Decreased Emergency Department Length

of Stay by Application of a Computerized

Consultation Management System
Suck Ju Cho, MD, Jinwoo Jeong, MD, Sangkyoon Han, MD, Seokran Yeom, MD,
Sung Wook Park, MD, Hyung Hoi Kim, MD, and Seong Youn Hwang, MD

Abstract
Objectives: Consultation difficulty occurs in hospitals located in many countries, and it is understood
that strategies to improve the emergency department (ED) consultation process are needed. The authors
constructed a computerized consultation management system in the ED of a tertiary care teaching hos-
pital to improve the consultation process and evaluate the influence of the consultation management sys-
tem on ED length of stay (LOS) and the throughput process.
Methods: Consultation management system software was developed and integrated into the hospital
information system. The development process took place between June 2008 and May 2009. Before the
development of the consultation system, ED personnel contacted on-call physicians of the specialty
department, who are usually residents, by cellular phones. After the system had been developed, ED
personnel selected the department and on-call physician in the specialty department using the consul-
tation management software and activated the automatic consultation process when specialty consulta-
tion was necessary. If the treatment plan had not been registered for 3 hours, all of the residents in
the specific department are notified of the delay in the treatment plan with a short message service
(SMS) message. If an admission or discharge order had not been made in 6 hours, all of the residents
and faculty staff in the specific department receive SMS messages stating the delay in disposition. ED
patient data were collected from the hospital information system for 40 days before the system was
developed (June 1, 2008, to July 10, 2008) and 40 days after the system was implemented (June 1, 2009,
to July 10, 2009).
Results: The median ED LOS decreased significantly, from 417.5 minutes (interquartile range
[IQR] = 178.8–1,247.5 minutes) in the presystem period to 311.0 minutes (IQR = 128.0–817.3 minutes) in
the postsystem period (p < 0.001). Also, the median time to disposition order decreased significantly,
from 336.0 minutes (IQR = 145.0–943.0 minutes) to 235.0 minutes (IQR = 103.0–21.5; p = 0.001). No signifi-
cant reduction was observed in the interval between the time of disposition decision and the time when
the patients left the ED. Significant reductions of ED LOS were observed after implementing the system
(p < 0.001) regardless of whether the visit occurred during the weekday daytime (09:00–17:00 hours), hol-
iday and weekend daytime (09:00–17:00 hours), or nighttime (17:00–09:00 hours next day).
Conclusions: This study found decreased ED LOS by implementation of a computerized consultation
management system in a tertiary care teaching hospital. The automated consultation and monitoring
process formalized communication between physicians providing ED patient care in the academic ED
with high consultation and admission rates.
ACADEMIC EMERGENCY MEDICINE 2011; 18:398–402 ª 2011 by the Society for Academic Emergency
Medicine

From the Department of Emergency Medicine and Medical Research Institute, Pusan National University Hospital (SJC, JJ, SH,
SY, SWP), Busan; the Department of Laboratory Medicine and Biomedical Informatics Unit, Pusan National University School of
Medicine (HHK), Busan; Busan U-city forum, City of Busan (HHK), Busan; and the Department of Emergency Medicine, Samsung
Changwon Hospital, Sungkyunkwan University School of Medicine (SYH), Busan, South Korea.
Received July 20, 2010; revision received September 19, 2010; accepted October 1, 2010.
The study was supported by a Medical Research Grant from Pusan National University.
The authors have no potential conflicts of interest to disclose.
Supervising Editor: Mark Hauswald, MD.
Address for correspondence and reprints: Jinwoo Jeong, MD; e-mail: advanced@lifesupport.pe.kr.

