Critical Care Medicine

Issue: Volume 26(2), February 1998, p 252-259

Copyright: Copyright (C) by 1998 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights
Reserved.
Publication Type: [Clinical Investigations]
ISSN: 0090-3493
Accession: 00003246-199802000-00020
Keywords: ethics, medical, referral and consultation, decision-making, medical futility, critical illness, intensive care
units

A study of proactive ethics consultation for critically and

terminally ill patients with extended lengths of stay
Dowdy, Melvin D. PhD; Robertson, Charles MD; Bander, John A. MS
Author Information
From the Bon Secours St. Mary's Hospital, Richmond, VA.
Supported, in part, by a grant from the Bon Secours-St. Mary's Health Care Foundation and from the
Trigon Blue Cross and Blue Shield of Virginia.
Address requests for reprints to: Dr. Melvin D. Dowdy, Medical Arts Building, 890 Poplar Church Road,
Camp Hill, PA 17011.
Abstract
Abstract: Objective: To assess the effect of proactive ethics consultation on documented patient care
communications and on declsions regarding high-risk intensive care unit (ICU) patients.

Design: Prospective, controlled study.

Patients: Ninety-nine ICU patients treated with >96 hrs of continuous mechanical ventilation.

Interventions: Three groups were compared: a) a baseline group enrolled in the study prior to the
establishment of the hospital's ethics consultation service; b) a control group where ethics consultation
was at the option of the care team; and c) a treatment group where the ethics service intervened
proactively after patients received >96 hrs of continuous mechanical ventilation. Patient care planning,
for subjects in the proactive group, was reviewed with physicians and with the care team using a
standardized set of prompting questions designed to focus discussion of key decision-making and
communication issues for critically and terminally ill patients. Issues and concerns were identified and
action strategies were suggested to those in charge of the patient's care. Formal ethics consultation,
using a patient care conference model, was made available upon request.

Measurements and Main Results: Post discharge chart reviews of the three groups indicated no
statistically significant differences on important demographic variables including age, gender, and acuity.
Comparisons of survivors and nonsurvivors for the three groups indicated, at statistically significant
levels, more frequent and documented communications, more frequent decisions to forgo life-sustaining
treatment, and reduced length of stay in the ICU for the proactive consultation group.
Conclusion: Proactive ethics consultation for high-risk patient populations offers a promising approach to
improving decision-making and communication and reducing length of ICU stay for dying patients. (Crit
Care Med 1998; 26:252-259)

Among the most challenging situations for which ethics is the difficult decision to forgo life-sustaining
treatment for the critically and terminally ill. These end-of-life decisions have been complicated by a
technological imperative which often drives decisions to continue treatment, despite the considerable
burdens of prolonging suffering and death, of emotional and family debilitation, and of financial
expense [1]. When the prognosis is considered dismal, some physicians may avoid talking with patients
or families about forgoing life-sustaining treatments; others may inhibit dialogue by cloaking
recommendations in objective sounding medical jargon; and a few may circumvent dialogue by
asserting their authority to unilaterally withhold or withdraw treatment, justifying their decision by appeals
to medical futility [2].

The largest study [3] examining the efforts to improve physician-patient communication and advance
planning for critically ill patients reported discouraging results. The study to understand prognoses and
preferences for outcomes and risks of treatments (SUPPORT) [3] provided corroborative evidence of
what many fear is typical of the treatment of hospitalized critically and terminally ill patients, namely,
poor communication between physician and patient, poor pain management, avoidance of discussions
about prognosis or about cardiopulmonary resuscitation, and failure to understand patient preferences.
However, the intervention designed by SUPPORT [3] investigators to alter physician behaviors, through
improved communication and understanding of prognoses and patient preferences, had no impact on
important dimensions of care for critically and terminally ill patients. In particular, it failed to reduce the
period of time that dying patients spent in the intensive care unit (ICU), in coma, or on mechanical
ventilation. Criticism of the study's intervention pointed to the physician's reliance on nurses and
computer-generated information as methods for improving physician communication and
understanding [4]. A more intimate, dynamic communication between patient and physician was not
prompted by the provision of more data, which many of the physicians failed to read.

In the present study, we focused on the use of proactive ethics consultation as a process-oriented
intervention for improving communication and decision-making between physicians, patients, and the full
care team. Zawacki and others [5-7] have argued that difficult decisions to forgo life-sustaining treatment
require close attention to the processes of communication surrounding a particular patient's plan of care.
These authors [5-7] proposed using a "counseling" or "educational" model of ethics consultation
provided proactively in anticipation of difficult choices. The focus of consultation is on increasing
dialogue between physicians and patients, heightening attention to whatever is perceived as
problematic, encouraging questions, helping persons to challenge one another, removing obstacles to
communication, and seeking the cooperation necessary to reach "a mutually beneficial outcome from
informed consensus or compromise" [3].

