
H
Reproduced, with permission, rom Kitchens CS, Kessler CM, Konkle BA,
e
not emergent. Previously, trans usion o red cells was considered
m
eds. Consultative Hemostasis and Thrombosis. 3rd ed. Philadelphia: Elsevier
a desirable intervention. However, the administration o red cells
a
Saunders; 2013.
t
was discovered to adversely a ect the morbidity and mortality o
o
l
o
hospitalized patients in a wide variety o clinical settings. Preopera-
g
tive anemia has to some extent been a silent risk actor or adverse
i
c
clinical outcomes due to unnecessary trans usion o red cells that
D
i
can increase perioperative complications. On the other hand, excess
s
o
number o red cells may cause hyperviscosity and lead to thrombo-
r
TABLE 54-2 Laboratory and Radiology Studies
d
e
sis. For instance, patients with polycythemia may need preopera-
r
s
All patients with hematologic disorders tive phlebotomy to reduce the risk, especially when it is caused by
• Complete blood count with di erential polycythemia vera. Common hematologic disorders that can cause
• Coagulation testing (PT, aPTT) anemia and polycythemia will be reviewed.
• Biochemistry pro ile
Nutritional anemias
• Liver unction tests
Patients with iron de ciency anemia should have their anemia cor-
• Type and cross
rected preoperatively with oral iron i there is suf cient time, or with
Hemostatic disorders* intravenous iron i a more urgent response is required. The cause
• Platelet unction tests o iron malabsorption or loss should be identi ed and any possible
• von Willebrand antigen and ristocetin co actor source o bleeding identi ed. Folic acid and vitamin B12 are nutri-
• Factor assay (eg, actor V, actor VIII, actor IX, actor XI, actor XII) ents required to promote proli eration o marrow cells to produce
red cells, as well as white cells and platelets. Once these nutrients
Hemolytic anemias
are replete, it may take several days to start to see a reticulocyte
• Peripheral blood smear response and several weeks to restore the blood count to normal.
• Reticulocyte count More commonly, some patients do not have a single actor con-
• Iron studies and erritin tributing to anemia and correction is o ten dif cult due to diseases
• Vitamin B12 and olate levels associated with renal insuf ciency and chronic in ammation that
• Lactate dehydrogenase (LDH)
can adversely a ect iron utilization, resulting in reduced red cell pro-
duction. In such circumstances, red cell trans usion may be required
Sickle cell disease
to maintain adequate tissue oxygen delivery during surgery.
• Same as in patients with hemolytic anemia
• Hemoglobin electrophoresis Hemolytic anemias
• Chest radiograph Anemia due to production o IgG or IgM autoantibodies is challeng-
ing and may require pre- and postoperative interventions. Patients
*
Order based on their speci ic hemostatic disorder. with autoimmune-mediated hemolytic anemia (AIHA) usually have
aPTT, activated partial thromboplastin time; PT, prothrombin time. positive Coombs tests, with pan-agglutination o their plasma with
353
donor red cells. The presence o warm IgG autoantibodies results limited pain crisis while others can have severe hemolytic anemia
in hemolysis o both the native and trans used red cells making complicated by pulmonary hypertension, liver disease, renal insu -
it dif cult to provide “crossmatch” compatible blood or surgery. ciency and osteonecrosis o hip and other bones. Strategies or
P
Despite this, when medically necessary, serologically incompat- risk reduction require a multidisciplinary approach with surgery,
A
ible but type-speci c red cells should be trans used i the patient’s anesthesia and medicine developing a customized care plan or
R
hemoglobin levels are causing hypotension and organ dys unc- the patient based on surgical risk and severity o sickle cell disease.
T
tion. Acutely decompensating patients with AIHA who received During the perioperative period, patients with sickle cell disease are
I
I
serologically incompatible blood trans usion do not experience at increased risk o pain ul crisis, acute chest syndrome, and cere-
trans usion-related alloimmunization or an increase in hemolysis. brovascular accidents as surgical procedures may be complicated
Thus, li e-saving red cell trans usion should not be denied due to by hypoxia, hypothermia or acidosis; all o which promotes red
serologically incompatible blood. cell sickling. The surgical and anesthesia team need to pay special
M
Patients with cold agglutinin disease produce IgM antibodies attention to perioperative hydration, oxygenation status and under-
e
d
that react in the cold and x complement to the red cells, promot- lying pulmonary, and cardiovascular disease. I the surgery involves
i
c
ing extravascular clearance by the reticuloendothelial system. Cold general anesthesia, it is strongly recommended that red blood cells
a
l
agglutinin disease can occur in response to mycoplasma in ec- be trans used (either through simple or exchange trans usion) in
C
o
tion, autoimmune disease and in lymphoproli erative disorders, patients with sickle cell anemia (Hb SS) to bring the hemoglobin
n
s
such as Waldenstrom’s macroglobulinemia. These antibodies can level to 10 g/dL as this has been shown to reduce perioperative
u
l
be detected with a positive Coombs test that is complement- mortality and complications. In patients who already have a hemo-
t
a
globin level higher than 8.5 g/dL without trans usion, are on chronic
t
mediated. Blood trans usion in patients with cold agglutinin disease
i
o
should be in used via a blood warmer. I the antibodies are actively hydroxyurea therapy or require high-risk surgery (eg, neurosurgery,
n
promoting hemolysis at room temperature, the patient should be prolonged anesthesia, cardiac bypass), a hematologist with experi-
kept warm to prevent hemolysis by increasing the room thermostat. ence in sickle cell disease management should be consulted or
Many patients with antibody-mediated hemolytic anemia may guidance on appropriate trans usion methods.
have had a surgical splenectomy as treatment or their hematologic As these patients are at high-risk or red blood cell (RBC) alloim-
disorder. Splenectomized patients are at increased risk or both bacte- munization, a presurgery type and screen is sent to Blood Bank
rial in ection and thromboembolism, both o which can a ect periop- ahead o time to identi y and characterize RBC antibodies (i any),
erative morbidity and mortality. All splenectomized patients should which can be a laborious e ort. I no antibodies are ound, typically
receive immunization against the ollowing organisms: Streptococcus leukocyte-reduced RBC units, which are phenotypically matched or
pneumoniae, Haemophilus inf uenzae type B, and Neisseria meningitidis. C, E, and Kell as well as ABO and D antigens, are trans used. I anti-
These vaccinations should be given at least 14 days prior to elective bodies are present, nding compatible phenotypically match units
splenectomy or immediately a ter an emergent splenectomy. In addi- can be time-consuming, hence the need to plan ahead.
tion, splenectomized patients are at increased risk or venous throm- Postoperatively, i the patient has a history o acute chest
boembolism (VTE) and should receive adequate VTE prophylaxis in syndrome, admission to the intensive care unit should be con-
the immediate postoperative period. sidered or close respiratory monitoring. Regardless o location,
Patients with atypical hemolytic-uremic syndrome (aHUS) or the primary team should pay particular attention to hydration
paroxysmal nocturnal hemoglobinuria (PNH) experience hemolysis and oxygenation status. Dehydration and low oxygen levels can
due to de ects in complement regulation. These patients may be precipitate erythrocyte sickling, leading to vaso-occlusive crisis. For
receiving treatment with the complement inhibitor, eculizumab. management o postoperative surgical pain, patients with chronic
Since this drug inhibits complement activation, a major mechanism pain may have opioid tolerance and require higher doses o nar-
to ght in ection, these patients are prone to meningococcemia. cotics. Reports o pain should be treated accordingly and narcotics
A raised awareness o the possibility o meningococcal in ection is should not be withheld or ear o addiction as stress rom acute
warranted in these patients. pain can trigger the onset o a vaso-occlusive crisis. On the other
Patients with hemolytic disorders should receive olic acid, hand, overhydration and excessive narcotic use can lead to pulmo-
vitamin B12 and iron supplementation to promote hemoglobin nary edema and respiratory depression, respectively. Hence, these
synthesis and red cell production. These nutrients should be replete patients require strict monitoring o hydration and oxygenation
prior to any elective surgery. Patients with coexisting renal ailure or status. An incentive spirometer should be prescribed to all patients
chronic in ammation may not produce adequate erythropoietin to with sickle cell disease postoperatively as this has been shown to
promote accelerated erythropoiesis in response to anemia. These reduce pulmonary complications. Patients with sickle cell disease
patients typically have red cell hal -lives that are reduced at least are also at increased risk o venous thromboembolism (VTE). Given
8- old and require reticulocyte levels >10% to maintain normal lev- that surgery is a known transient risk actor or VTE, appropri-
els o red blood cells. Exogenous erythropoietin injections may be ate VTE prophylaxis (either mechanical and/or pharmacological)
required to compensate or the shortened red cell survival. should be prescribed.
Hemolytic anemia can also occur due to metabolic de ects in red
Sickle cell disease and other hemoglobinopathies cells. Red cell membrane proteins are prone to oxidative damage.
There exists a substantial degree o variability in the clinical sever- The red cells must be capable o reducing oxidatively damaged
ity o all molecular orms o sickle cell disease. This is due to a wide cells or they will be removed rom the body, resulting in hemolysis.
variety o abnormalities in the hemoglobin molecule. The most Glucose-6-phoshate dehydrogenase de ciency is a common
common orm o sickle cell disease is due to a homozygous point X-linked metabolic disorder that can lead to hemolytic anemia.
mutation in the beta chain o hemoglobin that leads to the intracel- Patients with this disorder are usually well compensated but will
lular polymerization o soluble hemoglobin. The cells undergo a become symptomatic when exposed to drugs that lead to oxida-
change in their shape and orms cells that can be seen on blood tive stress. These patients must not be exposed to speci c drugs
smears as irreversible sickle cells. Sickle cell disease presents major that can lead to a major hemolytic crisis. The list o drugs that
challenges to the hospitalist during the pre- and perioperative care. causes hemolysis is published online at: http://g6pd.org/en/G6PD-
There can be variable degrees o anemia and end organ damage De ciency/Sa eUnsa e/DaEvitare_ISS-it and should be reviewed to
in sickle cell disease. Some patients experience mild anemia and minimize the hemolytic risk to these patients preoperatively.
354
■ DISORDERS OF STEM CELLS: MYELODYSPLASTIC increases the neutrophil count and decreases the severity and re-
SYNDROME AND MYELOPROLIFERATIVE DISORDERS quency o in ections.
C
Patients with acute hematologic malignancies (acute myeloid or
H
There are several hematologic disorders that can alter white blood
A
cell counts and qualitative unction that may lead to serious compli- lymphocytic leukemia) are o ten anemic, leukopenic and/or thrombo-
P
cation or patients should they require surgical intervention. There cytopenic. They are at a high perioperative risk or bleeding and in ec-
T
are a growing number o individuals over 60 years o age with tion risk. These cytopenias can occur as a clinical mani estation o the
E
underlying hematologic disorder or rom bone marrow suppression
R
myelodysplastic syndrome (MDS). This disease syndrome represents
a wide variety o de ects in the bone marrow stem cells that can rom treatment with antineoplastic agents. I a patient receiving treat-
5
ment or a hematologic malignancy requires surgery, the cytopenias
4
cause abnormalities in the production o red cells, white cells and/
or platelets. These patients may have varying degrees o blood cell secondary to antineoplastic agents or radiation therapy usually recov-
production abnormalities and be at risk or bleeding, in ection or ers 14 to 21 days a ter exposure to these marrow suppressive agents.
P
It may be prudent to delay elective surgery to allow or bone marrow
r
thrombosis and cardiopulmonary compromise due to anemia. They
e
o
may require support with red cells because they are severely anemic recovery to minimize both bleeding and in ection risk. I surgery is
p
imminent, red cells and platelet trans usion may be required along
e
and have a low reticulocyte count. Some patients with MDS are
r
a
responsive to exogenous erythropoietin but others require trans u- with aggressive antimicrobial therapy. The use o G-CSF may also be
t
i
v
sion support. The hematologist caring or these patients should considered to promote leukocyte recovery.
e
In addition, these patients are o ten deconditioned with poor
A
assist in providing guidance regarding marrow unction.
s
In patients with myeloproli erative disorders, such as polycy- unctional status rom the e ects o chemotherapy and prolonged
s
e
hospitalization, placing them at higher risk or postoperative com-
s
themia vera (PV) and essential thrombocythemia (ET), there is an
s
plications. Antineoplastic agents can also result in poor wound
m
increased perioperative thrombohemorrhagic risk. A high propor-
e
tion o these surgeries are complicated by vascular occlusion or by healing. Postoperatively, care ul attention should be paid to their
n
nutritional status and rehabilitation to minimize these risks.
t
major hemorrhage. Uncontrolled PV leads to marked increase in
o
blood viscosity and stasis, reduced capillary blood ow, and subse-
■ DISORDERS OF PLATELETS: CONGENITAL PLATELET
P
quently tissue hypoxia. The risk o cardiovascular death and major
a
DISORDERS AND THROMBOCYTOPENIA
t
i
thrombosis is reduced in patients with PV when the hematocrit
e
n
is maintained at less than 45% with the use o phlebotomy, with For patients with congenital platelet disorders with a known bleed-
t
s
or without cytotoxic chemotherapy. However, it is unclear i an ing history, trans usion o single-donor platelets is usually recom-
w
acute reduction in hematocrit results in a decrease in perioperative mended prior to surgery. The most common platelet disorder is
i
t
h
thrombohemorrhagic risk. A large retrospective study o patients storage pool disease, but there are several other abnormalities that
H
with myeloproli erative disorders undergoing surgery ound that a ect platelet aggregation. Postoperatively, depending on the type
e
m
despite adequate control o blood count with phlebotomy and o surgery, additional platelet trans usion maybe recommended
a
administration o standard VTE prophylaxis, these patients had prophylactically or i clinically indicated to manage postoperative
t
o
bleeding. In patients in whom platelet trans usion is ine ective due
l
increased bleeding risk (7.3%) and high rates o symptomatic VTE
o
to alloantibody development, recombinant actor VIIa (rFVII) has
g
(7.7%) a ter surgery.
i
c
In patients with ET, the uncontrolled production o platelets and also been used with some success. In certain individuals with mild
D
megakaryocyte proli eration can o ten lead to a dramatic increase in bleeding phenotype, the consulting hematologist may recommend
i
s
o
platelet count over one million. Platelet unction may also be abnor- the use o desmopressin (DDAVP). This is given daily or twice daily
r
d
mal with de ects in both adhesion and/or aggregation responses. or about 3 days be ore it becomes ine ective due to tachyphylaxis.
e
r
The bleeding risk seen in patients with ET can be attributable to DDAVP is known to cause uid retention leading to hyponatremia,
s
either dys unctional platelets or acquired von Willebrand disease. thus strict uid balance is required to avoid electrolyte imbalances
Patients with acquired von Willebrand disease and ET are more likely postsurgery.
to su er rom hemorrhagic events than thrombotic events. The risk Another common acquired hemostatic disorder is thrombocyto-
o both thrombosis and hemorrhage is higher in older patients and penia, which can be due to decreased production (eg, bone marrow
in those who have had prior events. Lowering o hematocrit and suppression rom hematologic malignancies) or increased platelet
platelet count should ideally be done over several weeks to months. destruction (eg, immune thrombocytopenia). With the availability
Elective surgery should be delayed to allow the hematologist to o platelet trans usion, even high-risk surgeries can be per ormed
treat the elevated platelet count or increased red cell mass. I surgery in the severely thrombocytopenic patients. A platelet count o
is imminent, repetitive phlebotomy in PV patients to lower the red >100,000/µL is usually adequate or all surgical procedures, includ-
cell count or platelet pheresis to rapidly lower platelet count in ET ing high-risk procedures (eg, neurosurgical or spinal surgery). For
are also therapeutic options. low- to moderate-risk surgical procedures, trans usion to increase
the platelet count to >50,000/µL is usually recommended. Serial
monitoring o the platelet count is important during the postopera-
■ DISORDERS OF WHITE BLOOD CELLS: BENIGN AND tive period as platelet survival is usually shortened and additional
HEMATOLOGIC MALIGNANCIES platelet support may be needed. In cases o coexisting liver disease
Hospitalists sometimes encounter a patient that has a low neu- due to cirrhosis or other disease process, the patient may have an
trophil count but is not symptomatic and no history o recurrent enlarged spleen that will sequester platelets, resulting in variable
in ections. Benign ethnic neutropenia is due to a polymorphism degrees o thrombocytopenia. Furthermore, patients with hepatitis
seen in several ethnic groups o A rican and Middle East descent. C and cirrhosis may have low thrombopoietin (TPO) levels. These
These patients usually do not have an increased risk o in ection patients may respond to TPO-mimetic agents (romiplostim and
and no speci c intervention is required perioperatively. In contrast, eltrombopag) and can be treated or a short duration i needed to
patients with severe congenital neutropenia are prone to in ec- optimize management. There is a thrombotic risk with the use o
tions and have a much lower absolute neutrophil count, usually TPO-mimetic agents and this treatment should only be considered
<500/mm 3. About hal o all cases o severe congenital neutrope- when other approaches have ailed and bleeding risk is high.
nia are caused by mutations in the ELANE gene. These patients In patients with immune thrombocytopenia (ITP), prophylactic
may require granulocyte-colony stimulating actor (G-CSF), which platelet trans usion prior to surgery is usually ine ective. Instead,
355
glucocorticoids and/or intravenous γ-globulin are prescribed to
increase the platelet count. Since prednisone and other gluco- TABLE 54-3 Treatment Options for Coagulation Factor
corticoids can inter ere with wound healing, other regimens are Deficiencies
P
pre erred in the patient that needs surgery. Other options include Deficiency Treatment Options
A
RhoGAM (anti-Rh therapy i the patient is Rh antigen positive) and
R
Fibrinogen Fibrinogen concentrate, cryoprecipitate, FFP
TPO-mimetic agents to increase the platelet count. This can take
T
several days to weeks to observe a response. Factor II PCC, FFP
I
I
Factor V FFP, platelets (contains actor V)
Factor VII rFVIIa, PCC containing actor VII, FFP
■ DISORDERS OF HEMOSTASIS Factor X PCC, FFP
Patients with inherited disorders o coagulation (eg, hemophilia A, Factor XI FFP, actor XI concentrate *
M
hemophilia B) are at increased risk o intra and postoperative bleed-
e
Factor XIII Factor XIII concentrate, FFP, cryoprecipitate
d
ing. These patients o ten have an established outpatient hematolo-
i
c
gist that can de ne the severity o the patient’s disease and risk or
a
*
Not available in the United States. FFP, resh rozen plasma; PCC, prothrombin
l
bleeding. The risk should not be underestimated as bleeding into
C
complex concentrates; rFVIIa, recombinant VIIa concentrate.
o
deep tissues, vital organs or along tissue planes can lead to severe
n
s
hypotension, organ damage and death. A wide variety o hemo-
u
l
static agents are available to reduce bleeding risk. Issues that assist
t
a
Similar principles apply or patients with rare actor de ciencies
t
in de ning the dose o hemostatic agent depend upon baseline
i
o
clotting actor level, nature o surgery and prior clinical response to ( brinogen, actor II, actor V, actor VII, actor X, actor XI, and actor
n
treatments. XIII). Table 54-3 lists the type o actor de ciencies and their treat-
Speci c guidelines have been published or management o ment options.
de ects in coagulation actors. All patients should have a treatment The wound healing process is dependent upon a provisional
plan de ning the nature o the product to be used to correct the matrix composed o brin. Hemostatic de ects that limit the orma-
de ect, how the treatment should be monitored and the duration o tion o a brin matrix may lead to wound healing de ects resulting
therapy. Elective surgery should not be per ormed at an institution in delayed bleeding or reduced rate o wound closure. Patients with
that does not have the availability to monitor and in use the recom- hemophilia or other congenital bleeding disorders display delayed
mended coagulation actor replacement. Perioperative recommen- bleeding a ter trauma or surgery i the initial clot that orms is not
dations should be detailed in the medical chart prior to surgery. In stable and resistant to breakdown. Furthermore, wound healing is
the event o an emergent surgery where no recommendations are a dynamic process. The initial clot is subsequently remodeled and
available, a hematologist should be consulted. In cases o an emer- the ability to re orm brin must be maintained until the initial brin
gency, it is recommended to in use either resh rozen plasma (FFP) matrix is replaced by the synthesis o extracellular matrix.
or cryoprecipitate to stabilize the patient and trans er to a tertiary Anti brinolytics such as aminocaproic acid and tranexamic acid
center that can manage these patients. FFP can be used in cases o are o ten prescribed as adjunct therapy postsurgery in patients with
actor IX or XI de ciency. Cryoprecipitate is enriched in vWF, actor congenital hemostatic disorders. These agents inhibit tissue brino-
VIII, and brinogen and can be used in cases o hemophilia A or von lysis, leading to clot stabilization. The prophylactic use o anti brino-
Willebrand disease. lytics has been shown to reduce postoperative bleeding in cardiac,
Generally, in patients with hemophilia A, B, or von Willebrand orthopedic and liver transplant surgery. In patients with hematuria
disease, speci c actor levels should be maintained as high and or o upper urinary tract origin, anti brinolytics are not recommended
as long as clinically indicated based on published guidelines and as it can cause intrarenal or ureteral obstruction through clot orma-
clinical response to surgery. The prescribed actor is usually in used tion. O note, there have been case reports o thrombotic complica-
as a bolus in the morning prior to surgery, with speci c actor levels tions associated with the use o anti brinolytics in patients that have
measured be ore anesthesia induction to con rm that the actor chronic DIC syndrome and cancer. Judicious use o these agents is
has been administered and the appropriate actor level achieved. recommended in patients who are known to be prothrombotic (eg,
Failure to achieve an adequate actor level may be a sign o an recent thrombotic events, known coronary artery disease, or cancer
inhibitor or inadequate dosing which may require that the surgery patients).
be postponed. I the patient has developed an inhibitor the patient The lupus anticoagulant represents a common preoperative
may need to be trans erred to a center with signi cant experience issue that generates signi cant con usion and o ten leads to delays
in managing inhibitor patients. in surgery. These patients will have an abnormal aPTT and no his-
Postoperatively, the duration o maintenance therapy depends tory o bleeding, assuming there is no other de ect reported. The
on the type o surgery per ormed, ranging rom a ew days or airly lupus anticoagulant is a misnomer because patients that have this
minor surgical procedures (dental work, simple biopsies) to 2 weeks laboratory abnormality do not bleed and are in act at an increased
or more invasive surgery such as abdominal or orthopedic surgery. risk o thrombosis. Lupus anticoagulants are likely to be detected
Factor levels are monitored daily or the rst ew days and can usu- in routine pre-op screening o patients that do not have any bleed-
ally be lowered once surgical hemostasis is achieved (usually by ing history. Laboratory testing will o ten document a prolongation
the ourth or th day). Once the patient is sa e to discharge rom o the aPTT. A mixing study o the patient’s plasma with normal
a surgical standpoint, it is important to ensure that these patients plasma will not correct the de ect a ter incubating the plasma or
have an adequate care plan in place or home actor in usion 60 minutes. The only caveat that warrants urther consideration
(central line placement, actor delivered to home, sel -in usion vs is the lupus anticoagulant patient that has a bleeding history or
home health nurse in usion) as they will likely require additional a very prolonged PT. These patients may have a coexisting actor
days o actor in usion. I a care plan is not in place or not easible, inhibitor or very low prothrombin levels due to the presence o
the patient should remain in the hospital until completion o actor antiprothrombin antibodies that promote the clearance o pro-
maintenance therapy. Inadequate actor replacement therapy is one thrombin. Consultation with a hematologist will be needed to
o the main risk actor or surgical readmissions due to postoperative urther de ne the lab abnormality and provide medical clearance
bleeding and wound dehiscence in these patients. or surgery.
356
Patients with liver disease are o ten coagulopathic rom decreased
TABLE 54-4 Established Risk Factors for Perioperative hepatic synthesis o procoagulant and anticoagulant clotting ac-
C
Hematologic Complications tors, impaired hepatic clearance o brinolytics, malabsorption o
H
A
Patient specific vitamin K rom impaired bile salt recirculation, dys brinogenemia,
P
thrombocytopenia rom hypersplenism and impaired thrombopoi-
• Advanced age
T
etin production, and platelet dys unction. As both procoagulant and
E
• Other comorbidities anticoagulant activities are a ected, these patients are not “autoanti-
R
Liver disease coagulated.” Instead, these patients remain in a balanced hemostatic
5
Kidney disease state, albeit a more tenuous state, which is easily disturbed by any
4
Cardiovascular disease external stressors, such as surgery. As a result, patients with liver dis-
ease are at increased perioperative risk o bleeding and thrombosis.
Procedure specific
P
For patients with mild liver disease (prolongation o INR <2.0)
r
e
• Type o surgery
o
undergoing low- or moderate risk surgery, prophylactic interven-
p
• Choice o anesthesia tion is usually not required. For high-risk surgery or severe liver
e
r
a
disease, the correction o actors that are severely depressed may
t
i
v
aide hemostasis. However, aggressive therapy with FFP to correct
e
the INR is not easible and will lead to volume overload in these
A
■ FACTORS THAT INCREASE RISK OF
s
patients. Recent studies using thromboelastograms to guide ther-
s
e
HEMATOLOGIC COMPLICATIONS apy have demonstrated that correction o the INR can be replaced
s
s
by an approach that attempts to correct low brinogen, low platelet
m
There are several patient-speci c and procedure-speci c risk actors
e
that are known to increase hematologic complications with surgery count and low coagulation actor levels. These patients are o ten
n
undernourished or exposed to antibiotics. As these patients may be
t
(Table 54-4).
o
vitamin K de cient, vitamin K replacement therapy can be given to
P
■ PATIENT-SPECIFIC RISK FACTORS see i this helps to correct the coagulopathy.
a
t
Chronic kidney disease is associated with increased perioperative
i
e
Advanced age
n
morbidity rom impaired hemostasis secondary to acquired platelet
t
Hematologic complications are greatly in uenced by advanced age.
s
dys unction, leading to uremic bleeding. Both DDAVP and conju-
w
The incidence o venous thromboembolism increases sharply with gated estrogen have been used, either alone or concurrently, as
i
t
age. In individuals in the 25 to 30 years age group, the incidence is
h
prophylaxis prior to surgery or or treatment o acute bleeding. As
approximately 1 per 10,000 person-years. In contrast, the incidence
H
DDAVP has a limited clinical e ect due to tachyphylaxis, it is more
e
increases about 80- old with nearly 8 per 1000 person-year in the
m
commonly used prior to minor surgical procedures (eg, biopsies
85 years and older age group. Advanced age is also a risk actor
a
and endoscopies) whereas conjugated estrogen, which has a longer
t
o
or bleeding. The IMPROVE bleeding risk model which provides an duration o action o up to 10 days, can be used or elective surgery.
l
o
estimate o in-hospital bleeding rom the time o admission up to Depending on the severity o the platelet dys unction, platelet
g
i
14 days ollowing admission or an acute medical illness demon-
c
trans usion may be indicated. Patients with chronic kidney disease
D
strated that the probability o major in-hospital bleeding or men are also requently anemic rom reduced erythropoietin production.
i
s
was 0.1% in the <40 years, 0.2% in the 40 to 84 years, and 0.5% in
o
Besides tissue hypoxia, anemia can also cause platelet dys unction
r
the ≥85 years. The probability o a clinically important in-hospital
d
by several possible mechanisms. Thus, a decrease in hematocrit may
e
bleeding was 0.5%, 1.1%, and 2.2%, respectively.
r
urther enhance risk or bleeding and should be corrected prior to
s
The prevalence o anemia, a known risk actor or postoperative surgery. Correction o anemia by use o erythropoietin or trans usion
mortality, also increases with advancing age, especially a ter the may improve platelet unction and is an additional strategy to aide
th decade o li e and exceeds 20% in those 85 years and older. hemostasis in uremic patients.
Anemia in the elderly may be due to decreased red cell production Given the rising incidence o cardiovascular disease, majority o
rom age-related decline in normal bone marrow unction, nutri- the population are on aspirin and other antiplatelet agents, includ-
tional de ciencies (iron, B12 or olate) rom poor dietary intake or ing GPIIb/IIIa antagonists, which cause de ects in platelet aggrega-
decreased erythropoietin production rom chronic kidney disease. tion. The concomitant use o antiplatelet and anticoagulant agents
In ammatory response related to chronic disease conditions can increase the risk o bleeding perioperatively. Prior to surgery, these
also causes anemia o chronic in ammation, a disorder mediated by drugs should be discontinued or avoided in the perioperative
an increase in hepcidin production. Preoperative anemia can result period, i sa e to do so rom a cardiovascular perspective. I needed,
in increased morbidity and mortality, particularly in elderly patients, platelet trans usion may be indicated prior to surgery or in the case
as the surgery itsel stresses the cardiovascular system, resulting o excessive bleeding.
in tissue hypoxia rom reduced cardiac output and underlying
atherosclerosis. This risk can be minimized through the correction
■ PROCEDURE-SPECIFIC RISK FACTORS
o anemia using iron to correct the anemia or through exogenous
erythropoietin injection. However, many elderly patients may have Type of surgery
a primary marrow de ect and not be responsive to such therapy. The postoperative risk o bleeding, which is the highest contributor
Underlying MDS or other changes that in uence marrow response to morbidity and mortality in patients with malignant or nonma-
may a ect the elderly and the threshold or providing trans usion lignant hematologic disease, is greatly in uenced by the type o
support is dependent on the magnitude and extend o other organ surgery. A bleeding risk strati cation based on the type o surgical
dys unction such as pulmonary, cardiac, vascular, or neurologic or invasive procedures helps predict postoperative risk o bleeding
problems. and guide preoperative management (Table 54-5).
Other comorbidities Choice of anesthesia

The presence o other comorbidities such as liver and kidney disease The choice o anesthesia can in uence perioperative morbidity and
also increases the risk o perioperative hematologic complications. mortality. General anesthesia may reduce myocardial contractibility
357
hemolysis in patients with G6PD de ciency. An awareness o this
TABLE 54-5 Risk for Bleeding with Surgical or Invasive diagnosis should be communicated to the surgical and anesthetic
Procedures team and placed in the chart to avoid the use o drugs known to
P
Risk Type of Procedure Examples precipitate hemolysis.
A
R
Low Nonvital organs Lymph node biopsy,
T
involved, exposed dental extraction, cataract CONCLUSION
surgical site, limited extraction, most cutaneous
I
I
dissection surgery, laparoscopic As the risk o perioperative complications can vary considerably
procedures, coronary depending on the underlying hematologic disorder, it is important
angiography to identi y the disorder and its severity. A care ul assessment o the
Moderate Vital organs Laparotomy, thoracotomy, bene ts o surgery against the risk o perioperative complications is
M
involved, deep, or mastectomy, major essential. With appropriate perioperative evaluation and manage-
e
d
extensive dissection orthopedic surgery, ment strategies, these patients should be able to undergo the major-
i
c
pacemaker insertion ity o surgeries with risk reduced by managing known risk actors.
a
l
High Bleeding likely Neurosurgery, ophthalmic
C
o
to compromise surgery, cardiopulmonary
n
SUGGESTED READINGS
s
surgical result, bypass, prostatectomy or
u
bleeding bladder surgery, major
l
t
a
complications vascular surgery, renal Feely MA, Collins CS, Daniels PR, Kebede EB, Jatoi A, Mauck KF.
t
i
o
requent biopsy, bowel polypectomy Preoperative testing be ore noncardiac surgery: guidelines and
n
recommendations. Am Fam Physician. 2013;87(6):414-418.
Reprinted with permission rom Reding MT, Key NS. Hematologic prob-
Fellin FM. Perioperative evaluation o patients with hematologic
lems in the surgical patient: bleeding and thrombosis. In: Ho man R, Benz
disorders. In: Merli F, Weitz H, eds. Medical Management o the
RJ, Shattil S, et al, eds. Hematology: Basic Principles and Practice, 6th ed.
Philadelphia: Churchill Livingstone; 2013. Surgical Patient. Saunders Elsevier, Philadelphia 2008.
Kitchens CS, Lawson JW. Surgery and hemostasis. In: Kitchens CS,
and cardiac output. This urther decreases tissue oxygen-delivery Kessler CM, Konkle BA, eds. Consultative Hemostasis and Thrombosis,
in elderly patients with borderline anemia, and may potentially 3rd ed. Philadelphia: Elsevier Saunders; 2013.
lead to acute cardiac decompensation, resulting in myocardial Reding MT, Key NS. Hematologic problems in the surgical patient:
ischemia, in arction and/or stroke. Certain inhalational general anes- bleeding and thrombosis. In: Ho man R, Benz RJ, Shattil S, et al,
thetic agents can also inhibit the enzymatic activity o glucose- eds. Hematology: Basic Principles and Practice, 6th ed. Philadelphia:
6-phosphate dehydrogenase (G6PD) activity, which leads to acute Churchill Livingstone; 2013.
358
SECTION 4
Prevention, Assessment, and
Management of Common Complications
in Noncardiac Surgery
359
55
CHAP TER EPIDEMIOLOGY AND RISK FACTORS FOR
SURGICAL SITE INFECTIONS
It is estimated that over 40 million surgical procedures are per ormed
every year in the United States. Surgical site in ections (Table 55-1)
complicate approximately 2% to 5% o these procedures, represent-
ing 38% o nosocomial in ections occurring in surgical patients. Risk
actors or surgical site in ections (SSIs) can be classi ed as either
patient-related actors or surgical actors, and can be strati ed into
modi able, potentially modi able, and nonmodi able risk actors.
Antimicrobial Modi able risk actors include elective operations in the presence
o associated in ections, prolonged preoperative hospital stays,
Prophylaxis in seromas, dead space, oreign bodies, and routine drain use, among
others, and can be improved with the use o good surgical practice
Surgery
and speci c preventive strategies. Nonmodi able risk actors are
most commonly patient-related and have an important e ect on
the incidence o SSI or each individual patient. The wound class
(Table 55-2) is a relatively good predictor o SSI and has tradition-
ally been used to estimate the risk o SSI and as a benchmark or
Louisa W. Chiu, MD comparisons between institutions. However, with the better recent
E. Patchen Dellinger, MD understanding o SSI and its multi actorial risk actors, more sophis-
ticated predictive scores, such as the National Nosocomial In ection
Daniel A. Anaya, MD Surveillance (NNIS) score, have been developed to better estimate
the risk o SSI or each individual patient, a ter considering the inter-
action between di erent risk actors (Table 55-3). Speci c preven-
tive measures have been identi ed and are used to decrease the risk
o SSI. These include minimizing the presence o microorganisms
with prophylactic antibiotics and optimizing the patient’s ability to
ght those still present at the surgical site during the perioperative
period.
PRACTICE POINT
• Patient risk actors or surgical site in ections may include
obesity, diabetes mellitus, older age, malnutrition, prolonged
hospital stay prior to surgery, active in ection at another
site, cancer, immunosuppression, and tissue ischemia due
to irradiation or vascular disease. Although some o these
risk actors may not be reversible, hospitalists taking care o
preoperative patients should optimize glycemic control and
nutritional status, and encourage smoking cessation.
IMPACT OF SURGICAL SITE INFECTION

SSIs are associated with several adverse outcomes. Patients with
SSI are more likely to develop additional complications, including
wound dehiscence, hernias, and necrotizing so t tissue in ections.
Multiple large, single-center, multicenter, and population-level
analyses have revealed at least a two old increased risk o postop-
erative mortality in patients with SSI. Additionally, SSI is associated
with longer hospital stays (10-12 excess days), a higher risk o inten-
sive care unit (ICU) admission, and a ve old higher risk o hospital
readmission. Similarly, the treatment o patients with SSI results in
excess costs o over $5000 per patient, representing a US national
cost between $130 and $845 million per year. Given the availability
o multiple preventive measures, SSI is used as a health care quality
indicator.
361
TABLE 55-1 Criteria for SSI
Classification Definition
P
A
Super icial incisional • Within 30 d postoperatively
R
• Involves skin or subcutaneous tissue o the incision and at least one o the ollowing:
T
Purulent drainage rom the super icial incision
Organism isolated rom an aseptically obtained culture o luid or tissue rom the super icial incision
I
I
At least one o the ollowing signs or symptoms o in ection: pain or tenderness, localized swelling, redness,
or heat and incision is deliberately opened by surgeon and is culture-positive or not cultured
Diagnosis by surgeon or attending physician
Deep incisional • Within 30 d i no implant in place; within 1 y i implant
M
• Involves deep so t tissues and at least 1 o the ollowing:
e
d
Purulent drainage rom the deep incision but not rom the organ/space component
i
c
Deep incision spontaneously dehisces or is deliberately opened and is culture-positive or not cultured and
a
l
has at least one o the ollowing: ever or localized pain or tenderness
C
o
An abscess or other evidence o in ection ound on direct examination, during reoperation, or by
n
histopathologic or radiologic examination
s
u
l
t
a
Organ/space • Within 30 d i no implant in place; within 1 y i implant
t
i
o
• Involves any part o the body, excluding the skin incision, ascia, or muscle layers, that is opened or
n
manipulated during the operative procedure
• Has at least one o the ollowing:
Purulent drainage rom a drain that is placed through a stab wound into the organ/space
Organisms isolated rom an aseptically obtained culture o luid or tissue
An abscess or other evidence o in ection ound on direct examination, during reoperation, or by
histopathologic or radiologic examination
RATIONALE FOR ANTIMICROBIAL PROPHYLAXIS INDICATIONS FOR ANTIBIOTIC PROPHYLAXIS

