Continuing Medical Education Article

End-of-life care for the critically ill: A national intensive care unit survey*
Judith E. Nelson, MD, JD; Derek C. Angus, MD, ChB, MPH; Lisa A. Weissfeld, PhD;
Kathleen A. Puntillo, RN, DNSc, FAAN; Marion Danis, MD; David Deal, BS; Mitchell M. Levy, MD, FCCM;
Deborah J. Cook, MD, MSc (Epid), FRCPC; for the Critical Care Peer Workgroup of the Promoting Excellence in
End-of-Life Care Project
P
atients cared for in the inten-
sive care unit (ICU) are at high
risk of death (14), with one in
ve (500,000) Americans dy-
ing in hospital following ICU admission
each year (1). As the population ages and
advancing medical technology is applied
to older and sicker patients, this number
will likely increase. Data suggest that
current end-of-life care for the critically
ill and their families is inadequate, with
high prevalence of pain and other dis-
tressing symptoms (58), poor commu-
nication (911), and discordance among
patients, families, and clinicians about
care plans and goals (1214). ICU clini-
cians report awareness that treatment of
dying patients is suboptimal and recog-
nize their own deciencies in knowledge
and skills for end-of-life care (15, 16).
In response, the Institute of Medicine
has identied improved quality of care for
patients dying in the ICU as a national
health priority (17). A profusion of con-
sensus documents, position statements,
and practice guidelines have been pub-
lished on this topic (1821), and several
tutes of Health, Bethesda, MD (MD); Survey Director,
Abt Associates, Cambridge MA (DD); Director, Critical
Care Services, Medical Director, MICU, Rhode Island
Hospital, Professor of Medicine, Brown Medical
School, Providence, RI (MML); and Professor of Med-
icine & Biostatistics & Clinical Epidemiology, Academic
Chair, Critical Care, Canada Research Chair, Depart-
ment of Clinical Epidemiology, McMaster University,
Hamilton, Ontario, Canada (DJC).
Supported, in part, by the Robert Wood Johnson Foun-
dation. Dr. Nelson is the recipient of an Independent
Scientist Award (K02 AG024476) from the National
Institute on Aging. At the time of this work, Drs. Nelson
and Puntillo were Faculty Scholars of the Project on
Death in America. Dr. Cook is a Research Chair of the
Canadian Institutes of Health Research.
Address requests for reprints to: Judith E. Nelson, MD,
JD, Box 1232, Mount Sinai School of Medicine, 1
Gustave Levy Place, New York, NY 10029. E-mail:
judith.nelson@mountsinai.org.
Copyright 2006 by the Society of Critical Care
Medicine and Lippincott Williams & Wilkins
DOI: 10.1097/01.CCM.0000239233.63425.1D
LEARNING OBJECTIVES
On completion of this article, the reader should be able to:
1. Describe barriers to improved end-of-life care in the intensive care unit.
2. List strategies that are likely to improve end-of-life care.
3. Use this information in the clinical setting.
All authors have disclosed that they have no nancial relationships with or interests in any commercial companies pertaining
to this educational activity.
Lippincott CME Institute, Inc., has identied and resolved all faculty conicts of interest regarding this educational activity.
Visit the Critical Care Medicine Web site (www.ccmjournal.org) for information on obtaining continuing medical education credit.
Objective: One in ve Americans dies following treatment in an
intensive care unit (ICU), and evidence indicates the need to improve
end-of-life care for ICU patients. We conducted this study to elicit the
views and experiences of ICU directors regarding barriers to optimal
end-of-life care and to identify the type, availability, and perceived
benet of specic strategies that may improve this care.
Design: Self-administered mail survey.
Setting: Six hundred intensive care units.
Participants: A random, nationally representative sample of nurs-
ing and physician directors of 600 adult ICUs in the United States.
Interventions: Mail survey.
Measurements and Main Results: We asked participants about
barriers to end-of-life care (1 huge to 5 not at all a barrier),
perceived benet of strategies to improve end-of-life care, and avail-
ability of these strategies. From 468 ICUs (78.0% of sample), 590 ICU
directors participated (406 nurses [65.1% response] and 184 physi-
cians [31.7% response]). Respondents had a mean of 16.6 yrs (SD 7.6
yrs) of ICU experience. Important barriers to better end-of-life care
included patient/family factors, including unrealistic patient/family
expectations 2.5 (1.0), inability of patients to participate in discus-
sions 2.7 (0.9), and lack of advance directives 2.9 (1.0); clinician
factors, which included insufcient physician training in communi-
cation 2.9 (1.1) and competing demands on physicians time 3.0 (1.1);
and institution/ICU factors, such as suboptimal space for family
meetings 3.5 (1.2) and lack of a palliative care service 3.4 (1.2). More
than 80% of respondents rated 14 of 14 strategies as likely to improve
end-of-life care, including trainee role modeling by experienced cli-
nicians, clinician training in communication and symptom manage-
ment, regular meetings of senior clinicians with families, bereave-
ment programs, and end-of-life care quality monitoring. However,
few of these strategies were widely available.
Conclusions: Intensive care unit directors perceive important bar-
riers to optimal end-of-life care but also universally endorse many
practical strategies for quality improvement. (Crit Care Med 2006;
34:25472553)
KEY WORDS: intensive care; intensive care units; critical care;
healthcare surveys; palliative care; communication.