We selected this strategy during our hospital's participation in the Decisions-Near-the-End-of-Life

(DNEL) program which had provided a subset of our staff with ethics education and then challenged us
to develop concrete action steps to improve terminal care. The DNEL model [8] encourages
multidisciplinary communication among all members of the healthcare team, planning in advance for
crises, and a process-oriented approach to change which calls for active participation of inhouse
leadership teams. Our medical staff strongly urged us to design a prospective study aimed at changing
patterns of communication and decision-making among physicians and all involved in the patient's care.

We assumed that clinical situations at risk of becoming ethically problematic could be identified in
advance and that interventions could be made before indecision and frustrations have had time to
ferment. Early intervention in high-risk cases may help develop psychologically and ethically functional
discussions of what constitutes the limits of care for a particular patient when that patient's treatment
response raises issues about what is an appropriate level of care. In particular, we hypothesized that
ethics consultation provided proactively to high-risk mechanical ventilation patients would increase
documented communication between physicians, patients, and the clinical team regarding important
features of end-of-life decisions. Taking the perspective that communication is a critical process variable
affecting the timeliness of decision-making, we also hypothesized that increased communication would
promote decisions to forgo life-sustaining treatment for dying patients, for those without hope of
recovery, and for those patients whose preferences directed a lower level of care. If these end-of-life
decisions increased in frequency and timeliness, we expected to observe a reduction in average length
of stay in the ICU.

Back to Top
MATERIALS AND METHODS
The secondary use of patient data for this study and the proactive ethics consultation protocol were
approved as exempt by the Institutional Review Board of St. Mary's Hospital. Since the main
intervention did not involve formal problem-solving or direct contact with patients or surrogates, and
since the intervention was viewed by the Institutional Review Board as a form of process monitoring and
education, formal consent was not required. However, in those instances where the intervention evolved
into a direct effort to assist all parties in the resolution of an ethical issue, full consent for the consult was
sought and obtained before allowing the consultation to move forward. The study sequentially enrolled
99 patients treated with continuous mechanical ventilation for >96 hrs in the hospital's ICU between
June 1992 and October 1994. Preliminary review of ICU patients with extended lengths of stay at this
hospital indicated that >95% of those patients with stays extending to weeks or months received
continuous mechanical ventilation for >96 hrs. Since an International Classification of Diseases (ICD)-9
code existed for these ventilator patients, this procedure provided convenient selection criteria.
However, chronic ventilation alone provides no basis for predicting mortality or for making difficult moral
decisions, and other hospitals may have very different experiences regarding modal treatments for
patients with extended lengths of stay.

Data were collected by chart review following discharge on patient age and gender, primary and
secondary diagnoses, method of payment, admission and subsequent code status, index of acuity using
the Acute Physiology and Chronic Health Evaluation (APACHE II) scores [9], hospital and ICU lengths of
stay, decision-making capacity upon admission, advance directives prior to or during hospitalization, do-
not-resuscitate orders and other orders to forgo life-sustaining treatments, specific documentation
mentioning futility and determinations of the limits of care, formal (not proactive) ethics consultation, and
consultation with other allied health professionals. Hospital charges for each patient were obtained
through fiscal services.

Data collection occurred in three stages. A baseline group of 37 patients was enrolled in the study over
a 6-mo period prior to the establishment of the hospital's ethics consultation service. Patients were
identified by using ICD-9 code 96.72 and enrolled sequentially until the end of the initial 6 mos of the
study; although 40 patients were identified, only 37 charts had complete data.

A second group of 31 control patients was enrolled sequentially during a 6-mo period after the
establishment of the ethics consultation service and after an initial phase of ICU staff enrollment in the
DNEL program. The consult service had been operational for 4 mos before the enrollment of the control
group and was included on routine rounds of the entire ward. The consult services were publicized
through announcements to department heads, notifications through inhouse communications, and
postings in the ICU. Upon request, patients, families, and all members of the clinical team had access to
ethics consultation for control group patient care. Control patients were identified using the unit's clinical
database terminal which displays daily status reports on patients by ventilation procedure. During the
control group enrollment, two control group patients received ethics consultation at the request of the
physician; all other control group patients did not receive ethics consultation.