Antimicrobial prophylaxis is used when in ection is not present but John Burke rst demonstrated the value o prophylactic antimicrobi-
the risk o postoperative SSI is present. The goal o prophylaxis is not als in the 1950s with animal studies. Since then, many retrospective
necessarily to sterilize the surgical site, but to reduce microbial colo- studies, prospective randomized trials, systematic reviews, and
nization o the wound to levels that the patient’s immune system is meta-analyses have con rmed the e cacy o prophylactic antimi-
able to handle. The absolute bene t o antibiotic prophylaxis and crobials in decreasing the risk o SSI and established the basic prin-
the number needed to treat to prevent an SSI varies by the baseline ciples. The magnitude o this protective e ect relates directly to the
risk o in ection or each procedure. The decision to use prophylaxis magnitude o the risk o SSI. The majority o studies have ocused
should be based on an assessment o the risk o in ection and the on evaluating the e cacy o prophylactic antibiotics when used or
cost and morbidity o the in ections that might occur or that proce- speci c operations. This has resulted in high-level data supporting
dure, balanced against the cost o prophylaxis and its potential or their use or clean-contaminated and contaminated operations, and
adverse e ects, including allergies, superin ections, and the genera- more controversial data or their use in operations classi ed as clean.
tion o bacterial resistance by overuse. However, the risk o SSI can vary signi cantly between patients
TABLE 55-2 Surgical Wound Classification Based on Degree of Potential Contamination According to the Centers for Disease
Control and Prevention (CDC) Guidelines
Classification Description
Clean • An unin ected operative wound in which no in lammation is encountered and the respiratory, alimentary,
genital, or unin ected urinary tract is not entered.
• Operative incisional wounds that ollow nonpenetrating (blunt) trauma should be included in this category i
they meet criteria.
Clean contaminated • An operative wound in which the respiratory, alimentary, genital, or urinary tracts are entered under controlled
conditions and without unusual contamination.
• Speci ically, operations involving the biliary tract, appendix, vagina, and oropharynx are included in this
category, provided no evidence o in ection or major break in technique is encountered.
Contaminated • Open, resh, accidental wounds
• Operations with major breaks in sterile technique (eg, open cardiac massage) or gross spillage rom the
gastrointestinal tract
• Incisions in which acute, nonpurulent in lammation is encountered are included in this category
Dirty • Old traumatic wounds with retained devitalized tissue
• Involve existing clinical in ection or per orated viscera
• This de inition suggests that the organisms causing postoperative in ection were present in the operative ield
be ore the operation
362
Clostridium dif cile colitis, change in resistance patterns), and (4) low
TABLE 55-3 National Nosocomial Infection Surveillance cost. The majority o guidelines helping to choose the appropriate
C
(NNIS) System Classification for Determining agent are based on randomized trials in which a speci c antibiotic
H
the Risk of Surgical Site Infection
A
regimen has been tested or a speci c procedure. We support the
P
Risk Factors Points use o this evidence-based decision making. However, the vast
T
majority o procedures are prone to SSI with similar species o bac-
Procedure duration ≥ 75th 1
E
teria, and similar types o antibiotics have been tested. I clinical trial
R
percentile o duration or that
speci ic operation data are not available or a speci c procedure, it is reasonable to
5
generalize rom trial data on procedures with comparable bacterial
5
Contaminated or dirty wound 1
f ora and risk.
American Society o 1 The rst consideration is to determine the class o wound and the
A
Anesthesiology score III-V
most common bacteria causing SSI in each case. For clean wounds,
n
t
Final Score Risk of Surgical Site Infection gram-positive skin f ora, including Staphylococcus aureus and coag-
i
m
0 1.5% ulase-negative staphylococci, are the most common pathogens
i
c
r
1 2.9% causing SSI. For these operations, ce azolin is the most commonly
o
b
2 6.8% used antibiotic, ollowing the our main principles already outlined.
i
a
Alternatives include other rst- or second-generation cephalospo-
l
3 13.0%
P
rins, oxacillin, and clindamycin. For patients with documented and
r
o
clinically relevant beta-lactam allergies, clindamycin or vancomycin
p
h
is an adequate alternative. Most clean-contaminated and contami-
y
l
a
undergoing similar surgical procedures. Various studies have shown nated operations are expected to result in exposure to gram nega-
x
i
tive and anaerobic bacteria. Appropriate regimens in this setting
s
that using a more comprehensive predictive tool (such as the NNIS
i
include ertapenem, ce otetan, and ce oxitin as single agents, or
n
score), the risk o SSI can in act be higher or some patients with
S
less contaminated wounds, depending on the presence o other multiple-agent regimens such as an aminoglycoside or a quinolone
u
r
plus clindamycin or metronidazole (Table 55-4). Antimicrobials
g
risk actors. This has been supported by randomized trials and
e
meta-analyses demonstrating a clinical bene t o using prophylac- agents with the narrowest spectrum o activity required or e cacy
r
y
tic antibiotics or breast, cardiac, orthopedic, and vascular surgery, in preventing in ection are recommended.
as well as other clean operations. Antimicrobial prophylaxis may be
reasonable or some clean operations, especially when the potential
PRACTICE POINT
consequences o SSI may be dire, as a ter prosthetic joint replace-
ment or a coronary artery bypass gra ting (CABG). Selection of antimicrobial agent
Based on existing data and the oregoing considerations, prophy-
1. Determine the class o wound and the most common bacteria
lactic antibiotics are indicated or the ollowing:
causing SSI in each case.
• Clean-contaminated and contaminated operations 2. Consider local resistance data rom your hospital’s surveillance
• Clean operations with a high risk o SSI systems.
NNIS score ≥ 1 +/– other associated important risk actors • A speci c agent may be recommended or an operation
Immunocompromised host based on national guidelines. However, i microbiologic
Other clean operations with known increased risk o SSI, data derived rom local surveillance programs reveal that
such as groin incisions and mastectomy expected f ora have a high resistance rate to that agent,
• Clean operations or which SSI has signi cant clinical another antibiotic must be used.
consequences
• Trending patterns o resistance to the di erent agents used
Examples: craniotomy, cardiac and vascular operations, and can help identi y emergence o resistance early and can
operations involving placement o prosthetic material, such guide local protocols.
as joint replacements and mesh hernia repairs
PRACTICE POINT Recent concern in the media regarding methicillin-resistant

Staphylococcus aureus (MRSA) in ections has led to calls to use
• Antimicrobial prophylaxis is indicated or contaminated vancomycin routinely as a prophylactic agent or clean operations.
procedures, such as laparotomy or bowel per oration, and Per 2013 ASHP/IDSA guidelines, routine use o vancomycin prophy-
clean-contaminated procedures, such as elective intestinal laxis is not recommended or any procedure. However, it may be
resection, lung resection, or vaginal hysterectomy. It is included in the regimen o choice when cluster cases are detected
indicated or clean procedures only in procedures with a at an institution, or patients with known MRSA colonization or at
signi cant inherent in ection risk, such as mastectomy or high risk or colonization. One recent study randomized cardiac
operations with groin incisions, or when the consequences o patients to vancomycin versus ce azolin in a hospital with a high
SSI are severe, as in prosthetic joint replacement, craniotomy, rate o MRSA, and ound more MRSA in ections in the ce azolin
and cardiac surgery. arm but more methicillin-sensitive Staphylococcus aureus (MSSA)
in ections in the vancomycin arm, with no di erence in overall
rates between the two groups. Thus, the use o vancomycin alone
may decrease the risk o MRSA SSI, but does not convey protection
SELECTION OF ANTIMICROBIAL AGENT against other common gram positive bacteria (including MSSA);
The characteristics o the ideal agent or e ective antibiotic prophy- its routine use may simply result in a shi t rom MRSA to MSSA SSI,
laxis include (1) bactericidal e ect on microorganisms expected to without a signi cant impact on the overall reduction o SSI. These
be present at the surgical site, (2) adequate biodistribution in the observations have led to use o two-drug regimens, with vancomy-
surgical site, (3) low risk o potential side e ects (allergic reactions, cin combined with ce azolin to cover both MRSA and MSSA.
363
TABLE 55-4 Recommended Antibiotic Prophylaxis Regimens for Different Types of Surgical Procedures Based on Published
Guidelines and Most Current Available Data
P
Type of Surgical Procedure Recommended Antibiotic Regimen Alternative Regimens for b-Lactam Allergies
A
R
Cardiac procedures Ce azolin Clindamycin
T
Ce uroxime Vancomycin
I
Thoracic procedures Ce azolin Clindamycin
I
Ampicillin-sulbactam Vancomycin
Vascular operations Ce azolin Clindamycin
Vancomycin
M
e
Gastrointestinal clean procedures and those Ce azolin Clindamycin
d
involving entry into the upper gastrointestinal
i
Vancomycin + cipro loxacin or
c
a
tract aztreonam or aminoglycoside
l
C
Small intestine obstruction Ce azolin + metronidazole Metronidazole + aminoglycoside or
o
n
Ce oxitin luoroquinolone
s
u
l
Ce otetan
t
a
t
Biliary tract operations Ce azolin Clindamycin
i
o
n
Ampicillin/sulbactam Vancomycin + cipro loxacin or
Ce otetan aztreonam or aminoglycoside
Ce oxitin Metronidazole + aminoglycoside or
cipro loxacin
Ce triaxone
Appendectomy Ce oxitin Cipro loxacin or an aminoglycoside +
Ce otetan metronidazole or clindamycin
Ampicillin/sulbactam
Ce azolin + metronidazole
Piperacillin/tazobactam
Ertapenem
Cipro loxacin or an aminoglycoside +
metronidazole or clindamycin
Colorectal Ertapenem Cipro loxacin or an aminoglycoside +
Ce oxitin metronidazole or clindamycin
Ce otetan
Ce azolin + metronidazole
Cipro loxacin or an aminoglycoside +
metronidazole or clindamycin
Ampicillin-sulbactam
In conjunction with mechanical bowel prep Oral neomycin + erythromycin or
oral metronidazole
Hip or knee arthroplasty Ce azolin Clindamycin
Vancomycin
Hysterectomy Ce azolin Cipro loxacin or an aminoglycoside +
Ce otetan metronidazole or clindamycin
Ce oxitin
Ce uroxime
Head and neck operations Ce azolin ± metronidazole Clindamycin
Ce uroxime + metronidazole
Urologic operations Ce azolin Cipro loxacin
Neurosurgery Ce azolin Vancomycin
Clindamycin
Liver transplantation Piperacillin-tazobactam Clindamycin or vancomycin +
Ce otaxime + ampicillin aminoglycoside or aztreonam or
luoroquinolone
Plastic surgery Ce azolin Clindamycin
Ampicillin-sulbactam Vancomycin
Indications may vary based on the speci ic operation and standard principles o prophylaxis.
364
Another approach is to screen or MRSA preoperatively, as MRSA vancomycin or f uoroquinolones, which both require slow in usion
colonization may increase the risk o SSI by 2- to 14- old. This allows and have longish hal -lives. In usion o these agents may begin
C
or targeted use o vancomycin (and double coverage when indi- 2 hours prior to the incision time, to avoid undesired antibiotic-
H
cated) or patients at risk. Some data also exist to support the use o
A
related side e ects. In addition, when the operation involves place-
P
mupirocin nasal ointment and chlorhexidine baths be ore surgery ment o a proximal tourniquet, the antibiotic in usion must be
T
to eradicate staphylococcal colonization. Further studies are needed completed be ore the tourniquet is inf ated, to allow or adequate
E
to de ne the role o routine or selective vancomycin prophylaxis, tissue levels at the surgical site. I a patient was previously receiving
R
and measures to eradicate the MRSA carrier state. therapeutic antimicrobials or an in ection be ore surgery, an extra
5
Although still controversial, recent studies have shown that or dose should be given in the 60 minutes prior to incision.
5
colorectal operations, mechanical bowel preparation (MBP) accom-
panied by oral prophylactic antibiotics decreased the incidence
A
o SSI. Oral regimens typically include a combination o neomycin ANTIBIOTIC DOSING
n
t
sul ate plus erythromycin or metronidazole, and should be given The dose o antibiotic should achieve adequate tissue levels
i
m
in addition to standard preoperative IV prophylaxis. The oral throughout the whole operation. The most accurate way to assure
i
c
r
antimicrobials should be given as three doses over approximately this is to use doses based on body weight. Most weight-based rec-
o
b
10 hours, starting the a ternoon and evening be ore the operation ommendations are derived rom the pediatric literature, making this
i
a
and a ter the MBP. approach somewhat cumbersome, and hence standard doses that
l
P
t within recommended parameters are commonly used or the
r
o
majority o adult patients. However, weight-based recommenda-
p
h
ANTIBIOTIC TIMING tions are particularly help ul when using prophylactic antibiotics in
y
l
a
patients with extreme body weights. Speci cally, the standard dose
x
i
may need to be decreased to avoid antibiotic toxicity in extremely
s
i
PRACTICE POINT thin or pediatric patients, particularly when using antibiotics with a
n
S
narrow therapeutic window. More requently, however, appropriate
u
Timing, dosage, and duration of prophylactic antibiotics
r
doses must be selected to achieve adequate tissue levels in obese
g
e
1. Administer antibiotics within 1 hour prior to incision time. patients. Forse and colleagues demonstrated that in obese patients
r
y
2. The chosen dose should accomplish adequate tissue levels undergoing bariatric surgery, a dose o ce azolin twice the standard
throughout the whole operation. dose (2 g vs 1 g) was necessary to achieve adequate serum and tis-
• Higher than standard doses are recommended or obese sue levels, and was associated with a lower rate o surgical site in ec-
patients. tion. Based on this study and others, higher than standard doses are
• Current guidelines recommend redosing o antibiotics or recommended or obese patients. Given the low cost and avorable
procedures lasting two or more hal -lives o the speci c sa ety pro le o ce azolin, it is reasonable to increase the dose to 2 g
agent used. or patients weighing more than 80 kg, and to 3 g or those weigh-
3. No studies have shown the superior e cacy o longer courses ing over 120 kg.
o prophylactic antibiotics in the postoperative period. In prolonged operations, it may be necessary to repeat the anti-
• Speci cally, or patients undergoing colorectal, other biotic dose, given the need or adequate antibiotic tissue levels
abdominal and gastrointestinal procedures, vascular, throughout the whole operation. Randomized trials, large single-
cardiac, gyn0ecologic, urologic, orthopedic, and head and and multiple-institution analysis, and the NNIS score have repeat-
neck operations, there is evidence to support the use o edly identi ed duration o operation as an important independent
short (<12-24 hours) regimens over longer ones, and in risk actor or SSI. The speci c mechanisms behind this have not
the majority o cases, signi cant evidence to recommend a been completely clari ed. However, subtherapeutic antibiotic tissue
single-dose strategy. levels at the end o long operations can at least partially explain this
association. Studies evaluating redosing o antibiotics or cardiac
• Prolonged courses o prophylactic antibiotics are not and gastrointestinal procedures longer than 3 to 4 hours have
associated with clinical bene t and may be attended with
shown a lower incidence o SSI when ollowing this practice. Given
signi cant harms, such as an increased risk o C. dif cile colitis
variation in the pharmacokinetic properties o di erent antibiotics,
and the acquisition o antibiotic-resistant pathogens.
current guidelines recommend redosing o antibiotics or proce-
dures lasting two or more hal -lives o the speci c agent used, or i
there is excessive blood loss (>1500 mL). Adjustments to this rule
The importance o antibiotic administration be ore incision was need to be considered or patients with renal insu ciency, in which
the rst key act determined about prophylaxis. John Burke rst the hal -li e o most antibiotics is prolonged (Table 55-5).
described the direct relation between timing o prophylaxis (in rela-
tion to incision time) and e ectiveness, and identi ed the decisive
period to be as close to the incision time as possible. More recently, DURATION OF ANTIBIOTIC PROPHYLAXIS
other investigators have corroborated these ndings, showing a Although adequate intraoperative antibiotic tissue levels are neces-
linear increase in the rate o SSI as the time o prophylaxis admin- sary to decrease the risk o SSI, the same is not true or the postoper-
istration gets urther rom the incision time. All studies show a ative period. Despite an extensive body o evidence supporting the
dramatic rise in SSI when antibiotics are rst given a ter incision. use o a single preoperative dose (and intraoperative redosing when
This probably results rom both the serum and the tissue antibiotic indicated) over a multiple-dose strategy (ie, one preoperative dose
levels at the time o operation, with the best and most persistent with subsequent postoperative doses), this is perhaps one o the
levels achieved when the antibiotic is given close to the incision most di cult principles to translate into real-li e practice. A recent
time. Based on these principles, current guidelines recommend analysis demonstrated that in current practice, antibiotics are con-
administering prophylactic antibiotics within 1 hour prior to inci- tinued or over 24 hours a ter the operation in over 50% o patients.
sion time. This is a more speci c time rame than the previously Multiple studies in di erent populations o patients have evaluated
recommended time “at induction o anesthesia.” The exceptions are the role o single- versus multiple-dose prophylaxis. Speci cally, or
365
TABLE 55-5 Characteristics of Selected Antibiotics More Commonly Used for Surgical Site Infection Prophylaxis
Recommended Standard Recommended Redose

P
Antibiotic Weight-Based Dose Recommended Dose Half-Life (in Hours)
*
Interval (in Hours)
A
R
Ampicillin-sulbactam 50 mg/kg 3g 0.8-1.3 2
T
Ce azolin 20-30 mg/kg 1-2 g 1.2-2.5 2-5
I
I
3 g i > 120 kg
Ce oxitin 1-2 g 20-40 mg/kg 0.5-1.1 2-3
Ce otetan 1-2 g 20-40 mg 2.8-4.6 3-6
Ce uroxime 50 mg/kg 1.5 g 1-2 3-4
M
e
Ce otaxime 50 mg/kg 1g 0.9-1.7 3
d
i
c
Ce epime - 1-2 g 2.2 4
a
l
Ce triaxone 50-75 mg/kg 2g 5.4-10.9 NA
C
o
Cipro loxacin 400 mg 400 mg 3.5-5 4-10
n
s
Clindamycin 3-6 mg/kg 600-900 mg 2-5.1 3-6
u
l
t
Ertapenem - 1g 4 8
a
t
i
Fluconazole 6 mg/kg 400 mg 30 NA
o
n
Gentamicin 3-5 mg/kg 300 mg 2-3 6
Levo loxacin 10 mg/kg 500 mg 6-8 NA
Metronidazole 15 mg/kg 500-1000 mg 6-14 12
Piperacillin/tazobactam 3.375 g 0.7-1.2 2-4
Vancomycin 10-15 mg/kg 1g 4-6 6-12
Oral antibiotics or colorectal
surgery in conjunction with
mechanical bowel prep
Erythromycin base 20 mg/kg 1g 0.8-3 NA
Metronidazole 15 mg/kg 1g 6-10 NA
Neomycin 15 mg/kg 1g 2-3 NA
Reported values applicable to adults.

*
Hal -li e calculated or normal renal unction.
patients undergoing colorectal, other abdominal and gastrointesti- IMPLEMENTATION PRACTICES/QUALITY IMPROVEMENT
nal, vascular, cardiac, gynecologic, urologic, orthopedic, and head Given the relative high incidence o SSI, its clinical impact, and the
and neck operations, evidence supports the use o shorter (<12- signi cant burden to the health care system, preventive strategies
24 hours) regimens over longer ones, and in the majority o cases, have been receiving more attention by physicians, health care
signi cant evidence to recommend a single-dose strategy. There institutions, and policymakers. One strategy to ocus e orts on the
are no studies demonstrating superior e cacy o longer courses o e ective delivery o known preventive strategies is to classi y all SSIs
prophylactic antibiotics. McDonald and colleagues reported results as potentially preventable (not all prevention measures used or used
o a systematic review based on 28 randomized trials totaling over incorrectly) or not apparently preventable (all known preventive mea-
9000 patients and reported no di erence in the overall rate o SSI sures used appropriately). Several local and national guidelines, spe-
with either approach. Based on these ndings, the authors recom- ci c implementation strategies, and collaborative group approaches
mended the use o a single-dose practice and emphasized the have been recently developed with the objective o improving
equivalent bene t and decreased cost o this strategy, with the addi- current practice targeted to minimize the number o potentially
tional bene t o decreasing the emergence o in ections caused by preventable SSIs. In 2002, the Centers or Medicare and Medicaid
resistant pathogens. There are no data to support the continuation Services (CMS) and the Centers or Disease Control and Prevention
o antimicrobial prophylaxis until all indwelling drains and intravas- (CDC) implemented the National Surgical In ection Prevention (SIP)
cular catheters are removed. project, with the aim o decreasing undesired outcomes and costs
derived rom potentially preventable SSIs. The appropriate use o
prophylactic antibiotics was identi ed as a crucial measure in pre-
SPECIFIC PROCEDURES venting SSI, and three practice per ormance measures were devel-
Di erent single- and multiple-antibiotic regimens have been evalu- oped: (1) antibiotic dose given within 60 minutes o incision time,
ated or various procedures. Based on these data, the expected (2) appropriate selection o antibiotic, based on national guidelines,
microorganisms at the surgical site o each speci c procedure, and (3) discontinuation o antibiotics within 24 hours o the end o
and the similar pharmacologic characteristics o these regimens, the procedure.
di erent alternatives are recommended. Table 55-4 lists some o An examination o over 11,000 operations per ormed on Medi-
these alternatives by surgical type. While this is a help ul guideline care patients around the United States revealed that only 56% had
or choosing antibiotics, the decision o which antibiotic to use antibiotic prophylaxis given within 1 hour o incision time, and only
is dynamic, and should also incorporate local and system-related 40% had the antibiotic discontinued within 24 hours a ter the opera-
aspects such as cost and antibiotic resistance data. tion ended. Multiple e orts have ocused on improving practice
366
patterns or the use o prophylactic antibiotics. Overall, practice o SSI such as in clean-contaminated and contaminated operations,
patterns have improved over time, with more recent assessments as well as or selected clean operations, based on the presence o
C
revealing better—although still inadequate—compliance with the other risk actors and the impact o SSI or each speci c patient
H
three SIP per ormance measures. Various interventions have been population. With ew exceptions, antibiotics must be given within
A
P
used to help improve substandard practice and generally ocus 1 hour prior to incision, using generous (but sa e) doses, redosed or
T
on education/ eedback interventions and detailed changes in the longer operations (usually at two to three hal -lives), and should be
E
process o care. A recent study led by the CMS and the SIP collab- discontinued within 24 hours o the end o the operation, i not at
R
orative group, involving 56 di erent hospitals, demonstrated that the end o the operation. When compliance with these principles
5
an intervention based on educational/ eedback sessions, models improves, the incidence o SSI decreases.
5
o improvement, and methods o implementing and measuring
changes resulted in signi cant improvement in the timing and SUGGESTED READINGS
A
duration o antibiotics as well as improved per ormance with other
n
t
nonpharmacologic preventive measures, and led to a 27% drop in
i
Bowater RJ, Stirling SA, Lil ord R. Is antibiotic prophylaxis in surgery
m
the rate o SSI when comparing the be ore and a ter study periods.
i
a generally e ective intervention? Testing a generic hypothesis
c
r
Other studies implementing educational and process-o -care inter-
o
over a set o meta-analyses. Ann Surg. 2009;249:551-556.
b
ventions have shown similar ndings, emphasizing the importance
i
Bratzler DW, Houck PM, Richards C, et al. Use o antimicrobial pro-
a
o in ormation dissemination and culture change. Some investiga-
l
phylaxis or major surgery: baseline results rom the National
P
tors have ocused on more speci c interventions, such as shi ting
r
o
Surgical In ection Prevention Project. Arch Surg. 2005;140:174-182.
the responsibility o providing the antibiotic rom the preoperative
p
h
nurse to the anesthesiologist, developing checklists and standard- Bull AL, Worth LJ, Richards MJ. Impact o vancomycin surgical anti-
y
l
biotic prophylaxis on the development o methicillin-sensitive
a
ized printed preoperative prophylaxis orders, and using automated
x
staphylococcus aureus surgical site in ections: report rom Australian
i
systems to improve practice o this strategy. All such interventions
s
Surveillance Data (VICNISS). Ann Surg. 2012;256:1089-1092.
i
have resulted in equivalent success.
n
S
Gagliardi and colleagues recently published a systematic review Bratzler DW, Dellinger EP, Olsen KM, et al. Clinical practice guidelines
u
r
assessing the actors likely to inf uence appropriate antibiotic pro- or antimicrobial prophylaxis in surgery: rom the American Society
g
e
phylaxis. Their results support the implementation o interventions o Health-System Pharmacists, the In ectious Diseases Society o
r
y
such as development o multidisciplinary pathways ocused on America, the Surgical In ection Society, and the Society o Health
inf uencing individual physicians’ knowledge, attitudes, and belie s, Care Epidemiology o America. Surg In ect. 2013;14:73-156.
enhancing team communication, and promoting the use o written Forse RA, Karam B, MacLean LD, Christou NV. Antibiotic prophylaxis
and computerized order sets. However, they also emphasize the or surgery in morbidly obese patients. Surgery. 1989;106:750-756.
need or urther research to better characterize the bene t o these
Gagliardi A, Fenech D, Eskicioglu C, Nathens AB, McLeod R. Factors
strategies. These uture e orts should ideally be carried on within a
inf uencing antibiotic prophylaxis or surgical site in ection pre-
ramework o regional or national collaborative groups, which have
vention in general surgery: a review o the literature. Can J Surg.
the advantage o overcoming di culties in translating knowledge
2009;52:481-489.
to practice at a single-institution level.
Kirkland KB, Briggs JP, Trivette SL, Wilkinson WE, Sexton DJ. The
CONCLUSION impact o surgical-site in ections in the 1990s: attributable mortal-
ity, excess length o hospitalization, and extra costs. In ect Control
Surgical site in ection is the most common complication ollow-
Hosp Epidemiol. 1999;20:725-730.
ing surgical procedures. It is associated with signi cant impact on
patient outcomes and health care resource utilization. The use o McDonald M, Grabsch E, Marshall C, Forbes A. Single- versus
prophylactic antibiotics is a proven strategy to reduce this risk. Cur- multiple-dose antimicrobial prophylaxis or major surgery: a sys-
rent guidelines or the appropriate use o antibiotic prophylaxis are tematic review. Aust N Z J Surg. 1998;68:388-396.
well supported by level I data. Although di erent antibiotics are rec- Steinberg JP, Braun BI, Hellinger WC, et al. Timing o antimicrobial
ommended or di erent operations, general principles or the use prophylaxis and the risk o surgical site in ections: results rom
o prophylaxis can help guide the use o this preventive measure. the trial to reduce antimicrobial prophylaxis errors. Ann Surg.
Prophylactic antibiotics are indicated when there is a signi cant risk 2009;250:10-16.
367
CHAP TER
56 WHAT IS THE RISK FOR VENOUS THROMBOEMBOLISM
IN PATIENTS REQUIRING NONORTHOPEDIC SURGERY?
■ EPIDEMIOLOGY
Each year, surgeons in the United States per orm more than 51 million
inpatient surgeries, the majority o which are nonorthopedic surger-
ies. Patients undergoing nonorthopedic surgeries are a heteroge-
neous group in terms o surgery type, comorbidities, and associated
risk or venous thromboembolism (VTE), which comprises deep vein
Venous
thrombosis (DVT), and pulmonary embolism (PE). There have been
a number o risk strati cation models proposed to guide VTE pro-
phylaxis in surgical patients; they are limited by complexity, and lack
Thromboembolism o rigorous prospective validation. However, in general, patients are
considered to be at very low risk or VTE i they undergo minor surgi-
(VTE) Prophylaxis cal procedures lasting <30 minutes, have no medical comorbidities,
and are immediately mobile ollowing surgery. Their estimated
baseline risk o VTE, i no prophylaxis is given, is estimated to be
for Nonorthopedic <0.5%. All other surgical patients are considered to be at moderate
or high risk or VTE and merit some consideration or prophylaxis.
Surgery VTE in the patient undergoing nonorthopedic surgery can cause

signi cant morbidity and mortality and is a common cause o read-
mission to the hospital.
Menaka Pai, MD, MSc, FRCPC ■ PATHOPHYSIOLOGY

Many actors contribute to VTE a ter nonorthopedic surgery
James D. Douketis, MD, FRCP(C), FACP, FCCP (Tab le 56-1).
Trauma and surgery both contribute to venous injury and activa-
tion o the coagulation system. Postoperatively, patients may have
persistently reduced mobility, which causes stasis o blood ow in
the deep venous system. Patients undergoing certain types o sur-
gery may also have independent risk actors or VTE, such as obesity
in the bariatric surgical patient.
As in the orthopedic surgery setting, most episodes o postopera-
tive DVT in nonorthopedic surgery are clinically silent. These unno-
ticed clots usually resolve spontaneously without administration o
antithrombotic therapy. However, 25% to 50% extend and cause
symptomatic DVT or PE.
WHICH PATIENTS UNDERGOING NONORTHOPEDIC

SURGERY NEED VTE PROPHYLAXIS?
■ DOES THIS PATIENT UNDERGOING GENERAL
SURGERY NEED VTE PROPHYLAXIS?
CASE 56-1
A 32-year-old mother o two comes to the emergency room
with abdominal pain and nausea. An ultrasound conf rms acute
appendicitis, and the general surgeon at your center eels this
patient should have an appendectomy within the next 6 hours.
He eels she is at low operative risk, since she has no past
medical history and is taking no medications, apart rom an oral
contraceptive pill. The general surgery resident phones you to ask
i the patient needs VTE prophylaxis.
Data rom studies done more than 20 years ago involving

patients who did not routinely receive VTE prophylaxis ound that
rates o asymptomatic DVT in patients having general surgical
procedures were between 15% and 30%, while atal PE occurred
368
TABLE 56-1 Factors that Increase Risk for Venous • General anesthesia poses a greater risk o VTE than spinal or
C
Thromboembolism in Surgical Patients epidural anesthesia and the duration o anesthesia irrespective
H
o the type o anesthesia in uences VTE risk, with >3.5 hours
A
Surgical Factors associated with the highest risk.
P
Antecedent trauma (as reason or surgery) •
T
Postoperative complications may urther increase the risk.
E
General anesthesia (compared with regional/local anesthesia)
R
Arthroplasty
5
Abdominal surgical approach (compared with vaginal approach) Patients at very low risk o VTE do not require any speci c VTE pro-
6
Open surgical approach (compared with laparoscopic approach) phylaxis, but should be mobilized postoperatively. All other patients
should be evaluated or VTE prophylaxis. Patients at a low risk or
Use o the lithotomy position intraoperatively
V
bleeding should receive pharmacologic prophylaxis with LDUH,
e
n
Extrinsic venous compression intraoperatively LMWH, or ondaparinux. There have been no trials that directly
o
u
Extended duration o surgery (>1 h) compare the two most popular dosing regimens o subcutaneous
s
LDUH, 5000 units every 8 hours and 5000 units every 12 hours. A
T
Postoperative in ection
h
meta-analysis o 51 randomized controlled trials compared LMWH
r
Central venous catheterization
o
and LDUH in general and abdominal surgery patients, and ound
m
Postoperative immobility (con ined to bed, needing assistance
b
to ambulate) that the risk o clinically evident VTE was ~30% lower in the LMWH
o
group. However, this di erence disappeared when the analysis
e
Nonsurgical Factors
m
was restricted to high-quality (blinded, placebo-controlled) trials.
b
Increasing age The meta-analysis also ailed to show any di erence in PE, death
o
l
Pregnancy and the puerperium
i
rom any cause, major bleeding, or wound hematoma between the
s
m
Comorbid medical illness (eg, congestive heart ailure, LMWH and LDUH groups. Our practice is to use the lowest recom-
(
obstructive lung disease, acute myocardial in arction, mended dose o heparin or prophylaxis, to minimize bleeding com-
V
T
in lammatory bowel disease) plications, and reduce the number o injections that a patient must
E
)
Recent ischemic stroke receive. Further, we pre er LMWH as it carries a lower risk o heparin
P
r
induced thrombocytopenia than LDUH. Parental anticoagulants
o
Cancer (active or occult)
p
can be given at various points in the patient’s hospitalization: 0 to
h
Sepsis
y
2 hour preop (i the patient does not receive an epidural and does
l
a
Previous VTE
x
not undergo a very high bleed risk procedure, such as liver resec-
i
s
Prior pelvic radiation tion); the evening o the day o surgery; or the morning a ter surgery.
o
Inherited or acquired thrombophilia Timing o administration is dependent on bleeding concerns, and
r
N
Obesity (BMI > 25 kg/m 2) should be a shared decision between all o the patient’s health care
o
providers.
n
Drugs (eg, chemotherapy, hormonal therapy, erythropoeisis
o
stimulating agents) The selective actor Xa inhibitor ondaparinux has also been
r
t
h
evaluated or major abdominal surgery. There does not appear to
o
be any signi cant di erence in PE, or non atal symptomatic VTE,
p
e
or death when ondaparinux is compared with LMWH. There was a
d
i
trend to increased non atal major bleeding in the ondaparinux, but
c
in 0.2% to 0.9% o patients. Current surgical practices, including
S
better perioperative care, rapid postoperative mobilization, and there was no demonstrable increase in atal bleeding or bleeding
u
r
requiring reoperation.
g
greater use o regional anesthesia have likely reduced these gures.
e
However, general surgery patients are still considered to be at Direct-acting oral anticoagulants (eg, rivaroxaban and apixaban,
r
y
moderate to high risk o VTE. Numerous randomized clinical trials actor Xa inhibitors, and dabigatran etexilate, a actor IIa inhibitor)
and meta-analyses have shown that VTE prophylaxis with low dose have shown ef cacy in VTE prophylaxis as well. These agents are
un ractionated heparin (LDUH) or low molecular weight heparin appealing to clinicians and patients, as they are ingested orally, in
(LMWH) reduces the risk o symptomatic VTE by 40% to 70% in a xed dosing schedule. They have not yet been approved or use
these patients. However, both o these agents—and others cur- a ter nonorthopedic surgery in the United States.
rently used or VTE prophylaxis—can increase the risk o postopera- Mechanical VTE prophylaxis is an option in patients with a high
tive bleeding. risk o bleeding. However, graduated compression stockings (GCS)
and intermittent pneumatic compression devices (IPC) are not
as e ective as pharmacologic prophylaxis, and do not appear to
PRACTICE POINT reduce the risk o proximal DVT or symptomatic PE. Though many
use them as an “add-on intervention” in general surgery patients
The risk of thrombosis
who are at particularly high risk o VTE, such as those with cancer,
• The type o surgery and the type o anaesthesia are the there is little evidence that they add to the protective e ect o
primary determinants o VTE risk in nonorthopedic surgical pharmacologic prophylaxis. I mechanical VTE prophylaxis is used,
patients. bleeding risk should be reassessed regularly, and pharmacologic
VTE prophylaxis should be started as soon as it is acceptably low.
The approximate DVT risk without prophylaxis is based on
The risk o VTE appears to be elevated or at least 12 weeks ollow-
objectively con rmed rates o DVT in asymptomatic patients
ing inpatient surgery. However, the evidence or extended VTE pro-
who did not receive prophylaxis.
phylaxis in general surgery is not as robust as in orthopedic surgery.
There is a range o DVT risk o approximately 10% to 40% A multicenter, randomized, blinded, placebo-controlled trial studied
in general surgical patients, depending on the speci c extended 21-day prophylaxis with bemiparin (a LMWH) in patients
procedure, complications, and traditional risk actors. who underwent abdominal or pelvic surgery or cancer. The primary
outcome was a composite o any DVT (including asymptomatic and
369
distal events), non atal PE, and death rom any cause. Though the There is limited evidence or VTE prophylaxis in vascular surgery
bemiparin group’s had a 24% lower risk o the composite outcome patients, with ew well-designed randomized clinical trials in this
and an 88% lower risk o proximal DVT than the placebo group, area. Studies have not shown a clear bene t or prophylaxis, with no
P
neither group had and symptomatic, non atal VTE events. The signi cant di erence in rates o DVT detected by routine ultrasound
A
results also showed no di erence in bleeding. At this time, it is not screening or rates o major bleeding. Why does VTE prophylaxis
R
recommended that all general surgery patients receive extended not appear to have a bene t in vascular surgery patients? One
T
VTE prophylaxis. However, cancer patients who are at particularly potential reason is that vascular surgery patients requently receive
I
I
high risk o VTE should be considered or post discharge prophylaxis antithrombotic agents such as intravenous heparin, and antiplatelet
or 4 weeks. agents such as aspirin and clopidogrel, to prevent arterial occlusion
a ter vascular reconstruction. These agents lower the risk o VTE,
making routine use o additional anticoagulants redundant (and
M
PRACTICE POINT possibly harm ul). However, the methodologic limitations o exist-
e
d
A risk assessment does not have to be complicated ing evidence are also a concern; vascular surgery patients do have
i
c
an increased risk o VTE postoperatively, and these patients were
a
•
l
The number o risk actors determines whether a patient is low, included in validation studies o surgical risk assessment models. At
C
moderate, or high risk, not just the surgical procedure itsel .
o
this time, it is recommended that vascular surgery patients should
n
• Once a patient exceeds two risk actors, the patient is at least
s
receive pharmacologic VTE prophylaxis in a similar ashion to
u
moderate risk.
l
general surgery patients. Caution should be used with mechanical
t
a
t
prophylaxis; GCS and IPC are relatively contraindicated in vascular
i
o
surgery patients undergoing lower limb bypass.
n
■ DOES THIS PATIENT UNDERGOING GYNECOLOGIC OR There is also limited evidence or VTE prophylaxis in cardiac sur-
UROLOGIC SURGERY NEED VTE PROPHYLAXIS? gery patients. Like vascular surgery patients, these individuals com-
monly receive antithrombotic agents such as intravenous heparin,
CASE 56-2 and antiplatelet agents such as aspirin and clopidogrel. Patients
who undergo cardiac valve replacement generally receive ull-dose
A 78-year-old man is undergoing a radical prostatectomy or anticoagulation postoperatively, making pharmacologic VTE pro-
prostate cancer. His past medical history is signif cant or type phylaxis redundant. The incidence o symptomatic VTE a ter cardiac
2 diabetes mellitus, peripheral neuropathy, and a below-knee surgery is thought to range rom 0.5% to 1.1%. Patients at highest
amputation per ormed 10 years ago. He has no history o bleeding. risk are those on prolonged bed rest, those with prolonged hospi-
The urologist asks you i you are aware o any intraoperative or talization be ore surgery, those with postoperative complications,
postoperative strategies to reduce this patient’s VTE risk. and those with congestive heart ailure. Coronary artery bypass
gra ting (particularly i done o -pump) carries a higher risk o VTE
than valve surgery. The risk o heparin-induced thrombocytopenia
There are ewer randomized clinical trials o VTE prophylaxis in (HIT) also in uences decision making regarding VTE prophylaxis.
gynecologic or urologic surgery. Existing risk assessment models have Approximately 20% o patients undergoing coronary artery bypass
also not been validated speci cally in these patients. However, the gra ting (CABG) who develop a PE are diagnosed with HIT. HIT
rates o DVT, PE, and atal PE in major gynecologic surgery are similar has been shown to be more common when LDUH is used, versus
to those in general surgery. For this reason, VTE prophylaxis recom- LMWH. It is uncertain i VTE prophylaxis should be administered
mendations are similar. There is data that cancer, use o hormones
and the pregnancy state are associated signi cantly with thrombosis
in gynecologic surgery, and must be considered in risk assessment.
VTE is an important problem in major urologic surgery, with rates
o postoperative symptomatic VTE between 1% and 5%. However,
there has been only one methodologically rigorous randomized
clinical trial in the last 20 years in this area. Again, VTE prophylaxis
recommendations are similar to those in general surgery. It is known
that open procedures (vs transurethral procedures) and the use o
the lithotomy position are associated with increased VTE risk. Com-
munication with the surgeon and the anesthetist can modi y these
intraoperative actors and help reduce the patient’s VTE risk.
■ DOES THIS PATIENT UNDERGOING CARDIAC OR

