497227

2013

PMJ28310.1177/0269216313497227Palliative MedicineHsu et al.

Original Article

After-hours physician care for patients

with do-not-resuscitate orders: An
observational cohort study

Palliative Medicine
2014, Vol. 28(3) 281287
The Author(s) 2013
Reprints and permissions:
sagepub.co.uk/journalsPermissions.nav
DOI: 10.1177/0269216313497227
pmj.sagepub.com

Nin-Chieh Hsu1,2,3, Ray-E Chang3, Hung-Bin Tsai1,2, Yu-Feng Lin1,2,

Chin-Chung Shu1,2, Wen-Je Ko1 and Chong-Jen Yu2

Abstract
Background: Medical care at night for patients with do-not-resuscitate orders and the practice patterns of the on-call residents have
rarely been reported.
Aim: To evaluate the after-hours physician care for patients with do-not-resuscitate orders in the general medicine ward.
Design: Observational study.
Setting/participants: This study was conducted at an urban, university-affiliated academic medical center in Taiwan. The night shift
nurses consecutively recorded every event that required calling the duty residents. Patients with and without a do-not-resuscitate
order were compared in demographics, reasons for calling, residents response, and nurses satisfaction. A standard report form was
established for the nurses to record events.
Results: From October 2009 to September 2010, 1379 inpatients contributed to 456 after-hours calls. do-not-resuscitate
patients accounted for 256 (18.7%) of all inpatients, and 160 (35.1%) of all after-hours calls. The leading reason for calls was
abnormal vital signs, which was significantly higher for patients with do-not-resuscitate orders compared to patients without a
do-not-resuscitate order (64.4% vs 36.1%, p < 0.001). The pattern of residents responses showed a significant difference with
more bedside visits for patients with do-not-resuscitate orders (p < 0.001). The nurses were usually satisfied with the residents
management of both groups.
Conclusion: Abnormal vital sign, rather than symptom, was the leading reason for after-hours calls. The existence of do-notresuscitate order produced different medical needs and physician workload. Patients with do-not-resuscitate orders accounted for
one-third of night calls and nearly half of bedside visits by on-call residents and may require a different care approach.
Keywords
After-hours care, do not resuscitate, palliative care

Introduction
The night shift is a time when on-site faculty supervision
and nurse staffing ratios are reduced resulting in a different
system of care at night compared to the day.1,2 There is some
evidence that the medical events which occur at night are
associated with poor clinical outcomes.3 Studying nursing
and beeper calls at night allows for researchers to understand the demand and need of inpatients at night.
Unfortunately, such studies are rare in the literature. Studies
focusing on paging patterns4,5 showed that interns were
interrupted frequently during both sleep and patient encounter and that nurses perception was not in agreement with
residents responses.6 Another study investigating paging
times showed that on-call days generate nearly five times of

workload to physicians.7 However, these studies focusing

on the occurrence of paging provide incomplete information
1Division

of Hospital Medicine, Department of Traumatology, National

Taiwan University Hospital, Taipei, Taiwan
2Department of Internal Medicine, National Taiwan University Hospital,
Taipei, Taiwan
3Institute of Health Policy and Management, College of Public Health,
National Taiwan University, Taipei, Taiwan
Corresponding author:
Wen-Je Ko, Division of Hospital Medicine, Department of
Traumatology, National Taiwan University Hospital, No. 7, Chung-Shan
South Road, Taipei 100, Taiwan.
Email: kowj@ntu.edu.tw

Downloaded from pmj.sagepub.com at NATIONAL TAIWAN UNIV LIB on March 5, 2015

282

Palliative Medicine 28(3)

regarding physician workload. The unanswered questions

regarding on-call residents workload are the sources of paging as well as workload produced after paging. To answer
these questions, the demand of paging and behavior of oncall residents should be observed.
Palliative care (PC) is a continuous process, which typically begins with discussing advance care planning and
end-of-life (EOL) care.8 Patients who have do-not-resuscitate (DNR) orders are a particular population in a hospital,
and their physical, social, and mental needs may be very
different.9,10 With the progress in palliative medicine,
patients are well cared for by specialist PC teams during the
daytime. However, the treatment for inpatients may be suboptimal at night and outside the PC unit. In the general
medicine setting, the medical needs of inpatients with DNR
orders may be different from other inpatients and are less
satisfactorily met at night with the current off-hours care
system. Our study aims to investigate the physician care for
inpatient at night and allows to compare between patients
with and without a DNR order. It is hypothesized that
patients with DNR order would generate different workload for after-hours physicians.

