Australian Critical Care (2010) 23, 13—19

RESEARCH PAPER

End of life management of adult patients in an

Australian metropolitan intensive care unit: A
retrospective observational study
Melissa Jane Bloomer RN, BN, MNP, MPET ∗,
Ravindranath Tiruvoipati MCh, MSc., FRCSEd,
Michael Tsiripillis MBBS,
John A. Botha M Med, FCP (SA), FRACP, FJFICM

Peninsula Health, Continuing Education and Development Unit, Hastings Road,

Frankston, VIC 3199, Australia

Received 2 March 2009 ; received in revised form 31 July 2009; accepted 12 October 2009

KEYWORDS Summary
Intensive care; Background: Death in the intensive care unit is often predictable. End of life manage-
Death; ment is often discussed and initiated when futility of care appears evident. Respect
End of life; for patients wishes, dignity in death, and family involvement in the decision-making
Communication
process is optimal. This goal may often be elusive.
Purpose: Our purpose was to review the end of life processes and family involvement
within our Unit.
Methods: We conducted a chart audit of all deaths in our 10 bed Unit over a 12-month
period, reviewing patient demographics, diagnosis on admission, patient acuity,
expectation of death and not-for-resuscitation status. Discussions with the family,
treatments withheld and withdrawn and extubation practices were documented.
The presence of family or next-of-kin at the time of death, the time to death after
withdrawal of therapy and family concerns were recorded.
Results: There were 70 patients with a mean age of 69 years. Death was expected
in 60 patients (86%) and not-for-resuscitation was documented in 58 cases (85%).
Family discussions were held in 63 cases (90%) and treatment was withdrawn in
34 deaths (49%). After withdrawal of therapies, 31 patients (44%) died within 6 h.
Ventilatory support was withdrawn in 24 cases (36%). Family members were present
at the time of death in 46 cases (66%). Family concerns were documented about the
end of life care in only 1 case (1.4%).

∗ Corresponding author. Tel.: +61 0402 472 334/397847732; fax: +61 397847213.
E-mail address: smakb@ozemail.com.au (M.J. Bloomer).