VASCULAR SURGERY NEED VTE PROPHYLAXIS?
CASE 56-3
An 82-year-old woman is undergoing emorodistal bypass or
longstanding peripheral arterial disease. She is an ex-smoker, and
has a history o hypertension and hyperlipidemia. The patient
is currently on aspirin and has been told that she will receive
additional blood thinners at the time o her operation to keep
her arteries rom getting blocked. She has read about deep vein
thrombosis, and wonders i she should receive any extra care to
prevent this complication o surgery. Figure 56 1 Computed tomography pulmonary angiography showing
emboli in both pulmonary arteries.
370
to all cardiac surgery patients; however, in CABG patients who do
not receive ull-dose therapeutic anticoagulation postoperatively TABLE 56-2 Contraindications to VTE Prophylaxis with
C
most physicians elect to use prophylactic-dose heparin or bilateral Anticoagulants
H
mechanical VTE prophylaxis (i the patient has not had saphenous
A
Excessive active bleeding (beyond that expected a ter surgery)
P
vein gra ting). Because the risk o HIT is high in cardiac surgery,
High risk or bleeding that precludes anticoagulants (eg, brain
T
LMWH is pre erred over LDUH. Attention must, however, be paid
lesion)
E
to bleeding risk; this risk goes up with concomitant antiplatelet use,
R
older age, renal insuf ciency, and longer bypass time. I bleeding is Recent serious bleeding (within 1 mo)
5
a concern, mechanical VTE prophylaxis is pre erred over pharmaco- Coagulopathy (eg, INR > 1.5, aPTT > 40)
6
logic VTE prophylaxis. Thrombocytopenia (platelets < 75 × 109/L)
V
e
n
■ DOES THIS PATIENT UNDERGOING NEUROSURGERY
o
WHAT PHARMACOLOGIC AND NONPHARMACOLOGIC
u
NEED VTE PROPHYLAXIS?
s
STRATEGIES SHOULD BE USED FOR VTE PROPHYLAXIS?
T
Patients undergoing major neurosurgery are at a moderate risk
h
Physicians have a number o options available when choosing
r
or VTE, with rates o proximal DVT as high as 5% postoperatively.
o
the type o VTE prophylaxis in nonorthopedic surgery. Mechanical
m
Those with malignant brain tumors are at particularly high risk. One methods o prophylaxis, which include graduated compression
b
prospective study o more than 250 patients with gliomas showed
o
stockings and intermittent pneumatic compression devices, are a
e
that 31% had symptomatic, venographically con rmed DVT within
m
sa e option in patients with an increased bleeding risk (Table 56-2).
5 weeks o their surgery. However, a major barrier to optimal VTE pro-
b
However, they are in erior to pharmacologic prophylaxis and are
o
phylaxis in neurosurgery patients is their risk o bleeding. Intracranial
l
insuf cient protection against VTE in patients at high risk o throm-
i
s
bleeding can have devastating clinical consequences, and or this
m
bosis. All patients with an increased risk or bleeding should be reas-
reason, preoperative and early postoperative pharmacologic pro- sessed regularly so pharmacologic prophylaxis can be started when
(
V
phylaxis should be used with caution in craniotomy patients. Rates
T
the bleeding risk decreases to an acceptable level.
E
o intracranial hemorrhage appear to double when postoperative
)
Pharmacologic prophylaxis should be the thromboprophylaxis
P
LMWH is compared to mechanical or no VTE prophylaxis (approxi-
r
strategy o choice in nonorthopedic surgery. Re er to Table 56-3 or
o
mately 2%-6% vs 1%-3%). Most o these bleeds occur within the rst
p
the recommended dose regimens. Note that these doses may be
h
2 days a ter surgery. A reasonable approach that balances the risks inadequate in the bariatric surgery population, and that LMWH dos-
y
l
o bleeding and thrombosis is to start mechanical VTE prophylaxis
a
ing in obese patients remains controversial. Two studies have shown
x
with properly tted IPC at the time o neurosurgery. Perioperative
i
s
that enoxaparin 40 mg every 12 hours subcutaneously ( or bariatric
use o IPC is highly e ective, reducing the risk o VTE by more than
o
surgery patients with BMI ≤50 kg/m2) or 60 mg every 12 hours ( or
r
two-thirds. I a care ul clinical assessment is stable and postoperative BMI >50 kg/m2) appears to be an e ective VTE prophylaxis strategy.
N
CT scan does not show bleeding at 24 to 48 hours, LMWH or LDUH
o
n
can be added to urther protect the patient rom VTE. There is no
o
r
evidence or extended prophylaxis in neurosurgery patients. PRACTICE POINT
t
h
o
• Pharmacologic prophylaxis should be the thromboprophylaxis
p
e
■ DOES THIS PATIENT UNDERGOING LAPAROSCOPIC strategy o choice in nonorthopedic surgery.
d
OR BARIATRIC SURGERY NEED VTE PROPHYLAXIS? •
i
c
Mechanical methods o prophylaxis, which include graduated
S
Laparoscopic surgery is becoming an increasingly popular alterna- compression stockings and intermittent pneumatic
u
r
tive to conventional open surgical procedures, due to decreased compression devices, are a sa e option in patients with an
g
e
tissue trauma and aster recovery times. However, there are some increased bleeding risk.
r
y
unique eatures o laparoscopic surgery that increase thrombosis • However, they are in erior to pharmacologic prophylaxis, and
risk, including longer intraoperative time, the reverse Trendelenburg are insuf cient protection against VTE in patients at high risk o
position, and pneumoperitoneum (which creates venous stasis in thrombosis.
the lower extremities). Nevertheless, rates o symptomatic VTE ol-
lowing laparoscopic surgery are lower than in general surgery, less
than 0.5% in most series. Rates o asymptomatic VTE are thought Other studies have tried to address this issue by using anti-Xa
to be lower than 1%. For this reason, routine VTE prophylaxis is not levels as a primary outcome instead o clinical events. Prospective
recommended in laparoscopic surgery. nonrandomized studies have questioned the validity o a “maxi-
Bariatric surgery, which includes Roux-en-Ygastric bypass, gastric mum LMWH dose” by showing that increasing the total daily dose
banding, vertical-banded gastroplasty, and biliopancreatic diver-
sion, is also a growing eld. More than 100,000 bariatric surgeries or
morbid obesity are per ormed in the United States every year. Obe-
sity is a known risk actor or VTE and puts bariatric surgery patients
TABLE 56-3 Pharmacologic Prophylaxis Options in
in a unique risk group. Rates o symptomatic VTE vary depending
Nonorthopedic Surgery
on the study quoted, anywhere rom 0.8% to 2.4%. However, there
is still insuf cient high-quality evidence to make clear recommenda- Fondaparinux (Arixtra) 2.5 mg SC once daily
tions regarding VTE prophylaxis in bariatric surgery. Early ambulation Dalteparin (Fragmin) 2500 units or 5000 units SC once daily
and mechanical prophylaxis (either IPC or GCS) are widely accepted
Enoxaparin (Lovenox) 40 mg SC once daily or 30 mg SC every 12 h
components o postoperative care in bariatric surgery patients. The
American College o Chest Physicians’ most recent guidelines also Nadroparin (Fraxiparine) 1900-3800 anti-Xa units SC once daily
recommend LMWH, LDUH, or ondaparinux be routinely used. Con- Tinzaparin (Innohep) 3500 or 4500 anti-Xa units SC once daily
sultation with a pharmacist is important, as patients who are obese Heparin 5000 units SC every 12 h or every 8 h
who may require higher drug doses.
371
o LMWH by 50% or patients with a BMI >50 kg/m 2 can e ectively SUGGESTED READINGS
increase patients’ anti-Xa levels to the target prophylactic range
(0.2-0.4 IU/mL). However, these studies were not all suf ciently Agnelli G, Bergqvist D, Cohen AT, et al. Randomized clinical trial o
P
powered to detect clinically relevant bleeding. Higher doses o postoperative ondaparinux versus perioperative dalteparin or
A
anticoagulant do seem necessary to achieve an appropriate level o prevention o venous thromboembolism in high-risk abdominal
R
prophylaxis or obese patients, though physicians must be cautious surgery. Br J Surg. 2005;92:1212-1220.
T
o bleeding complications as they increase LMWH dosing. Anti-Xa Bergqvist D, Agnelli G, Cohen AT, et al. Duration o prophylaxis
I
I
levels, though not a surrogate marker or bleeding risk, can provide against venous thromboembolism with enoxaparin a ter surgery
a use ul adjunct to close clinical monitoring. or cancer. N Engl J Med. 2002;346:975-980.
Borkgren-Okonek MJ, Hart RW, Pantano JE, et al. Enoxaparin throm-
M
WHAT ARE SOME PRACTICAL MANAGEMENT ISSUES boprophylaxis in gastric bypass patients: Extended duration,
e
FOR THE HOSPITALIST IN VTE PROPHYLAXIS? dose strati cation, and anti actor Xa activity. Surg Obes Relat Dis.
d
i
■ CONSIDERATIONS UPON DISCHARGE 2008;4(5):625.
c
a
FROM HOSPITAL
l
Caprini JA. Risk assessment as a guide or the prevention o the
C
o
There is no evidence that patients undergoing nonorthopedic sur- many aces o venous thromboembolism. Am J Surg. 2010;199:S3.
n
gery bene t rom routine DVT screening using Doppler ultrasound
s
Clagett GP, Reisch JS. Prevention o venous thromboembolism
u
or venography. This strategy may pick up asymptomatic venous
l
in general surgical patient: results o meta-analysis. Ann Surg.
t
a
thrombosis in a small number o patients, but does not appear
t
1988;208:227-240.
i
o
to reduce the rates o clinically signi cant events. There is also no
n
Collins R, Scrimgeour A, Yusu S. Reduction in atal pulmonary
robust evidence that VTE prophylaxis should be extended a ter
embolism and venous thrombosis by perioperative administra-
hospital discharge. Patients at particularly high risk o thrombosis
tion o subcutaneous heparin: Overview o results o randomized
(eg, major abdominal or pelvic surgery, cancer surgery, prolonged
trials in general, orthopedic, and urologic surgery. N Engl J Med.
immobility, and multiple other risk actors), may be candidates
1988;318:1162-1173.
or extended prophylaxis or our additional weeks, generally with
parenteral anticoagulants. The hospitalist can acilitate this transi- Goldhaber SZ, Schoep UJ. Pulmonary embolism a ter coronary
tion by arranging nursing care at home, instructing the patient (or a artery bypass gra ting. Circulation. 2004;109:2712-2715.
caregiver) on injection technique and sa e disposal o used needles, Gould MK, Garcia DA, Wren SM, et al. Prevention of VTE in nonortho-
and clari ying who to call i bleeding develops. All patients should pedic surgical patients: Antithrombotic therapy and prevention o
be educated about the signs and symptoms o VTE during their thrombosis, 9th ed. American College o Chest Physicians evidence-
hospital stay and at the time o discharge. based clinical practical guidelines. Chest. 2012;141:e227S.
Mismetti P, Laporte S, Darmon JY, Buchmüller A, Decousus H.
■ QUALITY IMPROVEMENT INITIATIVES TO Meta-analysis o low molecular weight heparin in the preven-
OPTIMIZE VTE PROPHYLAXIS tion o venous thromboembolism in general surgery. Br J Surg.
VTE prophylaxis in nonorthopedic surgery patients is a critical public 2001;88(7):913-930.
health issue. In 2009, the US Surgeon General cited VTE as one o Scholten DJ, Hoedema RM, Scholten SE. A comparison o two di er-
the most preventable hospital-acquired illnesses, and issued a call ent prophylactic dose regimens o low molecular weight heparin
to action to optimize its prevention. Though VTE prophylaxis in in bariatric surgery. Obes Surg. 2002;12(1):19.
nonorthopedic surgery is based on scienti c evidence, the quality
Simone EP, Madan AK, Tichansky DS, et al. Comparison o two low-
and number o trials is limited compared to the orthopedic sur-
molecular-weight heparin dosing regimens or patients under-
gery population. There is also signi cant room or improvement in
going laparoscopic bariatric surgery. Surg Endosc. 2008;22(11):
compliance rates with the American College o Chest Physicians’
2392-2395.
guidelines or in-hospital VTE prophylaxis. Concerted system-wide
and local e orts must be made to increase the appropriate use o
VTE prophylaxis in nonorthopedic surgery.
372
57
Over 13.8 million units o red blood cells are trans used to over
5 million patients in the United States in 2011; blood trans usion is
one o the most common procedure codes recorded at discharge
or hospitalized patients. It is estimated that 60% to 70% o these
trans usions occur in relation to surgical procedures. For over 40 years
it was generally assumed that patients bene tted rom trans usions
whenever the hemoglobin ell below 10 g/dL or i the hematocrit ell
below 30% (the so-called “10/30 rule”). Controversy remains regarding
Postoperative Blood the appropriate selection o patients requiring trans usion, as well as
the recognition and management o adverse trans usion reactions.
Transfusion POSTOPERATIVE ANEMIA

Trans usion in the postoperative period generally occurs due to
Olga S. Chajewski, MD anemia. The prevalence o postoperative anemia is di cult to assess
accurately, but it is undoubtedly common, particularly in critically ill
Jerry E. Squires, MD, PhD patients. Hemoglobin concentrations on admission to the intensive
care unit (ICU) are on average 11.0 to 11.3 g/dL in two large studies,
with overall trans usion prevalence rates o 37% to 44%. Postopera-
tive anemia can result rom several actors, but acute and chronic
blood loss is most requently cited. While intraoperative or traumatic
blood loss is generally replaced during the surgical procedure, a
signi cant number o patients still leave surgery with some degree
o anemia. Patients admitted to the ICU ollowing surgery on aver-
age have hemoglobin levels ranging rom 10.8 to 11.5 g/dL. Perhaps
more insidious is the chronic blood loss that can occur during the
postoperative period itsel . Causes or this include ongoing bleed-
ing rom the surgical site (eg, chest tube drainage) and repeated
blood collections or laboratory testing. It has been suggested that
40 to 60 mL o blood is routinely collected rom ICU patients daily.
While this daily blood loss may seem minimal, it is aggravated by
the decreased production o erythropoietin (EPO), resistance to the
e ects o EPO, and an inability to utilize iron in red cell production,
all o which have been documented in the postoperative period.
In patients who display symptoms o anemia in the postoperative
period, there are limited alternatives to trans usion. EPO and iron
therapy are o ten ine ective in the postoperative period and are o
little use in the patient with an acute need or blood.
PRACTICE POINT
Arguments in avor o conservative management o postoperative
anemia without the use o blood products include:
• Risks associated with blood trans usion.
• Accumulating evidence that trans usion can negative impact
patient outcomes.
Steps that may be taken include:
• Minimize blood draws to tests that would inf uence
management.
• Assess the clinical need or trans usion in each individual
patient and only trans using patients likely to bene t rom
trans usion (ie, generally those with a hemoglobin <7 g/dL or
those who are hemodynamically unstable with a hemoglobin
<10.0 g/dL).
373
POSTOPERATIVE TRANSFUSION All trans usion reactions should be promptly reported to the blood
bank or trans usion service, as additional testing may be required; it
Several well-documented randomized controlled trials have dem-
may also a ect the management o other blood components col-
onstrated that a hemoglobin o 7.0 g/dL, when used as a trigger
P
lected rom the donor o the implicated product. Acute trans usion
or the trans usion o red blood cells, is at least as sa e (nonin-
A
reactions are listed in Table 57-2.
erior) as a hemoglobin o 10.0 g/dL. Trans usion has also been
R
T
associated, in a number o reports, with increased risk o in ection,
I
morbidity, mortality, length o hospital stay, and length o ICU
I
stay. A number o pro essional organizations, such as the College PRACTICE POINT
o American Pathologists, American Society o Anesthesiologists,
and the American College o Physicians, have published guide-
• It is critical that trans usion reactions be promptly recognized
so that appropriate management o the patient and
lines or the trans usion o red blood cells in nonbleeding anemic
M
investigation o other blood components rom the implicated
e
patients. In general, each o these sets o guidelines recommend
d
donor can be initiated.
that at hemoglobin levels less than 7.0 g/dL red blood cell trans-
i
c
a
usion will be necessary, while at hemoglobin levels greater than
l
C
10.0 g/dL trans usion is rarely required. These guidelines all stress
o
n
the priority o assessing the clinical need or trans usion in each
s
u
individual patient. These guidelines suggest that a slightly higher TRANSFUSION-TRANSMITTED INFECTIONS (TTI)
l
t
hemoglobin trans usion threshold o 8.0 g/dL may be more appro-
a
The current risk or trans usion transmitted in ection is low. See
t
i
priate or patients with acute coronary syndrome. A red blood cell
o
Table 57-3.
n
trans usion practice guideline rom the American Association o In addition, with the recent FDA approval o pathogen-reduction
Blood Banks (AABB) makes our recommendations as outlined in technology or platelet and plasma components, these risks (and
Table 57-1. Patients with evidence o poor tissue per usion (eg, the risks associated with new pathogens in the blood supply) will
hypotension, tachypnea, tachycardia, chest pain) may require a be urther reduced.
red blood cell trans usion despite a hemoglobin value above the
trigger threshold o 7 g/dL. In adult patients with anemia who are
not actively bleeding, each unit o red blood cells should increase PATIENTS REQUIRING BLOOD PRODUCT IRRADIATION
the hemoglobin level by approximately 1.0 g/dL.
Certain patients, usually those who are immunocompromised,
require cellular blood products (red blood cells, platelets, white
ADVERSE TRANSFUSION OUTCOMES blood cells) that are irradiated in order to prevent trans usion-
Many adverse outcomes are preventable. One national hemo- associated gra t versus host disease (GvHD). See Table 57-4.
vigilance program reported adverse reactions occur at a rate o
10/100,000 components trans used, and ~70% o these events are
PATIENTS REQUIRING CYTOMEGALOVIRUS (CMV)
“preventable.” NEGATIVE BLOOD
While CMV in ections or most individuals cause mild, i any, symp-
NONINFECTIOUS ACUTE TRANSFUSION REACTIONS toms, some patients require the trans usion o “CMV-sa e” blood
Trans usion reactions are o ten classi ed as either acute (occur products. Leukoreduced blood products are considered to be
within 24 hours o trans usion) or delayed. The prompt recognition CMV-sa e and provide a more readily alternative to the use o blood
and management o acute reactions is critical to the patient’s care. products collected rom CMVseronegative donors.
FILTERS
TABLE 57-1 American Association of Blood Banks (AABB) All blood products, including red blood cells, platelets, plasma, and
Red Blood Cell Transfusion Practice Guidelines cryoprecipitate, must be trans used through a lter. The standard
Recommendation Evidence/Grade blood lter has a pore size o 170 to 260 microns.
Clinicians should adhere to a Strong recommendation;
restrictive trans usion strategy high-quality evidence
(ie, hemoglobin 7-8 g/dL) in INFUSION TIME
hospitalized, stable patients Blood product in usions must be completed with 4 hours o the
Clinicians might consider a slightly Weak recommendation; product being dispensed rom the blood bank. In general, a red blood
higher trans usion threshold moderate-quality cell trans usion should be started slowly (approximately 2 mL/min)
trigger (ie, hemoglobin 8 g/dL) evidence and the patient monitored or the rst 15 minutes. Most serious
in hospitalized patients with pre-
reactions occur within the rst 10 to 15 minutes o the trans usion.
existing cardiovascular disease
A ter the initial 15 minutes the rate o trans usion can be increased
Recommendations cannot be Uncertain to approximately 4 mL/min.
made or or against a liberal or recommendation; very
No medications or solutions other than 0.9% sodium chloride
restrictive trans usion threshold low-quality evidence
or hospitalized, hemodynamically should be in used with blood components. Solutions containing
stable patients with acute coronary dextrose have caused lysis o red blood cells and solutions contain-
syndrome ing calcium, such as Lactated Ringer’s solution, can cause clotting
Trans usion decisions should be Weak recommendation; o the blood component. There are a limited numbers o other
in luenced by symptoms as well as low-quality evidence products that are compatible with blood, but the package insert
hemoglobin concentration should be consulted to insure that that have been approved to be
compatible with blood or blood components.
374
TABLE 57-2 Acute Transfusion Reactions
C
H
Risk Incidence Etiology Signs/Symptoms Management
A
Allergic (urticarial) 1:100-1:33 Antibody in recipient to donor Urticaria SUSPEND trans usion
P
reactions plasma protein
T
Flushing Administer
E
Itching antihistamines
R
Report to blood bank
5
Anaphylactic reactions 1:20,000-1:50,000 Antibody to donor plasma Hypotension STOP trans usion
7
proteins (o ten anti-IgA in IgA Urticaria Maintain blood pressure
de icient recipient)
Bronchospasm Report to blood bank
P
o
s
Local edema
t
o
Febrile reactions 1:100 Antibody to donor white blood Fever STOP trans usion
p
e
cells or presence o cytokines in Chills Administer antipyretics
r
a
blood product
t
i
v
e
Acute hemolytic 1:76,000 Red cell incompatibility (o ten Fever STOP trans usion
B
reactions ABO incompatibility)
l
Chills/rigors Manage hypotension
o
o
Hypotension Maintain renal per usion
d
T
Back/ lank pain Report to blood bank
r
a
n
Pain at injection site
s
u
Hemoglobinemia
s
i
o
Hemoglobinuria
n
Renal ailure/oliguria
Disseminated
intravascular coagulation
Trans usion-related acute 1:5000 Donor HLA or white cell Respiratory distress STOP trans usion
lung injury (TRALI) antibodies Fever Provide respiratory
Tachycardia support
Hypotension Mechanical ventilation
Chest x-ray with bilateral
in iltrates
Trans usion-associated 1:100 Excess volume Respiratory distress STOP trans usion
circulatory overload Orthopnea Diuretic therapy
(TACO)
Cough Respiratory support
Hypertension
Bacterial contamination RBC: 2.6:100,000 Bacterial contamination o blood Fever STOP trans usion
Plts: 1:75,000 product Chills/rigors Maintain blood pressure
Hypotension Administer antibiotics
Nausea/vomiting Obtain blood cultures
TABLE 57-3 Most Common Transfusion-Transmitted Viral Infections
Agent Estimated Risk with Test-Negative Blood Product

Human Immunode iciency Virus (HIV) 1:2135,000
Hepatitis C Virus (HCV) 1:1935,000
Hepatitis B Virus (HBV) 1:205,000-488,000
375
TABLE 57-4 Patients Requiring Irradiation of Cellular
SUGGESTED READINGS
Blood Products
Carson JL, Terrin ML, Noveck H, et al. Liberal or restrictive trans-
P
Well-Documented Intrauterine trans usions usion in high-risk patients a ter hip surgery. N Engl J Med.
A
2011;365:2453-2462.
R
Low-birth-weight premature in ants
T
Congenital immunode iciencies Hebert PC, Wells G, Blajchmann MA, et al. A multicenter randomized,
controlled clinical trial o trans usion requirements in critical care.
I
I
Hematologic malignancies
N Engl J Med. 1999;340:409-417.
Peripheral blood or marrow
transplantation Liumbruno GM, Bennardello F, Lattanzio A, et al. Recommenda-
tions or the trans usion management o patients in the peri-
Fludarabine therapy
M
operative period. III. The post-operative period. Blood Transfus.
e
Granulocyte products 2011;9:320-335.
d
i
HLA-matched platelet products
c
Murphy GJ, Reeves BC, Rogers CA, et al. Increased mortality, post-
a
l
Blood components donated by blood operative morbidity, and cost a ter red blood cell trans usion in
C
relatives
o
patients having cardiac surgery. Circulation. 2007;107:2544-2552.
n
Potential Malignancies treated with cytotoxic
s
Stramer SL. Current risks o trans usion-transmitted agents. Arch
u
Indications agents
l
Pathol Lab Med. 2007;131:702-707.
t
a
Generally NOT Patients with human immunode iciency
t
i
o
Indicated virus
n
Full-term in ants
Nonimmunosuppressed patients
376
58
The importance o providing adequate nutrition as an adjunct to
medical care was identi ed as early as the era o Hippocrates. The
prevalence o malnutrition in the hospitalized patient was largely
ignored until 1974 when Butterworth published his landmark
paper entitled “The Skeleton in the Hospital Closet.” 30% to 50% o
hospitalized patients are malnourished upon admission, and they
tend not to improve nutritionally, and requently worsen, while
hospitalized.
Nutrition and Malnutrition is associated with increased in ection rates, longer
hospital length o stay, increased hospital costs, and mortality.
Metabolic Support There ore, the nutritional status o most patients should be assessed
throughout their hospitalizations.
This chapter will provide the hospitalist with a ramework or
assessing nutritional status, identi ying patients at increased risk
Nicole M. Bedi, RD, CNSC o malnutrition, and determining the most appropriate nutritional
prescription. This chapter will not cover nutrition at the end o li e.
Malcolm K. Robinson, MD See Chapter 216 (Palliation o Common Symptoms).
NUTRITION EVALUATION AND SCREENING

Usually a dietitian or nurse uses a nutrition screening questionnaire
to assess patients or malnutrition at the time o admission. The
assessment process includes a combination o anthropometric
measurements, and the history and physical examination. Most
validated nutrition screening tools evaluate a combination o body
mass index (BMI), recent unintentional changes in weight, recent
changes to oral intake, and severity o present illness.
Nutritional screening may identi y patients who are under-
weight and there ore at signi cant risk or developing nutritionally
related complications during treatment o acute illness. However,
nutritional screening may ail to identi y other patients at increased
risk based simply on weight. Patients admitted with volume over-
load due to heart ailure, cirrhosis, or renal insu ciency may have
their true dry weight masked by excessive uid retention. Acutely
ill patients who have received aggressive uid resuscitation gain
several kilograms above their true dry weight. Patients whose dry
weight is in the obese range may have malnutrition due to protein
degradation and loss o lean body mass rom rapid weight loss
during acute illness.
■ ANTHROPOMETRIC MEASUREMENTS AND

PHYSICAL ASSESSMENT
Anthropometric measurements are a set o noninvasive, quantita-
tive techniques or determining an individual’s body at composi-
tion by measuring speci c dimensions o the body, such as height
and weight, triceps skin- old thickness, mid-upper arm circum er-
ence, and bodily circum erence at the waist, hip, and chest. I such
measurements all below standard norms then one is considered
malnourished. For example, health care pro essionals can compare a
patient’s weight to the ideal body weight (IBW) ound in tables rom
various sources, such as the 1952 Metropolitan Li e tables. Alterna-
tively, one can use the Hamwi “rule o thumb,” which was developed
rom the Metropolitan table.
377
PRACTICE POINT PRACTICE POINT
Every patient should have a nutrition assessment when admitted • Unintentional weight loss has consistently been shown to be
to the hospital. a reliable measure o malnutrition, especially in older adults. A
P
loss o greater than 5% o one’s usual body weight (UBW) in
A
One assessment is the Hamwi “rule o thumb,”which can be used
R
to compare a patient’s IBW to actual body weight: 1 month or 7.5% o UBW over 3 months is considered clinically
T
• IBW or a woman is 100 pounds or the rst 5 eet plus an signi cant, even i one is initially obese.
I
additional 5 pounds or every inch over 5 eet.
I
• IBW or a man is 106 pounds or the rst 5 eet plus an Typically, albumin and prealbumin rise as patients recover rom
additional 6 pounds or every inch over 5 eet.
their in ammatory insult. However, during ongoing in ammatory
Another assessment is the body mass index (BMI):
illness (eg, cancer, wound in ection, abscess, and necrotic tissue),
M
• Underweight is a BMI less than 18 kg/m2
e
albumin and prealbumin will be persistently low despite adequate
• Overweight is a BMI greater than 25 kg/m2
d
nutrition.
i
c
• Obese is a BMI greater than 30 kg/m 2
a
Many non-nutritional actors may alter the level o circulating
l
C
Pit alls to relying on body weight as an indicator o nutritional nutritional proteins used to monitor nutritional status. C-reactive
o
status include
n
protein (CRP), an acute phase reactant, will usually be elevated
s
• Fluid overload states, which alsely elevate weight in a
u
due to the presence o in ammation. I the CRP is normal and the
l
t
malnourished patient. prealbumin is low, this likely represents malnourishment. I the CRP
a
t
• Obesity, which may mask the relevance o recent weight loss.
i
is high and the prealbumin is low, one will not be able to reliably
o
n
distinguish between malnourishment and in ammation. Although
such levels should be monitored, physicians must use caution when
BMI is independent o gender and body rame size. However, interpreting these markers.
those who meet the BMI criteria or overweight and obese may
still have signi cant depletion o nutrient stores, particularly when THE NUTRITION PRESCRIPTION
acutely ill. Thus, the BMI alone is insu cient to classi y an individual
Patients who are well nourished and are not at risk or malnourish-
as nutrient replete. Highly trained and muscular athletes with dis-
ment on admission do not need a plan, but should be reassessed
proportionately high lean body mass may not have a high body at
every 3 to 5 days. All patients who are malnourished on admission,
percent despite a BMI in the overweight or obese category.
or who are at risk o developing malnutrition, should have a ormal
In 2012, the Academy o Nutrition and Dietetics (AND) and the
nutrition plan developed.
American Society or Parenteral and Enteral Nutrition (ASPEN) pro-
duced a consensus statement or standardizing diagnostic criteria o
malnutrition. As no single parameter is de nitive or adult malnutri- ■ DETERMINATION OF NUTRIENT REQUIREMENTS
tion, the identi cation o two or more o the ollowing six character- Both under- and over eeding calories is detrimental. Excess carbohy-
istics is recommended or diagnosis: drate administration can result in hyperglycemia, and excess carbon
• Insu cient caloric (energy) intake (estimated rom dietary dioxide production, which is o particular concern in patients with
recall) lung disease (who may retain CO2 and have di culty weaning rom
• Weight loss (based on objective measurement o weight) the ventilator). Long-term over eeding may lead to hepatic steatosis,
• Loss o muscle mass (temporal/interosseus muscle wasting; ureagenesis, and immunosuppression (especially over eeding o lip-
clavicular prominence) ids). The goal is to determine nutrient needs precisely, and to avoid
• Loss o subcutaneous at (cheeks/orbital area, or space between both under- and over eeding.
thumb and ore nger) The “gold standard” or determination o caloric requirements is
• Localized or generalized uid accumulation, that may mask indirect calorimetry, or a “metabolic cart” study. The metabolic cart
weight loss (pitting edema) measures oxygen consumption and carbon dioxide production and
• Diminished unctional status (assessed by handgrip strength then calculates resting energy expenditure (REE) through utilization
with a dynamometer) o the Weir equation. Indirect calorimetry provides a result which
already includes the additional caloric expenditure related to dis-
Additional nutrition related physical assessment should include
ease stress, but not the caloric expenditure related to activity. Hence,
evaluation o skin, hair, mouth, and nails or signs o nutrient de -
the results o the metabolic cart study are increased by an activity
ciencies. Dry skin or rash, dry hair or hair loss, poor skin turgor, night
actor to calculate the nal daily caloric expenditure. Typically, activ-
blindness, glossitis, and muscle weakness may also point to vitamin,
ity actors range rom 0% to 5% in intubated patients to as high as
mineral, or atty acid de ciencies.
30% in ambulatory patients.
Metabolic cart testing remains the most reliable way to determine
■ LABORATORY DATA caloric requirements or patients with a very low or high BMI, or
There is no reliable laboratory marker to determine the presence those with amputations, and or those who have severe illness or
o malnutrition, or the response to adequate eeding in acutely injury such as multiple traumas or burns. Use o predictive equations
ill, hospitalized patients. Serum proteins, most requently albumin to estimate caloric requirements in such patients can lead to highly
(but also prealbumin, retinol binding protein, and trans errin) were inaccurate results.
historically used or this purpose; however, these proteins are all However, the use o predictive equations is the most common
acute phase reactants. The liver decreases their production during method or determining caloric requirements because metabolic
acute illness or ollowing surgery. Hyper- or hypovolemia, steroid cart studies are usually impractical (requires trained personnel and
administration, alcoholism, liver, and renal ailure can also alter cir- expensive equipment).
culating nutrition protein levels, independent o nutritional status. A • The Harris-Benedict equation (HBE) is based on studies o
decreased albumin is more accurately a marker o in ammation in healthy volunteers; it is the oldest and most widely used equation
the acute setting. or determining basal metabolic rate (BMR).
378
110 Burn S ize The actual body weight (ABW) is generally used or caloric
in tBSA assessment. In patients >120% o their ideal weight, IBW is recom-
C
mended. Some nutrition pro essionals will use an “adjusted weight”
H
100
A
80% or obesity, although no routine standardized adjustment has been
P
90 published (Figure 58-2).
T
For obese patients, “permissive under eeding” aims to eed the
E
70% patient at a target goal o 50% to 70% o estimated calorie require-
R
80
ments, providing an abundance o protein to preserve lean body
5
mass. With this, patients with a BMI 30 to 50 kg/m 2 should receive
8
70 60%
11 to 14 kcal/kg/d using ABW and those with a BMI >50 kg/m 2
S e ve re tra uma , s e ps is should receive 22 to 25 kcal/kg/d using IBW.
60
N
a nd re s pira tory fa ilure In addition to a caloric prescription, providers should determine
u
P e rc e nta g e 50%
t
their patients’ protein goals. Generally, hospitalized patients require
r
50
i
Cha ng e
t
i
between 1.2 and 1.5 g/kg/d, with burn patients requiring up to
o
40%
n
40 Multiple tra uma 2.0 g/kg/d. In critically ill obese patients, up to 2.5 g/kg/d (IBW) is
a
recommended.
n
S e ve re he a d injury
d
M
30 30%
e
PRACTICE POINT
t
Long bone fra cture
a
20
b
20% A quick “rule o thumb” or most hospitalized patients caloric and
o
P e ritonitis , pne umonia
l
i
protein needs (IBW):
c
10
• 25 to 30 calories/kg/d
S
u
P os tope ra tive • 1.2 protein g/kg/d
p
0
p
Norma l
o
r
t
– 10
Mild s ta rva tion
■ ROUTE OF FEEDING
Figure 58 1 Percent change in metabolic rate due to injury. The next component o the nutrition prescription is determining the
route o eeding (Figure 58-3). Oral nutrition is pre erred. Special-
ized nutritional support must be considered when patients cannot
sa ely or adequately meet their nutrient requirements through oral
• The Mi in-St. Jeor equation may be more accurate than
diet alone.
the HBE or determining caloric needs (when compared to
indirect calorimetry). This equation is recommended for obese
noncritically ill patients (although precision goes down with ENTERAL NUTRITION
increasing obesity). For the purpose o this discussion, enteral nutrition will re er to
• Penn State equation is the Mi in-St. Jeor equation modi ed eeding o patients via enteric tubes placed in the stomach or small
with age, body temperature, and minute ventilation or euca- bowel. The old adage “i the gut works, use it” guides decision mak-
loric eeding. This equation is recommended for obese critically ill ing or specialized nutritional support.
patients. Provision o nutrients into the GI tract preserves structural and
No equations actor in the increased caloric expenditure related unctional integrity and maintains gut-associated lymphoid tissue.
to disease. Hence, once the BMR is determined, an estimate o the Relative and absolute contraindications to enteral eeding include
additional caloric expenditure rom activity and metabolic stress major GI hemorrhage, peritonitis, severe ileus, bowel obstruction
( rom disease or acute illness) is actored in to determine the overall or stulae distal to enteral access site, intestinal ischemia, and mal-
daily caloric requirement. Stress actors can range rom 10% in rou- absorptive disorders with high-volume diarrhea (eg, short bowel
tine patients to 100% in severe burn patients (Figure 58-1). syndrome, radiation enteritis, gra t vs host disease o the GI tract).
Harris Be ne dic t Equatio n
Me n: [13.75 x we ight (kg)] + [5.00 x he ight (cm)] – [6.78 x a ge (y)] + 66.5

Wome n: [9.56 x we ight (kg)] + [1.85 x he ight (cm)] – [4.68 x a ge (y)] + 655.1
Mifflin-S t. Je o r Equatio n
Me n: (9.99 x we ight) + (6.25 x he ight) – (4.92 x a ge ) + 5

Wome n: (9.99 x we ight) + (6.25 x he ight) – (4.92 x a ge ) – 161
Example : A 45-ye a r-old ma n pre s e nts a dive rticula r a bs ce s s with ile us. He is 6 ft ta ll (182.9 cm) a nd 176 lbs (80 kg).
BMR (us ing Mifflin-S t. Je o r) = [9.99 x 80] + [6.25 x 182.9] – [4.92 x 45] + 5 = 1726
Calo rie Re quire me nt = 1726 (BMR) x 1.25 (ie , 25% ac tivity fac to r) x 1.10 (ie ,10% s tre s s fac to r) = 2473 c alo rie s /d
Figure 58 2 Equations for determining basal metabolic rate.
379
Cons ide r nutritiona l s upport if a ny of the following conditions a re pre s e nt:
• Pa tie nt ha s be e n without nutrition for ≥7-10 days.
• Expe cte d dura tion of illne s s >10 days.
P
A
• Pa tie nt is ma lnouris he d (we ight los s > 10% of us ua l we ight).
R
T
I
I
Initia te nutritiona l s upport only if tis s ue pe rfus ion is a de qua te
a nd p O 2 , p CO 2 , e le ctrolyte s , a nd a cid-ba s e ba la nce a re ne a r norma l
M
Is GI output ≥ 600 mL/24 h, ma s s ive GI he morrha ge , prolonge d ile us,
e
d
or othe r contra indica tion to e nte ra l fe e ding?
i
c
a
l
No Ye s
C
o
n
s
Initia te e nte ra l Adminis te r
u
l
t
fe e dings pa re nte ra l nutrition
a
t
i
o
n
Ente ra l Initia te TP N Initia te P P N
Ente ra l fe e ding
fe e ding
not tole ra te d
tole ra te d
Re a s s e s s ne e d for
PN
Figure 58 3 Determining route of nutritional support.
■ ENTERAL ACCESS ■ MONITORING ENTERAL FEEDING TOLERANCE

The physician must determine the most appropriate access route Feedings are typically started at a low rate, and gradually advanced
based on two major actors: (1) whether the patient requires short- to the in usion goal over a period o 24 to 48 hours.
or long-term eeding, and (2) whether the gastric status is normal. Patients are evaluated or other symptoms such as nausea, vom-
Re er to Chapter 45 (Surgical Tubes and Drains) on inserting and iting, diarrhea, abdominal pain, and bloating. Diarrhea commonly
maintaining enteral tubes. occurs due to gastrointestinal pathogens, bowel edema or in am-
mation, malabsorptive disorders, and medications (Figure 58-4).
Also see Chapter 82 (Diarrhea) or more on the assessment and
■ ENTERAL FORMULAS management o hospitalized patients with diarrhea.
Standard and specialized enteral ormulas can be used to provide I signs and symptoms o intolerance develop, the eedings are
nutrition. Since all o these ormulas have xed macronutrient held until resolution. Persistent intolerance requires an alternative
content, it may not be possible to deliver su cient protein without plan. Patients receiving gastric eedings should be considered or
over eeding calories. Hence, one usually delivers the eedings at a postpyloric eeding as this may improve most symptoms o intoler-
rate to provide appropriate calories and then supplements pro- ance with the exception o diarrhea.
tein with a protein modular. Table 58-1 describes typical enteric Checking gastric residuals has traditionally been the hallmark or
ormulas. Specialty ormulas or a variety o disease states such as monitoring tube eeding tolerance and avoiding tube- eed-associated
pulmonary, renal, and liver dys unction and diabetes should be aspiration. However, several well-designed studies have ailed to
selected with caution due to increased cost and low bene t to demonstrate a signi cant link to gastric residual volume (GRV) and
risk pro les. aspiration risk. Routine monitoring o gastric residuals interrupts
tube eeding with questionable proven bene t.
TABLE 58-1 Typical Enteric Formulas
Type Composition Indications Side Effects