2547 Crit Care Med 2006 Vol. 34, No. 10
professional organizations such as the
Robert Wood Johnson Foundation have
undertaken initiatives to improve the
quality of end-of-life care in the ICU (22).
However, nationally representative data
on the needs and attitudes of ICU care
providers are lacking.
The objectives of this study were to
elicit the views and experiences of ICU
directors across the United States about
current practices; barriers to optimal
end-of-life care; and type, availability, and
perceived benet of potential solutions to
improve this care. We surveyed ICU di-
rectors because they are often senior cli-
nicians, they serve as opinion leaders in
the critical care community, and they
have authority to effect change.
METHODS
Instrument Development
Item and Domain Generation. We elicited
domains of inquiry and potential items at a
meeting of the Critical Care Peer Workgroup
of the Robert Wood Johnson Foundations
Promoting Excellence in End-of-Life Care
Project (22). This work group included 43
international and interdisciplinary clinicians,
investigators, and educators with clinical and
research experience in end-of-life care. We
also performed computerized, bibliographic
literature review and examined personal les
to generate items for the instrument.
Instrument Format. We used closed-ended
questions to record respondent characteristics
(23). We included three domains to classify
possible barriers to the optimal care of pa-
tients dying in the ICU: patient/family factors
(seven items), clinician factors (14 items rat-
ing the roles of both physicians and nurses),
and institutional/ICU factors (11 items). We
asked respondents the extent to which these
factors were barriers (5-point scale ranging
from 1, a huge barrier, to 5, not a barrier at
all). In addition, we asked respondents to
share their experiences and views on practices,
programs, policies, and resources that may
improve end-of-life care (yes/no option to in-
dicate the availability of the strategy in the
respondents ICU, followed by a valuation of
the strategy to improve end-of-life care: 1, not
helpful; 2, somewhat helpful; 3, very helpful;
or available but havent used).
Pretesting. To ensure clarity, consistency,
and use of the full range of response options,
we pretested the instrument through semi-
structured interviews of ten critical care phy-
sicians and ten ICU charge nurses. We also
asked members of the Critical Care Peer
Workgroup to provide written and verbal feed-
back on the domains and items.
Clinical Sensibility Assessment. In the ab-
sence of a reference standard for testing ques-
tionnaire validity, we evaluated the clinical
sensibility of our instrument, assessing face
and content validity as well as clarity, utility,
redundancy, and discriminability (24). With
the survey instrument, we distributed a clini-
cal sensibility questionnaire addressing these
constructs to a) 20 ICU nurses and 20 ICU
physicians who had previously participated in
end-of-life questionnaire development; b) 21
nurse and physician members of the Critical
Care Peer Workgroup who were not part of the
investigative team; and c) ten nurse research-
ers with expertise in end-of-life care and in-
strument development. Using the clinical sen-
sibility ratings and additional suggestions, we
further modied the instrument.
Intrarater Reliability Testing. To evaluate
test-retest reliability, we administered the
same survey on two occasions separated by 2
wks to a) 40 ICU nurses and physicians; and b)
30 nurse and physician ICU directors, ad-
vanced practice nurses caring for ICU patients,
and ICU fellows in a large, academic medical
center in New York City. Items without strong
reliability (weighted Cohens 0.7) were
omitted or modied as appropriate.
Pilot Testing. We then mailed the ques-
tionnaire for self-administration to 60 nurse
and physician directors of 30 ICUs randomly
selected from our survey sample. We incorpo-
rated feedback from pilot testing in the survey
instructions, format, and length.
Sample Development
We surveyed nurse and physician directors
of a random, stratied, nationally representa-
tive database of adult ICUs across the United
States. With permission from the American Col-
lege of Chest Physicians, the American Thoracic
Society, and the Society of Critical Care Med-
icine, we used the ICU database created for the
Committee on Manpower for Pulmonary and
Critical Care Societies (COMPACCS) ICU Sur-
vey (25) to represent the full spectrum of ICUs,
including burn and trauma units. From the
COMPACCS database, which included approx-
imately 1,200 ICUs in 1,000 hospitals (strati-
ed by size and population of the city), we
selected a random, stratied sample of 600
ICUs in 546 hospitals that preserved the na-
tional representation and diverse composition
of the original sample. As in the original da-
tabase, neonatal, pediatric, and exclusively
cardiac ICUs were excluded. We telephoned
each hospital to determine the identity of the
current nursing and physician ICU directors.
Our nal sample of 600 ICUs represented ap-
proximately 10% of the total number of ICUs
in the United States.
Survey Administration
We mailed the survey for self-administra-
tion to the nurse and physician directors of
each ICU in our sample. To maximize the
response rate, we enclosed a postage-paid re-
turn envelope, and our survey research con-
sulting rm (Abt Associates, Cambridge, MA)
contacted nonresponders by telephone, re-
mailing the questionnaire as requested (26).
We closed the database in 2003.
Institutional Review Board
Approval
This study was approved by the Institu-
tional Review Board at the Mount Sinai School
of Medicine. Written informed consent was
waived since participation was voluntary and
condential.
Data Analysis
We summarize the clinical sensibility mea-
sures using mean (SD). We used weighted
Cohens to conduct the reliability analysis.