A proactive ethics consultation group of 31 patients was sequentially enrolled during the final 8 mos of
the study; two additional enrollment months were required to obtain a sample size equal to the control
group. The unit's clinical database terminal was used to identify when patients reached the 96-hr
criterion. Physician consent was obtained for participation in the study. No patients were lost from the
study due to refusal of consent. For all cases in the treatment group, proactive ethics consultation was
provided. When issues or conflicts arose which necessitated a problem-solving intervention involving all
decision-makers, consent was obtained from the patient, family, or surrogate in keeping with the
hospital's established protocol for this more traditional, formalized ethics consultation [10]. During the
proactive group period, there were three cases which received these formal ethics consultation services;
for all other patients, only the proactive ethics consultation was provided.

Proactive Ethics Consultation. Proactive ethics consultation was provided to the clinical team as a
consultee-centered consultation designed to increase the team's attention to key decision-making and
communication process issues as they arise in the care of critically and terminally ill patients [11].
Consultation was provided by a team of two clinicians trained in clinical ethics. The consultation had
both structured and unstructured dimensions. The structured dimension involved a series of standard
questions as follows: a) Has the patient created an advance directive? If so, is a copy of the advance
directive in the chart? b) Is the patient currently a capable decision-maker? If not, who is making
decisions on behalf of the patient? c) Is the decision-maker aware of the patient's current diagnosis,
prognosis, treatment objectives, and treatment alternatives? d) Do any of the treating physicians
anticipate major hurdles or threats to the patient's recovery? If so, what are these and what are the
indicators of response to treatment? e) If the patient's response to treatment has caused concern about
failure to recover, is there any discussion of the limits of treatment for this particular patient? f) Are there
other unaddressed issues regarding patient care, management, patient preferences, or communication?

After identifying a prospective patient for proactive ethics consultation, the two consultants contacted the
patient's attending physician to inform the physician of the study and to solicit enrollment. The first step
was to reach agreement that the role of the consultants was strictly to facilitate communication by
providing information, stimulating discussion among caregivers, increasing attention to the core set of
six questions mentioned above, and being available for more formalized consultation should problems or
significant conflicts arise. Attending physician approval for enrollment of a patient into the study included
physician consent for discussions with other physicians caring for the patient, review of the patient's
chart, discussions of the case with members of the care team, but not discussing the case with the
patient, surrogate, or family members. Consultants conformed to the hospital's ethics consultation policy
which requires consent of the patient or surrogate when consultation is sought for purposes of resolving
conflicts or facilitating decision-making among those with moral standing in a particular case. As long as
the consultation was confined to a consultee-centered, communication facilitation role, consent was
restricted to the physician.

The six core questions were first addressed by review of the patient's chart. If any caregiver seemed
unaware of information contained in the chart, the consultants provided that information (e.g., the
existence of an advance directive). If other questions arose from the chart review, these questions were
directed to the attending physician and the care team (e.g., conflicting information regarding the patient's
capacity, lack of complete disclosure regarding the patient's condition, or confusion regarding the
treatment objectives for this patient). The consultants met informally with nurses, chaplains, social
workers, and other allied health professionals involved in the patient's care to inform them of the
consultant's role and to encourage sharing of information and team communication. These efforts to
facilitate communication were conducted in the normal course of visiting the unit each day in the early
morning and afternoon hours. The same questions were repeated as necessary to each member of the
team at different times over the initial days of the consultation.

Since these queries evoked a wide range of information and perspectives, the unstructured dimension of
the consultation was an exploration of the issues that were raised. Consultants focused the attention of
team members on the need for communication, the importance of involving all concerned parties and
sources of additional information, and the value of identifying and addressing any perceived barriers to
collaboration and communication. Daily rounds by physicians provided opportunity for the consultants to
continue exploring any issues which remained unresolved. Thus, the ethics consultants routinely
provided information and other nondirective facilitation regarding how caregivers were reasoning and
communicating about a particular patient's care. For example, the consultants might indicate to a
physician who has agreed to a trial period of treatment that there is disagreement among the team
regarding what would constitute a positive response to treatment and that the trial period has not been
clearly specified. The consultants might inquire whether someone had explored with the patient or
surrogate what constitutes a burden or benefit from the patient's perspective. In any case, the
consultants sought to maintain neutrality through an even attention to each member of the care team
and adherence to the principle that shared information and two-way communication support good
decision-making.

When major hurdles intervened with the patient's care or when difficult decisions arose, the ethics
consultants were available as a sounding board to review the steps the physicians and care team were
taking to communicate with each other and with the patient or surrogate. When conflicts arose between
the physician and the patient or surrogate, the consultants were available for formal ethics consultation
pending the appropriate consents. In either case, physicians were encouraged to formulate with all
decision makers a plan for addressing these major hurdles, difficult decisions, or conflicts.