Polymeric (1.0, 1.2, Intact protein, ats, carbohydrates To meet daily requirements: or Concentrated eeding may
1.5, and 2.0 cal/mL) most patients 1-1.5 L/d cause diarrhea and may require
Concentrated eeding best i ree water supplementation
patient is volume restricted
Elemental or Peptide-based amino-acids or Maximizes absorption in May cause diarrhea
Semi-elemental di- or tri-peptides and simple sugars patients with malabsorption
disorders
Immune enhancing Forti ied with arginine, glutamine, May reduce in ection risk in Unclear bene it in medical
diets omega-3 atty acids and/or antioxidants surgical patients patients
380
Dia rrhe a > 600 cc/d
C
H
A
• Re duce tube fe e ding ra te
P
• Provide s oluble fibe r
T
E
R
5
De te rmine e tiology
8
Rule out a nd tre a t the following:
N
Ente ric pa thoge ns, e g Dis e a s e /infla mma tion, e g Offe nding me dica tions, e g
u
t
• C. d iffic ile • Pa ncre a tic ins ufficie ncy • Antibiotics
r
i
t
i
• S a lmone lla • Bile s a lt ma la bs orption • S orbitol-conta ining me dica tions
o
n
• S hig e lla • S hort bowe l syndrome • Unindica te d la ctulos e /la xa tive s
a
• Ca mpylob a c te r • Infla mma tory bowe l • Ma gne s ium-conta ining a nta cids
n
d
• Ye rs inia dis e a s e • Pota s s ium or phos phorus
M
• E. c oli • Bowe l wa ll e de ma s upple me nts
e
• Antine opla s tics
t
a
b
o
l
i
c
S
u
p
p
Dia rrhe a improve d <600 cc/d
o
Dia rrhe a pe rs is ts >600 cc/d
r
Gra dua lly incre a s e tube fe e ding
t
ra te to goa l a s dia rrhe a re s olve s
Antimotility me dica tions

Immodium (lope ra mide HCl)
Lomotil (diphe noxyla te HCl a nd a tropine s ulfa te )
Code ine
De odorize d tincture of opium
Dia rrhe a improve d <600 cc/d Dia rrhe a pe rs is ts >600 cc/d

Gra dua lly incre a s e tube fe e ding Cha nge to pe ptide -ba s e d or
ra te to goa l a s dia rrhe a re s olve s e le me nta l tube fe e ding formula
Dia rrhe a improve d <600 cc/d Dia rrhe a pe rs is ts >600 cc/d

Gra dua lly incre a s e tube fe e ding S top e nte ra l fe e dings, cons ide r
ra te to goa l a s dia rrhe a re s olve s TP N
Figure 58 4 Management of diarrhea in tube-fed patients.
PRACTICE POINT PARENTERAL NUTRITION

The American Society or Parenteral and Enteral Nutrition Patients who cannot tolerate EN should be considered or paren-
(A.S.P.E.N.) recommends monitoring o gastric residual volume teral nutrition (PN). PN is associated with higher rates o in ectious
(GRV) or the rst 48 hours o enteral nutrition initiation. and metabolic complications such as volume overload, hyper-
• I the GRVis greater than 250 mL on two residual checks, glycemia, and electrolyte abnormalities compared to EN. The risk
the physician should consider a motility agent such as o nutritionally related complication rates increases beyond 10 to
metoclopromide or erythromycin. 14 days o inadequate enteral eeding in well-nourished individu-
• I the gastric residual is greater than 500 mL, the eedings als. The decision o when to initiate TPN depends on the nutritional
should be held and the patient evaluated by a physician or status o the patient, the acute illness, the patient’s prognosis, and
abdominal distention, glycemic control, adequacy o motility the estimated duration o inadequate enteral eeding. ASPEN rec-
agents i not already ordered, or consideration o small bowel ommendations suggest around 7 days as an appropriate starting
eeding access. point or PN. PN is only justi ed i the need is anticipated or a
minimum o 5 days.
381
Indications or PN include, but are not limited to: 1 g/kg/d, and dextrose in usion should be less than 5 mg/kg/min.
• Severe malabsorptive disorders (short bowel syndrome, gra t Eventually, the PN solution can be cycled nocturnally (generally
vs host disease o the GI tract, radiation enteritis, in ammatory between 10 and 16 hours) in long-term PN patients, once glycemic
P
bowel disease) control is achieved. Once stable, patients can have lab monitoring
A
• High-output enterocutaneous stulae i distal eeding access decreased to once or twice weekly.
R
is not easible Patients receiving long-term PN may develop PN-associated
T
• Pancreatitis with enteral eeding intolerance liver dys unction (PNALD). However, alterations that occur within
I
I
• Bowel obstruction the rst 2 weeks o PN therapy generally resolve despite continu-
• Uncontrolled anastomotic leak ollowing GI surgery ation o TPN. Alternate causes o liver unction test (LFT) elevation
• Prolonged postoperative ileus should always be ruled out, such as hepatotoxic medications,
biliary obstruction, and sepsis. PNALD should be a diagnosis o
M
exclusion rather than the rst thought in the acutely ill patient that
e
■ PARENTERAL NUTRITION PRESCRIPTION
d
has just been started on PN. Excess provision o IVdextrose or lipid,
i
c
AND MONITORING and de ciencies o carnitine and choline may increase the likeli-
a
l
PN solutions contain carbohydrate in the orm o dextrose, protein hood o PNALD. Provision o the appropriate amount o dextrose
C
o
as crystalline amino acids, and lipids rom polyunsaturated long- and carbohydrate, repleting carnitine and choline, and administer-
n
s
chain triglycerides such as soybean oil or a sa ower/soybean oil ing even small amounts o enteral nourishment may decrease the
u
l
mixture. Vitamins, electrolytes, and trace elements are added to the risk o developing PNALD. Finally, decreasing PN cycle time to less
t
a
t
ormulation as needed. Electrolytes and other additives (eg, medica- than 24 hours each day may allow or hepatic “rest” and decrease
i
o
tions) must be added at the appropriate concentrations to avoid continuous insulin secretion, which may decrease atty deposits
n
“cracking” o lipid-containing PN. I cracking occurs, precipitates in the liver.
orm in the solution, and the lipid emulsion separates into layers,
making it unsa e or administration. ■ PARENTERAL ACCESS
Physicians must care ully monitor patients or metabolic changes
such as hyperglycemia or re eeding syndrome upon initiation o Central venous access is required or the administration o TPN
PN. Hyperglycemia may increase in ectious complications, hospi- because TPN solutions are hyperosmolar. Peripheral PN (PPN) solu-
tal length o stay, and cost. Re eeding syndrome is characterized tions, which are available in some hospitals, have limited dextrose
by electrolyte abnormalities that occur during the reinstitution and amino acid concentrations in order to keep osmolarity lower;
o carbohydrate calories to a starved patient. Serum phosphate, this improves tolerance o PN delivery via a peripheral vein, but low-
magnesium, and potassium depletion may develop and precipitate ers caloric density. PPN is generally not a good long-term (>3-5 days)
potentially li e-threatening cardiac arrhythmias or neuromuscular solution or most patients and only use ul or those patients who
complications. See Chapter 226 (Eating Disorders or more on the can tolerate high volumes o uid daily. The physician must also
re eeding syndrome). determine the most appropriate access device or PN. See Chapter
PN prescriptions typically provide rom 1 to 3 liters o uid per 120 (Vascular Access, or more in ormation about types and indica-
day depending on the physician’s assessment o maintenance uid tions o catheters).
requirements. A general rule o thumb is 30 mL/kg. Carbohydrates
generally make up about 50% to 60% o the caloric prescription, at SPECIAL CONSIDERATIONS
3.4 calories/g o dextrose. Protein generally provides about 15% to ■ PANCREATITIS
25% o the calories at 4.0 calories/g o amino acid. The current US
Food and Drug Administration (FDA)-approved lipid emulsions are Pancreatitis patients require no nutritional intervention in 80% to
primarily made up o omega-6-rich oils, which have been shown to 90% o cases as most will resume an oral diet within 7 days o admis-
be proin ammatory and potentially immunosuppressive. Hence, sion. EN remains the pre erred route o nutritional support or the
lipid provision should be limited to 20% to 25% o calories (at remaining patients because it is associated with a signi cant reduc-
10 calories/g o IV lipid). Novel IV lipids made up o olive oils or sh tion in in ectious morbidity, length o stay, and mortality compared
oil are available in other countries, and are currently under consid- to PN. EN can be appropriately achieved by eeding into the stom-
eration by the FDA. ach or jejunum depending on gastric emptying. O note, patients
It is advisable to start with a lower volume and concentration who are ed within the rst 24 to 48 hours o admission may tolerate
o dextrose when initiating PN to avoid metabolic complications. enteral eeding better than those who start eeding later, and such
One liter o a 10% dextrose solution is a good starting point. Subse- early eeding is associated with substantial outcome bene ts. These
quently the volume, and dextrose, lipid, and protein concentrations bene ts diminish in patients who start eeding as late as day 4.
are increased as needed in the metabolically stable patient to the PN should be considered in severe pancreatitis patients who can-
eventual goal caloric and uid provision (Figure 58-5). Blood sugar not receive enteral eeding or prolonged periods. PN may exacer-
levels should be closely monitored and maintained below 180 mg/ bate the in ammatory response to pancreatitis i initiated too early;
deciliter (dL), and abnormal electrolyte levels should be corrected, ASPEN guidelines suggest not starting PN or 5 days in those with
especially potassium, phosphate, and magnesium, be ore starting severe pancreatitis, even i enteral eeding is not possible.
or advancing to the goal solution. Additionally, any patient who is at
high risk or re eeding syndrome should receive IVthiamine replace- ■ RENAL DISEASE
ment prior to initiation o total parenteral nutrition (TPN) to prevent Dialysis patients o ten have poor oral intake, increased nutrient
development o Wernicke’s encephalopathy. Traditionally, patients losses in their dialysates, and increased catabolic stress. Thus
af ected with alcohol use were the hallmark high-risk patient or malnutrition is common in this patient population. The Kidney Dis-
Wernicke’s. However, those recovering rom weight loss surgery ease Outcomes Quality Initiative (K/DOQI) guidelines recommend
(eg, gastric bypass) with prolonged nausea and vomiting is a newly protein restriction or outpatient dialysis patients, but notably
recognized group o patients who are at risk or this complication. recommend that dialysis patients receive 1.2 to 1.3 g/kg o protein
The solutions should be administered using a volumetric pump set when hospitalized or acute illnesses. Many nutrition experts also
at a constant rate. In general, PN lipid in usion should not exceed believe that acutely ill patients who are “predialysis” should still
382
P a tie nt is a 45-ye a r-old ma n with dive rticula r a bs ce s s , a s in Figure 58-2. Adva nce me nt to a n ora l
C
H
die t ha s be e n uns ucce s s ful for 10 da ys , a nd TP N is re quire d.
A
A. Es tima te nutrie nt re quire me nts :
P
T
Ca lorie re quire me nt: 2373 ca lorie s /d (s e e Figure 58-2)
E
R
P rote in re quire me nt: 1.5 g/kg/d = 80 x 1.5 = 120 g/d
5
Fluid re quire me nt: 30 mL/kg body we ight = 80 x 30 = 2400 mL
8
B. De te rmine the s olution
N
u
1. P rote in conte nt
t
r
i
120 g prote in ÷ 2.4 L fluid = 50 g/L
t
i
o
n
50 g = 5% a mino a cids pe r lite r
a
4 ca lorie s /g of a mino a cids x 120 g = 480 kca ls of prote in
n
d
M
2. Lipid conte nt (a p p roxima te ly 20%-25% of nonp rote in kc a ls )
e
2373 tota l ca lorie s – 480 prote in ca lorie s from prote in = 1893 nonprote in kca ls
t
a
b
20% of 1893 ca lorie s = 380 fa t ca lorie s
o
l
i
Round to ne a re s t 100 = 400 kca ls lipid/d
c
S
2 kca ls /mL of 20% lipid s olution = 200 mL of 20% lipid
u
p
p
3. Ca rbohydra te (CHO) conte nt (a p p roxima te ly 50%-60% of nonp rote in c a lorie s )
o
r
t
2373 tota l ca lorie s – 480 prote in ca lorie s – 400 fa t ca lorie s = 1493 CHO kca ls
1493 CHO kca ls ÷ 3.4 kca ls /g of de xtros e = 439 g de xtros e
439 g de xtros e ÷ 2.4 L = 183 g de xtros e /L
183 g de xtros e = 18% de xtros e pe r lite r
4. Fina l s olution*
2.4 lite rs of 18% de xtros e , 5% a mino a cids , with 400 ca lorie s (40 g) lipid/da y.
*This a s s ume s a triple mix (ie , mix of ca rbohydra te , lipid, a nd a mino a cids in one ba g).
S ome ins titutions will ha ng lipid s e pa ra te ly. Othe r ins titutions do not a llow for cus tomize d
s olutions a nd ha ve a limite d s e le ction of s olutions from which the clinicia n choos e s to ge t
clos e to the ca lorie a nd prote in ne e ds of the pa tie nt.
Figure 58 5 Parenteral nutrition calculation.
receive 1.2 to 1.5 g/kg/d o protein, even i this precipitates the predicts adverse outcomes. Buzby and colleagues developed a
need or dialysis. measure o perioperative complication risk, which they termed
“nutrition risk index” (NRI) and is calculated as ollows:
■ HEPATIC DISEASE
NRI: 1.519 × the serum albumin level (in g/L) + 0.417
Malnutrition is also highly prevalent among patients with liver disease.
× (current weight/usual weight × 100)
Laboratory markers o nutrition status may be more re ective o
diminished hepatic synthetic capacity rather than nutrition. In addi- Patients were classi ed as borderline, mildly, or severely malnour-
tion, anthropometric measurements maybe altered by uid status. ished. The patients who met criteria or severe risk with an NRI <83.5
Previous thoughts regarding the bene ts o protein restriction in bene ted rom preoperative eeding with TPN because they had ewer
those with hepatic encephalopathy have not been supported by nonin ectious surgical complications (wound dehiscence, pressure
more recent studies. For example, Cordoba and colleagues ound ulcers, GI bleeding, cardiac arrest, renal ailure) compared to controls.
no dif erences in outcomes o encephalopathy between those on Mildly malnourished patients did not experience the same bene t;
normal and protein restricted diets, and that low protein diets in act rather they suf ered additional in ectious complications related to TPN.
caused breakdown o lean body mass. Although much research has Hence, one can conclude that the malnourished patient may bene t
been devoted to the evaluation o diets supplemented with branch rom preoperative eeding, while the well-nourished individual may be
chain amino acids (BCAAs), the routine use o BCAA-enhanced or- adversely af ected by this type o nutritional intervention.
mulas is only recommended or those with encephalopathy re rac- Less is known about orced enteral eeding (eg, tube eeds) in a
tory to luminal-acting antibiotics and lactulose therapy. preoperative patient. EN may be considered in patients who have
some degree o malnutrition but do not meet NRI criteria or pre-
■ PREOPERATIVE NUTRITION SUPPORT operative eeding. Nutrition support should be administered or a
Several papers indicate that malnourished patients suf er more post- minimum o 10 to 14 days prior to surgery to provide optimal bene t.
operative complications than those who are adequately nourished. Close consultation with the surgeon is required to determine both
Weight loss o greater than 10% o body weight most consistently the optimal route o eeding and the timing o the operation.
383
CONCLUSION McClave SA, Chang WK, Dhaliwal R, Heyland DK. Nutrition support
in acute pancreatitis: A systematic review o the literature. JPEN J
Malnutrition is prevalent in a signi cant number o hospitalized
Parenter Enteral Nutr. 2006;30:143-156.
patients. Because o the adverse ef ects o malnutrition on the host,
P
appropriate nutrition intervention may improve outcomes, decrease McClave SA, Martindale RG, Vanek VW, et al. Guidelines or the pro-
A
cost, and speed recovery. All patients should be screened or mal- vision and assessment o nutrition support therapy in the adult
R
critically ill patient: Society o Critical Care Medicine (SCCM) and
T
nutrition, have nutrition plans put in place to prevent deterioration
American Society or Parenteral and Enteral Nutrition (A.S.P.E.N.).
I
o nutritional status, and be treated or malnutrition when identi ed.
I
EN is the eeding route o choice, but parenteral nutrition is appro- JPEN J Parenter Enteral Nutr. 2009;33:277-316.
priate in those who cannot receive adequate nutrition by the enteral Mogensen K, Andrew B, Corona J, Robinson M. Validation o the
route or prolonged periods. Society o Critical Care Medicine and American Society or Parenteral
M
and Enteral Nutrition Recommendations or caloric provision to
e
SUGGESTED READINGS critically ill obese patients. JPEN J Parenter Enter Nutr. 2015 [epub
d
i
Apr 20, 2015].
c
a
Bankhead R, Boullata J, Brantley S, et al. Enteral nutrition practice
l
Perioperative total parenteral nutrition in surgical patients. The
C
recommendations. JPEN J Parenter Enteral Nutr. 2009;33:122-167.
o
Veterans Af airs Total Parenteral Nutrition Cooperative Study
n
Choban P, Dickerson R, Malone A, et al. A.S.P.E.N. Clinical guidelines: Group. N Engl J Med. 1991;325:525-532.
s
u
nutrition support o hospitalized adult patients with obesity. JPEN
l
Pittiruti M, Hamilton H, Bi R, MacFie J, Pertkiewicz M. A.S.P.E.N.
t
a
J Parenter Enteral Nutr. 2013;37:714-744.
t
Guidelines on parenteral nutrition: Central venous catheters
i
o
(access, care, diagnosis and therapy o complications). Clin Nutr.
n
Clinical practice guidelines or nutrition in chronic renal ailure. K/DOQI,
National Kidney Foundation. Am J Kidney Dis. 2000;35:S1-S140. 2009;28:365-377.
Correia MI, Waitzberg DL. The impact o malnutrition on morbidity, White J, Jensen G, Scho eld M, et al. Consensus statement: Academy
mortality, length o hospital stay and costs evaluated through a o Nutrition and Dietetics and American Society or Parenteral and
multivariate model analysis. Clin Nutr. 2003;22:235-239. Enteral Nutrition: characteristics recommended or the identi ca-
Gramlich L, Kichian K, Pinilla J, Rodych NJ, Dhaliwal R, Heyland DK. tion and documentation o adult malnutrition (undernutrition).
Does enteral nutrition compared to parenteral nutrition result in JPEN J Parenter Enteral Nutr. 2012;36:275-283.
better outcomes in critically ill adult patients? A systematic review
o the literature. Nutrition. 2004;20:843-848.
384
59
Over the past two decades, much had been learned about periop-
erative cardiac issues and the resultant short- and long-term com-
plications. This chapter addresses their demographics, risk actors,
and management. It is divided into sections o the most commonly
encountered postoperative cardiac complications; postoperative
myocardial in arction, congestive heart ailure, atrial brillation, and
ventricular arrhythmias.
Cardiac POSTOPERATIVE MYOCARDIAL INFARCTION

■ BACKGROUND
Complications a ter Postoperative myocardial in arction (PMI) is a distinct clinical entity
rom nonpostoperative myocardial in arction (MI). Compared with
Noncardiac Surgery patients presenting to emergency departments with MI, patients
with PMI die nearly twice as o ten. The incidence o PMI is depen-
dent upon patient risk actors and the type o surgery and the
technique used. Reported rates o PMI and associated mortality
Brian D. Wol e, MD have varied widely in the literature because o di erences in patient
selection, changes over time in surgical and anesthetic techniques,
Jef rey J. Glasheen, MD di erences in screening strategies and de nitions o PMI, and the
ever-increasing sensitivity o biomarkers or myocardial necrosis.
Nonetheless, PMI remains a common and devastatingly morbid
postoperative complication.
■ PATHOPHYSIOLOGY
There are three major conditions associated with the rise o cardiac
biomarkers in the perioperative period. Two o these are de ned by
Third Universal De nition o Myocardial In arction Consensus, pub-
lished in 2012, which describes Type I PMI as in arction secondary to
plaque rupture and intracoronary thrombus ormation and Type II
PMI as an in arction secondary to a mismatch between the supply o
oxygen and the metabolic demands o the myocardium. The third
cause o biomarker rise has generated several names, most recently
described as nonischemic myocardial injury with necrosis.
Type 1 PMI
Traditional MIs result rom an acute coronary syndrome (ACS).
More to the point, a vulnerable plaque experiences a spontaneous
rupture, ssuring, erosion, or other event leading to intracoronary
thrombus and downstream in arction. In the perioperative period,
a number o contributing mechanisms exist including an increase
in in ammation with an associated hypercoagulable state and an
increase catecholamine secretion leading to increase shear stress
on preexistent coronary plaques. The end result is intracoronary
plaque rupture, which initiates the coagulation cascade, leading to
thrombus and in arction. Electrocardiographically there may be ST
segment elevations, Q waves, or less speci c ST segment depres-
sions or T-wave changes.
Type 2 PMI
Type 2 PMI is associated with myocardial oxygen supply and
demand imbalances and occurs in operative and nonoperative set-
tings. The main driver or type 2 PMI is tachycardia secondary to ac-
tors such as increased adrenergic tone, postoperative pain, systemic
vasodilation, hypovolemia, anemia, and the withholding o chronic
β-receptor blocking medications. Patients with xed coronary
disease and increased le t ventricular mass possess increased risk
385
or this mechanism o injury. Electrocardiographically this presents troponin rises. This issue is urther complicated by the advent o
with ST-depression, nonspeci c T-wave changes or even a normal ever more sensitive troponin assays, which certainly increases the
electrocardiogram (ECG) much more commonly than ST-segment number o patients meeting the PMI de nition and others with
P
elevation. asymptomatic troponin elevations. While it is tempting to disregard
A
these “troponin leak syndromes,” even mild elevations in these ultra-
R
Nonischemic myocardial injury with necrosis sensitive assays are associated with poor outcomes as shown in the
T
Vascular Events in Noncardiac Surgery Patients Cohort Evaluation
I
This clinical entity is de ned by the presence o elevated cardiac-
I
(VISION) study.
speci c biomarkers without evidence o ischemia. It occurs in a
variety o clinical syndromes including sepsis, kidney ailure, heart
ailure and other conditions. One proposed mechanism o this myo- PRACTICE POINT
M
necrosis is rom circulating in ammatory mediators such as tumor
• Postoperative MI usually occur within 48 hours o surgery but a
e
necrosis actor.
d
small percentage may occur as late as 2 weeks postoperatively.
i
c
•
a
30-day mortality o PMI is higher than that o patients with
■ DIAGNOSIS AND PROGNOSIS
l
C
MI without recent surgery, but varies widely dependent on
o
Perioperative MI can be dif cult to diagnose as the key symptom o whether screening strategy is employed.
n
s
chest pain is o ten masked at least partially by anesthesia, analgesia, •
u
The underlying pathology determines optimal management.
l
and sedation. Several studies agree that approximately two-thirds
t
•
a
Type 1 PMI with true ST-elevation MI is generally treated as a
t
o patients with PMI have no ischemic symptoms. Importantly, in
i
traditional ACS. Risks o management (increased surgical site
o
one study mortality was unchanged by the presence or absence
n
bleeding) and bene ts (o restoring coronary blood ow) need
o ischemic symptoms. Beyond symptoms, the role or ECG in the
to be care ully weighed with the surgical team prior to initiation
diagnosis o PMI remains unclear due to its low sensitivity, especially
o therapy.
in the more common Type 2 PMI described above. In addition, the
association between ECG ndings with important outcomes has • Type 2 PMI is best treated by restoring the oxygen supply-
demand balance. Optimizing hemodynamics is the primary
recently been questioned.
goal in these patients.
As a result, PMI rates and prognosis depends heavily on how the
disease is de ned and how aggressively patients are screened or • Recognition and treatment includes management o
this entity. In most studies, PMI is de ned as evidence o myocardial arrhythmias, volume resuscitation, and trans usion.
necrosis plus symptoms consistent with in arction, ECG changes or
hemodynamic alterations or complications.
In the PeriOperative ISchemic Evaluation (POISE) trial cohort, ■ MANAGEMENT
postoperative biomarker measurements and ECGs were obtained The management o PMI should re ect its presumed pathophysiology;
in all patients as a screening test, and ound an incidence o PMI o however, the di erentiation between these entities is challenging
5% with an attendant 30-day mortality o 12%. While in the large and is not always possible. Type 1 PMI with true ST-elevation MI is
National Surgical Quality Improvement Program (NSQIP) cohort, relatively uncommon; when it occurs it necessitates a traditional
there were no speci c screening recommendations and only 0.65% approach to ACS with the goal o restoring ow in an acutely throm-
o the patients were ound to have a PMI, likely due to these physi- bosed artery, similar to the nonpostoperative setting. This manage-
cians ordering testing in only symptomatic patients; however, their ment is complicated by increased risks o surgical site bleeding but
30-day mortality rate was 61%. Interestingly, though these cohorts generally involves anticoagulation and antiplatelet therapy along
were di erent in size and in inclusion/exclusion criteria, relatively with coronary revascularization. These risks and bene ts need to
similar amounts o absolute patients (six per one thousand patients be care ully weighed with the surgical team prior to initiation o
vs our per one thousand patients) died at 30 days with the diagno- therapy. β-blockers and statin medications are also indicated.
sis o PMI in POISE versus NSQIP patient cohorts, respectively. This Type 2 PMI is best treated by relieving the oxygen supply-
begs the question o whether categorizing nearly eight times as demand imbalance. This includes reducing the adrenergic drive
many people with an aggressive screening strategy as having a PMI and associated tachycardia through correction o hypo- or hypervol-
was bene cial. emia, treatment o heart ailure, aggressive pain control, and treat-
Di erentiating between the previously described pathophysi- ment o ever. Other traditional therapies such as aspirin, β-blockers,
ological entities can be quite challenging, though this is a critical and statins require consideration on a case by case basis. In the
distinction as it heavily impacts complex treatment decisions. In POISE cohort, patients with PMI that were given aspirin had better
general, Type 1 PMI has the highest elevation in troponin values and outcomes than those not given antiplatelets. However, in the POISE
is most clearly characterized by ST-segment elevation. It typically 2 trial when perioperative patients were randomized to aspirin or
presents later in the postoperative course than Type 2 PMI. Mul- placebo, there was no di erence in mortality or other important
tiple studies show that most Type 2 PMI events occur in the rst endpoints associated with PMI. Thus, the ollowing recommenda-
24 to 48 hours postoperatively while Type 1 PMI events are more tions or Type 2 PMI are consistent with expert guidelines but lack
equally distributed over the rst two postoperative weeks. Type 2 trial data support.
PMI mismatch events are elt to occur in at least our out o ve Treatment o nonischemic myocardial injury with necrosis is
perioperative MI cases; however, there are two small postoperative ocused not upon the cardiac issue, but instead upon the sepsis,
autopsy trials completed over the last 30 years that would indicate renal ailure or other disease to which the necrosis was attributed. As
Type 1 PMI-indicative intracoronary thrombosis occurs in nearly 50% mentioned earlier, the di erentiation between these entities is di -
o PMI patients. These data highlight our incomplete understanding cult at times. Moreover, there are no interventional trials to support
o what pathophysiological mechanism predominates during the a speci c treatment regimen. Postoperative hypotension and asso-
perioperative period. ciated tachycardia due to a combination o volume depletion and
Di erentiating Type 2 PMI rom nonischemic myocardial injury anesthetic e ect are commonly associated with demand ischemia
with necrosis is particularly di icult as both can be associated and Type 2 PMI. This can be treated with isotonic uids. Hyperten-
with tachycardia, ever, changes in blood pressures and modest sion-induced Type 2 PMI should be treated with antihypertensives
386
and diuretics as needed. Volume overload is an underrecognized Trial do not include postoperative HF in their combined cardiac end-
cause o hypertension and thereby increased myocardial demand point. Nonetheless, it is an important postoperative complication
C
and should be addressed aggressively. Primary tachyarrhythmias that leads to increased morbidity, mortality, length o stay, and cost.
H
such as atrial brillation with rapid ventricular response should be
A
P
rate controlled or cardioverted as the situation dictates. Initiation o ■ DEFINITION AND DIAGNOSIS
T
β-blockers in the setting o Type 2 PMI is thought to be appropriate
E
when demand ischemia is due to hypertension and nonhypovole- There is no ormal de nition o postoperative HF. However, in trials
R
mic tachycardic syndromes (eg, ever). However, β-blockers should that have included HF as a major adverse cardiac event, it is de ned
5
be held in the setting o hypotension and anemia-induced ischemia as pulmonary edema. This is a reasonable de nition because pul-
9
as they will likely worsen these processes and clinical outcomes. monary edema, with its resulting dyspnea and hypoxemia, is the
Postoperative anemia management to prevent and treat PMI is most common presenting syndrome. That said, all postoperative
C
controversial as both postoperative anemia and liberal trans usion pulmonary edema is not HF. Noncardiac causes o airspace disease,
a
r
including aspiration, acute respiratory distress syndrome (ARDS),
d
have been shown to worsen outcomes. For example, a retrospec-
i
and pneumonia should also be considered. Once airspace disease
a
tive study ound preoperative hemoglobin (hgb) <13 g/dL were
c
has been identi ed, the evaluation should ocus on volume status,
C
associated with increased cardiac complications and mortality.
o
The Functional Outcomes in Cardiovascular Patients Undergoing including the amount o uid in used in the perioperative period,
m
Surgical Hip Fracture Repair (FOCUS) trial evaluated trans usion examination o the neck veins or distension, and nonpulmonary
p
edema, particularly in the sacrum or patients who have been bed-
l
i
cuto s o 10 g/dL versus 8 g/dL or symptomatic. The trial clearly
c
a
showed that the lower cuto was superior and not associated with ridden. A patient with pulmonary edema and such signs o volume
t
i
overload likely has postoperative HF.
o
provoking higher rates o myocardial ischemia; however, in that trial
n
Risk actors or postoperative HF include preexisting cardiac
s
nearly one-hal o the trans usions in the restrictive strategy group
a
were given or hypotension, tachycardia and chest pain. Until more disease, particularly recent MI or unstable angina, diabetes, signi -
t
cant intraoperative hemodynamic changes (mean arterial pressure
e
evidence is orthcoming using a trans usion threshold o 8 g/dL
r
increase or decrease >40 mm Hg rom preoperative baseline), and
N
appears reasonable with close consideration o volume status and
o
other hemodynamic indicators. abdominal aortic aneurysm repair. A case series o atal postop-
n
c
erative pulmonary edema in largely healthy patients ound that the
a
r
■ SURVEILLANCE average positive uid balance was 67 mL/kg/d positive, or about
d
i
5 liters in a 70-kg patient. Interestingly, another prospective evalua-
a
c
There are ew data to guide the use o perioperative surveillance tion o the risk o development o postoperative HF ound that lower
S
or PMI. Due to the relatively asymptomatic nature o PMI, screen-
u
volumes o administered uid and negative net uid balance were
r
ing with troponins and ECG greatly enhances the detection o
g
associated with higher rates o postoperative HF. These ndings
e
PMI. Given the predominance and accepted pathophysiology o
r
may be explained by the lower use o intravenous uids in patients
y
Type 2 PMI, it ollows that early intervention to improve supply and with a history o cardiac disease. Certainly, the amount o uids
decrease demand could alter outcomes; however, there are no data given in the perioperative period is an important historical act, and
to support this interventional approach. Moreover, there are con- despite these counterintuitive ndings, it is reasonable to consider
cerns about the potential negative impacts o some interventions more intraoperative IV uids to be a risk actor or postoperative
such as antiplatelet or anticoagulant medication administration pulmonary edema.
and revascularization procedures. The ACC/AHA guidelines have
evolved over time now calling into question any surveillance ECG
■ EVALUATION AND MANAGEMENT
or troponin strategy in asymptomatic patients. Other authors and
societies recommend the screening o high-risk patients, as de ned Once it has been determined that a patient has HF, initial evalu-
by severe underlying coronary disease or active cardiac symptoms. ation should ocus on early identi cation o myocardial ischemia.
All agree that ischemic symptoms should prompt ECG and troponin While all postoperative HF is not ischemic in nature, postoperative
assessment, but as outlined previously this represents the minority MIs o ten have atypical presentations requiring a high degree o
o patients. Hope ully, interventional trials will be orthcoming to aid suspicion. There ore, electrocardiography, cardiac monitoring, and
in the management o these complex issues. serial cardiac enzymes are prudent in the workup o postoperative
HF. Interestingly, patients with ischemia-induced postoperative HF
have increased risk o subsequent cardiac events, whereas those
■ CONCLUSION with nonischemic postoperative HF do not. Patients with ischemia-
Postoperative MI is common, driven by the increasing requency o induced postoperative HF should be managed according to ACC/
surgical procedures in an aging, comorbid population. Despite pru- AHA guidelines (re er to the section on PMI).
dent preoperative assessment, PMI still occurs and is o ten masked Postoperative HF typically occurs within the rst 36 hours a ter
by the operative state. PMI more o ten occurs rom oxygen supply- surgery. Pulmonary edema occurring immediately a ter extubation,
demand imbalances than ACS and as such is most o ten treated particularly in the postanesthesia care unit, could be a result o
with a return to homeostasis. “negative pressure pulmonary edema.” This is a poorly characterized
clinical syndrome associated with postextubation laryngospasm.
This leads to increased use o accessory muscles and greater nega-
POSTOPERATIVE HEART FAILURE
tive intrathoracic pressures, which in turn causes pulmonary edema.
■ BACKGROUND The classic patient is a young healthy male undergoing surgery o
Heart ailure (HF) ollowing noncardiac surgery is a common postop- the aerodigestive tract. Treatment is supportive, with diuretics, oxy-
erative cardiac complication. The incidence depends on the patient gen, continuous positive airway pressure (CPAP) and, occasionally,
population because it occurs a ter less than 5% o major surgeries reintubation. Although one series revealed that nearly hal o the
but as requently as 25% in patients with known cardiac disease. In patients with this entity required reintubation, the prognosis or a
a single-center series, the incidence o postoperative pulmonary complete recovery is excellent.
edema was 7.6% with a mortality rate o almost 12%. Many o the Nonischemic postoperative HF warrants urther evaluation. ACC/
landmark perioperative cardiac outcomes trials such as the POISE AHA guidelines recommend, at a minimum, a complete history
387
and physical as well as laboratory, ECG, chest radiography, and
transthoracic echocardiography (TTE). Consideration or noninvasive • For patients with a known or suspected accessory pathway or
or invasive coronary angiography should be reserved or patients heart ailure, amiodarone is the rst-line agent.
• For those patients who do not have resolution o their
P
presenting with ischemic HF. While these guidelines are not speci c
A
to the postoperative setting, it is reasonable to apply them to this arrhythmia, they will need basic cardiac evaluation according
R
scenario. to the American College o Cardiology/American Heart
T
Historical in ormation should include preoperative symptoms o Association (ACC/AHA) guidelines or atrial brillation.
I
•
I
heart ailure or angina, use o stimulant drugs and alcohol, as well Stroke prophylaxis with anticoagulation hinges on balancing
as chemotherapeutic agents that increase the risk o development the risk o stroke and risk reduction rom anticoagulation
o HF. Laboratory evaluation should include assessment o hepatic, against the risk and consequences o surgical bleeding.
thyroid, and renal unction as well as complete blood counts,
M
glycohemoglobin, and lipid testing. An exhaustive search or less
e
The risk actors or postoperative AF are similar to those ollow-
d
common causes o HF such as amyloidosis should be reserved or
i
c
those in which the clinical scenario is suggestive o such an etiology. ing cardiac surgery. Patient-speci c risk actors include advanced
a
l
Initial management o postoperative pulmonary edema consists age, male sex, premature atrial contractions on preoperative ECG,
C
valvular heart disease, American Society o Anesthesiologists class III
o
o diuresis and discontinuation o IV uids coupled with uid and
n
or IV, CHF, hypertension, and preoperative hypokalemia. Abdominal
s
sodium restriction. Supplemental oxygen can be used as needed. I
u
aortic aneurysm repair, thoracic surgery, and abdominal surgery all
l
le t ventricular systolic dys unction is discovered during the workup,
t
a
carry increased risk o postoperative AF in comparison to orthopedic
t
patients should be managed similarly to those in a nonoperative
i
o
setting. This includes initiation o an angiotensin-converting enzyme and other surgeries.
n
inhibitor, β-blocker medication among others. Statin medications Preoperative statin use prior to noncardiac thoracic surgery
are indicated in most o these patients irrespective o their lipid lev- has been shown to reduce the risk o postoperative AF, an e ect
els based on 10-year estimations o cardiac risk. Aspirin as secondary that has been attributed to the pleiotrophic or anti-in ammatory
prevention o ischemic events is indicated, although surgical hemo- e ect o statins. Perioperative β-blocker use is associated with an
stasis needs to be considered in the timing o antiplatelet initiation. increased risk o death in low-risk patients and an increased risk
In summary, HF is a relatively common and serious postoperative o death and stroke when started immediately preoperatively in
cardiac complication that is best de ned as cardiogenic pulmonary moderate- to high-risk patients and is there ore not recommended
edema. Risk actors are preexisting cardiac disease, diabetes, kidney solely or prophylaxis o postoperative AF (see Chapter 50: Preop-
disease and intraoperative hemodynamic changes. Large amounts erative Cardiac Risk Assessment and Perioperative Management).
o perioperative IV uids most likely present an additional risk actor, Amiodarone has been shown in some thoracic surgery studies to
although data are discordant on this issue, and “threshold” levels reduce postoperative atrial brillation rates, but it remains unclear
or the development o pulmonary edema are not well established. that this improves outcomes or is worth the risks that accompany
Ischemic and nonischemic HF should be distinguished because this amiodarone administration.
di erentiation has the largest impact on subsequent management
and prognosis. Ischemic HF should be managed per current societal ■ EVALUATION AND MANAGEMENT
guidelines, weighing the bene ts o anticoagulation and antiplate- The initial step in management o atrial arrhythmias is to veri y the
let agents against the risks o surgical site bleeding. The evaluation speci c diagnosis by ECG. AF is characterized by the absence o p
and management o nonischemic postoperative HF is similar to waves and an irregularly irregular R-R interval. Initial management
those encountered in other clinical settings and should ollow cur- depends on the clinical status o the patient. Hemodynamically
rent guidelines. unstable patients, characterized by hypotension or signs o poor
per usion such as loss o consciousness, need immediate direct
current electrical cardioversion. This is an uncommon clinical sce-
POSTOPERATIVE ATRIAL FIBRILLATION nario, and this management should be reserved or li e-threatening
■ INCIDENCE, RISK FACTORS, AND PREVENTION hemodynamic collapse.
Atrial brillation (AF) is the most common atrial arrhythmia ollowing While the need or cardioversion is rare, most patients will require
surgery. While it is a well-known complication o cardiac surgery, rate control. Initial rate control is best achieved via parenteral medica-
a ecting as many as 40% o coronary artery bypass procedures, it tions that slow conduction through the atrioventricular (AV) node.
is also common ollowing noncardiac surgery with an aggregate First-line agents include nondihydropyridine calcium channel block-
incidence o about 3%, though much higher a ter thoracic surgery. ers or β-blockers. For patients with a known or suspected accessory
pathway, AV blocking agents can lead to aster ventricular response
and are thus contraindicated. In these cases, amiodarone is the rst-
line agent. Likewise or patients with concomitant heart ailure, amio-
PRACTICE POINT
darone is the pre erred agent. While postoperative AF may convert
Peak incidence o postoperative atrial brillation occurs on spontaneously or ollowing initial rate control, patients with AF that
postoperative day 2 and it typically lasts 1 to 4 days. persists beyond a ew minutes will likely need long-acting oral medi-
• In 20% to 30% o cases it resolves without speci c intervention cations. It is reasonable to use the oral equivalent o the agent that
and less than 20% o patients will remain in AF at the time o was success ul with the initial rate control. It is important to remember
discharge. that AF is o ten a mani estation o an underlying perturbation such
• While the need or cardioversion is rare, most patients will as in ection or venous thromboembolism (VTE), and these diagnoses
require rate control. Initial rate control is best achieved via should be considered be ore escalating AVnodal blockade.
parenteral medications that slow conduction through the Anticoagulation is challenging in postsurgical patients with new
atrioventricular (AV) node. onset AF where the risk o stroke and risk reduction rom anticoagu-
• First-line agents include nondihydropyridine calcium channel lation must be balanced against the risk and consequences o surgi-
blockers or β-blockers. cal bleeding. In patients with atrial brillation it is recommended
to use the CHADS2 or CHA2DS2-VASc risk scoring systems to decide
388
upon the long-term risk o thromboembolism and whether antico- general population. It typically occurs early in the postoperative
agulation would be bene cial. There are no speci c guidelines or course. Preoperative statin therapy is associated with a lower risk
C
the timing or method o starting a new anticoagulant in the postop- o developing AF in some surgical populations, but there are insu -
H
erative setting. Many experts would use war arin without a heparin/
A
cient data to recommend this practice routinely, particularly in
P
low-molecular-weight heparin run-in phase to avoid raising the lower-risk patients. β-blockers are not recommended solely or
T
risk or postoperative bleeding. Others would recommend heparin prophylaxis o postoperative AF. Initial management o postopera-
E
or low-molecular-weight heparin to prevent the exceedingly rare tive AF commences with veri cation o the diagnosis with an ECG.
R
skin necrosis that can occur in patients with protein C de ciency. Hemodynamically unstable patients require emergent cardioversion.
5
In either case, the surgical site, risk o bleeding and other co- Patients with rapid ventricular responses require IVadministration o
9
morbidities must be considered. Most protocols would recommend nodal blocking agents ollowed by an oral equivalent. The decision
starting war arin 24 to 72 hours a ter hemostasis has been achieved; to anticoagulate must balance stroke risk in the acute setting with
C
however, there are no randomized data to support this approach. the risk o operative bleeding. Patients with persistent AF will need
a
r
d
More than 80% o new postoperative AF will resolve by discharge. adequate rate control and, should they remain symptomatic, consid-
i
a
Patients who do not have resolution o their arrhythmia will need eration o cardioversion and subsequent rhythm control. A CHADS2
c
C
basic cardiac evaluation. The American College o Cardiology/ score o 2 or greater is an indication or chronic anticoagulation in
o
American Heart Association (ACC/AHA) guidelines or AF list the most patients with persistent AF at the time o discharge.
m
minimum initial evaluation to include history and physical, ECG,
p
l
i
Transthoracic Echocardiogram (TTE), and laboratory tests o thyroid,
c
POSTOPERATIVE VENTRICULAR ARRHYTHMIAS
a
renal, and hepatic unction. The goal o this evaluation is to delin-
t
Ventricular arrhythmias are uncommon ollowing noncardiac sur-
i
o
eate whether this is paroxysmal AF has been present in the past or
n
gery. Because o this, data are sparse regarding risk actors and
s
is likely to recur in the uture. The TTE is help ul or determining le t
outcomes. In the thoracic surgery population, one study ound the
a
ventricular unction, atrial size, and valvular unction— actors that
t
incidence o nonsustained ventricular tachycardia (NSVT) to be 15%,
e
will help guide decisions regarding stroke prophylaxis.
r
but there were no episodes o sustained ventricular arrhythmias, nor
N
I a patient’s symptoms resolve with rate control, the employed
o
did any o these episodes lead to hemodynamic instability. Impor-
strategy should be continued. In this approach, no attempt is made
n
tantly, nonischemic ventricular tachycardia (VT) ollowing noncar-
c
at restoration o sinus rhythm, and the patient’s ventricular rate is
a
diac surgery is not associated with worse long-term outcomes.
r
d
controlled with AVnodal blockade alone. Rhythm control strategy is
The most common scenario described in the literature or
i
a
aimed at restoration o sinus rhythm and can be accomplished via
c
serious ventricular arrhythmias is ollowing postoperative MI in
chemical or electrical cardioversion and subsequent antiarrhythmic
S
patients with underlying heart disease. Fatal ventricular arrhythmias
u
medications. The rhythm control strategy mandates anticoagulation
r
g
are included in the outcomes o major trials such as CARP and
such that patients whose surgery precludes therapeutic anticoagu-
e
DECREASE-V, but the contribution o ventricular arrhythmias to the
r
y
lation are not candidates or cardioversion. Despite the theoretical
endpoint is not reported.
bene t o rhythm control, it is not associated with lower rates o
When con ronted with a ventricular arrhythmia, the rst decision
stroke and death, and there are inconsistent reports o improved
point is to ascertain the patient’s hemodynamic status. Unstable
quality o li e. (See Chapter 132 [Supraventricular Tachyarrythmias].)
patients, de ned as having hypotension or loss o consciousness,
need emergent treatment as described in Advanced Cardiac Li e
■ ASSOCIATED CONDITIONS Support (ACLS) protocols. Hemodynamically stable patients with
Postoperative AF may occur in isolation or be associated with ventricular arrhythmias need urgent evaluation or acute ischemia
underlying systemic conditions. More than hal o patients with AF including 12-lead ECG and serial cardiac enzymes. Should this
a ter noncardiac surgery have a major underlying condition, with workup reveal myocardial ischemia, treatment according to pub-
in ection having the strongest association. For example, patients lished guidelines should commence. Signi cant electrolyte abnor-
with postoperative AF have a relative risk o 7.4 or bacterial pneu- malities, notably hypokalemia and hypomagnesemia, are reversible
monia and 6.2 or sepsis compared to those without arrhythmia. For causes o ventricular arrhythmias and thus warrant prompt evalu-
patients in the surgical intensive care unit (ICU), the leading cause ation and repletion. In the setting o torsades de pointes, or poly-
o death in patients with new-onset AF is sepsis. In general, the inci- morphic ventricular tachycardia with prolonged QT interval, empiric
dence o sepsis in patients with new postoperative arrhythmias is magnesium should be given. As QT prolongation is o ten a result
20% to 30%, although this gure is primarily rom surgical ICU data. o medication use, the patient’s medication list should be reviewed
Primary cardiac events are less commonly antecedent to new and potentially o ending medications discontinued. Hemodynami-
AF. The relative risk o MI in patients with a new arrhythmia is 4.2. cally stable but sustained VT requires antiarrhythmic drugs.
In colorectal surgery patients, both pulmonary edema and overall In summary, postoperative ventricular arrhythmias are an uncom-
complications are more common in patients with new arrhythmias. mon cardiac complication, but with potentially serious conse-
Other postoperative complications associated with new-onset AF quences. Hemodynamically unstable patients will need emergent
include pulmonary embolism, gastrointestinal bleed, cerebrovascu- de brillation. Patients with ventricular arrhythmias need urgent
lar accident, hypokalemia, and anastomotic leakage. evaluation or cardiac ischemia. Fortunately, nonischemic VT does
Given the requency o serious underlying etiologies, new AF not appear to negatively impact long-term prognosis.
ollowing noncardiac surgery should be regarded as a possible
harbinger o li e-threatening postoperative complications. Thus, SUGGESTED READINGS
appropriate evaluation o new postoperative arrhythmia warrants a
complete evaluation o the patient with consideration o in ectious, Amar D. Prevention and management o perioperative arrhyth-
cardiac, and thrombotic complications. mias in the thoracic surgical population. Anesthesiology Clinics.
2008;26:325-335.
■ SUMMARY Amar D, Zhang H, Heerdt PM, et al. Statin use is associated with a
AF is the most common arrhythmia ollowing noncardiac surgery, reduction in atrial brillation a ter non-cardiac thoracic surgery
with patient-speci c risk actors similar to those or AF in the independent o C-reactive protein. Chest. 2005;128:3421-3427.
389
Botto F, et al. Myocardial injury a ter noncardiac surgery: a large, Fuster V, Rydén LE, Cannom DS, et al. ACC/AHA/ESC 2006 guidelines
international, prospective cohort study establishing diagnos- or the management o patients with atrial brillation. Circulation.
tic criteria, characteristics, predictors, and 30-day outcomes. 2006;114:700-752.
P
Anesthesiology. 2014;120:564-578. Go AS, Hylek EM, Chang Y, et al. Anticoagulation therapy or stroke
A
Carson JL, et al. Liberal or restrictive trans usion in high-risk patients prevention in atrial brillation. JAMA. 2003;290:2685-2692.
R
a ter hip surgery. N Engl J Med. 2011;365:2453-2462.
T
Gupta PK, et al. Development and validation o a risk calculator
I
Devereaux PJ, Yang H, Yusu S, et al. E ects o extended-release or prediction o cardiac risk a ter surgery. Circulation. 2011;124:
I
metoprolol succinate in patients undergoing non-cardiac 381-387.
surgery (POISE trial): a randomized controlled trial. Lancet. Landesberg G, Beattie WS, Mosseri M, Ja e AS, Alpert JS. Periopera-
2008;371:1839-1847. tive myocardial in arction. Circulation. 2009;119:2936-2944.
M
Fleisher LA, Fleischmann KE, et al. ACC/AHA 2014 guidelines on peri-
e
operative cardiovascular evaluation and management o patients
d
i
c
undergoing noncardiac surgery. JACC. 2014;64:e77-e137.
a
l
C
o
n
s
u
l
t
a
t
i
o
n
390
60
Hospitalists are requently called upon to provide perioperative care
to a broad spectrum o surgical patients, in either a consultative or
a comanagement role. Although historically much emphasis has
been placed on postoperative cardiac complications, postoperative
pulmonary complications are known to occur with equal or greater
requency and contribute substantially to morbidity, mortality, and
health care costs. Broadly de ned, postoperative pulmonary compli-
cations are conditions a ecting the respiratory tract that adversely
Management of in uence the clinical course o patients a ter surgery. The Con eder-
ate general, Thomas “Stonewall” Jackson, wounded in the Battle o
Postoperative Chancellorsville in 1863, was perhaps the earliest recorded victim