ISSN 1069-6563 ª 2011 by the Society for Academic Emergency Medicine

398 PII ISSN 1069-6563583 doi: 10.1111/j.1553-2712.2011.01039.x
ACAD EMERG MED • April 2011, Vol. 18, No. 4 • www.aemj.org
E
mergency department (ED) overcrowding is a
worldwide problem.1–5 The issues arise in three
independent stages of the ED process: input,
throughput, and output.6 Specialty consultation is an
important aspect of the throughput process, and there
are different types of consultations: consultations for
admission, for opinion only, for special procedures, for
transfer of care, and for outpatient referrals.7 Regardless
of the type of consultation, the patient is likely to stay in
the ED until the consultation process is finished. Consul-
tation difficulties and delayed response by specialty
departments contribute to disposition delays in hospitals
in many countries,7–10 and it is understood that strategies
to improve the ED consultation process are needed.10,11
We constructed a computerized consultation manage-
ment system in the ED of a tertiary care teaching hospital
to improve the consultation process and evaluate the
influence of the consultation management system on ED
length of stay (LOS) and the throughput process.
METHODS
Study Design
This was a pre–post observational study designed to
assess the effect of change on the consultation process
in an ED. The study was deemed exempt from review
and informed consent by the institutional review board
at the hospital site because of the observational nature
of the study.
Study Setting and Population
The study was conducted at a university hospital ED
that included an emergency medicine (EM) residency
program. The study hospital is the largest tertiary care
hospital located in Busan, an urban city in the Republic
of Korea, with more than 1,100 inpatient beds that are
occupied approximately 90% of the time. The study
hospital is designated as a ‘‘wide regional emergency
center,’’ which is the highest level of emergency care
facility in the Republic of Korea.12 The hospital uses a
custom-built computerized provider order entry (CPOE)
system, but electronic medical records have not been
applied. The CPOE system includes a database of ED
patients, which was designed according to the National
Emergency Department Information System standard.
The interface was developed by hospital staff. The data
in the database include characteristics of ED patient
visits, initial vital signs, sex, and age of the patients;
the time of ED arrival; the time of disposition decision;
the time when the patient left the ED; and the results of
the ED care. The time of arrival is registered when a
patient arrives at the triage area and the personal infor-
mation of the patient is entered into the system. The
time of disposition decision is registered when the
treating physician makes an admission or discharge
order.
The ED treats about 25,000–30,000 patients annually,
and 35% of the ED visits are admitted to inpatient beds.
The ED patients are cared for primarily by emergency
physicians, EM residents, and intern doctors on ED
rotation. Because of the severity of illness and injuries
of the patients, specialty consultations play an important
role in our ED care. Specialty department consultation
399
is required for a decision to admit a patient to an
inpatient bed. On-call physicians in the specialty
departments are primarily residents with two to three
postgraduate years of experience, but the patient
disposition decision is often made by senior residents
or faculty. The decision-making process in a specific
department is sometimes complex because of the sub-
specialty of the faculty staff on call or for various other
reasons usually unknown to ED staff.
Study Protocol
Consultation management system software was devel-
oped and integrated into the hospital information sys-
tem. The development process took place between June
2008 and May 2009. Before the development of the con-
sultation system, ED personnel contacted specialty
department on-call physicians by cellular phone. After
the computer consultation system was implemented,
ED personnel selected the department and on-call phy-
sician using the system and activated the automatic
consultation process when specialty consultation was
necessary. The system transmits the name, age, sex,
vital signs, chief complaint, ED LOS, and patient status
to the cellular telephone of the specified physician via
short message service (SMS). The physician who
receives the consultation visits the patient in the ED
and registers the patient visit in the system using either
a computer terminal or radiofrequency identification
tag. When the treatment plan is developed, the consult-
ing physician inputs the plan into the system in simple
text. The order to admit or discharge the patient is for-
matted in a paper document and the ED nurse enters
the disposition in the hospital information system when
the admission or discharge form has been received.
If the consulting physician does not register the ED
patient visit in the system within 20 minutes, the con-
sultation system automatically transmits a SMS mes-
sage to another specified physician in the consulting
department. If the treatment plan has not been regis-
tered for 3 hours, all of the residents in the specific
department are notified of the delay in the treatment
plan with an SMS message. If an admission or dis-
charge order has not been made in 6 hours, all of the
residents and faculty in the specific department receive
SMS messages stating the delay in disposition. When
only simple opinions are needed and the specialty
department physicians are not required to visit the ED,
ED personnel contact on-call physicians via cellular
phones without activating the computerized process.
Measurements
Patient data were collected from the hospital informa-