The treatment team was encouraged to discuss with the patient, surrogate, or family, the medical
rationale for the patient's current level of care, ethical issues which physicians, family members, or
surrogates were considering as relevant to the patient's current level of care, and any foreseeable
changes in the patient's capacity to recover which could place limits to the patient's level of care based
on advance directive information or substituted judgment. This plan of care was developed prior to
physicians seeking specific consent to withhold or withdraw life-sustaining treatment.
The hospital's do-not-resuscitate policy, guidelines for forgoing life-sustaining treatment, and ethics
consultation policy direct physicians to seek formal ethics consultation in situations where conflicts
among decision-makers persist; formal consultation requires both the consent of the physician and the
patient or surrogate. These policies proscribe unilateral decision-making by physicians, but also
recognize that physicians have moral standing in the care of a patient and may reject a course of
treatment which they view as either morally or medically unjustified. Whether supported by formal
consultation or by the study's proactive approach, the physician is accountable by hospital policy for
decisions to forgo life-sustaining treatment. In particular, these policies affecting end-of-life decisions
require the patient's or surrogate's informed consent, determination of the patient's values and
preferences, participation by the patient or surrogate in such decisions, and a determination that the
treatment(s) in question does not offer a reasonable hope of benefit or entail excessive burden.

Measurement of Communication. An observational tool was developed to gather documentation in the

patient's medical record of communications regarding treatment decisions. A team of three physicians, a
nurse clinical ethics consultant, a social worker, a quality assurance administrator, and a clinical ethicist
developed a check list covering the types of information which physicians and other caregivers should
enter into the medical record in the course of caring for critically and terminally ill patients. These
medical record entry items were used as a "quality of communication index" which included the
presence of an advance directive, a do-not-resuscitate order and accompanying physician notes, orders
to withhold or withdraw life-sustaining treatment and accompanying physician notes, specific
documentation regarding the limits of care, and consultations requested of pastoral care, social services,
and the pain management service. A score of 1 was given each time an entry appeared in the record,
except for accompanying notes and allied professional consultations which were scored a .5 for each
entry. The interest in communication was intentionally restricted to documented information relevant to
decisions near the end of life. Content validity was reviewed by the hospital's medical executive
committee and accepted for the purposes of this study.

Statistical Methods. The multivariate relationships on quality of communication and length of stay were
compared between groups for survivors and nonsurvivors using analysis of variance. Separate
comparisons of do-not-resuscitate and other life-sustaining treatment decisions were made using the
chi-square test. Step-wise regression analysis was used to determine variables which best predicted
hospital charges.

Back to Top
RESULTS
Patient Characteristics. The average age for the Table Opens a Opens a
baseline, control, and proactive groups were 1Opens a popup popup popup
63.4, 60.4, and 65.7 yrs, respectively. Mean window window window
APACHE II scores for the three groups were
20.5, 20.1, and 21.7, respectively. The
APACHE II scores are consistent with the
assumption that the patients in the study were
at considerable risk, although no attempt was
made during the hospitalization to predict
mortality [12]. Table 1 displays additional
patient characteristics by group. Only two
patients had advance directives at the time of
their admission to the hospital. Tests for the
presence of trends in these characteristics were
conducted because the groups were formed
sequentially. A chi-square trend test was used
for categorical variables and regression for
continuous variables. Groups were assigned
integer scores and significance was achieved at
p = .05. There was evidence of a trend in
mortality. There was also a trend regarding
capacity at the time of admission favoring more
subjects with capacity in the proactive group.
An analysis of diagnostic categories showed no
trend in patient diagnosis for the entire sample,
nor were there any statistically significant
differences between the three groups for
diagnosis.

Main Results. Preliminary analysis showed no Opens a Opens a

significant differences in quality of popup popup
Table
communication by age, gender, or acuity. The window window
2Opens a popup
means and standard deviations for quality of
communication by group and discharge status
window
(i.e., survived or died) are presented in Table
2. A two-way analysis of variance confirms the
relationships suggested by the data (F = 16.17,
degrees of freedom = 5,93, p < .01). Those
patients who died had significantly higher
communication scores than those discharged
alive, a finding we expected since the focus on
communication was restricted to end-of-life
decisions. A strong interaction between
discharge status and group was found. Among
those patients who died, the ethics proactive
group had significantly higher communication
scores than the other two groups which were
not significantly different.