o a postoperative pulmonary complication, dying o pneumonia
Pulmonary 8 days a ter the success ul amputation o his le t arm. It is estimated

that over 1 million patients undergoing nonthoracic surgery in the
United States annually experience postoperative pulmonary compli-
Complications cations. Pulmonary complications produce the highest attributable
costs among common categories o postoperative complications
and can result in a ve old increase in the median cost o an opera-
tion. The presence o pulmonary complications a ter major surgery
William I. Levin, MD increased 30-day mortality rom 2% to 22%, and 1-year mortality
rom 8.7% to 45.9% based on data rom the National Surgical Quality
John J. Reilly, Jr., MD Improvement Program (NSQIP). The most important postoperative
pulmonary complications are atelectasis, pneumonia, respiratory ail-
ure, and exacerbation o underlying chronic lung disease, although
earlier studies have also included transient and sel -limited clinical
ndings. A general principle is that the closer the operative site is to
the diaphragm, the higher the likelihood o postoperative pulmo-
nary complications. Interventions to reduce the incidence o these
complications depend on the aggressive application o preventive
measures to high-risk patients. Obstructive sleep apnea in particular
has received greater recognition as a requently undiagnosed and
prevalent condition in surgical patients that increases pulmonary
risk. Programs such as the NSQIP allow institutions to track their
per ormance and engage in quality improvement in this area. Antici-
pation, early diagnosis and prompt e ective therapies orm the next
line o de ense in treating postsurgical complications once they
occur. Studies o hospital mortality associated with inpatient sur-
gery suggest the variation between institutions is explained more
by their ability to “rescue” patients rom complications when they
occur rather than di erences in incidence. This chapter ocuses on
the pathogenesis, early recognition, and evidence-based treatment
o common postoperative pulmonary complications.
PRACTICE POINT
• A general principle to predicting the risk o postoperative
pulmonary complications is that the closer the operative site is
to the diaphragm, the higher the likelihood o complications.
ATELECTASIS
Atelectasis, or reversible alveolar collapse, is a common periopera-
tive phenomenon and occurs in 90% o patients receiving general
anesthesia. Computed tomographic (CT) studies have demonstrated
collapse o 15% to 20% o the lung volume near the diaphragm.
Dr William Pasteur, a Swiss physician practicing in England in the
early part o the last century, wrote extensively on the postoperative
391
lung and noted, “when the true history o postoperative lung com- sign can be positive, with obliteration o adjacent boundaries.
plications comes to be written, active collapse o the lung rom de - Posteroanterior (PA) and lateral images o the chest are pre erred,
ciency o inspiratory power will be ound to occupy an important and the ability o plain radiographs to detect atelectasis in recum-
P
position among determining causes.” Most atelectasis appearing bent critically ill patients is less certain. CT is sensitive in detecting
A
during general anesthesia resolves within 24 hours a ter surgery in areas o collapse, and may also reveal other pathology. MRI and
R
normal subjects and is o little clinical signi cance. Atelectasis can bedside ultrasound can also be use ul.
T
persist or 2 days or longer a ter major surgery, including abdominal
I
I
and thoracic surgery, and is thought to represent the starting point ■ TREATMENT
in a cascade o events that leads to the more serious complications Treatment o postoperative atelectasis centers on lung expan-
o pneumonia and acute respiratory ailure. sion techniques, shi ting rom supine position when possible, and
adequate postoperative analgesia. The FRC has been identi ed as
M
■ PATHOPHYSIOLOGY the single most important postoperative lung volume parameter,
e
d
The ormation o perioperative atelectasis can be understood by and e orts to restore normal pulmonary mechanics are bene cial.
i
c
a
considering the e ect o surgery on normal respiratory mechanics A simple posture change rom supine to seated will increase FRC
l
C
as well as the mechanisms involved in alveolar collapse. The induc- by 0.5 to 1.0 liters. Standing and early ambulation are also help ul
o
tion o anesthesia alters the distribution and timing o neural drive when tolerated.
n
s
to the respiratory muscles, inter ering with coordination o activity. The goal o lung expansion maneuvers is to produce a large and
u
l
t
The supine position and use o positive pressure ventilation alter the sustained increase in transpulmonary pressure that distends the
a
t
distribution o ventilation and lead to hypoventilation o dependent lung and re-expands the collapsed lung units. Techniques include
i
o
n
areas. Surgical trauma can produce re ex inhibition o the phrenic incentive spirometry, deep breathing exercises, chest physical ther-
nerve rom stimulation o the viscera, mechanical disruption o the apy, intermittent positive-pressure breathing (IPPB), and continuous
intercostal or abdominal respiratory muscles, and voluntary limita- positive airway pressure (CPAP). A systematic review ound that or
tion o respiratory motion rom postoperative pain. The character- patients undergoing abdominal surgery, any type o lung expansion
istic postoperative mechanical abnormality is a restrictive pattern intervention improved outcome, with no one modality being supe-
with severely reduced inspiratory capacity, vital capacity (VC), and rior. Incentive spirometry was the least labor intensive. IPPB is the
unctional residual capacity (FRC), clinically demonstrated by rapid most costly and was associated with unacceptable abdominal dis-
shallow respirations. tension in a signi cant number o cases. Low-quality evidence sug-
Pulmonary atelectasis occurs by three mechanisms: compression gests that use o incentive spirometry does not improve outcome,
atelectasis, absorption (resorption) atelectasis, and loss o sur actant. but it has been included in preventative care bundles. Another
Compression atelectasis results when the transmural pressure dis- systematic review also ound evidence that use o CPAP in patients
tending the alveolus is reduced, allowing the alveolus to collapse. who underwent abdominal surgery led to lower rates o postopera-
During anesthesia, change in diaphragmatic unction and chest tive atelectasis and pneumonia.
geometry causes pressure rom the abdomen to be transmitted The e ect o di erent types o analgesia in decreasing postop-
into the thorax, resulting in compression o lung tissue. Resorption erative atelectasis has been examined. Studies have been heteroge-
atelectasis describes collapse o alveoli related to absorption o gas neous and small, but a recent meta-analysis ound a trend toward
rom occluded or hypoventilated areas o the lung. Since oxygen is decreased postoperative atelectasis and pneumonia with the use o
absorbed more rapidly than nitrogen, air with high inspired FiO2 will postoperative epidural analgesia in patients undergoing abdominal
be absorbed more rapidly, resulting in collapse. Sur actant unction, surgery. Postoperative epidural and patient-controlled intravenous
important in stabilizing the alveoli, may be disrupted by anesthesia analgesia both seem superior to on-demand delivery o opioids in
and mechanical ventilation. The physiologic consequence is preventing postoperative pulmonary complications. The potential
ventilation-per usion (V/Q) mismatch resulting in hypoxemia. bene t o epidural anesthesia must be weighed against bleeding
There is strong interest in re ning anesthetic technique to ide- risk rom deep vein thrombosis (DVT) prophylaxis.
ally deliver a patient with no atelectasis and open lungs to the post
anesthesia care unit. Key areas o investigation include modulation ■ COMPLICATIONS
o FiO2 to optimize oxygen delivery but minimize resorption atel- Mild hypoxemia rom atelectasis is usually well tolerated, but more
ectasis, use o recruitment maneuvers, PEEP, and CPAP during the severe hypoxemia can a ect end organs. Atelectrauma re ers to the
course o the patient’s perioperative care. mechanism by which atelectasis can trigger acute lung injury, and
includes overexpansion o adjacent aerated lung tissue along with
■ DIAGNOSIS—DOES THIS PATIENT HAVE repetitive sheer orces rom the opening and closing o alveoli that
ATELECTASIS? results in the release o local in ammatory cytokines. Le t untreated,
atelectasis likely predisposes to the development o pneumonia and
Atelectasis is recognized by the nding o persistent postoperative
respiratory ailure, including acute respiratory distress syndrome.
hypoxemia in the absence o other plausible diagnoses. The patient
demonstrates dyspnea or tachypnea, and physical ndings can
include basilar rales and decreased breath sounds in the a ected POSTOPERATIVE PNEUMONIA
area. Atelectasis is o ten cited as a cause o postoperative evers, Pneumonia ranks as the third most common postoperative in ec-
but studies have demonstrated no association between atelectasis tion behind urinary tract in ection (UTI) and wound in ection. The
and ever and suggest that early postoperative evers are more likely incidence o pneumonia ollowing major abdominal surgery ranges
due to the in ammatory response to surgery. Atelectasis is detected between 2% and 19% and is a principal actor in increased mortality.
radiographically by opaci cation o a lobe or lobar segment and evi- Development o hospital-acquired pneumonia is associated with a
dence o volume loss. The most reliable sign is displacement o the 30% to 50% increased risk o developing acute respiratory ailure
interlobar ssure, but other signs include elevation o the hemidia- requiring mechanical ventilation and increases hospital stays by an
phragm, mediastinal shi t, and compensatory overin ation o adja- average o 7 to 9 days at an excess cost o $40,000 per patient.
cent aerated segments. There may be linear opacities (“plate-like”) Postoperative pneumonia is a subset o hospital-acquired pneu-
in the parenchyma in dependent portions o the lungs. Silhouette monia (HAP), which is pneumonia occurring 48 hours or more a ter
392
TABLE 60-1 Risk Factors for Multidrug-Resistant Pathogens TABLE 60-2 Early-Onset Hospital-Acquired Pneumonia
C
without MDR Risk Factors
H
Antimicrobial therapy in the preceding 90 d
A
Current hospitalization o 5 d or more Potential Pathogen Recommended Antibiotic
P
Streptococcus pneumoniae Ce triaxone
T
High requency o antibiotic resistance in the community
E
Presence o risk actors or health-care-associated pneumonia Haemophilus in luenzae or
R
(HCAP) Methicillin-sensitive Levo loxacin, moxi loxacin, or
6
Staphylococcus aureus cipro loxacin
0
Hospitalization or 2 d or more in the preceding 90 d
Residence in a nursing home or extended care acility Antibiotic-sensitive enteric or
Gram-negative bacilli Ampicillin/sulbactam
M
Home in usion therapy
Escherichia coli
a
Chronic dialysis within 30 d or
n
Klebsiella pneumoniae
a
Home wound care Ertapenem
g
e
Enterobacter species
m
Family member with multidrug-resistant pathogen
Proteus species
e
Immunosuppressive disease and/or therapy
n
Serratia marcescens
t
o
P
o
s
t
o
admission and not incubating at the time o admission. The major ■ DIAGNOSIS—DOES THIS PATIENT HAVE
p
early management goal or postoperative pneumonia is to provide
e
POSTOPERATIVE PNEUMONIA?
r
a
appropriate antibiotics in adequate doses based on the best predic-
t
The goal o diagnosis is to identi y which patients have a pulmonary
i
tion o suspected pathogens and resistance pattern.
v
e
The timing o onset o HAP is an important epidemiologic vari- in ection so that antibiotics are not delayed but that patients with
P
nonin ectious etiologies are not exposed unnecessarily to the anti-
u
able. Early-onset HAP, less than 5 days into admission, is more
l
biotics. All patients should undergo a comprehensive history and
m
likely to be caused by antibiotic-sensitive bacteria unless other
o
physical exam, chest x-ray (pre erably PA and lateral), measurement
risk actors or multidrug-resistant (MDR) pathogens are present.
n
o arterial O2 saturation, complete blood count (CBC), electrolytes,
a
Late-onset HAP, 5 or more days a ter admission, is more likely to
r
y
liver unction tests, and blood cultures. All ventilated patients should
be associated with MDR pathogens. Additional risk actors or MDR
C
have lower respiratory cultures obtained, ideally prior to starting
o
pathogens are included in Table 60-1. Hospital- and unit-speci c
m
antibiotics, to guide de-escalation o therapy. Other processes that
microbiologic data are also very important in selecting appropriate
p
produce similar symptoms including congestive heart ailure, atel-
treatment.
l
i
c
ectasis, pulmonary thromboembolism, drug reactions, pulmonary
a
The 2005 American Thoracic Society/In ectious Diseases Society
t
hemorrhage, and acute respiratory distress syndrome (ARDS) need
i
o America (ATS/IDSA) guideline emphasizes ventilator-associated
o
to be considered.
n
pneumonia (VAP) because it is more readily studied, but suggests
s
The Centers or Disease Control and Prevention (CDC) criteria or
it is reasonable to extrapolate the conclusions regarding risk actors
diagnosis o nosocomial pneumonia in adults require radiologic as
or in ection with speci c pathogens to nonintubated, nonventi-
lated HAP patients.
■ PATHOPHYSIOLOGY
TABLE 60-3 Late-Onset Hospital-Acquired Pneumonia or HAP
The sequence o events in HAP begins with colonization o the with MDR Risk Factors
oropharynx with pathogens, which can occur within 48 hours o
admission. Sources o pathogens include contaminated health Recommended Antibiotic
care devices, the environment, and trans er rom other patients Potential Pathogens Combination
or sta . These pathogens must be aspirated rom the oropharynx Streptococcus pneumoniae Antipseudomonal
into the lower respiratory tract, and then overwhelm the natural Haemophilus in luenzae cephalosporin (ce epime,
host de ense mechanisms. Microaspiration is known to occur in up ce tazidime)
Methicillin-sensitive
to 45% o healthy subjects during sleep and can be worsened in Staphylococcus aureus or
postsurgical patients by decreased gag re ex, ine ective coughing, Antibiotic-sensitive enteric Antipseudomonal carbapenem
sedation, supine posture, especially during enteral eeds, and rou- Gram-negative bacilli (imipenem or meropenem)
tine (rather than selective) use o nasogastric (NG) tubes. The host or
Escherichia coli
de enses are also a ected in multiple ways by general anesthesia, β-lactam/β-lactamase inhibitor
including mechanical impairment o normal mucociliary transport Klebsiella pneumoniae
(piperacillin-tazobactam)
and inter erence with unction o alveolar in ammatory cells, includ- Enterobacter species
plus
ing polymorphonuclear leukocytes, macrophages, lymphocytes, Proteus species
cytokines, antibodies, and complement. Antipseudomonal
Serratia marcescens luoroquinolone (cipro loxacin
The microbiology o early-onset HAP without MDR risk actors Plus or levo loxacin)
tends to mirror community-acquired pneumonia and includes
Pseudomonas aeruginosa or
Streptococcus pneumoniae, Haemophilus in luenzae, methicillin-
sensitive Staphylococcus aureus, and antibiotic-sensitive Enterobac- Klebsiella pneumoniae (ESBL) Aminoglycoside (Amikacin,
teriaceae (Table 60-2). Pathogens in late-onset HAP or the presence Acinetobacter species gentamicin, or tobramycin)
o MDR risk actors also include methicillin-resistant S. aureus (MRSA), Methicillin-resistant plus
Pseudomonas aeruginosa, extended-spectrum β-lactamase (ESBL)- Staphylococcus aureus Linezolid or vancomycin
producing Klebsiella, and Acinetobacter baumannii (Table 60-3).
393
The patient should be reevaluated in 48 to 72 hours, and antibiot-
TABLE 60-4 Centers for Disease Control and Prevention ics narrowed or discontinued based on culture results and clinical
Criteria for Diagnosis of Nosocomial Pneumonia response. Therapy lasting 7 to 8 days has been shown to be equally
P
Radiology Signs/Symptoms/Laboratory e ective to longer courses in patients receiving an appropriate ini-
A
tial antibiotic regimen with a good clinical response. Pseudomonas
R
Two or more serial chest At least one o the ollowing:
and MRSA are the exception due to signi cant rates o recurrence
T
x-rays with at least one o Fever (>38°C) with no other
the ollowing: and should still be treated with a longer course o antibiotics.
I
source
I
New or progressive Leukopenia (<4000 WBC/µL) or ■ COMPLICATIONS
in iltrate leukocytosis (>12,000 WBC/µL)
Consolidation Patients who ail to respond or worsen should be reevaluated at
Mental status changes with no 48 to 72 hours. Possible reasons or lack o response include the
M
Cavitation other cause in adult >70-y old
presence o a complication (empyema, lung abscess, drug ever),
e
d
And at least two o the ollowing: an alternate site o in ection (Clostridium dif cile colitis, line-related
i
c
a
New onset o purulent sputum in ection), wrong diagnosis (atelectasis, pulmonary embolism (PE),
l
C
New-onset cough, dyspnea, ARDS, pulmonary hemorrhage, neoplasm), or wrong organism
o
tachycardia (clinically unrecognized immunosuppression).
n
s
u
Rales or bronchial breath
l
t
sounds (BS) RESPIRATORY FAILURE
a
t
i
Worsening gas exchange Acute respiratory ailure is de ned as the requirement or mechani-
o
n
cal ventilation longer than 48 hours postoperatively or unplanned
reintubation or cardiac or respiratory ailure. It can be considered
the most severe o the clinically signi cant postoperative pulmonary
complications based on its impact on morbidity, mortality, and cost.
well as clinical and laboratory ndings (Table 60-4). The ATS/IDSA Mortality data rom the Department o Veterans A airs National Sur-
guideline describes two diagnostic approaches to HAP: clinical and gical Quality Improvement Program showed an increase in 30-day
bacteriologic. The clinical approach bases the diagnosis on a new mortality rom 2.3% to 29.1% in patients with respiratory ailure, and
in ltrate on chest x-ray plus clinical evidence o in ection. The bac- an increase in 1-year mortality rom 9.3% to 55.9%. Another study
teriologic strategy is based on lower respiratory tract samples and is ound the rate o postoperative respiratory ailure in patients under-
more suited to ventilated patients where there is ready access to the going general and vascular surgery to be 3%, with 30-day mortality
lower respiratory tract. The presence o a new or progressive radio- o 26.5%. Pulmonary complications raised median hospital costs to
graphic in ltrate plus two o three clinical eatures ( ever, leukopenia $62,704 compared to $5015 when the complications were absent,
or leukocytosis, and purulent secretions) yields a sensitivity o 69% more expensive than thromboembolic, cardiovascular, or in ectious
with speci city o 75% and represents the most accurate clinical complications.
criteria or starting antibiotics.
■ PATHOPHYSIOLOGY
■ TREATMENT Postoperative acute respiratory ailure results rom the onset over
Treatment strategies seek to balance the need to provide early, minutes to hours o impaired pulmonary gas exchange severe
appropriate empiric antibiotic therapy with avoidance o exces- enough to cause organ dys unction or to threaten li e. Respiratory
sive antibiotic exposure in both spectrum and duration. Delay in ailure can be broadly categorized as hypoxemic respiratory ailure
initiating appropriate antibiotics in VAP patients with severe sepsis (respiratory insuf ciency) or hypercapnic respiratory ailure (ventila-
has been shown to increase mortality, and initial inadequate drug tory ailure), and the two orms may coexist. Mechanisms underly-
selection, even when later adjusted, is also associated with worse ing hypoxemic respiratory ailure, de ned as an arterial pO2 o less
outcome. Aggressive early therapy combined with de-escalation o than 60 mm Hg, include decreased FiO2, hypoventilation, impaired
initial antibiotics based on clinical or microbiologic data is encour- di usion, V/Q mismatch, and right-to-le t shunt. Most hypoxemic
aged. Empiric drug selection takes into account the epidemiologic hospitalized patients have some combination o V/Q mismatch and
timing and presence o MDR risk actors, previous antibiotic expo- right-to-le t shunt. A structural-anatomic classi cation that localizes
sure, and the institution-/unit-speci c antibiogram. the primary pathology to the alveoli, interstitium, cardiopulmonary
For early-onset HAP without MDR risk actors, single-agent ther- vasculature, airways, or pleura may be help ul when trying to make
apy is reasonable (Table 60-2). For early-onset HAP with MDRrisk ac- a speci c diagnosis. In the postoperative patient, the alveoli and
tors or late-onset HAP, a three-drug regimen is recommended (see interstitium can be a ected by pulmonary edema, acute lung injury/
Table 60-3). Considerations in antibiotic selection might also include ARDS, atelectasis, and pneumonia. The vasculature can be a ected
pharmacodynamic properties and mechanism o action. Fluoro- by pulmonary embolism or develop pulmonary hypertension due
quinolones and linezolid achieve high concentration in bronchial to hypoxic vasoconstriction and/or elevated le t atrial pressures.
secretions. Aminoglycosides and uoroquinolones are bactericidal The airways can be a ected by exacerbations o COPD, asthma, and
in a concentration-dependent ashion, whereas vancomycin and mucous plugging, and the pleura may be a ected by pneumotho-
β-lactams are bactericidal in a time-dependent ashion. Aminogly- rax or pleural e usion.
cosides and the quinolones also have a postantibiotic e ect and Hypercapnic respiratory ailure, characterized by a pCO2 greater
suppress antibiotic growth even a ter concentrations all. A num- than 45 mm Hg and respiratory acidosis, can be classi ed as drive
ber o new antibiotics have become available or treating hospital ailure or pump ailure. Drive ailure results when the patient’s ven-
acquired pneumonia, including doripenem, telvancin, ce tobiprole, tilatory e ort is insuf cient and can be caused by drug overdoses,
and avibactam. The most appropriate use o these alternatives general anesthesia, central nervous system (CNS) disease, and
remains to be established. Aerosolized delivery o antibiotics is obesity hypoventilation syndrome. The most common contribu-
another mode o treatment that may have bene t or patients who tors in the perioperative setting are residual sedation rom general
have ailed IVtherapy or have MDR organisms. anesthesia or the e ects o opioid analgesics on respiratory drive
394
and level o consciousness. Pump ailure results when ventilatory
demand exceeds the patient’s capability and can be caused by TABLE 60-5 Noninvasive Ventilation—Indications and
C
prolonged e ect o neuromuscular blocking agents, underlying Contraindications
H
neuromuscular disorders, electrolyte abnormalities and metabolic
A
Indications Contraindications
P
disturbances, pleural disorders, chest wall abnormalities, and respi-
Clinical observations Absolute
T
ratory muscle atigue. It can be aggravated by increased CO2 pro-
E
duction in the setting o a hypermetabolic postoperative state; this • Moderate to severe dyspnea • Respiratory arrest
R
is especially true in patients with underlying loss o parenchyma • Tachypnea (>24 or • Unable to it mask
6
(emphysema) who have a decreased alveolar sur ace area available hypercapnic, >30 or Relative
0
or gas exchange. hypoxemic)
• Medically unstable
• Accessory muscle use or
• Unable to protect airway
M
■ DIAGNOSIS—WHAT IS CAUSING THIS PATIENT’S abdominal paradox
a
• Excessive secretions
n
RESPIRATORY FAILURE? Gas exchange
a
• Agitated, uncooperative
g
Diagnostic workup should begin with the ABCs—airway, breathing, • Acute ventilatory ailure:
e
• Recent upper gastrointestinal
m
and circulation—and treatment initiated concurrently. Supplemen- PaCO2 > 45 mm Hg, pH < 7.35
or airway surgery
e
tal oxygen should be provided and intravenous (IV) access obtained, • Hypoxemia: PaO2/FiO2 < 200
n
t
as well as cardiac monitoring and pulse oximetry. A ocused history • Multiple-organ ailure
o
and physical exam will yield clues to the presence or absence o
P
underlying cardiac and pulmonary disease. The hypoxemic respira-
o
s
tory ailure patient will appear tachypneic and tachycardic, perhaps
t
o
supports the use o NIV to treat COPD exacerbations, cardiogenic
p
with central cyanosis. The hypercapnic respiratory ailure with
e
decreased ventilatory drive will appear hypopnic or apneic, in no pulmonary edema, and to acilitate weaning rom the ventilator in
r
a
patients with COPD. Level 2 evidence supports its use in postopera-
t
respiratory distress. The ventilatory pump ailure patient will appear
i
v
tive respiratory ailure, community-acquired pneumonia with COPD,
e
in respiratory distress with rapid shallow ine ective respirations. All
P
patients should receive a chest x-ray, electrocardiogram (ECG), and asthma, extubation ailure, and in do-not-intubate-status patients
u
(Table 60-5). Noninvasive ventilation (NIV) should be initiated early
l
routine bloodwork including CBC and serum chemistries. An arterial
m
when indicated to take advantage o a therapeutic window o
o
blood gas should be obtained in order to calculate an A—a gradi-
n
ent, establish whether ventilatory ailure is present, and determine opportunity. Improvement is expected over the rst one to two
a
r
hours. I hypoxemia persists, then there may be a concomitant com-
y
acid–base status. Further diagnostic workup will be guided by the
C
results o these initial studies. I the etiology is not apparent on ini- plication such as aspiration or pneumonia. A ter initial stabilization, a
o
decision should be made regarding appropriate level o care. Deter-
m
tial evaluation, urther testing might include CT angiography o the
p
chest and echocardiography. minants o this will include ability to manage the airway, method o
l
i
oxygen delivery to maintain adequate O2 status, and the intensity o
c
a
t
nursing care required. Intubation should not be delayed in patients
i
■ TREATMENT
o
who ail an NIVtrial.
n
s
The goal or management o acute postoperative respiratory ailure
is to quickly and correctly identi y the underlying pathophysiologic ■ COMPLICATIONS
process and provide targeted treatment that will avoid the need Complications o postoperative respiratory ailure include increased
or intubation and mechanical ventilation. Hypoxemic respiratory
risks related to intubation, ventilator-induced acute lung injury pro-
ailure has a broad di erential diagnosis, and initial supportive care
gressing to ARDS, VAP, GI bleeding, and DVT.
should be ollowed by treatment o the speci c disease process
identi ed. Pneumonia or other in ection should be treated with COPD EXACERBATION
prompt initiation o appropriate antibiotics a ter obtaining cultures.
Volume overload and pulmonary edema can be treated with diuret- Patients with COPD have an elevated risk o developing postop-
ics and additional cardiac evaluation. Bronchospasm should be erative pulmonary complications, including an exacerbation o
aggressively treated with inhaled β-agonists and anticholinergics obstructive lung disease. Ideally these patients will have had good
with systemic corticosteroids as indicated. Suspected pulmonary preoperative risk assessment, with optimization o medical manage-
embolism should be expeditiously evaluated and treated as the ment and initiation o preventive measures such as lung expansion
postoperative conditions allow. maneuvers and smoking cessation 4 or more weeks prior to the
With hypercapnic respiratory ailure, the most common cause o procedure (when applicable). Postoperative management consists
insuf cient ventilatory drive is medication e ect. Opioids are the o continuing the preventive measures as well as maintenance o
most potent suppressor o both hypoxic and hypercapnic ventila- home medications. Patients with COPD have a higher prevalence o
tory drive, but other sedatives and hypnotics also cause respiratory comorbid conditions such as congestive heart ailure and coronary
depression. Unrecognized obstructive sleep apnea may also present artery disease (CAD), so they must be care ully evaluated should
as postoperative respiratory depression. CPAP or positive pressure they become short o breath. There is also a high rate o concomi-
ventilation can be initiated i needed, and precautions can be taken tant PE with COPD, up to 20%, so this also needs to be ruled out.
to minimize the use o opioids, patient controlled anesthesia, and
sedative hypnotics. Respiratory arrest, inability to protect the airway ■ PATHOPHYSIOLOGY
and severe respiratory acidosis mandate intubation. Noninvasive A COPD exacerbation is de ned as an event in the natural course
ventilation (NIV) is an important tool that can be used both to pre- o the disease characterized by a change in the patient’s baseline
vent and also to treat postoperative acute respiratory ailure. When dyspnea, cough, and/or sputum that is beyond normal day-to-day
success ul it has been shown to decrease the need or intubation, variations, is acute in onset, and may warrant a change in regular
the rate o complications, and the intensive care unit (ICU) length o medication. Exacerbations are thought to be an in ammatory event,
stay. NIVis indicated when there is a demonstrated need or ventila- and the airway mani estations include edema, bronchospasm, and
tory support and no contraindications (Table 60-5). Level 1 evidence increased sputum production. Exacerbations are heterogeneous
395
events caused by complex interactions between the host, respira- Incentive spirometry, Coughing and deep breathing, Oral care, Under-
tory viruses, airway bacteria, and environmental pollution, but standing (patient and amily education), Getting out o bed, and
in approximately one-third no cause is identi ed. Hyperin ation Head o bed elevation. The intervention resulted in a decrease in the
P
and bronchospasm lead to increased work o breathing. Progres- number o cases o pneumonia and also o unplanned intubations,
A
sive hypoxemia develops, producing a downward spiral that can and served to illustrate the bene t o a multidisciplinary e ort. Patient
R
progress to acute ventilatory ailure when the respiratory muscles and amily education were eatured prominently in the bundle, and
T
atigue. COPD is associated with other comorbid conditions, includ- encouragement to continue the intervention post discharge could
I
I
ing ischemic heart disease (HD), pneumonia, and diabetes mellitus certainly produce some additional bene t. Hospitalists can also help
(DM), as well as venous thromboembolism (VTE), and one o these assure adequate discharge planning or ollow-up o medical prob-
can contribute. lems as well as communication with primary care to minimize the
risks o readmission.
M
■ DIAGNOSIS—DOES THIS PATIENT HAVE A
e
d
POSTOPERATIVE EXACERBATION OF COPD? SUGGESTED READINGS
i
c
a
Typical symptoms include increased breathlessness accompanied
l
C
by wheezing and chest tightness, increased cough and sputum American Thoracic Society. Guidelines or the management o adults
o
production, change in color and tenacity o sputum, and ever. with hospital-acquired, ventilator-associated, and healthcare-
n
s
Routine evaluation should include chest x-ray, routine labs, arterial associated pneumonia. Am J Respir Crit Care Med. 2005;171:
u
l
t
blood gas, and ECG to di erentiate COPD rom other causes given 388-416.
a
t
the requent comorbidities.
i
Cassidy MR, et al. I COUGH: Reducing postoperative pulmonary
o
n
complications with a multidisciplinary patient care program.
■ TREATMENT JAMASurg. 2013;148(8):740-745.
Treatment o postoperative COPD exacerbation does not di er Duggan M, Kavanagh BP. Pulmonary atelectasis: a pathogenic peri-
rom typical treatment, except that other medical complications operative entity. Anesthesiology. 2005;102:838-854.
prevalent in the postoperative period, including PE and congestive
Ferreyra G, Long Y, Ranieri VM. Respiratory complications a ter major
heart ailure, need to be care ully considered. Lung unction can be
surgery. Curr Opin Crit Care. 2009;15:342-348.
optimized with inhaled short-acting β-adrenergic and anticholiner-
gic agents along with systemic glucocorticoids. A short course o Global strategy or the diagnosis, management, and prevention o
antibiotics may decrease duration o exacerbation. There is currently chronic obstructive pulmonary disease. GOLD Executive Summary.
no role or methylxanthines or chest physical therapy (PT). NIV is Am J Respir Crit Care Med. 2013;187:347-365.
the pre erred method o ventilatory support and has been shown Johnson RG, Arozullah AM, Neumayer L, Henderson WG, Hosokawa P,
to improve outcomes in COPD exacerbations. Invasive mechanical Khuri SF. Multi-variable predictors o postoperative respiratory ail-
ventilation is reserved or patients who have not responded to NIV. ure a ter general and vascular surgery: results rom the patient’s
sa ety in surgery study. J Am Coll Surg. 2007;204:1188-1198.
■ COMPLICATIONS Lawrence VA, Cornell JE, Smetana GW. American College o Physicians.
Complications o COPD exacerbation include pneumonia, progres- Strategies to reduce postoperative pulmonary complications a ter
sion to acute respiratory ailure, and pneumothorax. non-cardiothoracic surgery: systematic review or the American
College o Physicians. Ann Intern Med. 2006;144:596-608.
Liapikou A, et al. Pharmacotherapy or hospital-acquired pneumo-
POSTACUTE CARE
nia. Expert Opin Pharmacother. 2014;15(6):775-786.
Postoperative pulmonary complications continue to lurk even
Nava S, Hill N. Non-invasive ventilation in acute respiratory ailure.
beyond hospital discharge. Among Medicare bene ciaries who were
Lancet. 2009;374:250-259.
rehospitalized within 30 days a ter a surgical discharge, 70.5% were
rehospitalized with a medical condition, and pneumonia was the Qaseem A, Snow V, Fitterman N, et al. Clinical Ef cacy Assessment
second most requent reason a ter heart ailure. Speci c risk actors or Subcommittee o the American College o Physicians. Risk assess-
readmission with pulmonary complications have not been studied, ment or and strategies to reduce perioperative pulmonary com-
but risk is likely to be reduced by e ective care transitions that include plications or patients undergoing noncardiothoracic surgery: a
adequate patient education on continuing preventive measures at guideline rom the American College o Physicians. Ann Intern
home as well as adequacy o postdischarge ollow-up. The I COUGH Med. 2006;144:575-580.
trial described a multidisciplinary intervention that aimed to decrease Restrepo RD, Braverman J. Current challenges in the recognition,
the incidence o postoperative pneumonia by standardizing use o prevention, and treatment o perioperative pulmonary atelectasis.
a “bundle” o simple postoperative interventions. These included Expert Rev Respir Med. 2015;9(1):97-107.
396
CHAP TER
61 INTRODUCTION
The kidneys are responsible or several vital homeostatic processes,
including the excretion o nitrogenous waste products, the regula-
tion o uid volume and electrolytes, acid–base balance, and the
production o hormones important or blood pressure regulation,
erythropoiesis, and bone metabolism. They are requently a ected
by disease, both acute (occurring over days to weeks) and chronic
(occurring over months to years). Acute kidney injury (AKI), ormerly
known as acute renal ailure, has become an increasingly common
Assessment and cause o hospitalization, with an incidence o 5% to 7% among hos-
pitalized patients. Chronic kidney disease (CKD) reportedly a ects
Management o 13% o adults in the United States, and is associated with signi cant
morbidity, mortality, and expense. The recent advent o automatic
Patients with Renal reporting o estimated glomerular ltration rate (eGFR) with serum
creatinine by hospital laboratories has resulted in more patients
being identi ed as having impaired renal unction. In order to
Disease provide the highest level o care or patients presenting with acute
or CKD, the clinician should have a strong understanding o the
undamental issues relevant to their evaluation and management.
Albert Q. Lam, MD EVALUATION OF THE RENAL PATIENT