Methods
Study setting
This study was conducted at the National Taiwan University
Hospital (NTUH), a 2000-bed, university-affiliated tertiary
referral medical center in Taiwan. A hospitalist acute general medicine service was established in October 2009, and
general medicine patients are admitted to this hospitalist
ward from the emergency department (ED). The performance in caring for general medicine patients in this ward
was demonstrated in our previous study.11
Three working shifts, day, bridge, and night, were designed
for the hospitalist ward and remained unchanged during the
study period. Patients newly admitted from the ED were
assigned to both day and bridge hospitalists. In general, all
beds were fully occupied in the evening. The night shift was
from 11 p.m. to 8 a.m. the following day, with the staff taking
handoffs from the bridge shift (from 1 p.m. to 11 p.m.) and
covering the whole ward overnight. Nurse practitioners were
assigned to the day and bridge shift hospitalists, and residents
were assigned to the night shift hospitalist. Residents were the
first to whom night shift nurses reported inpatient problems
and data. Each hospitalist, resident, and nurse practitioner had
a mobile phone to communicate with each other.

Night shift call recording

A standard night shift call record form was designed by
our study group. The night shift nurses (working from 11
p.m. to 8 a.m.) were responsible for recording every
event that required calling the duty residents. In order to

minimize the bias of underreporting or overreporting,

night shift nurses who participated in the study were
informed that the record forms aimed to monitor the residents management and notify the day staff of important
events. The on-call residents were blinded to the study
design. To minimize the observation effect, the
Institutional Review Board (IRB) of our institution
agreed that only the night shift nurses required the
informed consent.

Population and data collection

The IRB of NTUH approved the study (201006028R).
This study was performed by prospective data collection
with retrospective analysis. A standard night shift call
record form was used for quality control purpose, and it
consisted of time of call, classified call reasons, vital signs
at calling, subsequent management of the physician,
and the nurses satisfaction on the whole management
process.
Advance care plans are scarcely noted in hospitalized
patients in Taiwan. DNR is typically discussed after admission and is usually initiated by physicians when the
patients clinical status becomes life-threatening or irreversible. If a patient has a DNR order that has been signed
by himself/herself (or families) and by in-charge physicians, it is legal for physicians to withhold cardiopulmonary resuscitation in our country since 2000. Patients who
had DNR codes during hospitalization in the hospitalist
ward were included as the population of interest. Patients
with DNR orders were labeled as DNR in green in the
health information system of our hospital, whereas the
remaining patients were not labeled in green. We were thus
able to easily identify DNR patients when the calls were
placed.
In order to analyze the night shift events and patterns,
the management of the residents was classified as telephone
order, immediate bed visit, or delayed bedside visit. Drug
prescriptions could be made by the residents via the electronic prescription system anywhere in the hospital. The
residents did not have to go to the nurse station for simple
medication requests such as sleeping pills or drug refills.
On-call residents, who were blinded to the study, decided
by themselves whether an immediate (within 15 min) bedside visit was required. A bedside visit more than 15 min
after a call was defined as a delayed bedside visit, as used
in a previous study.6 The actual time lag between the call
and visit was recorded by the nurses.
For analysis, the reasons for the night call were classified into six categories by the night shift nurses(1) abnormal vital signs; (2) original symptom/problem; (3)
new-onset symptom/problem; (4) need for physicians
evaluation, prescription, or procedure; (5) need for explanation/communication; and (6) othersin descending
order of priority with the category of highest priority being

Downloaded from pmj.sagepub.com at NATIONAL TAIWAN UNIV LIB on March 5, 2015

283

Hsu et al.
Table 1. Classification of call reasons with definitions and examples.
Call reason category

Definition

Example

Abnormal vital signs

Abnormal blood pressure, heart rate,

respiratory rate, body temperature, oxygen
saturation, or consciousness
An existing symptom or problem which has
been handed over from the previous shift
A new symptom or problem which was not
noticed in the previous shift
Events that nurses think the physician should
evaluate, prescribe orders, or perform medical
procedures
Situations on which the nurses think the
physician should answer questions or say
something to the patients or relatives
The physician should be informed but no
need for direct evaluation

Hypotension, arrhythmia, fever, or

hypothermia

Original symptom/problem
New-onset symptom/problem
Need for physicians evaluation,
prescription, or procedure
Need for explanation or
communication
Others

recorded. The classifications, definitions, and examples are

shown in Table 1.
The night shift nurses who participated in our study
were requested to complete informed consent process by
the IRB of NTUH. The nurses satisfaction was measured
using a Likert scale including five levels of satisfaction:
very satisfied, satisfied, unsure, dissatisfied, and very dissatisfied. The night shift nurses completed the satisfaction
scale within their shift.

