Article, Palliative Care

Emergency department mortality: Fair and square

a b s t r a c t

Objective: This study explored the therapeutic approaches used for end-of-life (EOL) patients admitted to the emergency department (ED) and examined whether the decision to perform life-extending treatment (LET) or to allow natural death (AND) depends on patient characteristics, medical staff variables, and ED setting.

Methods: A retrospective archive study was conducted from January 2015 to December 2017 in the ED of a ter- tiary hospital. The study sample were 674 EOL patients who had died in the ED. For each patient, data were col- lected and measured for dying process (LET vs. AND), patient characteristics, ED-setting variables, and medical- staff characteristics.

Results: The proportion of EOL patients undergoing LET increased from 18.1% in 2015 to 25.9% in 2016 and to 30.3% in 2017 (p = .010), and a quarter of them were treated by emergency medical services. Males tended to receive LET more than females (p b .001). An association was found between Jewish physicians and nurses and AND (p = .001). Heavier workload in the ED and greater severity of the triage classification predicted more LET (OR-1.67, CI = 1.05-1.76, p = .003 and OR = 1.42, CI-0.60-0.81, p b .001, respectively). Receiver operating characteristic analysis showed that patient characteristics contributed most crucially to the therapeutic ap- proaches (C statistic 0.624-0.675, CI-0.62-0.71).

Conclusions: The therapeutic approach used for EOL patients in the ED depends on variables in all three treatment layers: patient, medical staff, and ED setting. Applicable national programs should be developed to ensure that no external factors influence the dying-process decision.

(C) 2018

  1. Introduction
    1. Background

Emergency teams are trained to provide critical treatments for emergency medical conditions that are aimed at protecting life [1]. Alongside emergency teams, emergency departments (EDs) are in- creasingly used for end-of-life (EOL) patients [2,3]. A growing number of patients at the EOL are admitted to EDs and receive increasingly inva- sive care [4-6]. The characteristics of patients at EOL cover a wide range of demographic, clinical, and psychosocial factors in relation to time and place of death [2,7,8]. They also encompass a wide range of ages [9,10] and conditions [6,8,9]. These patients pose a challenge in the ED because most appear not to have access to palliative-care options or lack of awareness as to the status of EOL patients [2,5].

? The research has not been presented.

?? No financial support was received for this research.

? All authors attest to meeting the ICMJE.org authorship criteria.

* Corresponding author at: The Cheryl Spencer Department of Nursing, The Faculty of Social Welfare and Health Sciences, University of Haifa, Israel.

E-mail address: [email protected] (M. Saban).

In an acute setting, such as ED, three main variables affect treatment: the patient, the medical staff, and the ED setting.

Previous studies investigated patients and team factors associated with the decision of whether to allow natural death (AND) or to provide life-extending-treatment (LET) [4]. Patients and clinicians may ap- proach EOL discussions with different expectations and preferences, in- fluenced by religion, race, culture, and ED characteristics. Hamel et al.

[11] found that older patients preferred less aggressive care than youn-

ger patients, but that many older patients wanted cardiopulmonary re- suscitation (CPR) and care focused on life extension. Barnato et al. [12] found that more blacks and Hispanics than whites wanted LET. Ehlenbach et al. [4] found the incidence of CPR to be higher among black and other nonwhite patients. Mebane et al. [13] investigated whether physicians’ preferences for EOL decision making differs between races and gender. In the scope of quality of life vs. length of life, the au- thors found a significant preference for length of life among black physi- cians (p b .001). Additionally, significant differences between attitudes of male and female physicians (F(3,427) = 6.71 (p b .05)) and between attitudes of white and black physicians (F(3,425) = 49.71 (p b .001)) were found with regard to the issue of tube-feeding. Fifty-eight percent of white physicians, compared with 28% of black physicians, agreed that tube-feeding in terminally ill patients is “heroic.” In addition, 42.4% of

https://doi.org/10.1016/j.ajem.2018.08.036

0735-6757/(C) 2018

male physicians agreed that tube-feeding is “heroic” compared with only 28.9% of female physicians. Age was not significant, and there were no sig- nificant interactions.

Few studies, if any, have investigated the influence of ED setting, such as workload, on the decision to perform LET or to AND among EOL patients. Moreover, the relative contribution of each of the three factors has not been fully elucidated. Thus, the aim of this study was to explore the therapeutic approaches used for EOL patients admitted to the ED, and to examine whether and to what extent the decision to perform LET or to AND depends on each of these factors.

Importance

The volume of EOL patients admitted to the ED increase annually while the percentage of EOL patients who receive LET is at an all-time high, resulting in an unsustainable number of patients who are being kept alive by artificial means.