In ve instances, more than one individual
responded to a questionnaire intended for only
one of them. Because demographic and other
information provided by respondents was
unique, we retained all such responses, to to-
tal 1,205. We report survey results using mean
and SD. For response rate calculations by dis-
cipline of ICU directors, we determined the
discipline of the respondent on the basis of the
intended nurse and physician ICU directors to
whom the survey questionnaire was originally
mailed; otherwise, we used information pro-
vided in response to the questionnaire. We
compared response rates using a likelihood
ratio chi-square statistic. To compare perspec-
tives of physician directors and nurse direc-
tors, we analyzed pairs of responses (n 85)
from physicians and nurses directing the same
ICUs using paired Students t-tests. We com-
pared responses about clinician-related barri-
ers for nurses and physicians using two-
sample Students t-tests. We considered p
.05 signicant in statistical analyses.
RESULTS
The clinical sensibility of the instru-
ment was rated highly, on a 05 scale from
least to most sensible for each component
of sensibility, mean (SD): discriminability
4.1 (0.7), clarity 4.3 (0.7), utility 4.2
(0.7), face validity 4.6 (0.4), content va-
lidity 4.1 (0.8), and redundancy 3.6 (1.3).
The intrarater reliability was also high
(weighted Cohens 0.7 for each item).
From 468 of 600 (78%) ICUs in 48
states plus the District of Columbia, we
received responses from 590 of 1,205
(49.0%) ICU directors. With nurse direc-
tors and physician directors comprising
406 (68.8%) and 184 (31.2%) of these 590
directors, respectively, response rates by
discipline were 65.1% (406 of 624) for the
nurses and 31.7% (184 of 581) for the
2548 Crit Care Med 2006 Vol. 34, No. 10
physicians (p .0001). Respondents in-
cluded ICU directors from 428 of 546
hospitals (78.4%) represented by the 600
ICUs we sampled. Detailed characteristics
of the respondents, and the ICUs and hos-
pitals in which they worked, are shown in
Table 1. Respondents had considerable
ICU practice experience, and most of the
physician directors were trained in criti-
cal care medicine. The ICUs represented
varied widely in case-mix, care model,
and size, whereas the hospitals also var-
ied in size, location, and teaching status.
As such, the sample represents a diverse,
nationally representative cohort.
Response rates did not vary signicantly
(p .24) according to the population of the
hospitals surrounding metropolitan area.
Response rate was signicantly higher,
however, for directors of ICUs in hospitals
with 500 beds compared with those di-
recting ICUs in smaller hospitals (p
.0004) (Table 1).
Perceived Barriers
More than 85% of our respondents
identied at least one large or huge bar-
rier to optimal end-of-life care for criti-
cally ill patients, and 80% identied
multiple barriers of this magnitude (Ta-
ble 2). Barriers were recognized fre-
quently in all categories (Table 2). In the
category of barriers related to patients/
families, where we observed the lowest
mean scores (i.e., greatest magnitude of
barriers), half the directors reported that
unrealistic [patient/family] expecta-
tions and inability [of patients] to par-
ticipate in discussions were large or
huge barriers (Table 2). Important barri-
ers in the clinician category (which in-
cluded separate ratings with respect to
physicians and to nurses for each item)
related to communication by physicians,
with one third of directors reporting in-
sufcient training [of physicians] in com-
munication about end-of-life issues and
inadequate communication between the
ICU [physicians] and patients/families
about appropriate goals (Table 2).
Potential Solutions
More than 80% of respondents rated
all 14 strategies as helpful (Table 3). Al-
most universally endorsed were a family
bereavement program, regular pastoral
care visits, role modeling and close su-
pervision of trainees by clinicians experi-
enced in end-of-life care, and training of
ICU clinicians in relevant communica-
tion skills. However, the availability of
almost all of these potentially helpful
strategies for participating ICUs was lim-
ited (Table 3). For example, less than a
third of respondents reported availability of
bereavement programs or the incorpora-
tion of end-of-life care within ICU, hospital,
and health system quality monitoring pro-
grams. A third of directors reported reg-
ular meetings of a senior ICU physician
and nurse with families, whereas less
than half reported clinician communica-
tion training and role modeling. Most
common were pastoral care and clinical
ethics consultants.
Nurse/Physician Comparison
Our analysis of pairs of responses from
physicians and nurses representing the
same ICUs (n 85) showed no statisti-
cally signicant differences with respect
to either barriers or solutions for end-
life-care. Both groups considered nurses
and physicians to differ in the barriers
they pose. For eight of the 14 clinician
items, directors in both disciplines per-
ceived that barriers related to physicians
were signicantly greater than those re-
lated to nurses (Table 4). Responses dif-
fered most for insufciency of clinician
training in end-of-life communication
(mean difference [SD] 0.75 [1.0], p
.0001) and unrealistic expectations on
the part of clinicians about patient prog-
nosis or effectiveness of ICU treatment
(0.73 [1.0], p .0001) (Table 4).