A separate analysis showed that nearly twice as many patients (61%) in the proactive ethics group had
do-not-resuscitate and other life-sustaining treatment decisions made in the course of their ICU care
compared with baseline (32%) and control (38.7%) patients. The observed difference was significant
(chi squared = 6.13, degrees of freedom = 2, p < .05). Examination of decisions to withhold or withdraw
life-sustaining treatments other than cardiopulmonary resuscitation showed the proactive ethics group
had significantly more decisions in this area as well (chi squared = 8.05, degrees of freedom = 2, p <
.05). In the proactive group, this difference accounted for 48% of cases, compared to 18.9% of baseline
and 22.6% of control group patients. Combined with the analysis of variance results on quality of
communication, these analyses support the hypothesis that proactive ethics consultation yielded
improvements in both communication and decision-making.

The means and ranges of length of stay by Opens a Opens a

Table
group and discharge status are presented popup popup
3Opens a popup
in Table 3. In the regression, two variables window window
window
were significantly related to the length of ICU
stay: a) membership in the ethics proactive
group; and b) discharge status (F = 6.46,
degrees of freedom = 2,96, p < .01). These
variables accounted for 10% of the variability
in ICU stay. An equation that predicts ICU
lengths of stay based on these two factors
indicates that those who died had predicted
ICU stays of 7 days less than those patients
who survived. Members of the ethics proactive
group had predicted ICU stays of 6 days less
than the other two groups. These effects are
additive. For example, the average ICU length
of stay of those patients in the ethics proactive
group who died was 13 days shorter than a
survivor in either of the other groups.

The charge structure for hospital Opens a

Table
services changed three times during Table popup
4Opens a popup
the course of the study. It was 5Opens a popup window
window
necessary, therefore, to adjust all window
charges to reflect the charge structure
in place at the time of the baseline
group sample. The means and ranges
of adjusted charge by group and
discharge status are presented
in Table 4. For both survivors and
nonsurvivors, the ethics proactive
group had the lowest charges. A step-
wise regression analysis was
performed to determine which
variables best predicted adjusted
hospital charges. The step-wise
regression program selects from all
possible predictor variables those
explaining the greatest amount of
variance to achieve the most
parsimonious explanatory model. The
result presented in Table 5 was a
three-variable model which included
ICU length of stay, non-ICU length of
stay, and ICU discharge status. The
observed differences in adjusted
charges were in the expected
direction and may be practically
significant; that is, for proactive group
patients who died, there was a 16%
reduction in average charges
compared with baseline patients who
died and a 33% reduction in average
charges compared with control
patients. However, the strength of
these differences was not statistically
significant and, therefore, too weak to
raise expectations of proactive ethics
consultation as a reliable method for
reducing hospital charges.
Back to Top
DISCUSSION
Until this study was undertaken, a proactive ethics consult team was an unexamined intervention for
improving communication and decision-making in the care of critically and terminally ill patients. To date,
we are unaware of any study which has examined the benefits of ethics consultation, broadly speaking,
using a prospective control group design. There has been one other study [13] from a DNEL-affiliated
hospital which documented a reduction in length of stay through the introduction of ethics rounds in a
surgical ICU. Reductions in length of stay and increases in do-not-resuscitate orders have also been
observed [14] as a benefit of policy development supported by a strong ethics education program. The
benefits of proactive ethics consultation of communication, decision-making, and length of stay are
impressive, although we recognize the inherent limitations of a small sample size and the limited
examination of only those communications and decisions which could be documented in the patient's
medical record.

Many physicians find it easier to limit resuscitative efforts such as chest compression, defibrillation, or
antiarrhythmic drugs than to withdraw life-sustaining treatments such as ventilation, nutrition and
hydration, or antibiotic therapy. Although no valid moral distinction exists between with-holding and
withdrawing life-sustaining treatments [15], stopping treatment with the certain knowledge that the
patient will die evokes in many physicians an intuitive sense of causing the patient's death. It is precisely
these sorts of difficulties which a processoriented consultation approach hopes to address. Physicians
wrote twice as many do-not-resuscitate orders for patients in the proactive group, but the effectiveness
of the consultation may be appreciated more by their decisions to forgo other life-sustaining treatments
at frequencies significantly greater for those patients in the proactive group.