Julian L. Sei ter, MD ■ HISTORY AND PHYSICAL EXAMINATION
The evaluation o the patient with kidney disease begins with a thor-
ough history and physical examination. The clinician should identi y
whether the renal disease is acute or chronic. I the patient’s previ-
ous medical records are available, this can be determined by quickly
reviewing prior laboratory testing, with particular attention given to
serum creatinine, blood urea nitrogen, and urinalyses. Patients who
present with AKI should be questioned about recent symptoms (eg,
vomiting, diarrhea, edema, dif culty voiding, decreased appetite,
weight changes) and events (eg, changes in oral intake, new medi-
cations, nonsteroidal anti-in ammatory drug [NSAID] use, intrave-
nous contrast administration, recent colonoscopy) that may help
narrow the di erential diagnosis o AKI. Symptoms such as ever,
rashes, arthralgias, epistaxis, and hemoptysis suggest an underlying
in ammatory condition such as vasculitis. For patients who develop
AKI during their hospitalization, recent hospital events—including
episodes o hypotension, recent diagnostic and therapeutic proce-
dures, and initiation o new medications—should be reviewed. All
patients presenting with AKI or CKD should be questioned about
symptoms associated with uremia, including atigue, nausea, vom-
iting, pruritus, metallic taste, lethargy, and con usion, since these
symptoms may indicate the need or dialysis.
Patients should be asked whether they have a prior history o
kidney disease or other relevant systemic diseases, such as diabetes
and hypertension. In patients with CKD, who may or may not be pre-
senting with an acute kidney-related problem, the clinician should
establish the chronicity, severity, and cause o the underlying kidney
disease. In patients with end-stage renal disease (ESRD), in ormation
about the patient’s nephrologist, outpatient dialysis unit, and regu-
lar dialysis schedule (including the timing o the last dialysis session)
should be obtained and conveyed to the clinicians and other health
care providers who will be acilitating the patient’s dialysis during
the hospitalization. The clinician should also obtain a complete
and current list o the patient’s medications, including prescription
medications as well as all over-the-counter medications, herbal rem-
edies, and supplements. A amily history o kidney disease or other
systemic illnesses should also be documented.
397
The physical examination starts with a review o the patient’s essential or the diagnosis and management o most renal diseases.
vital signs. While ever should always raise suspicion or an in ec- In prerenal acute kidney injury, the presence o hypervolemia (eg,
tion, particularly in dialysis patients or immunosuppressed patients, elevated jugular venous pressure, pulmonary congestion, peripheral
P
it can also be observed in the setting o acute glomerulonephritis, edema) suggests decreased renal per usion rom congestive heart
A
vasculitis, and allergic interstitial nephritis. Blood pressure may be ailure or cirrhosis, whereas hypovolemia (postural pulse increase
R
elevated (eg, in acute nephritic syndrome, malignant hypertension, >30 beats/min, severe postural dizziness, dry axilla or mucous
T
scleroderma, long-standing kidney disease), normal, or low (eg, in membranes) would be more consistent with volume depletion
I
I
volume depletion, sepsis, cirrhosis, heart ailure). Fluid intake and rom bleeding or gastrointestinal losses. To best assess the jugular
output should be reviewed to help determine volume status and venous pulsation, the patient should be reclined with the head
the need or uid repletion, diuresis, or dialysis. elevated at 30° to 45°, and the elevation o the right internal jugular
Key aspects o the exam include the bedside determination o vein above the sternal angle should be measured. Certain physical
M
volume status and a search or physical signs associated with spe- ndings are associated with speci c renal diseases (Table 61-1).
e
d
ci c kidney diseases and uremia. Assessment o volume status is Palpable purpura may be observed in vasculitic processes such as
i
c
a
l
C
o
n
s
TABLE 61-1 History and Physical Examination Findings in Renal Disease
u
l
t
a
t
Renal Disease History Physical Exam Findings
i
o
n
Prerenal acute kidney • Volume depletion (hemorrhage, vomiting, diarrhea, • Orthostatic hypotension, dry mucous
injury diuretics, burns) membranes and axillae
• Heart ailure • Elevated JVP, +S3, lung rales, edema
• Cirrhosis (heart ailure)
• Medications (NSAIDs, ACE inhibitors/ARBs, • Jaundice, ascites, edema (cirrhosis)
cyclosporine, tacrolimus)
• Radiocontrast exposure
Intrarenal acute kidney • Gross hematuria or cola-colored urine • Fever, palpable purpura, arthritis
injury • Cough and hemoptysis (Goodpasture syndrome) (vasculitis)
Glomerular • Epistaxis, sinusitis, hemoptysis, arthralgias, • Saddle-nose de ormity (Wegener)
(Wegener, Churg-Strauss) • Oral ulcers, rash, arthritis, pericardial rub
• Rash, arthralgias (systemic lupus erythematosus) (SLE)
Interstitial • Recent respiratory in ection (postin ectious
glomerulonephritis, IgA nephropathy)
• Fever, arthralgias, rash
ATN • Medications (NSAIDs, antibiotics) • Fever, skin rash
• Episode o hypotension
Vascular • Medications (aminoglycosides, amphotericin B, cisplatin) • Hypotension
• Trauma, muscle necrosis (rhabdomyolysis) • Warm (early sepsis) or cold extremities
• History o multiple myeloma (late sepsis)
• History o atherosclerosis • Elevated BP
• Recent vascular intervention • Livedo reticularis, ischemic extremities
• Anticoagulation
• Flank pain (renal vein thrombosis)
Postrenal acute kidney • Urinary urgency, hesitancy, oliguria or anuria • Distended bladder
injury • Gross hematuria • Enlarged prostate
• Flank pain, renal colic • Palpable abdominal or pelvic masses
• History o nephrolithiasis
• Medications (acyclovir, indinavir, anticholinergics)
Nephrotic syndrome • Weight gain • Anasarca, ascites, edema
• Foamy urine
• Medications (NSAIDs, gold, penicillamine)
Uremia • Fatigue, lethargy, con usion, seizures • Asterixis
• Anorexia, nausea, vomiting • Pericardial or pleural riction rub
• Pruritus, metallic taste, bleeding • Hal -and-hal nails
• Dry and atrophic skin, pallor,
hyperpigmentation, ecchymoses,
uremic rost
ACE, angiotensin-converting enzyme; ARB, angiotensin receptor blocker; BP, blood pressure; JVP, jugular venous pressure; NSAID, nonsteroidal anti-in lammatory
drug; SLE, systemic lupus erythematosus.
398
granulomatosis with polyangiitis ( ormerly known as Wegener’s early detection o renal disease. Elevated serum BUN is sometimes
granulomatosis), microscopic polyangiitis, Churg-Strauss syndrome, attributable to nonrenal actors, such as high-protein intake, upper
C
or Henoch-Schönlein purpura. Abdominal bruits with re ractory gastrointestinal tract bleeding, and high catabolism states, such
H
hypertension and progressive renal ailure are suggestive o reno-
A
as ever, corticosteroids, and burns. Serum creatinine may also be
P
vascular disease. Funduscopic examination can reveal arteriolar a ected by many actors, including muscle mass and medications
T
narrowing, hemorrhages, exudates, or papilledema— ndings con- that impair tubular creatinine secretion, such as trimethoprim and
E
sistent with chronic hypertension. cimetidine. Though BUN and creatinine are the traditional primary
R
biomarkers o renal injury, their use may decrease in the uture in
6
avor o more sensitive and speci c biomarkers, including neutrophil
1
PRACTICE POINT
gelatinase-associated lipocalin (NGAL), kidney injury molecule-1
Key aspects o the physical examination include: (KIM-1), and cystatin C.
A
• Determination o the patient’s volume status;
s
s
• Identi cation o physical mani estations that suggest speci c
e
Estimated glomerular filtration rate
s
s
renal disease conditions;
m
All patients with kidney disease, both acute and chronic, should
• Search or signs o uremia.
e
have their kidney unction assessed by estimation o the glomerular
n
ltration rate (GFR). GFR may be estimated by measuring serum
t
a
creatinine, calculating the creatinine clearance, or using estimation
n
d
The physical ndings o uremia are highly variable. Uremic peri- equations such as the Cockcro t-Gault ormula, the Modi cation
M
carditis or pleuritis may be present, as mani ested by a pericardial o Diet in Renal Disease (MDRD) equation, or the CKD-EPI (Chronic
a
or pleural riction rub, respectively. The pericardial riction rub clas-
n
Kidney Disease Epidemiology Collaboration) equation. The normal
a
sically has three components, one systolic and two diastolic, and
g
GFR in a healthy adult is >90 mL/min. GFR decreases with age, at a
e
a scratchy or grating quality. Skin and nail changes may include rate o approximately 1 mL/min/y a ter age 35. Elderly patients may
m
uremic rost (the ne residue o excreted urea on the sur ace o the
e
also have lower-creatinine levels due to decreased muscle mass.
n
skin), skin hyperpigmentation, or hal -and-hal nails (sharp demarca- Measurement o serum creatinine is the most requently used sur-
t
o
tion between proximal and distal nail halves). Patients with uid rogate or GFR. As serum creatinine concentrations are a ected by
retention may have pulmonary congestion or peripheral edema.
P
muscle mass, dietary protein intake, and certain medications, it is not
a
Neurological ndings include con usion, coma, asterixis, and sen-
t
the most accurate method o estimating GFR. The Cockcro t-Gault,
i
e
sory de cits. The presence o these physical ndings, especially the
n
MDRD, and CKD-EPI equations take into account serum creatinine, as
t
pericardial riction rub and neurological abnormalities, may indicate
s
well as other de ned actors such as age, race, gender, and weight.
w
the need or dialysis. They were designed to estimate GFR in patients with established
i
t
h
CKD, and are most use ul or this purpose. While the per ormance o
R
■ LABORATORY TESTS these equations has been evaluated in a variety o di erent racial and
e
n
Serum electrolytes ethnic populations with and without kidney disease, they should still
a
l
be interpreted with caution in speci c patient populations that have
D
Serum electrolytes are essential to the evaluation o the patient with
yet to be well validated, including individuals with normal or near-
i
s
acute and chronic renal disease. Both hyponatremia and hyper-
e
normal renal unction, children, and elderly individuals. The CKD-EPI
a
natremia may be seen in patients with kidney disease. Impaired
s
equation may be superior to the MDRD equation in estimating GFR
e
renal unction decreases renal potassium excretion, and may lead
in patients with normal or mildly impaired (GFR >60 mL/min/1.73
to potentially li e-threatening hyperkalemia in oliguric or anuric
m 2) renal unction. Serum creatinine and the estimation equations
patients. The serum potassium concentration may not be an accu-
should only be used to approximate GFR in patients with stable kid-
rate indicator o total body potassium stores, since most o the total
ney unction (unchanging serum creatinine). I the clinician is uncer-
body potassium is con ned to the intracellular uid compartment.
tain about the accuracy o GFR estimation, a 24-hour urine collection
For example, in diabetic ketoacidosis, patients requently have
can be per ormed to calculate creatinine clearance.
elevated serum potassium levels despite diminished total body
potassium stores. Serum chloride and bicarbonate levels are use ul
in the assessment o volume and acid–base status. The serum anion
gap, used in the assessment o metabolic acidosis, can be calculated PRACTICE POINT
rom serum sodium, chloride, and bicarbonate concentrations (AG = Laboratory testing
Na+ – [Cl– + HCO3–]). Serum calcium, phosphorus, and magnesium • Serum creatinine and the estimation equations should only
levels yield important in ormation about renal tubular unction and be used to approximate GFR in patients with stable kidney
bone mineral metabolism. Hyperphosphatemia and hypocalcemia unction (unchanging serum creatinine).
are common in patients with acute and chronic renal disease, and • The examination o the urinary sediment by microscopy can
contribute to the development o secondary hyperparathyroidism. provide use ul diagnostic in ormation about both acute and
chronic kidney disease.
Blood urea nitrogen and creatinine
Urine particles lyse easily a ter collection, and there ore
Blood urea nitrogen (BUN) and creatinine are nitrogenous end urine samples should be examined within 2 to 4 hours o
products o metabolism that rise in the setting o renal disease. Urea acquisition.
is ormed rom ammonia derived rom protein breakdown, while The pathognomonic nding o ATN on urinary sediment is
creatinine is a byproduct o muscle creatine metabolism. Urea and the presence o coarse muddy brown granular casts, which
creatinine are reely ltered by the kidneys but handled di erently in represent extensive renal tubular epithelial cell injury.
the tubular system. Urea is partly reabsorbed in the proximal tubule • In acute kidney injury, a ractional excretion o sodium (FENa) in
and inner medullary collecting duct, while creatinine is secreted to combination with clinical history and other lab tests may help
a small extent by the tubules. Despite these con ounding e ects di erentiate between prerenal etiologies and acute tubular
o tubular handling, BUN and creatinine are still the most com- necrosis (ATN).
monly used biomarkers o renal unction. Neither test is ideal or the
399
the pathologic diagnosis rarely has any e ect on the management
While a FENa o <1% in the setting o AKI is generally or outcome.
thought to indicate prerenal azotemia, it can also be seen
in contrast-induced nephropathy, rhabdomyolysis, acute Abnormal urinalysis
P
A
glomerulonephritis, hepatorenal syndrome, early urinary The examination o the urinary sediment by microscopy can provide
R
obstruction, acute interstitial nephritis, and even ATN. use ul diagnostic in ormation about both acute and CKD. Urine par-
T
A FENa may be dif cult to interpret in the setting o diuretic ticles lyse easily a ter collection, and there ore urine samples should
I
therapy.
I
be examined within 2 to 4 hours o acquisition. White blood cells
• I the patient has a nonanion gap metabolic acidosis, the (pyuria), when greater than 2 per high-power eld, can be observed
urine anion gap (UAG) may help di erentiate between with upper or lower-urinary tract in ections, contamination rom
gastrointestinal losses o bicarbonate (eg, diarrhea) and renal genital secretions, or renal in ammation, as in interstitial nephritis or
M
tubular acidosis. It is calculated as ollows: (urine Na+ + acute glomerulonephritis.
e
urine K+) – urine Cl–.
d
Urinary casts are cylindrical aggregates o protein and/or cells
i
c
that orm in the lumen o the distal convoluted tubule or collecting
a
l
duct. Hyaline casts, the most common type o cast, are acellular and
C
Proteinuria
o
consist primarily o Tamm-Hors all mucoprotein produced by tubu-
n
s
Proteinuria, a hallmark o kidney damage, is most requently detected lar epithelial cells. They can be seen in the setting o dehydration
u
l
qualitatively by urine dipstick, which grades proteinuria on a scale o or vigorous exercise in normal patients who produce concentrated
t
a
t
concentration: trace, 1+ (30 mg/dL), 2+ (100 mg/dL), 3+ (300 mg/dL). urine, but can be seen in patients with proteinuria. Granular casts,
i
o
Normal urine may test slightly positive i very concentrated. The the second most common type o cast, are usually ormed rom
n
urine dipstick is only capable o detecting urinary albumin, which is degenerating cellular casts or protein-containing lysosomes, and
the most abundant protein seen with glomerular proteinuria. The can appear ne or coarse in texture. Muddy brown granular casts
presence o other proteins, such as immunoglobulin light chains, contain degenerating tubular epithelial cells, and are commonly
will not be detected by dipstick alone. Proteinuria by dipstick should seen in acute tubular injury. Fatty casts are hyaline casts that contain
prompt a more accurate quanti cation. This is done either by mea- lipid droplets and can be observed in patients with diseases causing
suring the urine protein and urine creatinine concentrations in a lipiduria, such as nephrotic syndrome. The approach to hematuria is
random urine sample to determine the urine protein-to-creatinine described above. When red blood cells leak through the glomerular
ratio, or by a 24-hour urine collection or protein and creatinine ltration barrier, they can orm red blood cell casts in the tubular
excretion rate. lumen, a nding that is consistent with acute glomerulonephritis.
White blood cell casts indicate renal in ammation or in ection, and
Hematuria can be seen in acute glomerulonephritis, interstitial nephritis, and
In the absence o gross bleeding, hematuria is most commonly acute pyelonephritis. Red blood cell casts and white blood cell casts
discovered on a urine dipstick (which detects the pseudoperoxidase are always pathologic, and should prompt urther evaluation o the
activity o hemoglobin) or urinalysis. False-positive dipstick results patient or the clinical entities already mentioned.
are seen in the setting o hemoglobinuria, myoglobinuria, men-
strual blood in the urine, vigorous exercise, and concentrated urine. Urine chemistries
I signi cant proteinuria or renal dys unction is also present, the Measurements o urinary sodium, potassium, chloride, and creati-
kidney should be considered the source o hematuria until proven nine can be use ul in the evaluation o a number o renal conditions.
otherwise, and a renal biopsy should be considered to establish a In acute kidney injury, the ractional excretion o sodium (FENa),
diagnosis. Microscopic hematuria in the absence o proteinuria and combined with clinical history and other lab tests, may help di -
renal dys unction is known as isolated hematuria. erentiate between prerenal etiologies and acute tubular necrosis
The di erential diagnosis o isolated microscopic hematuria can (ATN). The FENa can be calculated by the ollowing ormula: (urine
be divided into renal (glomerular) or extrarenal (nonglomerular) Na+ × plasma creatinine)/(plasma Na+ × urine creatinine) × 100. A
processes. Immunoglobulin A (IgA) nephropathy, thin basement FENa < 1% is commonly seen in prerenal causes o oliguria, and a
membrane disease, and Alport syndrome are three o the more FENa > 2% is usually indicative o ATN. However, a FENa < % 1 can
common causes o glomerular hematuria. Common etiologies o also be seen in contrast-induced nephropathy, rhabdomyolysis,
nonglomerular hematuria include urinary tract in ections, kidney acute glomerulonephritis, hepatorenal syndrome, early urinary
stones, urinary tract tumors, trauma, bladder polyps, polycystic kid- obstruction, acute interstitial nephritis, and even ATN. Furthermore,
ney disease, medullary cystic disease, and metabolic abnormalities a FENa may be dif cult to interpret in the setting o a patient taking
such as hypercalciuria and hyperuricosuria. Hematuria associated diuretics. In such cases, calculating the ractional excretion o urea
with exercise, especially running, is usually a benign condition in (FEUrea) may help to di erentiate prerenal AKI (FEUrea < 35%) rom
which the blood source is likely the renal pelvis. ATN (FEUrea 50%-65%).
Hematuria rom the glomerulus may or may not be associated In nonanion gap metabolic acidosis, one can calculate a urine
with ank pain, while ureteral conditions that obstruct the uri- anion gap (UAG) to help di erentiate between gastrointestinal
nary tract and cause bleeding can produce severe pain and renal losses o bicarbonate (eg, diarrhea) and renal tubular acidosis using
colic. Other causes o hematuria are usually painless. In extrarenal the ollowing ormula: (urine Na+ + urine K+) – urine Cl–. A negative
hematuria, the red blood cells typically appear normal on urinary UAG is consistent with gastrointestinal losses, whereas a positive
sediment, round and uni orm, whereas in glomerular hematuria, the UAG is requently seen with renal tubular acidosis.
red blood cells may appear dysmorphic due to distortion rom the
passage through the glomerular ltration barrier. Imaging studies Serum enzymes
are indicated to search or structural causes o hematuria. Detection Serum enzyme levels should be interpreted cautiously in patients
o persistent extrarenal hematuria should prompt urther workup with impaired renal unction. Cardiac enzymes, including cardiac
and urologic consultation to identi y the source o bleeding. In troponin T (cTnT), cardiac troponin I (cTnI), and the muscle/brain
older individuals, bladder cancer should be considered. In isolated (MB) isoenzyme o creatine kinase (CK-MB), are o ten elevated
glomerular hematuria, a renal biopsy is not typically indicated, since in acute or chronic kidney disease, even in the absence o acute
400
myocardial injury. A large percentage o alse-positive elevations in resistive indices (>0.80) in a stenotic kidney are suggestive o severe
cTnT and CK-MB are seen in patients with ESRD when these mark- parenchymal disease and a low likelihood o response to revas-
C
ers are used to diagnose acute myocardial in arction (MI). The use cularization. Given the enhanced toxicities o iodinated contrast
H
o cTnI is less likely to be associated with alse-positive elevations. agents or gadolinium in renal disease, Doppler ultrasonography has
A
P
Serial measurements o cTnI are currently the most speci c marker become widely used as the initial imaging study to evaluate renal
T
o myocardial damage in patients with renal ailure and suspected artery stenosis. The sensitivity o Doppler ultrasonography is highly
E
acute MI. operator dependent, and can be a ected by patient anatomy.
R
Liver and pancreatic enzymes can also be a ected in patients
6
with renal ailure. Serum aminotrans erase levels are requently Computed tomography
1
ound to be in the lower range o normal values in patients with CKD In the evaluation o the patient with suspected renal colic, noncon-
and ESRD. In the absence o liver disease, gammaglutamyl transpep- trast helical CT scanning is currently the gold standard or diagnos-
A
tidase (GGT) levels are most o ten normal, but may be elevated in a ing nephrolithiasis and can detect essentially all kidney stones, with
s
s
e
small percentage o patients. Serum alkaline phosphatase levels are the exception o indinavir stones. Noncontrast CT can also detect
s
s
o ten elevated in dialysis patients, usually rom coexisting bone dis- ureteric obstruction in acute kidney injury, which is particularly
m
ease. An isolated elevation in serum alkaline phosphatase may not help ul when intravenous (IV) contrast should be avoided due to
e
n
correlate well with hepatobiliary disease in ESRD patients; however, nephrotoxicity.
t
a
i a chronically elevated alkaline phosphatase level is accompa- The drawback to the use o iodinated contrast agents is poten-
n
d
nied by an elevation in serum GGT or 5′-nucleotidase, one should tial nephrotoxicity, especially in patients with preexisting renal
M
be more suspicious o an obstructive or in ltrative hepatobiliary impairment, diabetes, heart ailure, or hypovolemia (see below). In
a
process. patients with ESRD who have residual renal unction, administra-
n
a
Serum levels o both amylase and lipase can be elevated in tion o contrast dye can induce urther tubular damage and lead
g
e
patients with CKD and ESRD, even when acute pancreatitis is not to loss o the remaining renal unction. As preservation o residual
m
present. The levels o these enzymes in ESRD patients are commonly renal unction in patients with ESRD has been shown to correlate
e
n
three old to ve old higher than baseline, but are typically less with improved survival even a ter the initiation o dialysis, the use o
t
o
than three times the upper limit o normal. The elevations are due contrast in these patients should be avoided i possible. When the
primarily to decreased renal clearance, though in the case o serum risk o nephrotoxicity is not prohibitive, IV iodinated contrast is use-
P
a
lipase, the use o heparin during hemodialysis has also been ound ul or imaging o the renal parenchyma, acilitating the evaluation
t
i
e
to contribute to elevated levels. and detection o renal mass lesions such as renal cell carcinoma. CT
n
t
angiography can be used to diagnose suspected renal artery ste-
s
w
■ IMAGING STUDIES nosis or aneurysms. CT urography allows imaging o the collecting
i
t
system and can identi y lling de ects such as stones, blood clots,
h
Ultrasonography
and tumors.
R
e
Ultrasonography is a sa e, noninvasive, rapid, and inexpensive diag-
n
a
nostic imaging modality used to study the kidneys. Ultrasonography Magnetic resonance imaging
l
D
requires neither ionizing radiation nor a potentially toxic intravenous The primary role o renal magnetic resonance imaging (MRI) is the
i
s
contrast agent, which makes it a sa e initial imaging study, especially evaluation o renal masses. MRI can e ectively di erentiate benign
e
a
or patients with known renal insuf ciency. Renal ultrasonography versus malignant lesions in the kidney, especially when CT scanning
s
e
can provide valuable in ormation about kidney size, shape, and gross with intravenous iodinated contrast is contraindicated or i ultraso-
appearance. Normal adult kidneys are approximately 9 to 13 cm nographic and CT scans are nondiagnostic. MR angiography (MRA),
(4-5 in) in length and 5 to 7.5 cm (2-3 in) wide, and should not di er which involves the administration o intravenous gadolinium, has
by much more than 1 cm. With chronic injury, the renal parenchyma become the modality o choice in the evaluation o renovascular
is replaced with brotic tissue and the renal cortex becomes thinner, disease. According to one meta-analysis, gadolinium-enhanced
causing diseased kidneys to shrink. In patients with kidney disease MRA had a reported sensitivity o 97% and speci city o 85% or the
o uncertain duration, the nding o smaller kidneys on ultrasonog- detection o renal artery stenosis. However, the use o gadolinium-
raphy suggests longstanding renal disease. Enlarged kidneys may based contrast agents in moderate to severe CKD has been associ-
be seen in autosomal dominant polycystic kidney disease, urinary ated with the development o nephrogenic systemic brosis (NSF),
tract obstruction, HIVnephropathy, early diabetic nephropathy, and with debilitating brosis o the skin, joints, eyes, and other internal
in ltrative diseases such as amyloidosis or myeloma. Asymmetric organs. Patients with an estimated GFR <30 mL/min or requiring
kidneys may indicate unilateral kidney disease, and the clinician dialysis should not be given gadolinium-based contrast agents. In
must determine whether the smaller or larger kidney is abnormal. these patients, Doppler ultrasonography is a sa er alternative.
Increased renal echogenicity is common and nonspeci c nding,
usually denoting medical renal disease. Renal ultrasonography Radionuclide scans
can also identi y the presence o renal cysts, stones, or masses. In Radionuclide studies may be used to obtain unctional in ormation
patients presenting with acute kidney injury, renal ultrasonography about the kidneys. Static radionuclide scans employ a radiolabeled
can identi y obstructive uropathy, which usually mani ests as hydro- tracer (eg, technetium 99m-DMSA) that binds to renal parenchymal
nephrosis, although alse-negative results can be seen in patients cells, but is not excreted into the tubules. These studies are use ul in
with early obstruction (<3-4 days), volume depletion, or obstruction quanti ying the unctional cortical tissue o each kidney and deter-
due to retroperitoneal brosis or compression by retroperitoneal or mining the percentage contribution o each kidney to total renal
intraparenchymal tumor or blood. unction. Dynamic radionuclide scans use tracers (eg, technetium
99m-DTPA, technetium 99m-MAG3) that are taken up by nephrons
Doppler ultrasonography and then excreted into the collecting system. A diuretic such as
Doppler ultrasonography can provide in ormation about the pres- urosemide is o ten administered just prior to injection o the tracer
ence and ow o blood through the vessels o the kidney. High- to ensure high levels o diuresis during the study. Dynamic scans can
velocity or disorganized ow patterns can be seen in patients with be used to evaluate potential renal tract obstructions as well as the
hemodynamically signi cant renal artery stenosis. Elevated vascular response to treatment o the obstruction.
401
ACUTE KIDNEY INJURY depletion re ers to decreased e ective circulating volume in the set-
ting o normovolemia or hypervolemia, and can result rom marked
Acute kidney injury (AKI), ormerly termed acute renal ailure, is a
vasodilatation as seen in the setting o sepsis, heart ailure, cirrhosis,
sudden and sustained decline in renal unction with the ailure to
P
and third-spacing. Renal vasoconstriction is most o ten caused
excrete metabolic waste, maintain uid and electrolyte balance,
A
by medications, including angiotensin-converting enzyme (ACE)
and regulate acid–base homeostasis. AKI is an increasingly com-
R
inhibitors, angiotensin receptor blockers (ARBs), NSAIDs, intrave-
T
mon cause o hospitalization, with 1% o all patients reported to
nous iodinated contrast agents, and the immunosuppressant drugs
I
have AKI upon admission to the hospital and 2% to 5% o inpatients
I
cyclosporine and tacrolimus.
subsequently developing AKI during their hospitalization. In spite o
Patients with prerenal AKI usually present with an elevated BUN
advances in intensive care and dialysis support over the last 50 years,
and creatinine, and the ratio o BUN to creatinine is classically
the overall mortality rate o AKI remains high, ranging rom 20% to
greater than 20:1. Urinalysis o ten reveals an elevated speci c gravity
90% depending on illness severity and medical setting. The role o
M
without signi cant hematuria or proteinuria. The urinary sediment
e
the hospitalist is to diagnose common causes o AKI, to identi y and
d
is typically bland but may show hyaline casts. The kidneys, in an
treat reversible actors, to recognize when dialysis is required, and to
i
c
appropriate response to the reduction in renal per usion, maximize
a
know when to consult a nephrologist.
l
sodium and water reabsorption. Urine sodium is typically low, and
C
Two classi cation systems, the RIFLE and AKIN criteria, have
o
the FENa and urea are <1% and <35%, respectively. Patients with
n
de ned and strati ed AKI by stages o severity based on graded
s
prerenal AKI o ten respond avorably to volume resuscitation and
u
increases in serum creatinine and periods o decreased urine output
l
discontinuation o any o ending therapeutic agents.
t
(Table 61-2). The more recent AKIN criteria have proposed a de ni-
a
t
i
tion or AKI that incorporates the prognostic signi cance associated
o
Intrarenal
n
with small changes in serum creatinine. The diagnosis o AKI can
be established by (1) an abrupt (within 48 hours) absolute increase In all cases o intrarenal AKI, the primary abnormality is within the
in serum creatinine o ≥0.3 mg/dL rom baseline, (2) a percentage kidney. Intrarenal AKI can be subdivided into our anatomic cat-
increase in serum creatinine o ≥50%, or (3) oliguria o ≤0.5 mL/ egories: glomerular disease, interstitial disease, tubular disease, and
kg/h or > 6 hours. Although both the RIFLE and AKIN classi cation vascular disease. As in prerenal AKI, patients generally present with
systems have been validated in a variety o clinical settings, their an elevated BUN and creatinine, though the ratio is usually normal
utility at this time appears to be greater or research use than or (<20:1). The FENa may be variable and cannot reliably distinguish
the bedside. between the di erent causes o intrarenal AKI. The urinalysis is
AKI can be divided into three diagnostic categories based on requently abnormal, and ndings on the urinary sediment can
the anatomic location o injury: prerenal, intrarenal, and postrenal provide clues to the location o the kidney injury. In some patients,
(Table 61-3). It can be urther subdivided into oliguric (urine out- the clinical presentation and laboratory evaluation are insuf cient
put < 400 mL/d) and nonoliguric (urine output > 400 mL/d), with to establish a diagnosis, and a percutaneous renal biopsy may be
patients producing less than 100 mL urine/d considered to be indicated to better guide management.
anuric. These distinctions are important, given that epidemiological The most common cause o intrarenal AKI is ATN, which is
studies have ound that oliguria in the setting o AKI is an indepen- responsible or most cases o AKI in hospitalized patients. ATN may
dent predictor o mortality. Oliguric AKI is more characteristic o be caused by either ischemic or nephrotoxic injury. Ischemic ATN
prerenal etiologies and urinary obstruction, while nonoliguric AKI is is o ten associated with periods o prolonged hypotension and
commonly seen in intrarenal AKI. Anuria is uncommon and is usually markedly reduced renal per usion, which can be seen in the setting
associated with complete urinary tract obstruction, bilateral renal o heart ailure, sepsis, or cardiac surgery. Nephrotoxic ATN can be
in arction, renal vein thrombosis, cortical necrosis, or high-grade caused by either endogenous (eg, heme pigments) or exogenous
ischemic acute tubular necrosis. toxins (eg, aminoglycoside antibiotics, amphotericin B, cisplatin,
and iodinated contrast agents). While many patients typically expe-
rience an oliguric phase (onset within 24 hours o the renal insult
■ PRERENAL and duration o 1-3 weeks) ollowed by a diuretic phase (increase
Prerenal AKI is de ned as a reduction in GFR caused by hypoper u- in urine output that is indicative o renal recovery), some patients
sion o the kidney. In most cases o prerenal AKI, the kidneys are remain nonoliguric throughout. The pathognomonic nding on
morphologically normal. Prerenal AKI can be divided into conditions urinary sediment is the presence o coarse muddy brown granular
that cause volume depletion and conditions that induce renal vaso- casts, which represent extensive renal tubular epithelial cell injury.
constriction. True volume depletion may result rom hemorrhage or Due to impaired tubular sodium reabsorption, the urine sodium
gastrointestinal, urinary, or cutaneous uid losses. E ective volume is >40 mEq/L and the FENa is usually >2%. Both ischemic and
TABLE 61-2 De initions and Classi ication Systems or Acute Kidney Injury
RIFLE Increase in AKIN Increase in