DISCUSSION
In this national survey, we found that
ICU directors identied important short-
Table 1. Characteristics of directors (n 590), intensive care units (ICUs), and hospitals
Discipline, n (%)
Nursing 398 (70.6)
Medicine 166 (29.4)
Gender, n (%) Nurses Physicians
Female 320 19 339 (60.3)
Male 76 147 223 (39.7)
Age, yrs, mean (SD) 46.1 (6.9)
Specialty/subspecialty training of physicians, n (%)
Critical care medicine 106 (63.9)
Internal medicine 75 (45.2)
Pulmonary 82 (49.4)
Cardiology 10 (6.0)
Surgery 36 (21.7)
Anesthesia 11 (6.6)
Other 16 (9.6)
Duration of critical care practice, yrs, mean (SD) 16.7 (7.6)
Care model of ICU, n (%)
Open 352 (62.7)
Closed 64 (11.4)
Mixed 145 (25.9)
Type of ICU, n (%)
Medical/surgical 29 (32.2)
Medical 10 (11.1)
Surgical 13 (14.4)
ICU/cardiac 22 (24.4)
Burn, trauma, neurosurgery 6 (6.6)
Other 10 (11.1)
Size of ICU, n (%)
20 beds 84 (14.9)
10 to 20 beds 308 (54.6)
10 beds 172 (30.5)
Hospital, n (%)
Size
500 beds 484 (82.3)
500 beds 104 (17.7)
University afliation 244 (41.4)
Residency training program 293 (49.8)
Critical care medicine fellowship 115 (19.6)
n 590 for items not answered by all ICU directors in our respondent group. Percentages were
computed using the actual number of responses for particular items. Percentages exceed 100% for
items with multiple, nonexclusive response options. ICU care models were dened as follows: open
ICU, patients are cared for by their primary physicians with variable intensivist input; closed ICU,
intensivist assumes primary care responsibility. We asked ICU directors to provide the number of beds
in the largest single ICU administered by the respondent.
2549 Crit Care Med 2006 Vol. 34, No. 10
comings in current end-of-life care for crit-
ically ill patients and reported a number of
major barriers to better care. Physicians,
lacking skills and time for adequate discus-
sion, were perceived to be unable to estab-
lish realistic treatment goals on which pro-
viders, patients, and families could agree.
Patients were thought too ill to contrib-
ute to conversations themselves, and
families were poorly prepared, with unre-
alistic expectations. A background ICU
culture of the technological imperative,
coupled with lack of access to palliative
care services or appropriate meeting
rooms, was also cited as a factor contrib-
uting to suboptimal care. Accordingly,
better physician training and education
in end-of-life care and provision of ade-
quate meeting room space and support
services were perceived as potentially
useful solutions.
Table 2. Barriers to optimal end-of-life care
Survey Item
No. (%) Rating Large
or Huge Mean (SD) Rating
Patient/family barriers
Unrealistic patient and/or family expectations about prognosis or effectiveness of ICU treatment 272 (48.7) 2.5 (1.0)
Inability of many patients to participate in treatment discussions 259 (46.6) 2.7 (0.9)
Lack of advance directives 191 (34.2) 3.0 (1.0)
Disagreements within families about care goals 184 (33.0) 2.9 (1.0)
Absence of a surrogate decision maker for patients lacking decisional capacity 136 (24.3) 3.1 (0.9)
Disagreements between patients/families and the ICU team about care goals 91 (16.3) 3.4 (0.9)
Refusals by patients/families to forgo life-sustaining treatments for religious reasons 49 (8.8) 3.8 (0.8)
Institutional/ICU factors
Suboptimal space for meeting with families of ICU patients 142 (25.3) 3.5 (1.2)
Lack of a palliative care service to which a dying patient can be transferred 140 (25.0) 3.4 (1.2)
The technological imperative of the ICU 124 (22.3) 3.4 (1.1)
ICU admission of patients not predicted to survive critical illness 92 (16.6) 3.5 (1.0)
Insufcient recognition by colleagues or institutional leadership of the importance of optimal
end-of-life care
88 (15.7) 3.6 (1.1)
Inadequate support services for grieving families 73 (13.1) 3.7 (1.0)
Failure to locate existing advance directives 72 (12.9) 3.6 (0.9)
Lack of consultants with special expertise in management of symptoms that are distressing to
ICU patients
66 (11.8) 3.9 (1.1)
Insufcient continuity of care within the ICU due to physician stafng patterns 60 (10.9) 3.8 (1.0)
Insufcient continuity of care during transition into the ICU 32 (5.7) 4.0 (0.8)
Insufcient continuity of care within the ICU due to nursing stafng patterns 24 (4.3) 4.1 (0.8)
MD RN MD RN
Clinician factors
Insufcient clinician training in communication about end-of-life care issues 210 (37.2) 80 (14.2) 2.9 (1.1) 3.6 (0.9)
Competing demands for clinicians time 209 (37.1) 127 (22.5) 3.0 (1.1) 3.5 (1.0)
Inadequate communication between the ICU team and patient/families about
appropriate goals of care
164 (29.1) 69 (12.2) 3.2 (1.1) 3.8 (0.8)
Fear of legal liability for forgoing life-sustaining treatments 155 (27.5) 70 (12.4) 3.3 (1.2) 4.0 (0.9)
Unrealistic expectations by clinicians about patient prognosis or effectiveness of ICU
treatment
148 (26.2) 48 (8.5) 3.3 (1.1) 4.0 (0.8)
Insufcient clinician training in techniques for forgoing life-sustaining treatment
without patient suffering
141 (25.0) 64 (11.4) 3.3 (1.1) 3.8 (0.9)
Inadequate communication between the ICU team and other clinicians about
appropriate care goals
138 (24.5) 89 (15.8) 3.3 (1.1) 3.7 (1.0)
Clinicians reluctance to use opioids or sedatives because of concern about side effects 129 (22.9) 66 (11.7) 3.5 (1.1) 4.0 (1.0)
Limited reimbursement for time spent providing end-of-life care 127 (22.5) 76 (13.5) 3.7 (1.3) 4.2 (1.1)
Psychological/emotional stresses of providing care to dying patients 117 (20.7) 87 (15.4) 3.3 (1.0) 3.6 (0.9)
Fear of legal liability for administering opioids or sedatives to dying patients 114 (20.2) 74 (13.1) 3.6 (1.1) 4.0 (1.0)
Insufcient attention to diverse cultural norms and customs with respect to dying,
death, and grief
103 (18.3) 83 (14.7) 3.6 (1.0) 3.8 (0.9)
Insufcient clinician training in the management of symptoms that are distressing to
critically ill patients
97 (17.2) 60 (10.6) 3.5 (1.0) 3.7 (0.8)
Inadequate communication on the ICU team about appropriate goals of care 94 (16.7) 61 (10.8) 3.6 (1.0) 3.9 (0.9)
ICU, intensive care unit; MD, physicians; RN, nurses.