The design of proactive ethics consultation in this study differs in several important ways from the
general practice of clinical ethics consultation [10]. First, consultation was provided prior to decision-
making difficulties, that is, before an identified ethical problem or conflict. Because the aim of the
intervention is to prevent the negative outcomes often associated with medically futile situations,
especially breakdowns in communication and continued aggressive treatment provided to dying patients
with no reasonable hope of recovery, our approach could be understood as a form of "preventive
ethics" [16]. As a preventive ethics approach, proactive ethics consultation aims to improve
communication before problems develop. By focusing on a particular patient's care, it holds promise of
giving more precise, clinically sensitive suggestions than a more generalized educational approach. The
approach in our study appears to result in several possible outcomes: a) satisfactory outcomes for
patient care where there is a high risk of value conflict; b) relevant medical and ethical information
perceived in a more rational light prior to the experience of conflicting views; and c) enhanced quality of
patient care associated with improvements to communication before end-of-life decisions.

A second characteristic of proactive ethics consultation is a shift in the initial focus from "ethical
problems" to clinical communication essential to quality patient care and to sound ethical dialogue. Our
approach is largely an educational intervention similar to Caplan's consultee-centered case
consultation [11]. Reviewing the course of the patient's care, current indications for potential problems,
and options for a plan of care helps solicit new thoughts about the patient's needs, relevant contextual
factors, and ethical issues. More significantly, the consultee is given an opportunity to consider how he
or she is managing communication with all parties concerned with the patient's care. We sought to
improve communication by improving clinician knowledge and skill, as well as improving the clinician's
objectivity and readiness to address difficult decisions. Although an ethics consultant has the option to
become directly involved with all parties in a problem-solving modality, a proactive period of ethics
consultation may focus more on the consultee's capacity to anticipate problems, develop a baseline of
effective communication, and engage relevant ethical issues as they arise.

We believe ethics consultation provided in a consultee-centered format may actually diminish the need
for more formalized consultations involving all parties in a process of problem-solving and mediation. We
should also note that no presumption was made that a clinician or the clinical team were "having
problems" in the care of these patients, although one motivation to initiate a study was the concern
expressed by the medical staff that patient care planning for this population could be improved. Given
the consultee-centered focus, consent for the involvement of the ethics consultant was obtained from
each clinician and from the clinical team. Whenever the consultation extended to involvement of all
parties, patient and/or surrogate consent were obtained. We feel this was an appropriate consent
strategy consistent with general consultation practice.

The benefits of proactive ethics consultation extended to survivors in the proactive group as well as
those who died. Although proactive group survivors had no higher communication scores than their
nontreatment group cohorts, they had shorter lengths of stay. We can only speculate that this difference
in length of stay may be associated with a generalized improvement in the ICU culture. At the conclusion
of the study, we met with physicians, nurses, and allied professionals to discuss their views of the study
and to prepare a report for the administration. Anecdotal observations suggested that the staff perceived
a shift in attitude, one which placed greater value on collaboration among all members of the care team.
Members of the team commented that, following the study, discussions of ethical issues were not
postponed to addressing changes in levels of care but more often included considerations for involving
the patient and family in advance patient care planning. By the conclusion of the study, physicians and
nurses expressed greater conviction that patients and surrogates should be kept informed and involved
in decision-making and that they could be approached about difficult decisions without stimulating
defensiveness, fear, or loss of hope. Discussion of ethical issues between nurses and physicians also
appeared less defensive in tone; for example, the mere mention of an ethical issue no longer implied
wrongdoing.

Length of stay was significantly reduced for dying patients in the proactive group. Although we observed
that more proactive group patients died, we do not know how many patients in either control or proactive
groups died within a few days or weeks following their discharge. The study was prompted by concern
not to prolong the period dying patients spend in the ICU, and the intervention called attention to
questions or concerns about respecting patient values in situations of declining capacity to recovery. We
are not disturbed, therefore, to find that more patients died in the hospital within the proactive group.
This issue of mortality raises interest in the number of patients who were discharged as "survivors" from
the baseline and control groups. A pilot study of the baseline group indicated that [approximately]12% of
the "survivors" died within 6 mos; unfortunately, attempts to replicate this observation for the control and
proactive groups failed due to family relocation, inadequate next-of-kin address information, etc. A 6-mo
mortality rate of 20.6% was recently reported for one group of ICU patients [17]. Adjusting this figure for
those who died before discharge, an approximate mortality rate of 12% to 15% could be expected. A
future study should test the assumption that, when controlled for long-term survival, mortality rates for
patients in the proactive ethics consultation group are roughly equal to the mortality rates in the control
groups.