Statistical analysis
The data were analyzed using SPSS software (version 16;
SPSS Inc., Chicago, IL, USA). We compared the basic demographic data, reasons for the night call, residents response,
time to bedside visit, and nurses satisfaction score between
patients with and without a DNR order. Subgroup analysis
was done for all DNR night calls to compare characteristics
between cancer patients and noncancer patients. Intergroup
differences were compared using Pearsons chi-square test or
Fishers exact test for dichotomous variables and using the
MannWhitney U test for continuous variables.

Results
Demographic data
From October 2009 to September 2010, a total of 1379
patients were admitted to the hospitalist ward. Among
them, 258 patients (18.7%) had DNR orders during hospitalization. Table 2 depicts the demographic data and outcomes of all patients with DNR orders.

Reasons for night shift calls

Within the 1-year study period, a total of 456 night calls to
physicians were recorded. Table 3 shows the characteristics

Cancer pain breakthrough and ileus

with refractory vomiting
Chest pain, shortness of breath,
and oliguria
Hyperglycemia, difficulty in sleeping,
and Foley obstruction
Refuse protective constraints,
refuse treatment advice, and angry
patient or relative
Falling without obvious injury

Table 2. Demographics of the patients with DNR orders in the

whole study cohort.
DNR patients (N = 258)
Age (years)
Male
BMI (kg/m2)
CCI on admission
Barthel Index on admission
Cancer history
Organ failure
From hospitalization to
DNR (h)
Hospital LOS (days)
ICU admission
Hospital mortality
Outcome and disposition
Death
GHTD
Home
Nursing home
Other department/
institution

77.7 (36105)
137 (53.1)
20.8 (12.133.8)
3.0 (011)
10 (0100)
140 (54.3)
162 (62.8)
69.5 (01791)
9 (088)
12 (4.7)
117 (45.4)
99 (38.4)
18 (7.0)
85 (33.1)
26 (10.1)
29 (11.3)

BMI: body mass index; CCI: Charlson comorbidity index; DNR: donot-resuscitate; GHTD: go home to die; ICU: intensive care unit; LOS:
length of stay.
Data are expressed as median (minimummaximum) or number of
cases (%).

of the night calls with comparisons between patients with

and without a DNR order.
Patients with and without a DNR order accounted for
160 (35.1%) and 296 (64.9%) of all calls, respectively. For
both groups, the leading reason for the night calls was
abnormal vital signs, which was significantly higher for the
DNR group (64.4% vs 36.1%, p < 0.001). Compared to
patients without a DNR order, patients with DNR orders

Downloaded from pmj.sagepub.com at NATIONAL TAIWAN UNIV LIB on March 5, 2015

284

Palliative Medicine 28(3)

Table 3. Comparison of demographics, call reasons, and residents responses for night calls between patients with and without
DNR orders.

Demographics
Age (years)
Male
Classification of call reason
Abnormal vital sign
Original symptom/problem
New-onset symptom/problem
Need for physicians evaluation,
prescription, or procedure
Need for explanation/communication
Others
Residents response
Immediate bedside visit
Delayed bedside visit
Telephone order without visit
Time to bedside visit (min)
Nurses satisfaction score
Very satisfied
Satisfied
Unsure
Dissatisfied
Very dissatisfied

Calls from patients with

DNR (n = 160)

Calls from patients without

DNR (n = 296)

p value

77.5 13.3
78 (49.4)

69.8 14.3
143 (48.6)

<0.001a
0.883b

103 (64.4)
16 (10.0)
13 (8.1)
22 (13.8)

107 (36.1)
34 (11.5)
67 (22.6)
76 (25.7)

<0.001b

4 (2.5)
2 (1.2)

6 (2.0)
6 (2.0)