The survey instrument listed these 32 factors in three categories (patient/family factors, institutional/ICU factors, and clinician factors) as possible
barriers to optimal care of patients dying in ICUs in the United States. We asked respondents the extent to which these factors were barriers (5-point scale:
1, a huge barrier; 2, a large barrier; 3, a moderate barrier; 4, a minimal barrier; 5, not a barrier at all); the lowest scores represent the greatest perceived
barriers. For clinician factors, all respondents were asked to answer separately for both physicians and nurses. We present for each category of barriers the
number and percentage of respondents (ordered from largest to smallest within each of the three categories) who rated the barrier as large or huge
and the mean (SD) rating for each item. Barriers attributed to physicians and nurses are shown separately in the clinician factors category; ratings provided
by physician and nurse respondents are combined. As discussed in the results section of the text, most physician and nurse ICU directors responded
similarly, both groups perceived differences in barriers between those related to physicians and those related to nurses (for most items, respondents ratings
for nurses and for physicians were different), and in each case, the barrier was considered larger with respect to physicians than nurses.
2550 Crit Care Med 2006 Vol. 34, No. 10
Nurse and physician directors generally
agreed about the magnitude of barriers and
the potential for specic strategies to im-
prove end-of-life care, reinforcing the broad
concerns about challenges for good end-of-
life care. Nurse directors reported greater
concern than did physician directors about
the role and skill of physicians, but both
sets of directors believed that physicians
posed greater barriers than nurses, echoing
the widely held perception that the role of
physicians in end-of-life care requires par-
ticular attention. In particular, respondents
cited several common decits in physician
communication, including a lack of regu-
larly scheduled meetings of senior ICU cli-
nicians with families, a perception that
communication between physicians and
patients or families was inadequate, and a
perception that physicians received insuf-
cient training in the appropriate commu-
nication skills. These ndings are consis-
tent with other work documenting
deciencies in clinician-patient/family
communication during serious illness (9
11, 27). Although the vast majority of re-
sponding directors endorsed training of
ICU clinicians in communication skills as
well as role-modeling and close supervision
of trainees by clinicians with experience in
end-of-life decision making and care, most
ICUs represented in our survey lacked such
programs.
Our ndings suggest several priorities
for better end-of-life care of the critically
ill. First, ICU clinicians need training in
the knowledge and skills to communicate
with dying patients and their families,
including the ability to discuss complex,
distressing information clearly and sensi-
tively; to approach decision making in
the context of prognostic uncertainty; to
elicit patient preferences; to assist pa-
tients, families, and colleagues in estab-
lishing realistic and appropriate treat-
ment goals; and to resolve conicts. A
training program specically designed to
prepare clinicians for the unique chal-
lenges of end-of-life communication in
the ICU setting would provide a valuable
addition to existing programs, such as
EPEC (Education for Physicians in End-
of-Life Care) and ELNEC (the End-of-Life
Nursing Education Consortium) (2830).
Education that includes role-modeling
and close supervision of trainees by expe-
rienced clinicians is likely to be particu-
larly effective (31, 32). For ICU families,
who consistently rank communication at
the same high level of importance as clin-
ical skills (3335), improved communica-
tion may reduce anxiety and distress (36)
and enhance satisfaction with care (37).
Other potential solutions that could be
readily adapted from existing sources in-
clude training in symptom management,
use of formal scales for symptom assess-
ment and documentation, broader avail-
ability of bereavement services, accom-
modation of cultural diversity among
patients and families, scheduling of staff
to promote continuity of care for dying
patients, and mechanisms for providing
emotional support to staff.
Given the widespread perception among
our respondents and others (38) that unre-
Table 3. Strategies to improve end-of-life care in the intensive care unit (ICU)
Strategy to Improve Care No.