The use of continuous mechanical ventilation as a selection criteria in no way suggests an expectation
that prolonged ventilation alone predicts poor outcomes. We recognized from our initial baseline study
that this group of patients was at considerable risk the longer they remained on ventilation and the
greater the number of organ system failures. Among the survivors were many patients with good
outcomes. All nonsurviving patients in the proactive group developed >or=to3 organ system failures
prior to the time decisions were made to forgo life-sustaining treatments, although some of these
patients had only one organ system failure at the time of enrollment in the study. The development of
multiple system failure after 4 days on mechanical ventilation does not generally favor survival; rates in
this instance may range toward 89% mortality [18].

When physicians confront situations where continued treatment provides little or no chance of recovery,
they have a clear obligation to disclose the poor prognosis and to engage decision makers in an effort to
resolve conflicts. They do not have an obligation to offer treatments they consider futile [19]. However,
the legal system in this country has not yet resolved the issue of a physician's refusal of requested care
and courts have discouraged continued reliance on legal process to resolve such conflicts. Our study
does not encourage the conclusion that decisions should be made unilaterally but promotes the premise
that effective communication processes can be developed to support a culture of care which anticipates
problems and involves everyone in advance of major decisions.

On the other hand, our results suggest that proactive ethics consultation is not an effective means of
reducing costs. Another study [20] has also discouraged expectations of cost savings through
improvements in decision-making for terminally ill patients. A reduction of hospital charges by 16% may
translate into desirable savings for hospital administrators seeking to leverage every opportunity to
improve efficiency, but these savings are not a reliable outcome of proactive ethics consultation.
Reductions in charges are highly correlated with length of stay; however, among the many factors
affecting length of stay, proactive ethics consultation accounted for only 10% of length of stay in this
study's sample. Also, charge structures are inherently unstable data, and heterogeneous patient
populations vary widely in the types of resources needed for their care.

The results of the present study suggest that process-oriented, proactive ethics consultation merits
further investigation as an effective resource for improving patient care. The limitations of this study
derive mostly from the small sample size and the challenge of standardizing a process-oriented
intervention. A multi-institutional replication of this study with a larger sample should demonstrate that
the results are not an artifact either of one hospital's culture or of unintended bias associated with the
consultants' style. Process research is difficult in clinical settings, but additional systematic research is
needed to understand the key processes which account for the effect proactive consultations have on
communication and decision-making, as well as to understand how such consultations are perceived by
the clinical team. Future studies might expand the range of communication variables and clinical
decisions included as quantitative or qualitative data. We would also be interested in following survivors
for 6 mos to 1 yr to evaluate their long-term survivability and health status.

Back to Top
ADDITIONAL READING
Solomon MZ, et al: Decisions Near the End of Life Coordinating Center, Decisions Near the End of Life
Curriculum Materials. Vols 2-7. Newton, MA, Education Development Center, 1997

Jecker NS: Is refusal of futile treatment unjustified paternalism? J Clin Ethics 1995; 6:133-137

Purtilo RB: A coment on the concept of consultation. In: Ethics Consultation in Health Care. Fletcher JC,
Quist N, Jonsen AR (Eds). Ann Arbor, Health Administration Press, 1989

Rowan KM: Intensive Care Society's Acute Physiology and Chronic Health Evaluation (APACHE II)
study in England and Ireland. A prospective, multicenter, cohort study comparing two methods for
predicting outcome for adult intensive care patients. Crit Care Med 1994; 22:1392-1401

Troug RD, Brett AS, Frader J: The problem with futility. N Engl J Med 1992; 326:1560-1564
Wong DT: A comparison of the Acute Physiology and Chronic Health Evalutation (APACHE II) score
and the Trauma-Injury Severity Score (TRISS) for outcome assessment in intensive care unit trauma
patients. Crit Care Med 1996; 1642-1647

Back to Top
REFERENCES
1. Callahan D: The Troubled Dream of Life. New York, Simon & Schuster, 1993 [Context Link]

2. Solomon MZ: How physicians talk about futility: Making words mean too many things. J Law Med
Ethics 1993; 21:231-237 [Context Link]

3. The study to understand prognoses and preferences for outcomes and risks of treatments
(SUPPORT): A controlled trial to improve care for seriously ill hospitalized patients. JAMA 1995;
274:1591-1598 [Context Link]

4. Lo B: Improving care near the end of life: Why is it so hard? JAMA 1995; 274:1634-1636 Bibliographic
Links [Context Link]

5. Zawachi BE: The "futility debate" and the management of gordian knots. J Clin Ethics 1995; 6:112-
117 Bibliographic Links [Context Link]

6. Wear S & Logue G: The problem of medically futile treatment: Falling back on a preventive ethics
approach. J Clin Ethics 1995; 6:138-148 Bibliographic Links [Context Link]