1036-7314/ $ — see front matter © 2009 Australian College of Critical Care Nurses Ltd. Published by Elsevier Australia (a division of Reed International Books Australia Pty Ltd). All rights reserved.
doi:10.1016/j.aucc.2009.10.002
14
Conclusion: Our data suggests that death in our Unit was often predictable and that
end of life management was a consultative process.
© 2009 Australian College of Critical Care Nurses Ltd. Published by Elsevier Australia
(a division of Reed International Books Australia Pty Ltd). All rights reserved.
Introduction
Death in the intensive care unit (ICU) raises many
ethical, familial and clinical issues. Despite the fact
that the primary goal of intensive care treatment
is to assist patients to survive acute threats to
their health,1 death remains inevitable for some
critically ill patients, with approximately one in
five deaths in the United States now occurring in
the ICU.2 The development of life-sustaining treat-
ments has resulted in some patients remaining on
life support despite futility of care, and when death
seems inevitable.3
Societal expectations may focus on active treat-
ment in the ICU even when clinicians anticipate
that death is imminent and expected.4 Even though
death in the ICU may sometimes be perceived as a
failure by family members, transparent and sensi-
tive family discussions will assist families to accept
end of life care and enable palliation.4 Clinicians
in the ICU may guide the family towards a unified
acceptance of end of life care and eventual death
of their loved one.4
Whilst withdrawal of care may often be the
most humane approach to terminal patient care,
data suggests a variable approach to end of life
management in the ICU.3,5,6 Various factors have
been identified in the literature that may impact of
end of life management, including organisational
issues,3 the clinician preference for treatment
options,1 the language used by clinicians and per-
sonnel involved in the end of life care.1 Most
importantly, end of life management should ensure
that patient care and family communication needs
remain the priority.1,4
The Australian and New Zealand Intensive Care
Society (ANZICS) guidelines on withholding and
withdrawing treatment recommend that medical
consensus should be achieved, followed by accep-
tance from the patient’s family or next-of-kin
regarding the treatment plans for end of life care,7
an opinion well supported in recent literature.1,4,6,8
The guidelines also stress the value of establishing
local guidelines for managing end of life (EOL) in the
ICU. New South Wales Health released Guidelines
for end-of-life care and decision making in 2005,9
which specify several principles for end of life care
including respect for life and care in dying, appro-
priate withholding and withdrawal of life-sustaining
M.J. Bloomer et al.
treatment, a collaborative approach to care and
continuous improvement.9
Enhancing our understanding and improving end
of life care were the original motivators for this
review. We sought to establish if death in our ICU
was predictable and, if end of life management in
our ICU was a consultative process with substantial
dialogue between family or next-of-kin and attend-
ing ICU staff. The palliative process in the ICU was
furthermore an area that we considered required
analysis as part of our strategy to improve care of
the dying patient in ICU.
Methods
The study was a retrospective observational study
of all deaths in a 10 bedded ICU over a 12-month
period. The Human Research and Ethics Committee
of the Health Service exempted the study from for-
mal submission as it was a retrospective review and
data was de-identified. The ICU is a mixed medi-
cal/surgical ICU that admits critically ill patients
from the Mornington Peninsula region. The ICU does
not admit neurosurgical patients or patients who
have had cardiothoracic surgery. The consultant
staff consists of 3.2 effective full time equivalent
Intensivists who are all Fellows of the Joint Fac-
ulty of Intensive Care Medicine. There were 6 junior
medical staff members at the time of the review,
and over 55 nursing staff with varying levels of
qualifications and experience. Patient demographic
data relating to age, sex, reason for admission,
acute physiology and chronic health evaluation
(APACHE) III and simplified acute physiological score
(SAPS) II was recorded.
The reason for ICU admission, duration of ICU
stay, time of death and cause of death was noted.
Organ failure on admission was recorded. Doc-
umentation in the case records that death was
expected was noted. Where death was unexpected,
the performance of cardiopulmonary resuscitation
(CPR) was noted. In anticipated deaths, charts were
reviewed to ensure that not-for-resuscitation (NFR)