% Reporting
Available
% Rating
as Helpful
End-of-life care quality monitoring 425 26.5 85.4
Bereavement program or service 423 29.4 96.7
Regular meetings of senior ICU physician and nurse
with patients families
446 35.1 92.6
Training of ICU clinicians in end-of-life communication
skills
440 40.0 94.1
Role modeling and supervision of trainees by clinicians
experienced in end-of-life care
436 41.0 95.2
Formal mechanism for emotional support of staff caring
for dying patients
450 41.2 87.3
Access to palliative care consultants 461 45.9 88.3
Training of ICU clinicians in symptom management 461 53.7 93.5
Scheduling staff to promote continuity of care for dying
patients
472 60.3 92.6
Formal system for scaled assessment and charting of
patients symptoms
468 61.8 88.0
Method to help resolve differences about appropriate
care goals
480 63.1 90.0
Resources to accommodate diversity among patients/
families at the end-of-life
493 72.3 92.5
Access to clinical ethics consultants 517 83.8 81.9
Regular pastoral care visits to the ICU 526 88.7 96.4
For 14 strategies suggested as methods to improve the care of patients dying in ICUs, we asked
respondents to indicate availability in, or to, their ICUs (yes/no). For available strategies, respondents
rated the extent to which each had been helpful in meeting needs of staff, patients, and/or families in
end-of-life care in the ICU: not helpful, somewhat helpful, very helpful, or available but havent used.
For strategies they did not have available, respondents were asked to indicate the extent to which you
believe it would be helpful in meeting these needs. We present the number and percent of respondents
who rated each strategy as somewhat or very helpful and the availability of these strategies to their
ICUs.
Table 4. Comparison of physician barriers and nurse barriers to better end-of-life care
Difference (SD)
(RN-MD) p Value
Insufcient training in communication about end-of-life issues 0.75 (1.0) .0001
Unrealistic expectations about prognosis/effectiveness of ICU treatment 0.73 (1.0) .0001
Inadequate communication between ICU team and patients/families 0.69 (0.93) .0001
Fear of legal liability for forgoing life-sustaining treatments 0.66 (0.97) .0001
Limited reimbursement for time spent providing end-of-life care 0.58 (1.1) .0001
Competing demands for clinicians time 0.54 (0.99) .0001
Reluctance to use opioids/sedatives because of concern about side effects 0.51 (1.0) .0001
Insufcient training in techniques for forgoing life-sustaining treatments
without patient suffering
0.51 (0.90) .0001
RN, nurses; MD, physicians; ICU, intensive care unit.
For each item in the clinician factors category of the survey, we compared the rating by each
individual respondent regarding physicians with that individuals response regarding nurses. (As
discussed in the text, respondents rated the magnitude of each potential barrier separately with respect
to physicians and to nurses, answering parallel sets of response options for each discipline for each
item.) We computed the mean (SD) of the difference (rating with respect to nursesrating with respect
to physicians) using paired t-tests. Differences for all 14 clinician factors were statistically signicant
(p .0001). We present those of greatest importance, as indicated by a mean difference 0.50 on the
5-point scale.
2551 Crit Care Med 2006 Vol. 34, No. 10
alistic patient and family expectations are a
major barrier to high-quality end-of-life
care, public education also requires greater
attention. Wider availability of life-sustain-
ing treatments has made it difcult for pa-
tients and families to understand that
sometimes these fail to sustain life and
serve only to prolong death. Expectations
for favorable outcomes often do not take
account of the adverse consequences of
prior chronic illnesses, which affect an in-
creasing proportion of our ICU population.
Although society appears to value good
end-of-life care in the ICU (39), expecta-
tions of current ICU technology and treat-
ment may be unrealistic. Since death fol-
lowing admission to an ICU is more
common than was appreciated previously
(1), public educational programs and other
methods are needed to help citizens under-
stand critical illness and appropriate treat-
ment goals in the ICU.
Survey data can provide insights into
experiences, beliefs, and stated practice
patterns; however, such data are indirect
compared with actual observations. Al-
though we found that most ICUs do not
currently monitor end-of-life care quality
in a systematic way, our survey responses
could inform the design and implemen-
tation of instruments to evaluate im-
provement in the quality of care provided
to dying ICU patients. Valid and respon-
sive tools to measure the quality of care
(19), family satisfaction (40), and quality
of death and dying (19, 41, 42) now exist
and could be used for baseline informa-
tion, benchmarking, and formal evalua-
tion of quality improvement initiatives.
This study is also limited in that we ob-
tained a moderate response rate for phy-
sician ICU directors, which has also char-
acterized prior mail surveys of these
physicians (25); we recognize that nonre-
sponding physicians (and nurses) may
have perceived the issues differently and
represent a potential source of bias. How-
ever, we obtained responses from almost
80% of the ICUs we sampled and from two
thirds of the nursing directors. We sought
the perspective of ICU directors because
they typically have considerable clinical
experience, administrative knowledge,
and status as opinion leaders in the crit-
ical care community, with authority to
effect change. However, their views on
solutions to improve end-of-life care may
not improve patient-centered outcomes;
further studies are needed to test effect.
The perspectives of patients and families
are also critical.