38 (23.8)
43 (26.9)
79 (49.4)
7.9 8.7

56 (18.9)
35 (11.8)
205 (69.3)
7.7 13.0

58 (36.5)
64 (40.3)
34 (21.4)
2 (1.3)
1 (0.6)

128 (44.1)
116 (40.0)
44 (15.2)
1 (0.3)
1 (0.3)

<0.001b

0.881a
0.197c

DNR: do-not-resuscitate.
Data are expressed as mean standard deviation or number of cases (%).
aMannWhitney U test.
bPearsons chi-square test.
cFishers exact test.

had significantly fewer calls for symptoms or problems,

regardless of whether they were original or new.
Among 158 calls from patients with DNR orders, 70
(44.3%) were from patients with history of cancer. When
comparing cancer and noncancer patients, there was no significant difference between call reasons (Table 4).

On-call residents response and nurses

satisfaction
The percentages of bedside visits were 50.6% and 30.7%,
and the percentages of immediate visits within 15 min
were 23.8% and 18.9% for patients with and without a
DNR order, respectively. Telephone orders without visits
occurred in 49.4% of night calls for patients with DNR
orders but 69.3% for patients without a DNR order.
Overall, the pattern of the residents responses showed a
significant difference between patients with and without a
DNR order (p < 0.001), with more bedside visits for the
former.
Of all 172 bedside visits, patients with and without a
DNR order accounted for 81 (47.1%) and 91 (52.9%),
respectively. Regarding direct patient care workload, it was

similar between the DNR and non-DNR groups although

the former was a minor subpopulation.
A total of 449 nurse satisfaction reports were used for analysis (7 were excluded due to missing data). The nurses scored
very satisfied in 36.5% and 44.1% for patients with and
without a DNR order, respectively. In general, the nurses
were satisfied with the residents management, and there was
no significant difference between the two groups (Table 3).

Time distribution of night shift calls

Figure 1 depicts the distribution of the 456 calls on an
hourly basis throughout the night shift. Two peaks of the
night calls, at 13 a.m. and 68 a.m., were noted. The variation seemed to be more prominent for patients without a
DNR order than patients with DNR orders, although there
was no significant difference.

Discussion
Abnormal vital sign was the leading reason for after-hours
calls whether patients had DNR or not. To the best of our

Downloaded from pmj.sagepub.com at NATIONAL TAIWAN UNIV LIB on March 5, 2015

285

Hsu et al.
Table 4. Comparison of call reason and residents response for DNR patients with and without a diagnosis of cancer.

Demographics
Age (year)
Male
Classification of call reason
Abnormal vital sign
Original symptom/problem
New-onset symptom/problem
Need for physicians evaluation, prescription, or procedure
Need for explanation/communication
Others
Residents response
Immediate visit within 15 min
Telephone order with delayed visit
Telephone order without visit
Nurses satisfaction score
Very satisfied
Satisfied
Unsure
Dissatisfied
Very dissatisfied

Cancer (n = 70)

Noncancer (n = 88)

p value

70.1 13.7
38 (54.3)

83.4 9.5
40 (45.5)

<0.001a
0.270b

47 (67.1)
8 (11.4)
5 (7.1)
8 (11.4)
1 (1.4)
1 (1.4)

55 (62.5)
7 (8.0)
8 (9.1)
14 (15.9)
3 (3.4)
1 (1.1)

0.849b

11 (15.7)
23 (32.9)
36 (51.4)

26 (29.5)
20 (22.7)
42 (47.7)

0.092b

25 (35.7)
25 (35.7)
18 (25.7)
1 (1.4)
1 (1.4)

32 (36.8)
39 (44.8)
15 (17.2)
1 (1.1)
0 (0.0)

0.467c

DNR: do-not-resuscitate.
Data are expressed as mean standard deviation or number of cases (%).
aMannWhitney U test.
bPearsons chi-square test.
cFishers exact test.

Figure 1.Time distribution of night calls to physicians.