tion system for 40 days before the system was devel-
oped (June 1, 2008, to July 10, 2008) and 40 days after
the system was implemented (June 1, 2009, to July 10,
2009). The data included the total number of depart-
ments consulted, disposition, length of time from ED
visit to disposition order, and ED LOS for each ED
visit.
Data Analysis
The data analysis included medians and interquartile
ranges (IQRs). Mann-Whitney U- and Kruskal-Wallis
400
tests were used to compare the numeric data. The chi-
squared test was used to compare frequencies of the
categorical data. The PASW software, version 18 (SPSS
Inc., Chicago, IL) was used for the statistical analyses.
A p-value < 0.05 was deemed to indicate statistical sig-
nificance.
RESULTS
In total, 3,429 patients in the presystem days and 3,632
patients in the postsystem days were seen in the ED.
The median numbers of patients treated per day were
84.0 (IQR = 79.0–92.0) and 89.5 (IQR = 89.5–92.5),
respectively. There was no significant difference in the
proportion of inpatient admissions between the presys-
tem and postsystem periods (35.5% vs. 36.0%). How-
ever, there was a significant increase in the number of
daily ED visits in the postsystem period (p = 0.026).
The median ED LOS decreased significantly, from
417.5 minutes (IQR = 178.8–1247.5 minutes) in the pre-
system period to 311.0 minutes (IQR = 128.0–817.3 min-
utes) in the postsystem period (p < 0.001; Figure 1). Also,
the median time to disposition order decreased signifi-
cantly, from 336.0 minutes (IQR = 145.0–943.0 minutes)
to 235.0 minutes (IQR = 103.0–21.5 minutes; p = 0.001).
No significant reduction was observed in the interval
between the time of disposition decision and the time
when the patients left the ED.
The patient visits were categorized by time of visit:
weekday daytime (09:00–17:00 hours), holiday and
weekend daytime (09:00–17:00 hours), and nighttime
(17:00–09:00 hours next day). In all three time periods,
significant reductions of ED LOS were observed after
implementing the system (p < 0.001; Figure 2).
The number of consultations was only available for
the postsystem period, when precise recording of con-
sultations was made possible by the consultation sys-
tem software. Of 3,632 patients in the postsystem
period, 2,817 (77.6%) needed at least one consultation.
The number of consulted departments was associated
Figure 1. Boxplot of ED LOS in the pre- and postconsultation
management system periods. The central line is the median,
the box represents the IQR, and the whiskers extend to the 10th
and 90th percentiles. IQR = interquartile range; LOS = length of
stay.
Cho et al. • COMPUTERIZED CONSULTATION IN THE ED
with increased ED LOS. There was a significant differ-
ence in ED LOS between groups with none, one, and
two or more specialty consultations, for which the med-
ian values of ED LOS were 249.0 (IQR = 94.0–826.0),
397.5 (IQR = 195.0–1073.5), and 556.5 minutes (IQR =
272.5–1282.5 minutes), respectively (p < 0.001).
DISCUSSION
Although many factors contribute to ED overcrowding,
consultation difficulty is one of the most important.7 At
least one consultation is required for 24% to 60% of ED
patients.11 At our hospital, however, 75% of patients
require at least one consultation. This study also
showed a high rate of admission, 35.8% overall. The
high rates of consultation and admission suggest that
the ED of the study hospital cares for patients with
severe and complex problems. In fact, the medical
insurance system of Korea discourages patients with
minor problems from visiting EDs of wide regional or
regional emergency centers.
The number of consultations and ED LOS were cor-
related, which suggests that the relatively long ED LOS
at the study hospital might be partly attributed to the
heavy reliance on specialty consultations. Delayed
response to emergency consultation negatively influ-
ences the safety and quality of patient care. Problems in
the consultation process may result in life-threatening
situations.7 Sometimes ‘‘bouncing’’ of patients (i.e.,
attempts by inpatient teams to steer the patient to
another team) occurs.10 The complex hierarchy of the
decision-making process in a specific department may
also be unknown to emergency physicians, especially
because many of the first-line on-call physicians seen in
the ED are junior residents, who then notify senior phy-
sicians in their own departments.
Although the difficulties with consultation are well
known, there are only a few reports suggesting solu-
tions. A systematic review performed by Lee et al.7
found only two articles that recommended solutions
out of 15,000 articles reviewed.
We applied information technology to the consulta-
tion process as a possible method of involving respon-
sible doctors in the flow of ED patients. The
consultation management software recorded the time
taken at each stage of patient care and reported any
delay associated with the assigned physician on an
ascending scale. Thus senior doctors in the specialty
department were alerted to any delays in care and
intradepartmental communication was improved. The
ED LOS and time to admission or discharge order
decreased significantly, despite the increased number
of ED visits and the absence of any formal incentive or
punishment. These findings correspond to those of
Barry et al.,8 who reported a reduced median consulta-
tion time on an ascending scale of informing superiors,
including the dean, and concluded that an authoritative
intervention by a responsible leader was key to long-
term improvement. In the study hospital, the faculty
members of specialty departments were rarely involved
directly in the care of the ED patients. By use of the
consultation system, faculty were able to acknowledge
delays in the ED patient dispositions and had the
ACAD EMERG MED • April 2011, Vol. 18, No. 4 • www.aemj.org 401