This is the rst survey of ICU directors
in the United States about end-of-life
care, and the rst study of both nurses
and physicians perceptions of barriers to,
and solutions for, best practice. Although
previous surveys of ICU practitioners in
North America about knowledge, atti-
tudes, or practices about end-of-life care
have been conducted, most have focused on
forgoing of life-sustaining treatments or on
euthanasia and assisted suicide (13, 14, 43
45). A survey of critical care staff nurses
addressed certain obstacles and helpful be-
haviors for end-of-life care, focusing pri-
marily on issues related to patients, fami-
lies, and nurses (38). Our survey had a
broader scope, eliciting an interdisciplinary
perspective on prevailing end-of-life prac-
tices and a wide range of barriers and strat-
egies for optimal care. Other strengths of
this work include rigorous development
and testing of a reliable and clinically sen-
sible survey instrument. We used random-
ized trials in survey methodology to maxi-
mize the completeness and validity of data
collection (23) and the response rate (26)
We obtained perspectives from a diverse
group of nurse and physician ICU directors
working in nationally representative hospi-
tals, enhancing the generalizability of our
ndings. Respondents generated many
practical suggestions viewed by both nurse
and physician directors as likely to improve
the quality of end-of-life care in the ICU.
CONCLUSIONS
As the population ages and an increas-
ing number of Americans die in the ICU,
concern continues to mount about how
best to balance the use of invasive lifesav-
ing technologies with compassionate
end-of-life care. It is therefore essential to
understand how end-of-life care is cur-
rently provided in ICUs and how it might
be improved. Key to this understanding
are the views of relevant stakeholders. In
this national survey, physician and nurse
ICU directors expressed concerns about
the ability to provide ideal end-of-life
care, and they strongly endorsed a range
of practical potential solutions. Next
steps in this eld of research include ex-
ploring the viewpoints of other stake-
holders and formally evaluating the sug-
gested strategies for improving end-of-
life care in the ICU.
ACKNOWLEDGMENTS
We are grateful to the members of the
Critical Care Peer Workgroup of the Robert
Wood Johnson Foundations Promoting
Excellence in End-of-Life Care Project for
their contributions to this work. The
views expressed are those of the authors
and do not necessarily reect policies of
the National Institutes of Health or the
United States Department of Health and
Human Services.
REFERENCES
1. Angus DC, Barnato AE, Linde-Zwirble WT, et
al: Use of intensive care at the end of life in
the United States: An epidemiologic study.
Crit Care Med 2004; 32:638643
2. The Acute Respiratory Distress Syndrome
Network: Ventilation with lower tidal vol-
umes as compared with traditional tidal vol-
umes for acute lung injury and the acute
respiratory distress syndrome. N Engl J Med
2000; 342:13011308
3. Groeger JS, Lemeshow S, Price K, et al: Mul-
ticenter outcome study of cancer patients
admitted to the intensive care unit: A prob-
ability of mortality model. J Clin Onc 1998;
16:761770
4. Knaus WA, Wagner DP, Zimmerman JE, et
al: Variations in mortality and length of stay
in intensive care units. Ann Intern Med 1993;
118:753761
5. Nelson JE, Meier DE, Oei EJ, et al: Self-
reported symptom experience of critically ill
cancer patients receiving intensive care. Crit
Care Med 2001; 29:277282
6. Stein-Parbury J, McKinley S: Patients expe-
rience of being in an intensive care unit: A
select literature review. Am J Crit Care 2000;
9:2027
7. Puntillo KA: Pain experience of intensive
care unit patients. Heart Lung 1990; 19:
525533
8. Desbiens NA, Wu AW, Broste SK, et al: Pain
and satisfaction with pain control in seri-
ously ill hospitalized adults: Findings from
the SUPPORT research investigations. Crit
Care Med 1996; 24:19531961
9. Curtis JR, Wenrich MD, Carline JD, et al:
Patients perspectives on physician skill in
end-of-life care: Differences between patients
with COPD, cancer, and AIDS. Chest 2002;
122:356362
10. Azoulay E, Chevret S, Leleu G, et al: Half the
families of ICU patients experience inade-
quate communication with physicians. Crit
Care Med 2000; 8:30443049
11. Teno JM, Fisher E, Hamel MB, et al: Deci-
sion-making and outcomes of prolonged ICU
stays in seriously ill patients. J Am Geriatr
Soc 2000; 48:S70S74
12. Breen CM, Abernethy AP, Abbott KH, et al:
Conict associated with decisions to limit
life-sustaining treatment in intensive care
units. J Gen Intern Med 2001; 16:283289
13. Cook DJ, Guyatt GH, Jaeschke R, et al: De-
terminants in Canadian health care workers
of the decision to withdraw life support from
the critically ill. JAMA 1995; 273:703708
2552 Crit Care Med 2006 Vol. 34, No. 10
14. Asch DA, Hansen-Flaschen J, Lanken P: De-
cisions to limit or continue life-sustaining
treatment by critical care physicians in the
United States: Conicts between physicians
practices and patients wishes. Am J Respir
Crit Care Med 1995; 151:288292
15. Solomon MZ, ODonnell L, Jennings B, et al:
Decisions near the end of life: Professional
views on life-sustaining treatments. Am J
Public Health 1993; 83:1423
16. Vincent JL: European attitudes toward ethi-
cal problems in intensive care medicine: Re-
sults of an ethical questionnaire. Intensive
Care Med 1990; 16:256264
17. Field MJ, Cassel CK (Eds): Approaching
Death: Improving Care at the End of Life.