documentation had been completed. Where death
was anticipated, documentation that the patient’s
condition was discussed with the relatives was
reviewed. If no discussion had occurred, the rea-
End of life management of adult patients
sons charted in the folder were noted. Attendance
during end of life discussions by hospital chaplains,
social workers and the palliative care team was also
recorded. The documented frequency of end of life
meetings was also noted. Any documentation in the
hospital chart as to why consideration was given to
withholding or withdrawing therapy was recorded.
If treatment was withheld, the nature of treatment
withheld was recorded. If a decision had been made
to withdraw therapy, the type of therapy withdrawn
was noted. Specifically, withdrawal of mechanical
ventilation, inotropic drug therapy, renal replace-
ment therapy and nutrition was noted. The number
of patients extubated at the time of death was also
established. Analgesic and sedative drugs admin-
istered at the time of death were reviewed and
noted. Where the data was available, the time
elapsed from withdrawal of therapy to death was
noted, as was the time of death. If documented,
the presence of relatives at the time of death was
recorded. Available data on concerns raised by the
family during the end of life process in the ICU were
noted, and if a death was referred to the coroner
this was documented.
The aspects of end of life management selected
for evaluation in this review were chosen for their
clinical relevance to the process of palliation and
end of life care in this intensive care unit.
Results
Seventy patients died in the ICU during the 12-
month period with a mean age of 69.33 years. The
mean APACHE II score was 27.30 and the mean
APACHE III and SAPS II scores were 97.44 and 58.78,
respectively. The mean duration of stay in ICU was
3.67 days prior to death. Cardiac arrest (30%) and
sepsis (18.5%) were common reasons for ICU admis-
sion (Table 1). The number of organs failed, as
identified at the time of admission to ICU, was
also substantial with three or more organ failure
Table 1 Reason for ICU admission.
Reason for ICU admission n (%)
Cardiac arrest 21 (30)
Sepsis 13 (18.5)
Pneumonia 13 (18.5)
End stage chronic obstructive 3 (4.2)
pulmonary disease
Postoperative 8 (11.4)
Intracranial pathology 5 (7.1)
Other 7 (10)
Total 70 (100)
15
Table 2 Number of organs failed at admission.
Number of organs failed at admission n (%)
0 6 (8.6)
1 19 (27.1)
2 22 (31.4)
3 17 (24.3)
4 2 (2.9)
Missing data 4 (5.7)
Total 70 (100)
occurring in 17 (24.3%) of the deaths reviewed
(Table 2).
Documentation in the charts revealed that death
was expected in 60 cases (85.7%). Chart review con-
firmed that NFR procedures had been recorded in 58
(84.5%) of the deaths. Discussions with patient fam-
ilies had occurred in 63 of the 70 deaths (90%). The
maximum number of discussions that occurred with
the family of a patient prior to death was 7, with the
mean number of family discussions 2.3. Discussions
regarding end of life care did not occur in 7 deaths,
with no explanation provided in three of these
cases. The reasons documented in the remaining
4 deaths were unexpected death, estranged fam-
ily, no relatives and sudden clinical deterioration.
Members of the hospital chaplaincy were involved
in family discussions in 16 cases (22.9%) The pallia-
tive care team was involved in discussions in 3 cases
(4.3%) and the social workers in 3 cases (4.3%).
Members of both the hospital palliative care team
and hospital chaplaincy were involved in 2 discus-
sions (2.9%) where death was considered probable.
The reasons for withdrawal and withholding of
therapy were documented as futility of care in 42
cases (60.7%), followed by irreversible neurologi-
cal injury in 13 cases (18%), life support against the
patient’s wishes in 7 cases (9.8%) and disseminated
malignancy in 2 cases (3.3%). The most frequently
withheld therapy was mechanical ventilation in 7
cases (10%) followed by various combinations of
therapy withdrawal (Table 3). Inotropic drug ther-
apy alone was withdrawn in 21 deaths (30%) and in
other deaths, various treatments either individually
or in combination, were withdrawn (Table 4).
At the time that a decision had been made
to withdraw therapy and commence palliation, 4
patients (5%) were not intubated, and 24 of the
intubated patients (36%) were extubated. Review of
the charts confirmed that 46 patients (66.2%) were
receiving analgesia at the time of death. The most
frequently administered medication at the time of
death was a combination of morphine and midazo-
lam in 30 cases (34.3%). Morphine was used as a
single agent in 11 cases (15.7%) and midazolam in
16 M.J. Bloomer et al.