RIFLE Stages AKIN Stages Serum Creatinine Serum Creatinine RIFLE and AKIN Urine Output
Risk (R) 1 ≥150%-200% ≥0.3 mg/dL or ≥150%-200% <0.5 mL/kg/h × >6 h
Injury (I) 2 >200%-300% >200%-300% <0.5 mL/kg/h × >12 h
Failure (F) 3 >300% >300% or acute renal <0.3 mL/kg/h × ≥24 h
replacement therapy
Loss (L) Complete loss o kidney unction or >4 wk
End-stage kidney Need or renal replacement therapy or >3 mo
disease (E)
AKIN, Acute Kidney Injury Network; RIFLE, Risk-Injury-Failure-Loss-ESRD.
402
TABLE 61-3 Etiologies o Acute Kidney Injury
C
H
Prerenal Postrenal
A
Volume depletion Prostatic hypertrophy
P
T
• True volume depletion Obstruction
E
GI losses (vomiting, diarrhea) • Bladder outlet obstruction
R
Renal losses (diuretics, osmotic diuresis) • Stones
6
1
Skin losses (burns, sweating) • Crystals (acyclovir, indinavir)
• E ective volume depletion • Tumors
A
Congestive heart ailure • Clots
s
s
Cirrhosis • Retroperitoneal ibrosis
e
s
s
Nephrotic syndrome
m
e
Vasoconstriction
n
t
• NSAIDs
a
n
• ACE inhibitors/ARBs
d
• Iodinated contrast agents
M
a
• Cyclosporine and tacrolimus
n
a
Hepatorenal syndrome
g
e
Hypotension
m
e
Intrarenal
n
t
Glomerular • Nephrotoxic acute tubular necrosis
o
• Acute glomerulonephritis Aminoglycoside antibiotics
P
a
t
ANCA-associated vasculitis Amphotericin B
i
e
(granulomatosis with angiitis, microscopic
n
Cast nephropathy (myeloma kidney)
t
polyangiitis, Churg-Strauss syndrome)
s
Cisplatin
w
Anti-GBM disease (Goodpasture syndrome)
i
Iodinated contrast agents
t
h
Immune complex disease (lupus nephritis,
Pigment nephropathy (hemoglobin, myoglobin)
R
poststreptococcal glomerulonephritis,
e
Vascular
n
cryoglobulinemia, IgA nephropathy)
a
l
Interstitial • Large vessel
D
i
• Drug-induced (NSAIDs, penicillin analogues Bilateral renal artery stenosis
s
e
and cephalosporins, ri ampin, sul a drugs)
a
Renal vein thrombosis
s
e
• Autoimmune (SLE, Sjögren syndrome) Renal thromboembolism
• In ections (legionella, leptospirosis, • Small vessel
cytomegalovirus, streptococci)
Thrombotic microangiopathies (HUS, TTP)
Tubular
Cholesterol atheroembolism
• Ischemic acute tubular necrosis
Malignant hypertension
Hypotension
Scleroderma renal crisis
Sepsis
ACE, angiotensin-converting enzyme; GI, gastrointestinal; HUS, hemolytic uremic syndrome; NSAID, nonsteroidal
anti-in lammatory drug; RB, angiotensin receptor blocker; SLE, systemic lupus erythematosus; TTP, thrombotic
thrombocytopenic purpura.
nephrotoxic ATN resolve in most cases, but dialysis is sometimes tests such as complement levels, antistreptococcal antibodies, anti-
required when renal injury is severe. A recently reported cause o bodies against hepatitis B and C, antinuclear antibodies, antineutro-
renal injury is the use o sodium phosphate salts or bowel cleansing phil cytoplasmic antibodies, antiglomerular basement membrane
prior to colonoscopy, with phosphate precipitation in the kidney in antibodies, and cryoglobulins. De nitive diagnosis usually requires
volume-depleted patients or those with CKD. a renal biopsy.
The glomerular type o AKI involves acute in ammation o Acute interstitial nephritis (AIN) is de ined as in lammation o
the glomeruli or glomerular vessels. Acute glomerulonephritis the renal interstitium that results in AKI. AIN is most o ten caused
can be either renal-limited or associated with systemic illnesses by medications, such as antibiotics, NSAIDs, anticonvulsants,
such as in ections, such as poststreptococcal glomerulonephritis, and proton pump inhibitors, but can also be associated with
autoimmune disorders such as systemic lupus erythematosus, or in ections and autoimmune diseases. Classic symptoms include
vasculitides such as granulomatosis with angiitis. The urinalysis is ever, rash, and arthralgias. However, the classic triad o ever,
always abnormal and classically reveals evidence o damage to the maculopapular erythematous rash, and eosinophilia is observed
glomerular ltration barrier, with proteinuria, dysmorphic red blood in only 10% o cases o AIN. Laboratory testing may reveal a FENa
cells, and red blood cell casts. Urine sodium and FENa may be low. >1%, but this is not always reliable. Urinalysis may show mild pro-
The workup o acute glomerulonephritis should include serologic teinuria (<1 g/d), and the urinary sediment may reveal red blood
403
cells, white blood cells, and white blood cell casts. Occasionally, renal per usion, including ACE inhibitors, ARBs, NSAIDs, and
urine eosinophils are observed with a Wright or Hansen stain, calcineurin inhibitors, should be discontinued. Patients who
but this inding is neither highly sensitive nor speci ic or the are hypovolemic should be given volume accordingly, with
P
diagnosis o AIN and can be seen in other in lammatory condi- either crystalloids, colloids, or blood products. With ew
A
tions. De initive diagnosis can be established with a renal biopsy. exceptions, such as the setting o cirrhosis, the use o colloids
R
Treatment o AIN is primarily the identi ication and cessation o has not been proven to be more bene cial than crystalloids
T
the o ending agent. in AKI. Vasopressors or inotropes should be considered in
I
I
AKI can also be caused by acute vascular disease a ecting patients who remain hypotensive despite volume resuscita-
either the large or the small renal blood vessels. Large-vessel tion. In patients who are hypervolemic, the role o diuretics in
diseases involve the renal arteries and veins and include bilat- the treatment o AKI is controversial. Although loop diuretics
eral renal artery stenosis, renal thromboembolism, renal artery may be use ul to treat volume overload in an oliguric patient,
M
dissection, and renal vein thrombosis. As a general rule, large- conversion o oliguric to nonoliguric AKI with diuretics has
e
d
vessel disease must be bilateral in order to cause AKI, with the not been shown to improve survival or shorten the time to
i
c
exception o unilateral disease in the patient with a solitary renal recovery. At high doses, loop diuretics may also lead
a
l
kidney. Patients may present with symptoms o renal in arction, to ototoxicity. There ore, these medications should be used
C
o
complaining o acute lank pain and hematuria. Small-vessel judiciously in patients with AKI.
n
s
diseases that can cause AKI include malignant hypertension, 2. Close monitoring and management o renal unction,
u
l
scleroderma renal crisis, and cholesterol atheroembolic disease. acid–base status, and serum electrolytes. Serum BUN, cre-
t
a
t
Patients with cholesterol atheroembolic disease o ten have a atinine, and electrolytes should be monitored daily. I hyperka-
i
o
history o recent aortic instrumentation or surgical intervention lemia is present, medical treatment should be initiated, such as
n
or anticoagulation. Physical exam may reveal livedo reticularis on intravenous calcium gluconate, insulin, inhaled albuterol, and
the skin overlying the lower extremities, toe or oot discoloration, sodium polystyrene sul onate (kayexalate). Speci c treatment
or Hollenhorst plaques in the retina. The urinalysis in vascular AKI depends on the severity, urine output, and ECG abnormalities;
typically shows microscopic hematuria with or without protein- dialysis may be necessary i electrocardiographic abnormalities
uria. Eosinophilia, eosinophiluria, and hypocomplementemia can are present. Potassium intake via diet, medications, and intra-
also be seen in cholesterol atheroembolic disease. Imaging (eg, venous uids should also be eliminated. Hyperphosphatemia
CT, MRI, or radionuclide studies) is o ten required to con irm the can be treated with oral phosphorus binders such as calcium
diagnosis o large-vessel disease. acetate, calcium carbonate, sevelamer hydrochloride, and
sevelamer carbonate. Aluminum hydroxide is highly e ec-
Postrenal tive at lowering phosphorus levels in severe cases, but its use
In all patients presenting with AKI, urinary tract obstruction must should be limited to no more than 1 to 2 weeks due to the
be ruled out early, since timely intervention o ten improves or ully potential or aluminum toxicity. I acidemia is present, patients
restores renal unction. Obstruction to the ow o urine commonly can be treated with intravenous uids containing sodium
occurs at the level o the prostate, particularly in adult men, but bicarbonate.
can occur at any location along the urinary tract. Upper urinary 3. Appropriate adjustment o medication dosing. All medica-
tract obstruction (ie, at the level o the ureters or renal pelvis) must tions should be dosed to re ect the level o renal impairment,
be bilateral in order to cause AKI; the sole exception is unilateral based on estimated GFR, or the need or dialysis (Table 61-4).
obstruction in the patient with a solitary kidney. Common causes Since eGFR can only be calculated when the serum creatinine
o postrenal AKI include hypertrophy or cancer o the prostate, is stable, a GFR o <10 should be assumed or patients whose
obstructing kidney stones, urothelial tumors, and retroperitoneal serum creatinine is acutely increasing. Narcotics may accumu-
brosis or malignancies. Patients with bilateral obstruction may late in patients with renal impairment and should be used with
present with oliguria (partial obstruction), polyuria (a sign o great caution.
associated nephrogenic diabetes insipidus), or anuria (complete 4. Avoidance o nephrotoxins. Medications that are nephro-
obstruction), and may report symptoms o ank pain, abdominal toxic, such as NSAIDs and aminoglycosides, should not be
pain, renal colic, or hematuria. Ultrasonographic imaging usually given to patients with AKI. ACE inhibitors or ARBs taken on a
reveals hydronephrosis, though this may be absent in retroperito- chronic basis or hypertension or cardiovascular disease should
neal or in ltrative diseases that encase the ureters or kidneys. CT be stopped until renal unction has recovered. Intravenous
and dynamic radionuclide studies can also be used to diagnose iodinated contrast agents and gadolinium-containing agents
urinary obstruction. Treatment o postrenal AKI ocuses on relie should be avoided.
o the obstruction. 5. Management o uremic bleeding. Patients with severe AKI
and uremia may develop bleeding diatheses due to uremic
■ MANAGEMENT STRATEGIES platelet dys unction. This can be treated with synthetic arginine
vasopressin analogues (eg, intravenous DDAVP 0.3 mcg/kg ×
General principles
1-2 doses). Hemodialysis is the de nitive treatment, and should
Management o AKI should be ocused on treating and reversing be per ormed in cases o severe bleeding.
the speci c cause o injury. Patients with prerenal AKI, or example, 6. Nutritional supp ort. Malnutrition is highly prevalent in
should be given volume resuscitation to restore euvolemia. In patients with AKI. It is associated with higher risks o in-hospital
postrenal AKI due to urinary obstruction, relie o the obstruction can mortality, nosocomial complications, and p rolonged
improve and in many cases ully restore renal unction. Currently, hospitalization. Appropriate nutritional is thus essential to the
there are no e ective pharmacologic therapies or the treatment o management o AKI, and consultation with an experienced
AKI, and treatment ocuses more on supportive management. Basic dietitian may be bene cial. Hyperkalemia and hyperphos-
principles o management in AKI include: phatemia are common in patients with AKI, and a diet that
1. Op timization o volume status and hemod ynamic is low in potassium and phosphorus should be instituted.
parameters. Daily weights and intake and output should Critically ill patients with AKI are in a highly catabolic state
be monitored closely. Medications that can compromise and at high risk or severe protein energy wasting. Nutritional
404
TABLE 61-4 Dosing Adjustments or Commonly Prescribed Medications in Patients with Impaired Renal Function
C
H
GFR > 50 mL/ GFR 10-50 mL/min/
A
Drug Usual Dose min/1.73 m 2 1.73 m 2 GFR < 10 mL/min/1.73 m 2
P
Acyclovir (oral) 200-800 mg every 4-12 h 100% 100% 200 mg every 12 h
T
E
Allopurinol 300 mg daily 75% 50% 25%
R
Ampicillin/ 1.5-3 g IVevery 6-8 h 100% 1.5-3 g IVevery 12 h 1.5-3g IVevery 24 h
6
sulbactam (Unasyn) (GFR 15-29) (GFR 5-14)
1
Ce azolin (Ance ) 500 mg-1.5 g every 8 h 100% Every 12 h 50% every 24-48 h
Ce tazidime (Fortaz) 1-2 g every 8-12 h 100% Every 12-24 h Every 24-48 h
A
s
s
Ce triaxone 1-2 g every 24 h No adjustment
e
(Rocephin) needed
s
s
m
Cipro loxacin 400 mg IVor 500-750 mg 100% 50%-75% 50%
e
orally every 12 h
n
t
Enoxaparin Prophylaxis: 30 mg SC Usual dosage 30 mg SC daily (GFR < 30)
a
n
(Lovenox) every 12 h 1 mg/kg SC every 24 h (GFR < 30)
d
DVT treatment: 1 mg/kg
M
SC every 12 h or
a
n
1.5 mg/kg SC once daily
a
g
Fluconazole 200-400 mg every 24 h 100% 50% 50%
e
m
(Di lucan)
e
n
Gabapentin 300-600 mg Usual dosage 400-1400 mg/d 100-300 mg daily
t
(Neurontin) three times daily (divided twice daily) (GFR 30-59) Not recommended
o
200-700 mg/d (GFR 15-29)
P
a
Levetiracetam 500-1500 mg every 12 h Usual dosage 250-750 mg every 12 h (GFR 30-50)
t
i
e
(Keppra) 250-500 mg every 12 h (GFR < 30)
n
t
s
Hemodialysis: 500-1000 mg every 24 h, supplemental dose o
w
250-500 mg recommended a ter dialysis
i
t
h
Levo loxacin 250-750 mg Usage dosage 500 mg initial dose, 500 mg initial dose,
R
(Levaquin) orally/IVdaily then 250 mg every 24 h then 250 mg every 48 h
e
n
Met ormin 500-1000 mg twice daily Contraindicated in men with serum creatinine >1.5 mg/dL and women with serum
a
l
(Glucophage) creatinine >1.4 mg/dL or patients with GFR <60
D
i
s
Should be temporarily discontinued 24-48 h prior to administration o any
e
radiocontrast agents and not restarted or 48 h a terward due to the risk o
a
s
e
developing lactic acidosis
Metoclopramide 10-15 mg three Usual dosage 50% 25%
(Reglan) to our times daily
Piperacillin/ 3.375 g IVevery 6-8 h 100% 2.25 g IVevery 6 h (GFR 20-40) 2.25 g IVevery 8 h
Tazobactam (Zosyn) 2.25 g IVevery 8 h (GFR < 20)
Simvastatin (Zocor) 10-80 mg daily Usual dosage Usual dosage Start at 5 mg daily
Vancomycin 1 g IVevery 12 h 1 g IVevery 12 h Start with 1 g IVevery 12 h (GFR 40-60)
Start with 1g IVevery 24 h (GFR < 40)
Determine dose by serum level monitoring
DVT, deep vein thrombosis; GFR, glomerular iltration rate; IV, intravenous; SC, subcutaneous.
support, parenteral or enteral, is requently required in order PRACTICE POINT

to ensure adequate delivery o protein and energy, pre-
Indications or emergent nephrology consultation include:
vent urther metabolic derangements and complications,
improve wound healing, bolster the immune system, and • Volume overload in an oliguric or anuric patient
decrease mortality. • Hyperkalemia with serum potassium >5.5 to 6 mEq/L and/or
associated with changes on the electrocardiogram and other
It should be noted that, in spite o the available treatment modali- electrolyte abnormalities, especially in an oligoanuric patient
ties and advances in dialysis technology, mortality in patients with • Toxic overdoses that can be treated with hemodialysis,
AKI remains high, with a rate o approximately 50% to 80% in criti- including ethylene glycol, methanol, and lithium
cally ill patients. • Symptomatic or severe hyponatremia
• Hypertensive crises
When to consult a nephrologist • Rapidly progressive glomerulonephritis
There are a number o common clinical scenarios that are consid- • Microangiopathic hemolytic anemias, including thrombotic
ered nephrologic emergencies, and immediate evaluation by a thrombocytopenic purpura and hemolytic uremic syndrome
nephrologist should be requested.
405
Indications for renal biopsy aminoglycosides. Patients may present postoperatively with either
Percutaneous renal biopsy can be instrumental to the diagnosis o an acute elevation in serum creatinine or reduced urine output.
AKI. This procedure is typically per ormed under ultrasonographic A number o principles can guide the evaluation and manage-
P
guidance with local anesthesia, although CT-guided biopsy is an ment o postoperative AKI:
A
alternative in morbidly obese patients. In the setting o AKI, a renal 1. Identif cation o inciting actors. Perioperative records and
R
biopsy may be most help ul either when the diagnosis o acute owsheets should be thoroughly reviewed or evidence o
T
glomerulonephritis is suspected or when the cause o renal ailure hypotension, signi cant intraoperative or postoperative uid
I
I
is unknown. Other common indications or per orming a renal losses (eg, blood and intravascular uid losses, insensible
biopsy include unexplained glomerular hematuria, signi cant pro- losses, drainage losses, and third-spaced uid losses), and the
teinuria, and nephrotic syndrome. Absolute contraindications to administration o potentially nephrotoxic agents (eg, NSAIDs
percutaneous renal biopsy include uncontrolled moderate to severe or pain control or hydroxyethyl starches used or volume
M
hypertension, uncontrolled bleeding diathesis or severe anemia, resuscitation).
e
d
an uncooperative patient, and a solitary unctional kidney. Relative 2. Hemodynamic monitoring. Patients should have close peri-
i
c
a
contraindications include anatomic abnormalities o the kidney that operative hemodynamic monitoring, and i necessary, invasive
l
C
may increase the risk o the procedure, skin in ection overlying the monitoring with intra-arterial, central venous, or pulmonary
o
biopsy site, active renal or perinephric in ection, hydronephrosis, arterial catheters.
n
s
and the presence o multiple renal cysts or a renal tumor. The most 3. Maintenance o adequate renal per usion. Though no
u
l
t
common complication ollowing percutaneous renal biopsy is optimal mean arterial pressure (MAP) has been established to
a
t
bleeding, which usually occurs within 12 to 24 hours postbiopsy. ensure adequate renal per usion, maintaining a MAP o at least
i
o
n
Other complications include pain, gross hematuria, and in ection. >65 mm Hg and pre erably >75 to 80 mm Hg is recommended.
Chronic war arin anticoagulation is not a contraindication to 4. Optimization o volume status. Intravenous uid hydration
renal biopsy. However, the need and urgency or biopsy must be should be administered to optimize renal per usion in patients
weighed against the risk o thrombosis i anticoagulation is stopped. with volume depletion or hemodynamic instability. Patients
In patients chronically taking aspirin or other antithrombotic agents, who develop oliguria are o ten hypovolemic and should be
these medications should be held as soon as it is known that a given a uid challenge. I they respond avorably with an
biopsy will be per ormed (ideally 1-2 weeks prior to the procedure) improvement in urine output or hemodynamic parameters,
and should not be resumed until 1 to 2 weeks a ter the procedure. more uid challenges can be attempted.
Heparin should be stopped at least 6 hours prior to the biopsy, and 5. Avoidance o nephrotoxic agents. Concomitant use o
held or at least 12 to 24 hours postbiopsy. nephrotoxic medications is a risk actor or the development
o postoperative AKI. I iodinated contrast agents must be used
Indications for dialysis or diagnostic or therapeutic purposes, the smallest amount o
Dialysis is initiated to prevent and treat the li e-threatening compli- nonionic iso-osmolar volume o contrast should be used. Other
cations and uremic symptoms associated with severe AKI. Generally drugs such as NSAIDs, aminoglycosides, and amphotericin B
accepted indications to start dialysis in the setting o AKI include (1) should be avoided i possible. Patients who take ACE inhibitors
severe metabolic acidosis; (2) hyperkalemia, especially i electrocar- or ARBs on a chronic basis should discontinue these medica-
diographic abnormalities are present; (3) volume overload re ractory tions prior to surgery, since chronic ACE inhibition reportedly
to the use o diuretics; and (4) uremic signs and symptoms, such as increases the risk o postoperative AKI.
pericarditis, altered mental status, or seizures. The optimal timing 6. Pharmacologic agents. Several agents, including dopamine,
o dialysis initiation has not been well established. Although a ew enoldapam, atrial natriuretic peptide, mannitol, calcium-channel
retrospective and nonrandomized trials have ound that earlier ini- blockers, and loop diuretics, have been tested or their ability
tiation o dialysis may improve survival, these results have yet to be to prevent postoperative AKI. The results o these studies are
tested in a large prospective randomized clinical trial. inconclusive, and there is insuf cient evidence to recommend
their use at this time.
■ SPECIFIC SYNDROMES
Postoperative renal failure Hepatorenal syndrome
Postoperative AKI resulting in oliguria and an elevated serum cre- Hepatorenal syndrome (HRS) is a unctional orm o AKI that occurs
atinine is one o the most common and serious complications o primarily in patients with cirrhosis and ascites. The pathophysiology
surgery, representing 18% to 47% o all cases o hospital-acquired o HRS is thought to be due to nitric oxide–induced vasodilation o
AKI. It is associated with a higher risk or serious in ections and sepsis, the splanchnic circulation, leading to marked intrarenal arterial vaso-
greater costs o hospitalization, and increased mortality ollowing constriction and a reduction in GFR. There are two types o HRS: type
both cardiac and noncardiac surgery. Up to 30% o patients under- 1 HRS is the rapidly progressive orm o the disease characterized
going cardiovascular and thoracic surgeries develop postoperative by a doubling o the initial serum creatinine level to greater than
AKI, and up to 7% o these patients need renal replacement therapy. 2.5 mg/dL over a period o less than 2 weeks. The prognosis o
Furthermore, postoperative AKI that requires dialysis carries an inpatients with type 1 HRS without liver transplantation is generally
hospital mortality rate o 60% to 80%. very poor. Type 2 HRS is a more moderate orm o renal ailure
The most common cause o postoperative AKI is ischemic ATN characterized by serum creatinine levels between 1.5 and 2.5 mg/dL
resulting rom decreased renal per usion during surgery. Common and associated with a more indolent course and improved survival
surgical scenarios or the development o ATN include supra- or compared to type 1 HRS. Both type 1 and type 2 HRS can occur
in rarenal aortic cross-clamping in vascular surgery and cardiopul- spontaneously or develop a ter a precipitating event, most com-
monary bypass during cardiac surgery. Risk actors or the develop- monly a bacterial in ection such as spontaneous bacterial peritonitis
ment o postoperative AKI include preexisting renal dys unction, (SBP). Diagnostic criteria or HRS were recently revised by the Inter-
diabetes mellitus, advanced age (>65), major vascular surgery, car- national Ascites Club (IAC) in 2015, taking into account newer de ni-
diopulmonary bypass times greater than 3 hours, and recent expo- tions o AKI, and now include the ollowing: (1) cirrhosis with ascites;
sure to nephrotoxic agents including contrast dyes, NSAIDs, and (2) increase in serum creatinine o ≥0.3 mg/dL within 48 hours, or
406
≥50% increase in serum creatinine rom baseline, known or pre- shown less substantial or even insigni cant bene ts. Given
sumed to have occurred within the past 7 days; (3) no response its relatively benign side e ect pro le and low cost, however,
C
a ter two consecutive days with diuretic withdrawal and volume N-acetylcysteine is still o ten recommended as an adjunctive
H
expansion with albumin 1 g/kg body weight; (4) absence o shock; agent to IVhydration. In at-risk patients, it can be administered
A
P
(5) no current or recent treatment with nephrotoxic drugs; and as 600 or 1200 mg orally twice daily on the day be ore and the
T
(6) no macroscopic signs o structural kidney injury (proteinuria day o the procedure.
E
>500 mg/d, microhematuria with >50 red blood cells per high- 4. Diuretics. The use o diuretics, particularly mannitol and uro-
R
power eld, and/or abnormal renal ultrasonography). With proper semide, has not shown any bene t and may actually be harm-
6
medical treatment, HRS is potentially reversible. Type 1 HRS can be ul to patients.
1
treated with vasoconstrictors (eg, terlipressin, midodrine in combi- 5. Hemodialysis/hemof ltration. Although iodinated contrast
nation with octreotide, norepinephrine) combined with albumin. agents are removable by dialysis, there is currently no de nitive
A
Transjugular intrahepatic portal shunt (TIPS) may be considered in evidence to suggest that prophylactic hemodialysis or hemo l-
s
s
e
patients with type 1 HRS with either partial response (decrease in tration reduces the incidence o contrast-induced AKI.
s
s
serum creatinine to ≥0.3 mg/dL above the baseline value) or no
m
response (no regression in AKI) to medication. There is currently no
e
Drug toxicity
n
de nitive evidence demonstrating a bene t to using vasoconstric-
t
Therapeutic agents requently cause AKI in the hospital setting.
a
tors in patients with type 2 HRS. In patients being treated or SBP,
n
The clinician should suspect drug toxicity when there is an acute
d
prophylaxis with albumin is indicated, as this has been shown in one
rise in serum creatinine associated with the recent administration
M
randomized clinical trial to lower the incidence o HRS by 66%, with
o a drug. As with AKI, drug nephrotoxicity can be divided into
a
signi cant reductions in 30-day mortality rates. The suggested dose
n
prerenal, intrarenal, and postrenal mechanisms. The most common
a
o albumin is 1.5 mg/kg body weight on the rst day, ollowed by 1
g
mechanisms involve direct renal tubular injury resulting in ATN or
e
mg/kg body weight on the third day. Liver transplantation remains
m
renal interstitial in ammation leading to AIN. Other orms o injury
the treatment o choice or both type 1 and type 2 HRS.
e
include tubular obstruction due to drug precipitation, alterations
n
t
in intrarenal blood ow, and, less commonly, glomerular disease.
o
Contrast induced nephropathy Drugs that are commonly associated with nephrotoxicity and their
P
primary mechanisms o toxicity are listed in Table 61-5.
a
Contrast-induced nephropathy is one o the most common causes
t
i
e
o AKI in the hospital setting, with incidence rates ranging rom <5%
n
to >30%. Contrast-induced AKI is commonly de ned as an increase
t
s
in serum creatinine (either an absolute increase o 0.5 mg/dL or a
w
i
25% increase rom baseline) within the rst 24 hours a ter contrast TABLE 61-5 Nephrotoxic Drugs in Acute Kidney Injury
t
h
exposure. The mechanism o injury involves renal vasoconstriction,
R
Nephrotoxicity Drugs
e
impaired vasodilation, medullary hypoxia, and direct tubular cell
n
Prerenal ACE inhibitors
a
damage. Preexisting renal impairment (eGFR < 60 mL/min) and
l
D
diabetes mellitus are the most important risk actors or contrast- Angiotensin receptor blockers
i
s
induced AKI, though heart ailure, hypovolemia, nephrotoxic drugs, Cyclosporine
e
a
and hemodynamic instability are also signi cant risk actors.
s
IL-2
e
A number o preventive strategies have been studied in patients
Iodinated contrast agents
at risk or contrast-induced AKI:
NSAIDs
1. Type o contrast agent. The choice o contrast agent is
important, since higher-osmolar agents are associated with Tacrolimus
greater nephrotoxicity. In high-risk patients, nonionic iso- Intrarenal NSAIDs
osmolar (eg, iodixanol) and low-osmolar (eg, iohexol, ioversol, • Acute interstitial nephritis Penicillin analogues
iopamidol) contrast agents have been shown to have lower (na cillin, oxacillin)
nephrotoxicity. Cephalosporins
2. Volume expansion. Intravenous hydration is clearly bene cial Sul a drugs (sul amethoxazole,
in the prevention o contrast-induced AKI, though the optimal thiazide diuretics)
hydration uid has yet to be determined. The current evidence Ri ampin
indicates that isotonic uids (either normal saline or sodium
Cipro loxacin
bicarbonate) are more protective than hal -normal saline.
Although initial clinical trials showed a bene t to using sodium Proton-pump inhibitors
bicarbonate over normal saline, more recent evidence has not Aminoglycoside antibiotics
con rmed these ndings. The rate and timing o hydration are Amphotericin B
also unclear. I using normal saline, one possible regimen is • Acute tubular necrosis Cisplatin
1 mL/kg or 6 to 12 hours be ore the procedure, ollowed by Iodinated contrast agents
1 mL/kg or 6 to 12 hours a ter the procedure. Alternatively, i
HIVmedications (ade ovir,
using isotonic sodium bicarbonate (three 50 mL ampules each
ritonavir, teno ovir)
containing 50 mEq o sodium bicarbonate in 850 mL o 5%
dextrose in water), one possible regimen is a bolus o 3 mL/kg Postrenal Acyclovir
or 1 hour prior to the procedure, ollowed by an in usion o Analgesics
1 mL/kg or 6 hours a ter the procedure. Indinavir
3. N Acetylcysteine. Though requently used in the prevention Methotrexate
o contrast-induced AKI, N-acetylcysteine has had inconsis-
tent results in most clinical studies and meta-analyses. While ACE, angiotensin-converting enzyme; HIV, human immunode iciency virus;
some trials have reported signi cant protection, others have IL, interleukin; NSAID, nonsteroidal anti-in lammatory drug.
407
Drug-induced ATN is seen with the administration o medications administration o other nephrotoxic drugs or iodinated contrast
that are excreted primarily by the kidneys, including aminoglyco- agents, urinary tract obstruction, and underlying renal disease
side antibiotics, amphotericin B, and cisplatin. Aminoglycosides, increase the risk o chemotherapy-induced nephrotoxicity. Cis-
P
commonly prescribed or the treatment o Gram-negative bacterial platin can cause dose-related acute tubular necrosis and signi cant
A
in ections, cause dose-dependent ATN with a requency ranging hypomagnesemia due to renal magnesium wasting. AKI may be
R
rom 10% to 20%. Neomycin causes the greatest nephrotoxicity; reversible, though the repeated administration o cisplatin may
T
gentamicin, tobramycin, and amikacin cause intermediate neph- lead to chronic and irreversible kidney damage. Aggressive hydra-
I
I
rotoxicity; and streptomycin causes the least nephrotoxicity. It tion with intravenous uids, particularly isotonic normal saline, can
should be recognized that aminoglycoside toxicity may ollow oral increase urine volume and ow and reduce the risk o cisplatin tox-
administration o neomycin in cirrhotics, joint lavage a ter orthope- icity. Newer-generation platinum compounds such as carboplatin
dic procedures, and skin applications in burn patients. ATN typically and oxaliplatin are generally less nephrotoxic, but may also cause
M
develops 5 to 10 days a ter initiation o aminoglycoside treatment, acute tubular injury.
e
d
and is generally nonoliguric. The kidney injury is usually reversible Methotrexate is not nephrotoxic at low doses (<0.5-1.0 g/m 2)
i
c
with withdrawal o the drug, but renal replacement therapy may be but may have nephrotoxicity at higher doses (1-15 g/m 2). Metho-
a
l
necessary in some cases. trexate may precipitate in the tubules, causing direct tubular
C
o
AIN accounts or 3% to 15% o all drug-induced AKI. The most injury and urinary obstruction. Prophylaxis with intravenous uid
n
s
common o ending agents include NSAIDs, penicillins, cephalo- administration and urinary alkalinization reduces the potential or
u
l
sporins, sul onamides, ri ampin, cipro oxacin, and proton-pump toxicity. Dosing must be adjusted in patients with preexisting renal
t
a
t
inhibitors. While the onset o drug-induced AIN has been reported impairment.
i
o
as early as a ew days a ter a secondary exposure to a medication, Alkylating agents such as cyclophosphamide and i os amide
n
it usually occurs 7 to 14 days and as late as weeks to months a ter a are known to cause hemorrhagic cystitis and hyponatremia. I os-
primary exposure. AIN is typically reversible with withdrawal o the amide is more nephrotoxic than cyclophosphamide, and can cause
drug, though renal recovery may take weeks to months. Treatment signi cant proximal tubular dys unction, leading to a Fanconi-like
o AIN with steroids has an unclear bene t, though some case series syndrome with renal tubular acidosis and hypophosphatemia, and
suggest that a short course o prednisone (1 mg/kg/d or up to as well as distal nephron toxicity resulting in nephrogenic diabetes
4 weeks) may increase the rate o recovery. insipidus. Interleukin-2, o ten used to treat renal cell carcinoma
Drug-induced urinary obstruction generally results rom the pre- and metastatic melanoma, can cause reversible AKI by inducing a
cipitation o drugs within the renal tubules or ureters. Crystal-induced capillary leak syndrome that leads to interstitial edema and volume
AKI and nephrolithiasis may occur with acyclovir and indinavir. Certain depletion. Treatment is ocused on restoring intravascular volume
analgesics containing aspirin, phenacetin, and ca eine may cause and stabilizing hemodynamic parameters. Table 61-6 lists several
renal papillary necrosis, and with sloughing o the necrotic tissue that chemotherapeutic agents that commonly cause nephrotoxicity.
may lead to acute ureteral obstruction. Patients with drug-induced
urinary obstruction may present with symptoms o renal colic and Cardiorenal syndrome
acute urinary tract obstruction. Management involves hydration, pain Cardiorenal syndrome (CRS) describes a set o acute or chronic
control, and discontinuation o the medication, although invasive conditions involving the heart and the kidney in which dys unc-
removal o the stones may be required in severe cases. tion o one organ leads to dys unction o the other. Though it was
A number o medications are known to modulate renal hemo- previously thought that primary cardiac disease gave rise to renal
dynamics and cause a prerenal type o AKI. When renal per usion dys unction, evidence now suggests that renal impairment can
is decreased, regulation o GFR involves vasodilation o the a erent also lead to cardiac dys unction. A recently proposed classi cation
arteriole and vasoconstriction o the e erent arteriole. Drugs that system divides CRS into ve subtypes: (1) type 1 CRS (acute worsen-
inhibit these compensatory mechanisms urther impair renal per u- ing o cardiac unction leads to acute kidney injury), (2) type 2 CRS
sion and lead to AKI. These agents include NSAIDs, ACE inhibitors, (chronic abnormalities in cardiac unction lead to CKD), (3) type 3
ARBs, cyclosporine, tacrolimus, and iodinated contrast agents. NSAIDs CRS (acute worsening o renal unction causes acute cardiac dys-
inhibit the production o prostaglandins, which mediate a erent arte- unction), (4) type 4 CRS (CKD contributes to decreased cardiac unc-
riolar vasodilation. In patients with normal renal unction, this e ect is tion, ventricular hypertrophy, diastolic dys unction, and increased
largely inconsequential, but in those whose baseline renal per usion is risk o adverse cardiovascular events), and (5) type 5 CRS (a systemic
already impaired (eg, patients with heart ailure or volume depletion), condition causes both cardiac and renal dys unction). Type 1 CRS,
it can signi cantly reduce intrarenal blood ow and renal unction.
In contrast, ACE inhibitors and ARBs selectively block angiotensin
II-mediated vasoconstriction o the e erent arteriole. An increase
in serum creatinine o up to 30% is acceptable with ACE inhibitors TABLE 61-6 Chemotherapeutic Agents and Mechanisms
and ARBs, given the proven long-term renal protective e ects o o Toxicity
these medications, but more signi cant loss o renal unction may be Chemotherapeutic Agent Mechanism of Toxicity
observed in patients with decreased renal per usion or renovascular
Alkylating agents Tubular injury, renal
disease. Cyclosporine and tacrolimus, calcineurin inhibitors widely magnesium wasting
used as immunosuppressants, cause intense a erent and e erent • Cisplatin, carboplatin,
oxaliplatin Hemorrhagic cystitis,
arteriolar vasoconstriction. Most patients taking these medications
• Cyclophosphamide hyponatremia
experience a reduction in GFR within weeks to months o starting
therapy. As this e ect is generally reversible and thought to be dose • I os amide Proximal tubular dys unction
related, cyclosporine-, or tacrolimus-induced AKI can usually be man- Antimetabolites • Crystal-induced tubular
aged with dose reduction. • Methotrexate injury with high-dose
treatment
Chemotherapy induced nephrotoxicity Biological response modi iers • Capillary leak syndrome and
Several agents used to treat cancer are toxic to the kidneys • Interleukin-2 volume depletion
and may cause AKI. Intravascular volume depletion, simultaneous
408
which is a common occurrence, is most relevant to the discussion
o AKI. Patients with type 1 CRS present with acute heart ailure TABLE 61-7 Staging o Chronic Kidney Disease
C
that leads to the development o AKI, due to a reduction in renal
H
Stage Description GFR (mL/min/1.73 m 2)
per usion. AKI tends to be more severe in patients with acute heart
A
1 Kidney damage with ≥90
P
ailure with systolic dys unction compared to those with diastolic
normal or ↑ GFR
T
dys unction. The early diagnosis o type 1 CRS is dif cult, since at
E
the time when an elevation in serum creatinine is detected, kidney 2 Kidney damage with 60-89
R
injury has already occurred, and little can be done therapeutically. mild ↓ GFR
6
O ten, patients with type 1 CRS develop a decreased responsiveness 3 Moderate ↓ GFR 30-59
1
to diuretic therapy, and the use o higher doses or combinations o 4 Severe ↓ GFR 15-30
diuretics can worsen the AKI. Patients with volume overload who are 5 Kidney ailure <15 or dialysis
A
re ractory to diuretics may need uid removal through ultra ltration.
s
s
e
Potent vasodilating medications used in heart disease, such as GFR, glomerular iltration rate.
s
s
hydralazine and calcium channel blockers, may mani est as edema,
m
decreased urinary salt and water excretion, azotemia, and diuretic
e
n
resistance. This syndrome occurs primarily in the patient with CKD contacted upon admission and discharge. This promotes commu-
t
a
or renovascular disease and can be thought o as a renal “steal” nication and acilitates continuity o care, o ten providing the hos-
n
d
syndrome. Many drugs used in a cardiac setting are excreted by pitalist with the most current patient in ormation, including patient
M
the kidney and may reach toxic systemic levels in renal disease. history, medication regimen, vascular access history, and baseline
a
These include digoxin, procainamide, and morphine. Blood pres- parameters such as blood pressure and estimated dry weight. Sec-
n
a
sure reduction may have paradoxical e ects on cardiac and renal ond, admission orders should take into account the special needs
g
e
unction. For many reasons, management o patients with CRS is o patients with CKD and ESRD. Vital signs should include regular
m
challenging, and involvement o a multidisciplinary team consisting blood pressure measurements, daily weights, and accurate mea-
e
n
o nephrologists, cardiologists, critical care physicians, and cardiac surements o intake and output. Unnecessary phlebotomy should
t
o
surgeons is recommended. be avoided, particularly in patients with anemia, and routine blood
tests in dialysis patients can o ten be drawn at their dialysis sessions
P
a
Rapidly progressive glomerulonephritis just prior to initiation. Third, all measures should be taken to protect
t
i
e
the vascular access o ESRD patients who are receiving hemodialysis.
n
Rapidly progressive glomerulonephritis (RPGN) is characterized by
t
Blood pressures and blood draws should be per ormed in the arm
s
the acute onset o glomerular in ammation and progressive loss o
w
renal unction over a short period o time (days to weeks to months). contralateral to the one with the vascular access. I blood must be
i
t
drawn rom the ipsilateral arm, it should be taken rom the most
h
Crescent ormation within injured glomeruli is one o the pathologic
distal vein possible, pre erably rom the dorsum o the hand. Given
R
hallmarks o this disease process. Patients may present with hyper-
e
their increased risk o in ection, hemodialysis catheters should be
n
tension, azotemia, oliguria, proteinuria, and edema. The urinary
a
reserved or dialysis use only, and under no circumstances with the
l
sediment is typically active, with dysmorphic red blood cells and
D
red blood cell casts. RPGN is classi ed into three categories based exception o li e-threatening emergencies should a dialysis catheter
i
s
be accessed or other purposes.
e
on the cellular mechanism and immuno uorescence pattern: type
a
Patients with CKD requently have altered drug metabolism due
s
1 (anti-GBM disease, linear pattern o IgG staining), type 2 (immune
e
complex disease, granular pattern o IgG staining), and type 3 to changes in glomerular ow and ltration, tubular reabsorption
(pauci-immune disease, little or no immuno uorescent staining). and secretion, and renal bioactivation and metabolism. In addition,
Serological tests (eg, anti-GBM antibody, antineutrophil cytoplas- other actors such as drug absorption, bioavailability, distribution
mic antibodies [ANCAs], antinuclear antibody [ANA], complement volume, and protein binding may also be altered and can in u-
levels) should be ordered, though de nitive diagnosis requently ence the handling o medications. Inappropriate dosing can result
requires renal biopsy. The diagnosis o RPGN should be considered in either drug toxicity or ine ectiveness. On hospital admission,
a nephrologic emergency, and a nephrology consultation should the complete medication list should be care ully reviewed, and
be requested immediately to assist with renal biopsy and initiate particular attention should be given to medications that produce
appropriate treatment. long-lasting active metabolites in the setting o reduced renal
clearance. All medications, especially those that are initiated during
the hospitalization, should be appropriately dosed according to a
CHRONIC KIDNEY DISEASE
patient’s reduction in GFR.
CKD a ects approximately 13% o all adults in the United States. Nutrition is a vital part o the care o the hospitalized CKD patient.
The Kidney Disease Outcomes Quality Initiative (K/DOQI) program I available in the hospital, a registered renal dietitian can be o
de nes CKD in adults as either (1) evidence o structural or unc- tremendous value in providing dietary recommendations during
tional kidney abnormalities, such as albuminuria, abnormal urinaly- the hospitalization as well as in counseling patients on healthy
ses, abnormal renal imaging, with or without decreased glomerular eating habits ollowing discharge. According to the K/DOQI guide-
ltration rate (GFR); or (2) decreased GFR persisting or more than lines, patients with CKD stages 1 to 4 should be on a low-sodium
3 months. The National Kidney Foundation has strati ed CKD into (<2000 mg/d) diet, and potassium and phosphorus intake should
ve stages o severity (Table 61-7). CKD and ESRD are associated be adjusted according to lab values. Provided that urine output
with signi cant complications, including anemia, hypertension, is normal, there is no restriction on the amount o uid in these
bone disease, and acid–base and electrolyte disturbances, all o patients’ diets. Patients with ESRD should be placed on a diet that is
which are requently encountered in the hospital setting. low in potassium (2000–3000 mg/d), low in phosphorus (800-1000
mg/d), and low in sodium (<2000 mg/d). Fluid intake should be
■ GENERAL INPATIENT MANAGEMENT limited to 1.5 to 2 liters daily to prevent large increases in inter-
The management o a hospitalized patient with CKD or ESRD should dialytic weight gain. Certain water-soluble vitamins are lost during
be guided by a number o important general principles. First, the hemodialysis and can be replaced with a daily multivitamin such as
patient’s nephrologist and dialysis unit, i applicable, should be Diatx ZN (Pamlab, LLC), Dialyvite 3000 (Hillestad Pharmaceuticals),
409
Nephplex Rx (Nephro-Tech, Inc.), Nephrocaps (Fleming Company), setting. Patients with CKD or ESRD with hyperphosphatemia should
and Nephro-Vite Rx (Watson). be placed on low-phosphorus diets (<800-1000 mg/d) and coun-
seled to limit their intake o oods that are high in phosphorus, such
P
Hypertension as dairy products, meats, dried beans and peas, and cola drinks.
A
An estimated 50% to 75% o patients with a GFR <60 mL/min/1.73 m 2 Hyperphosphatemia that cannot be adequately controlled by
R
(CKD stages 3-5) have hypertension, and as renal unction declines, dietary modi cation alone should be treated with oral phosphorus-
T
hypertension becomes increasingly prevalent. Given the higher risk binding agents. Oral aluminum hydroxide, historically the rst agent
I
I
o cardiovascular morbidity and mortality associated with hyper- made available to treat hyperphosphatemia, is rarely used these
tension and CKD, the National Kidney Foundation Clinical Practice days because o its long-term risk o aluminum toxicity and osteo-
Guidelines or Hypertension recommend that in all patients with malacia. It has been largely replaced by the calcium-containing
hypertension and CKD, blood pressure should be targeted to a sys- (calcium acetate, calcium carbonate, and calcium citrate) and non–
M
tolic value o <130 mm Hg and a diastolic value o <80 mm Hg to calcium-containing phosphorus binders (sevelamer hydrochloride,
e
d
decrease the risk o cardiovascular events and delay the progression sevelamer carbonate, and lanthanum carbonate). When admin-
i
c
istered with meals, these medications inhibit the gastrointestinal
a
o CKD. ACE inhibitors and ARBs should be considered as rst-line
l
absorption o phosphorus; thus they are not e ective at lowering
C
therapy or hypertension in CKD, given their antiproteinuric e ects
o
and long-term renoprotective e ects. Diuretics, particularly loop serum phosphorus levels in patients not receiving any dietary intake.
n
s
diuretics, can be particularly use ul in optimizing blood pressure. Calcium acetate has been demonstrated in a number o studies to
u
l
be more cost-e ective than sevelamer. However, in patients who
t
Loop diuretic doses should be titrated upward as tolerated until
a
t
normalization o blood pressure is achieved or the patient develops develop extraskeletal calci cations or recurrent hypercalcemia rom
i
o
calcium-containing phosphorus binders, sevelamer and lanthanum
n
symptoms or signs o overly aggressive diuresis (eg, lightheaded-
ness, hypotension, rising BUN and creatinine). The e ectiveness o are suitable, though more expensive, alternatives. Calcium-containing
thiazide diuretics decreases in patients with a GFR <30 mL/min; phosphorus binders and sevelamer or lanthanum can also be used
however, these medications can be used synergistically with loop in combination to treat hyperphosphatemia that is dif cult to con-
diuretics to improve diuresis in patients with re ractory edema. trol with a single agent. Sevelamer hydrochloride has been associ-
Patients with ESRD on hemodialysis should have their morning ated with metabolic acidosis; in these patients, substituting with
doses o blood pressure medications held on dialysis days to pre- sevelamer carbonate may be o bene t, as this ormulation does not
vent intradialytic hypotension and acilitate volume removal during decrease serum bicarbonate levels.
dialysis.
Acid–base and electrolytes
Anemia Patients with CKD may have a metabolic acidosis due to impaired
acid secretion. Sodium bicarbonate should be administered to
Normocytic, normochromic anemia is a common complication
patients with serum bicarbonate concentrations <22 mEq/L to
o CKD and ESRD. It is primarily due to a de ciency in erythropoi-
prevent the complications o chronic metabolic acidosis, speci cally
etin production by the kidneys, though other contributing actors
bone disease and loss o lean body mass due to increased break-
may include iron de ciency, shortened red blood cell survival,
down o skeletal muscle. Sodium bicarbonate can be given as oral
uremic inhibitors o erythropoiesis, hemolysis, bleeding, loss o
tablets (650 mg [7 mEq] twice daily with meals) or alternatively
blood in hemodialysis circuits, and repeated blood draws. Anemia
in the orm o baking soda (1/2 to 1 teaspoon dissolved in water
becomes more common as GFR decreases to <60 mL/min/1.73 m 2.
or juice twice daily with meals). Patients may experience some
Treatment o anemia in CKD patients improves quality o li e and
abdominal bloating with bicarbonate treatment. Citrate salts should
decreases mortality. The K/DOQI guidelines recommend that the
not be used as alkalinizing agents in CKD, as they may increase alu-
hemoglobin target in dialysis and nondialysis patients with CKD is
minum absorption.
generally in the range o 11.0 to 12.0 g/dL. Treatment with an
Electrolyte disorders are also common in patients with CKD and
erythropoiesis-stimulating agent such as erythropoietin or darbe-
ESRD. When GFR decreases to <15 to 20 mL/min, renal potassium
poietin al a reduces the need or requent blood trans usions and
excretion is impaired and hyperkalemia may occur. In patients
is recommended in anemic CKD patients. In nondialysis patients,
who still produce adequate urine output, acute hyperkalemia can
levels above this target should be avoided, due to recent evidence
usually be managed medically with calcium gluconate (i elec-
demonstrating that these levels are associated with adverse cardio-
trocardiographic changes are present), insulin, inhaled albuterol,
vascular outcomes. There ore, the hemoglobin target in dialysis and
potassium-binding resins (eg, sodium polystyrene sul onate), and
nondialysis patients with CKD should not be >13.0 g/dL.
loop diuretics. Patients with ESRD and oliguria or anuria will o ten
To ensure that anemic patients will respond to treatment with
require dialysis therapy to treat hyperkalemia. Chronic management
erythropoietin or darbepoietin, iron stores should be monitored
o hyperkalemia involves dietary potassium restriction. Loop diuret-
regularly via the serum erritin concentration, serum iron concentra-
ics and potassium-binding resins are usually not necessary but can
tion, and total iron binding capacity. Iron de ciency in patients with
be use ul or long-term control. I resins are used, it should be noted
CKD is de ned as trans errin saturation (TSAT) <20% or serum erritin
that they can result in hypocalcemia, sodium overload, and malab-
<100 ng/mL. Patients who meet either o these criteria should be
sorption o other medications. When given as a retention enema,
given iron supplementation, either orally (eg, errous sul ate 325 mg
they can cause colonic ulceration. These resins should not be given
three times daily) or intravenously (eg, iron sucrose, iron gluconate)
with aluminum hydroxide gels. Hypokalemia is less common in
to maintain a TSAT >20% to 25% and serum erritin between 200
patients with CKD but can be caused by low-potassium intake,
and 500 ng/mL.
diuretic use, or gastrointestinal losses.
The ability o the kidney to properly concentrate or dilute urine
Bone metabolism is reduced as renal unction is progressively lost, and both hypo-
Renal phosphorus excretion is decreased in patients with CKD and natremia and hypernatremia are common in patients with CKD.
can result in elevated serum phosphorus levels and lower serum cal- Hyponatremia may be due to impaired ree water clearance or
cium levels due to increased binding o phosphorus. Serum calcium volume depletion through renal or extrarenal sodium losses. A care-
and phosphorus levels should be ollowed regularly in the inpatient ul assessment o volume status can guide appropriate treatment.
410
Patients who are euvolemic or hypervolemic will usually ben- SUGGESTED READINGS
e t rom ree water restriction and occasionally diuretics, whereas
C
patients who are hypovolemic may require administration o intra- Angeli P, Gines P, Wong F, et al. Diagnosis and management o acute
H
venous normal saline. Hypernatremia maybe due to impaired water
A
kidney injury in patients with cirrhosis: revised consensus recom-
P
intake, in patients with poor thirst mechanisms or decreased access mendations o the International Club o Ascites. Gut. 2015;64:531.
T
to water, or excessive renal or extrarenal water losses. Hypernatre- Bellomo R, Kellum JA, Ronco C. Acute kidney injury. Lancet.
E
mia may also accompany recovery rom AKI, during the osmotic
R
2012;380:756-766.
diuresis o high levels o urea. Patients should be given ree water,
6
Coresh J, Selvin E, Stevens LA, et al. Prevalence o chronic kidney
either orally or intravenously, to correct the water de cit.
1
disease in the United States. JAMA. 2007;298:2038-2047.
Lamiere NH, Bagga A, Cruz D, et al. Acute kidney injury: an increasing
A
PRACTICE POINT global concern. Lancet. 2013;382:170-179.
s
s
e
The management o a hospitalized patient with CKD or ESRD Perazella MA. Onco-nephrology: renal toxicities o chemotherapeu-
s
s
should be guided by the ollowing general principles tic agents. Clin J Am Soc Nephrol. 2012;7:1713-1721.
m
e
• Ronco C, Di Lullo L. Cardiorenal syndrome. Heart Fail Clin.
n
The patient’s nephrologist and dialysis unit, i applicable, should
t
be contacted upon admission and discharge. 2014;10:251-280.
a
n
• Admission orders should take into account the special needs o Sharp VJ, Barnes KT, Erickson BA. Assessment o asymptomatic micro-
d
patients with CKD and ESRD. scopic hematuria in adults. Am Fam Physician. 2013;88:747-754.
M
•
a
All measures should be taken to protect the vascular access o Tan KT, van Beek EJ, Brown PW, et al. Magnetic resonance angiogra-
n
a
ESRD patients who are receiving hemodialysis. phy or the diagnosis o renal artery stenosis: a meta-analysis. Clin
g
•
e
All medications, especially those that are initiated during the Radiol. 2002;57:617-624.
m
hospitalization, should be appropriately dosed according to a
e
n
patient’s reduction in GFR.
t
o
P
a
t
PRACTICE POINT
i
e
n
• In ESRD patients with little or no urine production, oral uid
t
s
intake should be closely monitored and restricted to 1 to
w
i
1.5 L/d. Large interdialytic weight gains (>4-5 kg) due to liberal
t
h
uid consumption or administration o intravenous uids and
R
e
medications can make volume removal during dialysis more
n
a
dif cult.
l
D
i
s
e
a
s
e
411
62
Neurologic and psychiatric complications are o ten encountered
in the postoperative period and may be very alarming. Hospitalists
must be able to recognize and initiate treatment or many o these
postoperative complications. The di erential diagnosis o common
problems (such as headache) may also be very di erent in the peri-
operative period. This chapter will cover common presentations,
risk actors and prevention techniques, and management o several
postoperative neurologic and psychiatric conditions, including:
Postoperative seizures, delirium, con usion, delayed emergence, muscle weakness
by anesthetic drugs, stroke, blindness, awareness under anesthesia,
Neurologic cognitive dys unction, headache, spinal cord injury, and peripheral
nerve injury.
and Psychiatric SEIZURES
Complications
Seizures in the postoperative are generally rare but require immediate
treatment. Dangers o seizures include hypoventilation, hypoxemia,
musculoskeletal injury, aspiration, and death.
■ RISK FACTORS
Catherine Dawson Tobin, MD There are many di erent causes o seizures in the perioperative
Michel J. Sabbagh, MD period, including hypoglycemia rom prolonged NPO status, local
John Scott Walton, MD anesthetic systemic toxicity (LAST) rom intravascular injection o
local anesthetic such as bupivacaine, and electrolyte abnormalities.
Tod A. Brown, MD Patients with intracranial structural lesions, patients having intracra-
nial surgery, and patients with traumatic brain injury (TBI) are also at
risk or seizures.
Most anesthetics drugs are antiepileptic. However, some anes-
thetic drugs can induce seizure oci or at least cause changes in
the EEG while in using. These include ketamine, methohexital, and
meperidine. O note, methohexital and etomidate are o ten used as
general anesthesia on patients prior to electroconvulsive therapy
(ECT) where a seizure is purposely induced.
■ MANAGEMENT
Stopping the seizure is the initial and most important treatment.
Airway management and protecting against head or body injuries
is also important. Pharmacologic adult intravenous therapy includes
benzodiazepines such as midazolam 1 to 5 mg, diazepam 5 to 10 mg,
propo ol 50 to 100 mg, and phenytoin 500 to 1000 mg (in used
slowly). I LAST is the cause, Intralipid (20% at emulsion) bolus
1.5 mL/kg over 1 minute (about 100 cc) ollowed by continuous
in usion o 0.25/kg/min is the treatment.
PRACTICE POINT
• In patients who have had a peripheral nerve block or who have
a continuous nerve catheter or pain management, LAST as the
cause o seizure must be considered.
• Intralipid 20% therapy is used to treat LAST.
Electrolyte abnormalities and hypoglycemia must be corrected in