Figure 2. Boxplot representing time to disposition decision of pre- or postconsultation management system periods categorized
by the time of patient visits. Pre- and postweekday, 09:00–17:00 hours on weekdays of the pre- and postsystem periods; pre- and
postholiday, 09:00–17:00 hours on weekends and holidays of the pre- and postsystem periods; and pre- and postnight, 17:00–
09:00 hours next day on weekdays, weekends, and holidays. The central line is the median, the box represents the IQR, and the
whiskers extend to the 10th and 90th percentiles. IQR = interquartile range.

chance to intervene in the ED process. Furthermore,
residents were observed to make decisions earlier to
avoid reprimands from their seniors.
The decrease in the ED LOS was almost the same as
that in the time to decision. However, the physician fac-
tor (time to decision) and the hospital resource factor
(time from decision to patient leaving, which includes
the administration process and assigning inpatient
beds) cannot be separated from the results of this study
because many physicians in the study hospital have a
tendency to make admission orders when they are sure
they have inpatient beds available, and transferring
patients to other hospitals is also an option. Figure 2
shows that the time to decision was shorter on week-
ends and holidays, when more inpatient beds were
unoccupied.
The system required little additional effort from the
ED personnel because most processes were automated,
except for the initial activation. Unlike under the previ-
ous system, specialty department physicians were not
required to stop other important activities to answer
cellular phone calls. The newly developed system uti-
lizes SMS, which does not require silencing every time
a contact is attempted. The system activates second-line
on-call physicians after 20 minutes, which is a highly
useful feature when the first-line on-call physician is
busy, cannot respond to the ED, and cannot activate
second-line physicians him- or herself. Otherwise, the
system notifies other physicians of a delay when there
is a complex problem that needs attention, such as
when patients are boarded in the ED and no disposition
decision has been made for more than 6 hours.
LIMITATIONS
This study was limited because the number of consulta-
tions could not be obtained from the presystem period
and compared between the periods, because consulta-
tion tracking was only instituted with the newly devel-
oped system. Furthermore, the person who analyzed the
data is an emergency physician working at the study
hospital and was therefore not blinded to the study
question. However, bias may have had little influence on
the result, as the data were directly downloaded from
the server, and the main outcome variable (i.e., ED LOS)
is objective numerical data. There was no formal
improvement process in the ED other than the consulta-
tion system during the study period, as the implementa-
tion and stabilization of a new consultation system took
considerable effort. However, there might have been
some unnoticed change in the ED process that may have
confounded the result. ED practice varies in different
hospital settings. The study hospital used in this study
has a very high rate of admission and consultations, and
the effect of the consultation system may have been
more substantial than in hospitals in different settings.
CONCLUSIONS
This study found decreased ED length of stay by imple-
mentation of a computerized consultation management
system in a tertiary care teaching hospital. The auto-
mated consultation and monitoring process formalized
communication between doctors providing ED patient
care in the academic ED with high consultation and
admission rates.
References
1. Fatovich DM, Hirsch RL. Entry overload, emer-
gency department overcrowding, and ambulance
bypass. Emerg Med J. 2003; 20:406–9.
2. Holroyd BR, Bullard MJ, Latoszek K, et al. Impact
of a triage liaison physician on emergency depart-
ment overcrowding and throughput: a randomized
controlled trial. Acad Emerg Med. 2007; 14:702–8.
3. Li G, Lau JT, McCarthy mL, Schull MJ, Vermeulen
M, Kelen GD. Emergency department utilization in
the United States and Ontario, Canada. Acad Emerg
Med. 2007; 14:582–4.
402
4. Olshaker JS, Rathlev NK. Emergency department
overcrowding and ambulance diversion: the impact
and potential solutions of extended boarding of
admitted patients in the emergency department.
J Emerg Med. 2006; 30:351–6.
5. Richardson S, Ardagh M, Hider P. New Zealand
health professionals do not agree about what
defines appropriate attendance at an emergency
department. N Z Med J. 2006; 119:U1933.
6. Asplin BR, Magid DJ, Rhodes KV, Solberg LI, Lurie
N, Camargo CA Jr. A conceptual model of emer-
gency department crowding. Ann Emerg Med.
2003; 42:173–80.
7. Lee RS, Woods R, Bullard M, Holroyd BR, Rowe
BH. Consultations in the emergency department: a
systematic review of the literature. Emerg Med J.
2008; 25:4–9.
Cho et al. • COMPUTERIZED CONSULTATION IN THE ED
8. Barry EB, Talmage MH, Donald DS, Charles WS,
Reece EA. Reducing specialty consultation times in
the emergency department [abstract]. Acad Emerg
Med. 2004; 11:463.
9. Drummond AJ. No room at the inn: overcrowding
in Ontario’s emergency departments. CJEM. 2002;
4:91–7.
10. Reid C, Moorthy C, Forshaw K. Referral patterns:
an audit into referral practice among doctors in
emergency medicine. Emerg Med J. 2005; 22:355–8.
11. Woods RA, Lee R, Ospina MB, et al. Consultation
outcomes in the emergency department: exploring
rates and complexity. CJEM. 2008; 10:25–31.
12. Hwang SO, Lee CC, Kim TM, Singer AJ. The cur-
rent status of the emergency medical system in
Korea. Am J Emerg Med. 2007; 25:846–8.