Table 3 Therapy withheld.

There were family concerns about the process of
death in the ICU in only 1 of the 70 deaths reviewed.
Therapy withheld n (%) These were subsequently addressed after consulta-
No therapy withheld 46 (65.7) tion with the medical staff. The treating medical
Inotropic drug therapy 1 (1.4) staff notified the coroner in 13 of the deaths and
Intravenous or oral nutrition 2 (2.9) autopsies were performed in 5 of the deaths.
Renal replacement therapy 6 (8.6)
Renal replacement therapy & Ventilatory 1 (1.4)
support
Surgical intervention 1 (1.4)
Ventilatory support 7 (10)
Discussion
Ventilatory support & inotropic drug 1 (1.4)
therapy Our data confirms that death in this ICU was seldom
Ventilatory support, inotropic drug 1 (1.4) unexpected and that the transition to palliation
therapy & renal replacement therapy was a consultative process. Furthermore the mea-
Ventilatory support, renal replacement 1 (1.4) sured predictors of mortality such as APACHE II and
therapy & surgical intervention III scores and SAPS II scores confirmed how criti-
Missing data 3 (4.3) cally ill the patients were on admission to the ICU.
Total 70 (100) The APACHE II scores of patients who died in our
ICU are consistent with a previous study that found
an APACHE II score of 25 ± 9 in ICU patients where
Table 4 Therapy withdrawn. life support was withdrawn.10 Furthermore the sub-
Therapy withdrawn n (%) stantial number of organs failed on admission to ICU
indicated a potentially high ICU mortality.
No therapy withdrawn 36 (51.4)
Antibiotic therapy & steroid 1 (1.4) This retrospective review indicated the value
therapy placed on communication by the staff of the inten-
Diuretic therapy 1 (1.4) sive care unit. Numerous previous studies have
Inotropic drug therapy 21 (30) highlighted the importance of communication with
Inotropic drug therapy & total 3 (4.3) the patient’s family or next-of-kin in end of life
parenteral nutrition discussions.1,4,9,11—15 In these studies regular dia-
Inotropic drug therapy & 2 (2.9) logue was essential to family satisfaction, but the
ventilatory support frequency and duration of family meetings was not
Inotropic drug therapy & renal 1 (1.4) specified. Whilst these studies confirmed the impor-
replacement therapy
tance of communication in this context, there was
Renal replacement therapy 2 (2.9)
conflicting data regarding the role of the family
Total parenteral nutrition 2 (2.9)
Ventilatory support 1 (1.4) and next-of-kin in decision-making. The study by
Cook et al.,16 for example, revealed a preference
Total 70 (100) for a shared responsibility in decision-making by
physician and family. This contrasted with data
2 cases (2.9%). Relatives were present at the time from Azoulay et al. that confirmed 85% of fam-
of death in 46 deaths (65.9%). After withdrawal of ily members in his study did not wish to take
therapy, the median time to death was 105 min, part in end of life decision-making.17 This may
with 20 deaths (28%) occurring within 1 h of therapy reflect cultural differences between North Amer-
withdrawal, 31 deaths (44%) occurring between 1 ica and Europe. The NSW Health End-of-Life and
and 6 h, 5 deaths (8%) occurring between 6 and 12 h Decision-Making Guidelines9 recommend a shared
and death after 12 h in only 14 cases (20%). There decision-making approach, which involves consen-
did not appear to be a particular time of day when sus from the treating team and the family, thus
death was most likely to occur in the ICU (Table 5). avoiding additional stress on the family through the
burden of decision-making. These guidelines are
consistent with recommendations of the American
Table 5 Timeframe of death.
College of Critical Care Medicine1 who also support
Timeframe of death n (%) family involvement in decision-making.
8 am to 2 pm 22 (31.4) The recognition that death was likely to occur
2 pm to 8 pm 24 (32.9) necessitated communication with family members,
8 pm to 8 am 25 (35.7) and documentation of NFR status where appro-
Total 70 (100) priate. Our review demonstrated that death was
expected in 60 cases (86%) and family discussions
End of life management of adult patients
were held in 63 cases (90%). However, due to the
retrospective nature of our review details as to the
precise nature of each discussion are not known.
Practices in this intensive care unit may differ from
those described by Cook et al.,16 and Azoulay et
al.,17 and possibly reflect practices in other Aus-
tralian intensive care units.
Our review confirmed that the number of dis-
cussions held with the family of a patient prior to
death varied, with up to 7 discussions held prior to
patient death. The frequency of meetings with fam-
ily members prior to a patient’s death suggests that
numerous meetings may be required to facilitate
acceptance of imminent palliation. The study by
Lautrette et al.12 confirms that intensive and proac-
tive communication is vital to the family member’s
experience.
Inadequate support services for grieving families
has previously been identified as a barrier to opti-
mal care,18 and having access to other experts such
as social workers and pastoral care is seen as a way
of improving end of life care.19 Our data confirmed
that chaplains were only involved in 16 (22.9%)
cases, and palliative care and social workers in
less than 4 cases. We were not able to establish
from this retrospective data why their involvement
was limited. There were 25 (35.7%) deaths that
occurred between 8 pm and 8 am, and the avail-
ability of support service personnel during this time
period may have contributed to under utilisation of
support services. As the median time to death from
withdrawal of therapy was 105 min, there may have
been cases where it was not possible to access sup-
port service personnel prior to death. Alternately,
the next-of-kin may have declined support services