DNR: do not resuscitate.

knowledge, this is the first report on after-hours calls to

physicians comparing patients with and without DNR. The
most valuable finding is that the reasons for the calls to the
residents at night were significantly different between

patients with and without a DNR order, with more urgent

calls for abnormal vital signs for the former. The behavior
patterns and workloads of the on-call residents were also
different, and these findings may be valuable in improving
resident care at night.
Patients with DNR orders were a minority of the patients
in the general medicine ward; however, they accounted for
one-third of night calls and half of direct patient care workload. This finding is important for researchers because we
usually assume that patients with DNR orders require fewer
medical interventions. After discussing DNR and EOL
issues, we also assume that patients and relatives have an
insight into terminal or near-terminal conditions and
demand conservative, noninvasive, and symptom-oriented
treatments after signing DNR consent. However, we may
be neglecting the need for mental support and bedside visits
just for the comfort and reassurance of continuing care.
When vital signs worsen, which accounted for 64.4% of
night calls in our study, the patients and relatives may still
need supportive visits. These psychological needs are similar to the patients without a DNR order. There were fewer
symptoms requiring night calls for patients with DNR
orders compared to those without a DNR order, which indicates fair pain and symptom management for the former
group in the study ward. However, if unexpected pain,
dyspnea, or delirium occurs, evaluation and prescriptions
by the residents are still mandatory. Finally, death can occur

Downloaded from pmj.sagepub.com at NATIONAL TAIWAN UNIV LIB on March 5, 2015

286

Palliative Medicine 28(3)

at any time. All the aforementioned needs contributed to the

reasons for the night calls for patients with DNR orders in
this study, and they were a significant source of night shift
workload for both nurses and residents.
The attitude toward the patients with DNR orders did
not appear to change because of DNR consent itself. In
addition, it is natural for nurses to call duty physicians
when the idea of calling is necessary for unstable conditions is so firmly rooted in their minds. The fact that twothirds of the calls for patients with DNR orders were due to
abnormal vital signs has important clinical implications
and suggests a way to decrease avoidable calls at night. It
appeared as though pain and symptom management were
performed well. However, our results may suggest that we
can discuss the issue of managing abnormal vital signs
more with the nursing staff. First, we should have a consensus on the use of fluid resuscitation, blood transfusions, and
vasopressors. We should even discuss whether bedside
monitoring is appropriate or what to do when a monitor
sounds an alarm. Is a fast heartbeat harmful or just a mirror
of stress and anxiety to which we should pay attention? Is a
slow heartbeat dangerous or a reflection of being comfortable? Should we do routine workup for fever or just give
antipyretics and relieve the associated symptoms? All these
points may be helpful when the nursing staff face changing
conditions. In addition, fewer routine checkups may alleviate the anxiety of the patients and relatives. In our study, the
two peaks of night call distribution shown in Figure 1
reflected the nursing round pattern at night.12 While routine
checkups are mandatory for general patients, is it necessary
to closely monitor patients with DNR orders during sleep?
If we communicate more and reach a consensus, the nurses
will be more likely to feel at ease without doing routine
round and checkups at night.
Second, we should facilitate communication between
nurses and physicians, especially during off-hours.13 Nurses
could be included in the PC and EOL discussions or, at
least, the nurses could be made aware of how we plan to
handle abnormal events. Without these steps, the patients
and relatives may become confused with the inconsistent
attitudes of the nurses and physicians. Such internal communication takes time but is always worthwhile.
With regard to the residents responses, 50% of afterhours calls for patients with DNR orders required immediate or delayed bedside visits by on-call residents. Again, the
workload for the residents appeared to be much higher than
for patients without a DNR order who needed only 30%
bedside visits. When nurses called the residents for a patient
with DNR order, it could only be handled by telephone
order or electronic prescription system half of the time.
When analyzing all bedside visits made by the residents,
half were for patients with DNR orders and half for patients
without DNR, although the former group accounted for
18.7% of the ward. From this point of view, it is reasonable

to assume that the workload for night staff will rise substantially if the number of patients with DNR orders increases.
In a previous survey in Japan, Morita et al.14 pointed out
that the number of nurses and the number physicians were
significant determinants of family satisfaction. A question
that researchers should study is how to predict the necessary level of manpower required to care for patients with
DNR orders.
This study has several limitations. First, the reasons why
the patients called the night shift nurses were not studied. In
a previous study, the reasons were different from those causing nurses to call doctors.15 The nursing need for patients
with DNR orders is also an important issue. However, we
only focused on the need to call an on-call resident, so the
response patterns could be studied. Second, we only surveyed the need for resident care at night for patients with
DNR orders, which cannot be generalized to the daytime.
Care at night aims to provide the best possible sleeping
environment for the patients.16 Night call studies are thus
valuable because they reflect the real needs for the patients
at night, without being confounded by daytime activities. To
improve nighttime problems and design effective care in a
hospital, it is critical to gather frontline information at night.2
Third, patients with DNR orders in a general medicine setting may be different from those in a specialist PC unit.
However, the subgroup analysis showed that cancer and
noncancer patients had similar reasons for night calls.
Therefore, we believe that the results are representative.
The medical need for the patients with DNR orders was
different at night, but we have not yet designed a patientcentered, after-hours PC system in the hospital. Therefore,
we can hardly say that PC at night is as good as that in the
day. In order to minimize the inequity of day-and-night PC,
potential solutions are to facilitate handoff communication,
to improve doctornurse communication, and to introduce
mandated 24-h coverage by specialist PC teams or hospitalists with PC training.17