7. Truog RD: Process in the futility debate. J Clin Ethics 1995; 6:128-132 Bibliographic Links [Context
Link]

8. Solomon MZ, Jennings B, Guilroy V, et al: Toward an expanded vision of clinical ethics education:
From the individual to the institution. Kennedy Institute of Ethics Journal 1991; 1:225-245 Bibliographic
Links [Context Link]

9. Knaus WA, Draper EA, Wagner DP, et al: APACHE II: A severity of disease classification system. Crit
Care Med 1985; 13:818-829 Ovid Full Text Bibliographic Links [Context Link]

10. Kanoti GA, Younger S: Clinical ethics consultation. In: Encyclopedia of Bioethcs. Vol I. Reich WT
(Ed). New York, Macmillan, 1995, pp 404-408 [Context Link]

11. Caplan G: Theory and Practice of Medical Health Consultation. Basic Books, New York,
1970 [Context Link]

12. Rogers J, Fuller HD: Use of daily Acute Physiology and Chronic Health Evaluation (APACHE) II
scores to predict individual patient survival rate. Crit Care Med 1994; 22:1402-1405 [Context Link]

13. Holloran S, Starkey G, Burke P, et al: An educational intervention in the surgical intensive care unit
to improve ethical decisions. Surgery 1995; 118:294-299 Bibliographic Links [Context Link]
14. Kanoti GA, Gombeski Jr WR, Gulledge AD, et al: The effect of do-not-resuscitate orders on length of
stay. Cleve Clin J Med 1992; 59:591-594 Bibliographic Links [Context Link]

15. President's Commission for the Study of Ethical Problems in Medicine and Biomedical and
Behavioral Research: Deciding to Forego. Life-Sustaining Treatment. Washington, DC, US Government
Printing Office, 1983, pp 77-82 [Context Link]

16. Forrow L, Arnold RM, Parker LS: Preventive ethics: Expanding the horizons of clinical ethics. J Clin
Ethics 1993; 4:287-294 Bibliographic Links [Context Link]

17. Konopad E, Noseworthy TW, Johnson R, et al: Quality of life measures before and one year after
admission to an intensive care unit. Crit Care Med 1995; 23:1653-1659 Ovid Full Text Bibliographic
Links [Context Link]

18. Zimmerman JE, Knaus WA, Wagner DP, et al: A comparison of risks and outcomes for patients with
organ system failure: 1982-1990. Crit Care Med 1996; 24: 1633-1641 Ovid Full Text Bibliographic
Links [Context Link]

19. Luce JM: Physicians do not have a responsibility to provide futile or unreasonable care if a patient or
family insists. Crit Care Med 1995; 23:760-766 Ovid Full Text Bibliographic Links [Context Link]

20. Emanuel EJ, Emanuel LL: The economics of dying: The illusion of cost savings at the end of life. N
Engl J Med 1994; 330:540-544 Bibliographic Links [Context Link]

Key Words:: ethics, medical; referral and consultation; decision-making; medical futility; critical illness;
intensive care units

Table 1. Group profile by gender, religion, Medicare, mortality rate, and capacity at
time of time of admission
A study of proactive ethics consultation for critically
and terminally ill patients with extended lengths of
stay.
Dowdy, Melvin; Robertson, Charles; Bander, John
Critical Care Medicine. 26(2):252-259, February 1998.
 A study of proactive ethics consultation for critically
and terminally ill patients with extended lengths of
stay.
Dowdy, Melvin; Robertson, Charles; Bander, John
Critical Care Medicine. 26(2):252-259, February 1998.
Table 2. Quality of communication by
group and discharge status (mean +/-
SD)

A study of proactive ethics consultation for critically Table 3. Mean, minimum, and maximum lengths of stay
and terminally ill patients with extended lengths of by group and discharge status
stay.
Dowdy, Melvin; Robertson, Charles; Bander, John
Critical Care Medicine. 26(2):252-259, February 1998.
A study of proactive ethics consultation for critically Table 4. Means, minimum, and maximum adjusted
and terminally ill patients with extended lengths of charges by group and discharge status
stay.
Dowdy, Melvin; Robertson, Charles; Bander, John
Critical Care Medicine. 26(2):252-259, February 1998.

A study of proactive ethics consultation for critically

Table 5. The r2 values for stepwise regression predicting
and terminally ill patients with extended lengths of
adjusted charges
stay.
Dowdy, Melvin; Robertson, Charles; Bander, John
Critical Care Medicine. 26(2):252-259, February 1998.