and this may not have been documented in the case
history.
There were numerous explanations for with-
drawal or withholding of therapy. Futility of care
was documented in 42 (60.7%) cases and this is con-
sistent with a previous study which found that in
68% of cases, the primary reason for withholding
or withdrawing therapy was failure to respond to
maximum therapy.14 Thirteen patients (18%) had
an irreversible neurological injury documented as
the reason for withdrawal of therapy. This would
be consistent with the number of patients admitted
to the ICU after cardiac arrest. This group would
often have sustained hypoxic brain injury follow-
ing cardiac arrest. Advanced directives were not
commonly encountered in our health service and
in the 7 cases (9.8%) where therapy was limited
to respect patient wishes, there were few written
directives. A decision to withhold or withdraw ther-
apy under these circumstances was the result of in
depth discussions with family and next-of-kin who
17
were able to express what the patient had previ-
ously verbalised.
The most frequently withheld therapy in iso-
lation was ventilatory support in 7 cases (10%),
followed by renal replacement therapy in 6 cases
(8.6%). It is probable that these were the most fre-
quently withheld therapies as they were considered
the most invasive forms of life support. In one case
surgery was withheld as this was also viewed as
aggressive management that was unlikely to change
the patient’s outcome.
Inotropic drugs were the most commonly with-
drawn therapy (30%) and this measure may have
been considered subtle and non-confrontational
once a decision had been made to withhold therapy.
Inotropic drug therapy was withdrawn with TPN in
3 cases (4.3%), with ventilatory support in 2 cases
(2.9%), and with renal replacement therapy in 1
case (1.4%). As many patients had multiple organ
failure, withdrawal of inotropic drug therapy sup-
port would have made death imminent. The nature
of therapy withdrawn and withheld was invariably
discussed with family members prior to interven-
tion and the likely consequences of such action
communicated. This facilitated the vigil process
for family members and forewarned them of the
patient’s imminent death.
The ICU did not have a protocol for sedative and
analgesic use during palliation but these agents are
used liberally during terminal care. Over two-thirds
of patients (66.2%) were receiving analgesia at the
time of death. This suggests that patient comfort
remained an integral management strategy dur-
ing palliation in the ICU. Morphine and midazolam
were the most frequently used analgesic/sedative
combination in patients who are ventilated in our
ICU. This combination was sustained during pal-
liation and was used for over a third (34.3%) of
patients after withdrawal of therapy. The judicious
use of sedation and analgesia is recognised and sup-
ported by the American College of Critical Care
Medicine.1
The vast majority of our patients (44%) died
within 1 and 6 h of treatment withdrawal. Whilst
the time of death is not always predictable these
data can have implications for staff and families.
Staff may need to communicate that if family wish
to be present at the time of death, then their pres-
ence during the first 6 h after treatment withdrawal
would be most appropriate.
Despite patients dying at all hours, the family
was present at more than two-thirds of the deaths
(65.9%). The need for family members to maintain
a vigil or presence is well documented20 and family
members of our patients were encouraged to be
present during end of life care and at death. It was
18
not clear from our study why family or next-of-kin
were not present at one-third of the deaths (34.1%).
This information is not routinely recorded in the
patient’s medical records and was not available due
to the retrospective nature of this study.
There were concerns raised by family mem-
bers in only 1 of the 70 ICU deaths reviewed.
This suggests that despite limitation of therapy in
many deaths reviewed, the process was considered
acceptable by the majority of family members.
The importance of communication in managing
the dying patient and their family in the ICU was
reflected in the frequency of family meetings held
(mean 2.3). Transparency and accountability for
end of life decision-making was maintained as the
coroner was notified in 13 of the 70 deaths.
Despite useful insights gained from this data
there are several limitations to this study. The
data represented are retrospective with all the
associated limitations of such a study design.
Regarding the specific content and communication
that occurred during family discussions, no firm
conclusions could be drawn. The details of staff
attending meetings and the degree to which family
or next-of-kin were involved in end of life deci-
sions is not known, as this information could not
be obtained retrospectively. Furthermore, this was
a single centre study that may not reflect the prac-
tice of all metropolitan ICUs in Australia. There
may be other factors that impact on end of life
management in our intensive care unit. The mean
age of patients admitted to our intensive care unit
is 64 and the mean age of patients whose deaths
were reviewed was 69. Data from the ANZCIS Out-
comes Database21 reveals a mean age of patients
in metropolitan Australian intensive care units of
59. Furthermore, our ICU may have admission cri-
teria and management practices that differ from
other units and this may influence outcomes. The
homogenous patient population on the Mornington
Peninsula of Australia may also not be reflective of
the management practices in other intensive care
units that admit a greater cultural and religious
diversity of patients.
Documentation regarding precise therapy with-
drawn or withheld, and referral to support service