the treatment o seizures. Sodium abnormalities leading to seizures are
most common in neurosurgical patients, and are most o ten seen in
conjunction with the Syndrome o Inappropriate Antidiuretic Hor-
mone (SIADH), Diabetes Insipidus (DI), or due to hypertonic therapy
412
to treat increased intracranial pressures. Seizures can present in antipsychotics are a better choice. Also cholinesterase inhibitors
any o those conditions but are more likely when sodium levels are should not be newly prescribed to treat delirium. For more in orma-
C
<120 mM or >158 mM. tion, see Chapter 81: Delirium.
H
Management includes ruling out other conditions that are not
A
P
seizures. I it is unclear i seizure activity is present, an electroen-
PRACTICE POINT
T
cephalogram (EEG) should be obtained in consultation with neurol-
E
ogy. Pseudoseizures or psychogenic non epileptic seizures must • Do not use benzodiazepines as a rst line treatment o
R
remain on the di erential in the management. Characteristics o agitation seen in delirium.
6
pseudoseizures include asynchronous episodes which last about •
2
Antipsychotics, such as haloperidol, risperidone, or olanzapine,
90 seconds o shaking, orced eye closure, and retained pupil at the lowest e ective dose or the shortest amount o time
response; autonomic mani estations are usually absent such as possible should be used i patient is agitated and at risk to
P
urination, de ecation, cyanosis, and tachycardia. Pseudoseizures are
o
harm themselves or others.
s
important to recognize because iatrogenic injury rom respiratory •
t
Cholinesterase inhibitors should not be newly prescribed to
o
p
depression o antiseizure drugs and unnecessary endotracheal intu- treat or prevent postoperative delirium.
e
bation can be harm ul. Postoperative shivering can also resemble a •
r
Pain control is important in treatment and prevention o
a
t
seizure; it may be treated with low-dose lorazepam, oxygen therapy
i
delirium.
v
e
and rewarming.
N
e
u
DELIRIUM CONFUSION, DELAYED EMERGENCE, AND MUSCLE
r
o
WEAKNESS BY ANESTHETIC DRUGS
l
Postoperative delirium may a ect about 25% to 30% o patients
o
g
greater than age 65. Two main subtypes o delirium are hypoactive Many anesthetic drugs used in the perioperative period can cause
i
c
(decreased motor activity and withdrawn behavior) and hyperactive con usion, delayed emergence, and muscle weakness. Although
a
n
(agitated and possibly aggressive behavior). Patients with delirium these symptoms can mimic a stroke, they are most likely caused
d
o ten have increased hospital length o stay and are at risk or harm by over sedation and reversal agents should be tried (naloxone or
P
s
to themselves or others. Postoperative delirium in the elderly is opioids and umazenil or benzodiazepines).
y
c
preventable in approximately 40% o cases. There ore, risk reduction Ketamine is a dissociate anesthetic which is commonly used in
h
i
strategies are important. anesthesia. It causes sedation and treats pain, but patient’s airway
a
t
r
re exes remain intact and the patient continues breathing. It has
i
c
■ RISK FACTORS AND PREVENTION properties similar to the drug phencyclidine (PCP). Patients o ten
C
o
Risk actors or postoperative delirium are older age, history o have emergence delirium or color ul dreams a ter this drug is used.
m
Neuromuscular blocking agents are commonly used in dur-
p
dementia, and prior hearing or visual de ects. Other contributing
l
ing surgery to aid in endotracheal intubation and also to keep
i
causes include ever, stress o surgery, pain, emesis, sleep depriva-
c
a
tion, and loss o regular routine. The incidence can be reduced by a patient still during surgical cases where movement could be
t
i
o
using lighter anesthesia, using regional versus general anesthesia, dangerous. Normally the degree o neuromuscular blockade can
n
be checked with a peripheral nerve stimulator placed over the
s
using “ ast track” anesthesia (early extubation and ambulation), hav-
ing good pain control, and avoiding benzodiazepines. ulnar or posterior nerve. A train o our or tetanus can be seen
Certain anesthetic drugs are best avoided, especially in the depending on the setting o the device. The depolarizing muscle
elderly, because they contribute to delirium. These drugs include relaxant succinylcholine is short acting and does not require rever-
anticholinergics (promethazine, oxybutynin, scopolamine), antihis- sal; however, the use o high-dose succinylcholine can result in a
tamines (diphenhydramine, hydroxyzine, histamine 2 blockers), and rare condition called a Phase II block where the patient is paralyzed
meperidine. or hours. Additionally, i the patient has a pseudocholinesterase
Anticholinergics such as atropine (used to treat bradycardia) and de ciency, succinylcholine can last hours. This condition is also
scopolamine patches (to prevent postoperative nausea) can cause rare and can be tested with a dibucaine test. In these situations, i
signi cant con usion, as both cross the blood brain barrier. Other a prolonged paralyzed state ensues, endotracheal intubation and
side e ects include dry mouth, tachycardia, pupil dilation, and pos- sedation are needed. More commonly, nondepolarizing blocking
sible hallucinations. I present, con usion/delirium rom these agents agents such as rocuronium, cisatracurium, or vecuronium are used
can be reversed by the anticholinesterase inhibitor physostigmine and require reversal with an anticholinesterase inhibitor, such as
at a dose o 0.01 to 0.03 mg/kg (average adult doses range rom neostigmine, paired with an anticholinergic such as glycopyrro-
1 to 5 mg). Glycopyrrolate is a drug used to treat bradycardia and late. I reversal drugs are not given or i a ull amount is not given,
decrease secretions, and can be used in reversal o neuromuscular patients are not ully reversed and can have a “ oppy sh” appear-
drugs along with neostigmine. It does not cross the blood brain bar- ance; they o ten are not ully able to move their body, or they try to
rier and may be a better choice than atropine in an elderly patient. talk but cannot, and they look in distress. In rst hour a ter surgery
it is important to place a peripheral nerve stimulator or make sure
■ MANAGEMENT the patient has had ull pharmacologic reversal o neuromuscular
blockers, i they were used.
Management o delirium should include an evaluation to rule out
an intracranial cause such as a stroke. Brain imaging may be indi-
■ MANAGEMENT
cated. I hypoactive delirium is seen one must consider oversedation
caused by narcotics or benzodiazepines, which can be reversed by The ollowing treatment algorithm or delayed emergence is
naloxone or umazenil respectively. However, pain control is very recommended:
important in the postoperative management o delirium so reversal 1. Check vital signs (heart rate, blood pressure, temperature, oxygen
o opioids should be care ully done. saturation)
Treatment o delirium should not include benzodiazepines as the 2. Per orm a neurologic exam including pupils and response to
rst choice. These drugs are commonly requested by nursing sta pain
and education o the care team is o ten needed. Instead low-dose 3. Check twitches with a peripheral nerve monitor
413
4. Check nger stick glucose PRACTICE POINT
5. Draw an arterial blood gas with electrolytes
1) Maximize preoperative health prior to surgery
6. Consider pharmacologic reversal with naloxone, umazenil,
2) Make sure all previous stokes and TIAs have been thoroughly
P
and physostigmine
A
7. Make arrangements or brain imagining (head CT) evaluated and treated
R
3) Consider delaying elective surgery in those with a recent TIA/
T
CVA (1-3 months)
I
STROKE 4) Develop a coordinated anticoagulation plan with surgery,
I
Stroke is an in requent but tragic complication o surgery. Periop- anesthesia and neurology that is well communicated to each
erative stroke is de ned as a cerebral in arction up to 30 days a ter service and the patient.
surgery. Perioperative stroke rates range rom 0.1% to 10% depend- 5) Rapid diagnosis (neurology consult and brain scan) o
M
ing on how the stoke is diagnosed and the type o surgery. Cardiac, postoperative ocal de cits enable the possibility or early
e
d
vascular and neurosurgery have the highest rates o neurologic endovascular therapy in the immediate (3-6 hour) period
i
c
consequence. Covert stroke (de ned by MRI) occurs much more 6) Many strokes do not mani est themselves immediately
a
l
requently than overt stroke (de ned by clinical symptoms). postoperatively
C
o
n
■ RISK FACTORS AND PREVENTION
s
u
l
t
Traditionally reducing stroke risk has largely ocused on maximizing POSTOPERATIVE BLINDNESS
a
t
a patient’s baseline neurologic health and de ning preoperative
i
Blindness a ter surgery is another rare but tragic complication. It
o
n
stroke risk prior to surgery. Risk actors change slightly depend- occurs in <0.2% o spinal surgeries, and 0.0186% o hip replacement
ing on which studies are cited but, advanced age, previous TIA or and knee arthroplasty procedures; in other procedures it is exceed-
stroke and renal ailure are consistently associated with increased ingly rare (1/10 to 1/2 this rate). There are three di erent types o
risk o perioperative stroke. Other associated risk actors are emale blindness a ter surgery with some additional rare variations. The
sex, cardiac disease, atrial brillation, hypotension, smoking, recent most common is ischemic optic neuropathy (ION), which has been
beta blocker cessation, recent statin cessation, anemia, dehydration, described most requently a ter prone positioned spine surgery and
hypercoagulable state, recent MI, and CHF. Modi able comorbidities cardiac surgery. ION appears to be a type o compartment syndrome
should be optimized prior to surgery. For example, patients with surrounding the optic nerve as a result o increased venous pres-
recent TIAs should undergo a complete evaluation with an MRI, sures in the head and neck. This seems to also explain how bilateral
echocardiogram, carotid scans and neurological consultation, which radical neck procedures and robotic prostatectomies (per ormed
can appropriately guide therapy during the impending preopera- with steep head down positioning) also can be associated with
tive period. Recent TIA or stroke (within 3 months) likely warrants a ION. Another orm o blindness is retinal artery thrombosis (retinal
delay o elective surgery until post in arction/ischemic impairment vascular occlusion). It is ar less common and likely caused by direct
o cerebral auto regulation improves. pressure on the globe o the eye during surgery. Cortical blindness,
The evaluation o anticoagulation therapy is imperative or which is equivalent to a stroke involving the visual pathways in the
patients with atrial brillation and a recent TIA/CVA. Certain surger- brain, is rare and is surprisingly more common in younger patients
ies (eg, neurosurgery and spine surgery) usually require withholding and children than older adults. Lastly, direct surgical trauma to the
anticoagulation while others (eg, cataract surgery) are usually sa e visual apparatus can occur in ophthalmic, brain and sinus surgeries.
to remain on normal anticoagulants. Between these two extremes,
it is o ten necessary or the preoperative, neurological and surgi- ■ RISK FACTORS AND PREVENTION
cal teams to con er in an e ort to establish an optimal risk bene t
ION, which comprises the vast majority o postoperative blindness
regime o anticoagulation.
cases, has the ollowing risk actors in noncardiac surgery: the use
Intraoperatively the availability o neuroprotective agents has
o a Wilson rame position head rest (in noncardiac surgery), male
been elusive; barbiturates, steroids, magnesium, and hypothermia
gender, obesity, raction o crystalloid to colloid in used (higher-
have limited or no proven bene t. Patients do bene t rom avoid-
percentage possibly protective), length o surgery (>6 hours) and
ance o hyperthermia and hyper or hypoglycemia. The maintenance
high-blood loss (>1 L). Given this, prevention e orts include evaluat-
o an adequate per usion blood pressure is also intuitively reason-
ing the type o table or rame and the height o the head in relation
able although a strict de nition o adequate has not been univer-
to the heart, and the percent o colloid versus crystalloid in used.
sally accepted. A commonly agreed upon suggestion is to maintain
Staging cases (two shorter surgeries as opposed to one long sur-
20% o the patients baseline blood pressure but the absolute mini-
gery) has been proposed as a prevention strategy, but is generally
mums o blood pressure and de nitions o baseline pressure have
impractical and rarely utilized, but it should at least be considered
not been adequately de ned.
when the surgery is anticipated to last greater than 6 hours. The use
o colloids remains controversial in light o the possible deleterious
■ MANAGEMENT renal e ects in critically ill patients, so is a rarely utilized preven-
Many perioperative strokes (about 92%) do not mani est themselves tion strategy. The risk o blindness should be part o the in ormed
in the PACU but present over the rst ew postoperative days. Very consent in cases with predictable blood loss greater than 1 liter and
ew preoperative CVAs are hemorrhagic; rather, most are occlusive length o surgery greater than 6 hours. More research needs to be
rom either emboli or thrombus. Diagnosis in the early PACU period done in this area but the rare incidence makes this dif cult.
is dif cult because patients may appear to have neurologic aber-
rations and even transient ocal de cits as they ully emerge rom ■ MANAGEMENT
anesthesia. Any suspicion should evoke immediate and ongoing I blindness is suspected in the PACU, an urgent ophthalmologic
neurologic evaluations. Management requires maintenance o consultation should ensue. I the cause is not obvious on oph-
adequate per usion pressure and possible early therapeutic interven- thalmologic examination, then neuroimaging is o ten required to
tion, such as endovascular techniques shortly a ter the presentation clearly de ne the etiology. At present, there are no good thera-
o symptoms (3-6 hours). pies or postoperative blindness like ION. Steroids, mannitol and
414
antiplatelet agents have been utilized, but have not been shown to method (such as used with emergency C-sections), or the use o
be bene cial. neuromuscular blockade. There are also patient-speci c risk actors,
C
including a history o drug use, signi cant comorbidities, a history o
H
A
awareness, or the presence o a dif cult airway.
PRACTICE POINTS
P
Cardiac and trauma surgery are typically per ormed on sick and
T
• Blindness can occur without obvious breaches in medical care ragile patients who may not tolerate similar levels o anesthetics as
E
and in young healthy individuals. their healthier counterparts without the use o vasopressors, volume
R
• Patients with many risk actors should have this included in replacement or both. These lower doses o anesthetics may contrib-
6
ute to the recall. Similarly, C-sections with general anesthesia and the
2
their in ormed consent.
use o total intravenous anesthesia are more prone to inadequate
• Blindness can be dif cult to diagnose in the PACU because
patients are not ully awake. anesthesia (not acknowledging the anesthetic needs or concerns that
P
the patient may not tolerate heavier doses o anesthetics). Neuromus-
o
• Postoperative patients with risk actors should have a simple
s
cular blockade, though typically considered to be an integral part o
t
o
vision exam in both eyes in the PACU; ophthalmology should
p
an anesthetic, increases the rates o awareness under anesthesia.
be immediately consulted i any concern is discovered.
e
Anesthetic gases are typically titrated to a certain expired concen-
r
•
a
Prognosis is very poor and treatment limited.
t
tration o a gas known as an end-tidal value. This type o anesthetic
i
v
e
is based on the minimum alveolar concentrations (MAC) o these
N
agents. A MAC o an inhalational anesthetic is de ned as the end-
e
AWARENESS UNDER ANESTHESIA
u
tidal concentration o a gas at which 50% o the population will not
r
o
Anesthesia is de ned as the presence o muscle relaxation, pre- move in response to a pain ul stimulus. This value can vary based on
l
o
served physiology, analgesia and unconsciousness. Awareness age and comorbidities and is adjusted as such. Titration o anesthet-
g
i
ics to MAC values has been shown to produce sa e anesthetics with
c
under anesthesia is one o the most eared complications o anes-
a
thesia by patients. Although most patients do not have signi cant low rates o recall.
n
d
complications rom it, posttraumatic stress disorder has been There are commercially available awareness monitors, the most
P
described in patients with traumatic experiences, especially those common o which is the bispectral index (BIS). It is the only device
s
y
reporting awareness with paralysis. The identi cation o this compli- approved by the FDA to monitor intraoperative awareness. This
c
h
cation typically requires systematic ollow-up and patient interviews device acquires raw EEG data continuously which is then converted
i
a
t
ollowing the anesthetics. Patients o ten do not understand the di - by a proprietary algorithm to a value that can be used to titrate the
r
i
c
erence between a sedation case and a general anesthetic, and most anesthetic. Studies on this method o monitoring, compared to
C
episodes o recall under anesthesia occur under sedation. MAC titration o anesthetics, have varied results and are considered
o
m
Awareness under anesthesia is de ned as the recall o speci c equivocal.
p
events during general anesthesia. Incidence varies depending on Finally, the use o a balanced anesthetic, which includes other
l
i
c
multiple actors including the method used to identi y recall, the sedative medications such as opioids, benzodiazepines, anticho-
a
t
linergic agents, and even alpha-2 agonists, may decrease MAC
i
de nition used or recall, as well as the process (quality assessment
o
n
versus prospective study). Detection methods during the case requirements by adding to the level o sedation and amnesia, and
s
include the presence o movement under anesthesia, and EEG can reduce the risk o awareness.
(more below). Published studies typically use a question set known
as the Brice test postoperatively, at varying intervals ranging rom POSTOPERATIVE COGNITIVE DYSFUNCTION
the immediate postoperative period to a month or two ollowing
Anesthesia alters brain unction, which can mani est as changes in
the anesthetic (Table 62-1).
attention, memory and reaction time in the immediate postoperative
The incidence o awareness under anesthesia tends to be higher
period. These unctions return to normal at varying rates ollowing
when reported in study conditions in patients being interviewed
anesthetics. These are due to rates o elimination, length o anesthet-
using the Brice questions (up to 0.1%), when compared to sel -
ics, type o anesthesia, type o surgery, patient characteristics, and
reporting or a quality assurance questionnaire (as low as 0.005%).
any adjuncts used. Postoperative cognitive dys unction (POCD) is
According to the Fi th National Audit Project (NAP-5) rom the
considered separate rom postoperative delirium and short-term
United Kingdom published in 2014, recall occurs in roughly 1 in
cognitive disturbance. POCD is described as a change in cognition
19,000 anesthetics. High-risk populations, such as patients undergo-
targeting memory, learning and concentration, ranging rom weeks
ing cardiac surgery or C-sections under general anesthesia, have a
to months. This should be assessed with a preoperative and a post-
much higher incidence o 0.012% and 0.15%, respectively.
operative cognition examination. These examinations are o ten lim-
ited or absent due to constraints in the perioperative period. Timing
■ RISK FACTORS AND PREVENTION
o these examinations is crucial to identi ying true cognition changes.
Factors associated with awareness include the type o surgery (eg, The preoperative examination should be thorough enough that a
cardiac surgery and trauma surgery), an intravenous anesthetic baseline mentation is determined. This examination cannot be so
thorough as to lose the interest o the patient or so close to the date
o surgery that the patient is overcome by anxiety and stress. Both o
TABLE 62-1 Brice Questions these scenarios can signi cantly alter testing and baseline cognition
assessment. Similarly, postoperative examinations should ollow sim-
1. What was the last thing you remember be ore you went to ilar guidelines, and those per ormed urther out in the postoperative
sleep? period (at least 4 weeks postoperatively) are more likely to identi y
2. What was the irst thing you remember a ter your operation? persistent and signi cant mentation changes.
3. Can you remember anything in between?
4. Did you dream during your operation? ■ RISK FACTORS AND PREVENTION
5. What was the worst thing about your operation? Published data associates a 25% incidence o POCD with major
noncardiac surgery a ter 1 week, and 10% a ter 3 months, compared
415
to 3% in the control group at both time points. Risk actors include a signi cant risk (4%-10%): repair o the descending aorta, correc-
advanced age, duration o anesthesia, respiratory complications, tion o kypho-scoliosis, or other major spinal repairs. In the case
second operation, low educational level, and prior stroke. Although o descending aortic repair, the arterial blood supply to the spinal
P
anesthetic type has been believed to play a major role in POCD, cord may be compromised. In the case o spinal corrective surgery
A
studies comparing regional to general anesthesia demonstrated a multiple mechanisms or spinal cord injury are present. When
R
signi cant di erence at 1 week but no di erence at 3 months. per orming these high-risk operations, the operative team should
T
consider multiple risk reduction strategies and consider intraopera-
I
I
POSTOPERATIVE HEADACHE tive monitoring o spinal cord integrity. Spinal cord integrity can be
Postoperative headaches are most commonly due to: postdural monitored intraoperatively using Somatosensory Evoked Potential
puncture headache, ca eine withdrawal headache, pneumocepha- (SSEP) and Motor Evoked Potential (MEP) monitoring. Both these
lus, or hyperper usion syndrome. Post dural puncture headache neurophysiologic monitoring moda

Mckean Sylvia Ross John Dressler Daniel D Scheurer Danielle

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Mckean Sylvia Ross John Dressler Daniel D Scheurer Danielle

Uploaded by

Copyright:

Available Formats

Principles and Practice

John J. Ross, MD, CM, FIDSA

Daniel D. Dressler, MD, MSc, SFHM, FACP

Danielle B. Scheurer, MD, MSCR, SFHM

This book was set in Myriad pro by Cenveo Publisher Services.

RR Donnelley was printer and binder.

Library o Congress Cataloging-in-Publication Data

PART I: THE SPECIALTY OF HOSPITAL

23 Building, Growing and Managing a

SECTION 1 Surgery PART III: REHABILITATION AND SKILLED

Perioperative Risk Assessment and

51 Perioperative Pulmonary Risk Assessment and 77 Evaluation of Anemia ............................... 533

Prevention, Assessment, and 84 Dizziness and Vertigo ............................... 587

106 Weakness: How to Localize the Problem . . . . . . . . . . . . . . 763

137 Inpatient Cardiac Arrest and Cardiopulmonary

SECTION 2 Radiology 144 Flushing and Urticaria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1105

122 Arterial Blood Gas and Placement of A-line . . . . . . . . . . . . 901

John J. Ross, MD, CM, FIDSA

Daniel D. Dressler, MD, MSc, SFHM, FACP

Danielle B. Scheurer, MD, MSCR, SFHM

Samer Abdel-Aziz, MD Daniel A. Anaya, MD

Patrick Avila, MD, MPhil, MPH Tom Baudendistel, MD, FACP

Arline D. Bohannon, MD Tod A. Brown, MD

Alexander R. Carbo, MD, FACP, SFHM Kenneth K. Chen, MD, FRACP

Alexandra Columbus, MD Sonye K. Danof , MD, PhD

Michael Klompas, MD, MPH Victoria D. Lackey, MD

Makeida B. Koyi, MD Lucia Larson, MD, FACP

Svetlana Krasnokutsky, MD, MS Jodi Layton, MD

Harold Kudler, MD Brian Leber, MDCM, FRCPC

William L. Lyons, MD Alayne D. Markland, DO, MSc

Andrew McFarlane, MLT, ART, FCSMLS(D) David Meltzer, MD, PhD

Jakob I. McSparron, MD Paul A. Monach, MD, PhD

Thomas A. Owens, MD Vihas Patel, MD

Menaka Pai, MD, MSc, FRCPC Jill M. Paulson, MD

Michaella Maloney Prasad, MD Kerry Reynolds, MD

Lynn Schlanger, MD Ann M. Sheehy, MD, MS

Saul N. Weingart, MD, PhD Eric S. Winer, MD

Irene M. Yeh, MD, MPH

The Value and Values of 22 The Role of Information Technology in

1 The Face of Health Care: Emerging Issues

5 Professionalism in Hospital Medicine . . . . . . . . . . . . . . . . . . 29 25 Teamwork in Leadership and Practice-Based

SECTION 2 Critical Decision Making at the

15 Principles of Patient Safety: Intentional

Care: Emerging approximately 15 clinicians and have complicated alternating sched-

Care for Hospitalists Value =

Hospital costs represent the single largest segment o the nearly

HOSPITAL COSTS IN THE NATIONAL SPOTLIGHT

Domain Weight Measures

Areas of Waste in Annual

PRACTICE POINT ■ IMPLEMENTING VALUE-BASED INITIATIVES

Disparities in include rates o vaccination, cancer screening, secondary prevention

Hospital Care and angioplasty. Disparities in health outcomes include cardiovas-

Historically, disparities in hospital care originated in the policy o

Orthopedic Patients involvement o a medical comanager may make practical and

REVIEW AVAILABLE LITERATURE ON COMANAGEMENT

Chronic Medical Conditions

TABLE 4-3 CCF Program Timeline

Figure 4 1 Cleveland Clinic orthopedic comanagement triage algorithm.

length o hospital stay coupled with decreased readmission rates

WHAT DOES “PROFESSIONALISM” MEAN IN

PROFESSIONALISM AND PATIENT SAFETY

Principles o o leadership and concludes by discussing some o the challenges

LEADERS VERSUS MANAGERS