Conclusion
Our study unveils the picture of medical care delivery at
night in the general medicine setting. The DNR status
increased medical need at night and on-call residents
workload. Patients with DNR orders were a minority of the
patients in the general medicine ward but accounted for
one-third of after-hours calls and nearly half of direct
patient care workload by on-call residents. The medical
need at night for patients with DNR orders may require a
different care model. Therefore, further studies into afterhours PC are warranted.
Declaration of conflicting interests
None.

Downloaded from pmj.sagepub.com at NATIONAL TAIWAN UNIV LIB on March 5, 2015

Hsu et al.

287

Funding

10. Fischbeck S, Maier BO, Reinholz U, et al. Assessing somatic,

psychosocial, and spiritual distress of patients with advanced
cancer: development of the advanced cancer patients distress
scale. Am J Hosp Palliat Care 2013; 30: 339346.
11. Shu CC, Lin JW, Lin YF, et al. Evaluating the performance of a hospitalist system in Taiwan: a pioneer study for
nationwide health insurance in Asia. J Hosp Med 2011; 6:
378382.
12. Campbell AM, Nilsson K and Pilhammar Andersson E. Night
duty as an opportunity for learning. J Adv Nurs 2008; 62(3):
346353.
13. Whitson HE, Hastings SN, Lekan DA, et al. A quality

improvement program to enhance after-hours telephone communication between nurses and physicians in a long-term care
facility. J Am Geriatr Soc 2008; 56: 10801086.
14. Morita T, Chihara S and Kashiwagi T. Family satisfaction
with inpatient palliative care in Japan. Palliat Med 2002; 16:
185193.
15. Tzeng HM. Perspectives of staff nurses of the reasons for and
the nature of patient-initiated call lights: an exploratory survey study in four USA hospitals. BMC Health Serv Res 2010;
10: 52.
16. Oleni M, Johansson P and Fridlund B. Nursing care at night:
an evaluation using the Night Nursing Care Instrument. J Adv
Nurs 2004; 47: 2532.
17. Muir JC and Arnold RM. Palliative care and the hospitalist:
an opportunity for cross-fertilization. Am J Med 2011; 111:
10S14S.

This research received no specific grant from any funding agency

in the public, commercial, or not-for-profit sectors.

References
1. Bellini L. Off-hours care: not so off. J Hosp Med 2011; 6: 34.
2. Shulkin DJ. Like night and dayshedding light on off-hours
care. N Engl J Med 2008; 358: 20912093.
3. Peberdy MA, Ornato JP, Larkin GL, et al. Survival from inhospital cardiac arrest during nights and weekends. JAMA
2008; 299: 785792.
4. Katz MH and Schroeder SA. The sounds of the hospital: paging patterns in three teaching hospitals. N Engl J Med 1988;
319: 15851589.
5. Harvey R, Jarrett PG and Peltekian KM. Patterns of paging
medical interns during night calls at two teaching hospitals.
CMAJ 1994; 151: 307311.
6. Beebe SA. Nurses perception of beeper calls. Arch Pediatr
Adolesc Med 1995; 149: 187191.
7. Patel SP, Lee JS, Ranney DN, et al. Resident workload, pager
communications, and quality of care. World J Surg 2010; 34:
25242529.
8. Larson DG and Tobin DR. End-of-life conversations: evolving practice and theory. JAMA 2000; 284: 15731578.
9. McIlfatrick S. Assessing palliative care needs: views of
patients, informal carers and healthcare professionals. J Adv
Nurs 2007; 57: 7786.

Downloaded from pmj.sagepub.com at NATIONAL TAIWAN UNIV LIB on March 5, 2015