personnel was inconsistent in the charts reviewed.
The frequency of discussions with patient family
and next-of-kin regarding care decisions and end of
life management was documented. In contrast, the
details regarding meeting attendance and the exact
content of these meetings was not well documented
or clear in the charts reviewed. Furthermore whilst
the review confirmed one death where concerns
were raised by the family, it is possible that there
were additional cases where family concerns were
M.J. Bloomer et al.
not documented. These observations reflect the
limitations of retrospective research.
Despite the reservations regarding the inter-
pretation of a retrospective study our data does
suggest that death in our ICU was often predictable
and that ongoing communication with family and
next-of-kin facilitated end of life management with
invariable planned palliation. The acceptance of
death in the ICU by most family members implies
that the process followed achieved limitation of
therapy in the absence of apparent conflict.
Recommendations
Whilst the difficulties of conducting retrospective
research are acknowledged and the results are
not necessarily representative of other Australian
intensive care units or populations, this review
has revealed a unique perspective of end of life
management in this particular intensive care unit.
Despite the limitations of the study, the results will
assist in the development of a multidisciplinary,
coordinated approach to the development of a local
protocol for end of life management of terminal
patients in our ICU. In particular such a protocol
will include a focus on family communication and
appropriate referral to support service personnel.
Whilst this data reveals that end of life is often
predictable, and various approaches are taken in
end of life care, further work is necessary to accu-
rately portray and define the nurse’s role in end
of life decision-making, management, and family
care. Nurses are uniquely placed to play an integral
part in assisting the acceptance of death by family
and loved ones, and it is therefore our intention
to proceed to further prospective studies regarding
the end of life communication, the decision-making
process and management in the intensive care unit.
Acknowledgement
We would like to thank D Hannah for her help in
data collection.
References
1. Truog R, Campbell M, Curtis J, et al. Recommendations
for end-of-life care in the intensive care unit: a consen-
sus statement by the American College of Critical Care
Medicine. Crit Care Med 2008;36(3):953—63.
2. Luce JM, Prendergast TJ. The changing nature of death in
the ICU. In: Curtis JR, Rubenfeld GD, editors. Managing
death in the intensive care unit: the transition from cure
to comfort. Oxford, UK: Oxford University Press; 2001. p.
19—29.
End of life management of adult patients
3. Glavan BJ, Engelberg RA, Downey L, et al. Using the
medical record to evaluate quality of end-of-life care
in the intensive care unit. Crit Care Med 2008;36(4):
1138—46.
4. Azoulay E, Metnitz B, Timsit J, et al. End-of-life in 282 inten-
sive care units: data from the SAPS 3 database. Intens Care
Med 2009;35:623—30.
5. Crighton MH, Coyne BM, Tate J, et al. Transitioning to end-
of-life care in the intensive care unit: a case of unifying
divergent desires. Cancer Nurs 2008;31(6):478—84.
6. Cook DJ, Guyatt GH, Jaeschke R, et al. Determinants in
Canadian health care workers of the decision to withdraw
life support from the critically ill. Canadian critical care
trials group. JAMA 1995;273:703—8.
7. Weidermann CL, Druml C. End of life decisions in Austria’s
intensive care units. Intens Care Med 2008;34:1142—4.
8. Australian and New Zealand Intensive Care Society.
Statement on withholding and withdrawing treatment.
Melbourne: ANZICS; 2004.
9. Baggs JG, Norton SA, Schmitt MH, et al. Intensive care
unit cultures and end-of-life decision making. J Crit Care
2007;22:159—68.
10. Health NSW. End-of-life care and decision making—–
guidelines. Sydney, NSW: Department of Health; 2005.
11. Hall RI, Rocker GM. End-of-life care in the ICU: treatments
provided when life support was or was not withdrawn. Chest
2000;118(1):1424—30.
12. Zib M, Saul P. A pilot audit of the process of end-of-life
decision making in the intensive care unit. Crit Care Resusc
2007;9(2):213—8.
Available online at www.sciencedirect.com
19
13. Lautrette A, Ciroldi M, Ksibi H, et al. End-of-life fam-
ily conferences: rooted in the evidence. Crit Care Med
2006;34(11):S364—72.
14. Steinhauser KE, Clipp EC, McNeilly M, et al. In search
of a good death: observations of patients, families, and
providers. Ann Intern Med 2000;132(10):825—32.
15. Collins N, Phelan D, Marsh B, et al. End-of-life care in the
intensive care unit: the Irish ethicus data. Crit Care Resusc
2006;8(4):315—20.
16. Levy MM. End-of-life care in the intensive care unit: state
of the art in 2006. Crit Care Med 2006;34(11):S306—8.
17. Cook D, Rocker G, Giacomini M, et al. Understanding
and changing attitudes toward withdrawal and withhold-
ing of life support in the intensive care unit. Improving
the quality of end-of-life care in the ICU. Crit Care Med
2006;34(11):S317—23.
18. Azoulay E, Pochard F, Chevret S, et al. Half the family mem-
bers of intensive care unit patients do not want to share in
the decision making process: a study in 78 French intensive
care units. Crit Care Med 2004;32(9):1832—8.
19. Nelson J, Angus D, Weissfeld L, et al. End-of-life care for
the critically ill: a national intensive care unit survey. Crit
Care Med 2006;34(10):2547—53.
20. Ciccarello GP. Strategies to improve end-of-life care
in the intensive care unit. Dimens Crit Care Nurs
2003;22(5):216—22.
21. Australian and New Zealand Intensive Care Society.
ANZICS outcomes database. Comparative ICU Patient
Age Distributions—–Frankston Hospital. http://www.anzics.
com.au accessed March 13, 2009.