Washington, DC, National Academy Press
(Institute of Medicine), 1997
18. Carlet J, Thijs LG, Antonelli M, et al: Chal-
lenges in end-of-life care in the ICU/
Statement of the 5th International Consen-
sus Conference in Critical Care: Brussels,
Belgium, April 2003. Intensive Care Med
2004; 30:770784
19. Clarke EB, Curtis JR, Luce JM, et al: Quality
indicators for end-of-life care in the intensive
care unit. Crit Care Med 2003; 31:22552262
20. Truog RD, Cist AF, Brackett SE, et al: Rec-
ommendations for end-of-life care in the in-
tensive care unit: The Ethics Committee of
the Society of Critical Care Medicine. Crit
Care Med 2001; 29:23322348
21. Danis M, Federman D, Fins JJ, et al: Incor-
porating palliative care into critical care ed-
ucation: Principles, challenges, and opportu-
nities. Crit Care Med 1999; 27:20052013
22. Promoting Excellence in End-of-Life Care.
Available at: http://www.promotingexcellence.
org. Accessed March 15, 2005
23. Grifth LE, Cook DJ, Guyatt GH, et al: Com-
parison of open and closed questionnaire for-
mats in obtaining demographic information
from Canadian general internists. J Clin Epi-
demiol 1999; 52:9971005
24. Feinstein AR: Clinical sensibility. In: Clini-
metrics. Feinstein AR (Ed). New Haven, CT,
Yale University Press, 1987, pp. 141166
25. Angus DC, Kelley MA, Schmitz RJ, et al:
Caring for the critically ill patient. Current
and projected workforce requirements for
care of the critically ill and patients with
pulmonary disease: Can we meet the require-
ments of an aging population? JAMA 2000;
284:27622770
26. Edwards P, Roberts I, Clarke M, et al: Increas-
ing response rates to postal questionnaires:
Systematic review. BMJ 2002; 324:11831192
27. Kutner JS, Steiner JF, Corbett KK, et al:
Information needs in terminal illness. Soc
Sci Med 1999; 48:13411352
28. The EPEC project. Available at: http://www.
epec.net/EPEC. Accessed March 15, 2005
29. American Association of Colleges of Nursing.
Available at: http://www.aacn.nche.edu/
ELNEC. Accessed March 15, 2005
30. End of Life/Palliative Education Resource
Center. Available at: http://www.eperc.mcw.
edu. Accessed March 15, 2005
31. Stevens L, Cook DJ, Guyatt G, et al: Educa-
tion, ethics and end-of-life decisions in the
intensive care unit. Crit Care Med 2002; 30:
290296
32. Prendergast TJ, Sullivan AM, Arnold RM, et al:
Fellowship education in end-of-life care. Am J
Respir Crit Care Med 2002; 165(Suppl):B4
33. Johnson D, Wilson M, Cavanaugh B, et al:
Measuring the ability to meet family needs in
an intensive care unit. Crit Care Med 1998;
26:266271
34. Hickey M: What are the needs of families of
critically ill patients? A review of the litera-
ture since 1976. Heart Lung 1990; 19:
401415
35. Molter N: Needs of relatives of critically ill
patients: A descriptive study. Heart Lung
1979; 8:332339
36. Pochard F, Azoulay E, Chevret S, et al: Symp-
toms of anxiety and depression in family
members of intensive care unit patients: Eth-
ical hypothesis regarding decision-making
capacity. Crit Care Med 2001; 29:18931897
37. Heyland DK, Cook DJ, Rocker GM, et al:
Decision-making in the ICU: Perspectives of
the substitute decision-maker. Intensive
Care Med 2003; 29:7582
38. Kirchhoff KT, Beckstrand RL: Critical care
nurses perceptions of obstacles and help-
ful behaviors in providing end-of-life care
to dying patients. Am J Crit Care 2000;
9:96105
39. Bryce CL, Loewenstein G, Arnold RM, et al:
Quality of death: Assessing the importance
placed on end-of-life treatment in the inten-
sive care unit. Med Care 2004; 42:423431
40. Heyland DK, Rocker GM, Dodek PM, et al:
Family satisfaction with care in the intensive
care unit: Results of a multiple center study.
Crit Care Med 2003; 30:14131418
41. Curtis JR, Patrick DL, Engelberg RA, et al: A
measure of the quality of dying and death:
Initial validation using after-death interviews
with family members. J Pain Symptom Man-
age 2002; 24:1731
42. Steinhauser KE, Bosworth HB, Clipp EC, et
al: Initial assessment of a new instrument
to measure quality of life at the end of life.
J Palliat Med 2002; 5:829841
43. Society of Critical Care Medicine Ethics
Committee: Attitudes of critical care profes-
sionals concerning forgoing life-sustaining
treatments. Crit Care Med 1992; 20:320326
44. Puntillo KA, Benner P, Drought T, et al:
End-of-life issues in intensive care units: A
national random survey of nurses knowledge
and beliefs. Am J Crit Care 2001; 10:216229
45. Asch DA: The role of critical care nurses in
euthanasia and assisted suicide. N Engl
J Med 1996; 334:13741379
2553 Crit Care Med 2006 Vol. 34, No. 10