Goals of this investigation

The goal of this study is to map the contributing factors to the deci- sion to perform LET or AND, enabling the development of a better and more appropriate decision-making process.

  1. Methods
    1. Study design and setting

A retrospective archive study in 2018 reviewed admissions from Jan- uary 2015 to December 2017 in the ED of a tertiary hospital, after insti- tutional review board approval. The ED contains 100 beds and serves about 130,000 patients, on average, over the age of 18 years per year. In cases of trauma all age groups are accepted in the ED.

Selection of participants

The study sample consisted of 226, 220, and 228 files of patients who had died in the ED during 2015, 2016, and 2017, respectively. Patient files were chosen from electronic medical records (EMR) using the ICD-9 code for mortality in ED as a filter (R79891, R7983, R79811, and R7982).

Procedure and measurements

For each patient, file data were collected and measured, by the sec- ond and the fourth authors, on the dying process (LET vs. AND), individ- ual characteristics (age, gender, ethnicity), morbidity and mortality variables, ED-setting variables (triage urgency classification according to the Canadian Triage and Acuity Scale [CTAS], shift, day and workload), and variables associated with the medical staff (position and ethnicity). We defined ethnicity as per our nation’s central bureau of statistics (Jewish, Arabs and others).

Interrater agreement assessed by having a sample of charts reviewed independently by the first and second author (? = 0.94).

The first and the third authors were blinded to the group assignment, and the third and fifth authors were blinded to the study hypothesis.

We defined EOL patients primarily based on existing medical records indicating that this was a EOL patient for whom no invasive intervention should be performed, called “hospice status.” The first and second au- thors performed database search in the EMR by several queries: “Do not resuscitate”; “DNR”, “AND”, “palliative care“, “hospice”.

We defined LET patients according to the American Heart Associa- tion Guidelines for Cardiopulmonary Resuscitation and Emergency Car- diovascular Care (2015). These patients may receive the following interventions: cardiopulmonary resuscitation include defibrillations and chest compressions, vasopressor agents, mechanical ventilation, blood products and antibiotics.

We defined AND patients according to the American Heart Associa- tion Guidelines for Cardiopulmonary Resuscitation and Emergency Car- diovascular Care (2015). These patients may receive the following alleviating care: Parenteral fluids, nutrition, oxygen, analgesia, sedation, antiarrhythmics, or vasopressors, unless these are included in the order. Where there were no records, the medical team asked the patient’s relatives about pre-instructions. In other cases, three senior ED physi- cians determined whether the diagnosis of the patient was acute, with poor prognosis (Life expectancy b12 h), due to the patient’s comorbidity, such as catastrophic hypovolemic shock attributable to multiple trauma

in older sick patients.

Outcomes

The coprimary outcomes measured were the proportion of patients who received LET or AND and the rate of contribution of each character- istic (patient, medical staff, ED setting) to the therapeutic approach.

Analysis

The statistical analyses were performed by the first author using de- scriptive data analysis, including ranges, means, medians, standard de- viations (SDs), and interquartile ranges for continuous variables, and frequencies and percentages for categorical variables. Comparisons of means were performed, using one-sample t-tests, between the gold standard mean and the study results. Comparisons of means between more than two groups were performed using One-way ANOVAs. Post- hoc comparisons were performed using the Bonferroni test.

In addition, three logistic regression models were performed for each year, to predict the effect of each factor category–patient, medical staff, and ED setting–on LET. Odds ratios (ORs) and 95% confidence in- tervals (CIs) were estimated for each predictor.

To test the additive value of each factor, we entered the variables into receiver operating characteristic curves one at a time: pa- tient characteristics (cause of death, age, male, Jewish), followed by team characteristics (physician ethnicity, physician position, nurse eth- nicity) and, finally, the ED-setting variables (shift, day, workload). The level of significance for all statistical analyses was 5%. We analyzed the data using the Statistical Package for Health & Welfare Science for Win- dows (SPSS, version 22.0, Chicago, IL, USA).

  1. Results

During the study period, 674 patients died in the ED (226 in 2015, 220 in 2016, and 228 in 2017); they were between the ages of 1 and 104 years (M = 73.58; SD = 16.18). No statistical differences for age were found during the study period. Most patients were males (52.1%) and Jewish (77.9%).

Most patients experienced AND (52.2%), although a decrease was ob- served in this rate during the study period (60.6% in 2015, 49.5% in 2016, 46.5% in 2017; p = .007). Simultaneously, the proportion of patients un- dergoing CPR increased, and there was also growth in the number of pa- tients who were dead on arrival (DOA) after receiving treatment by emergency medical services (EMS); from 18.1% in 2015, to 25.9% in 2016, to 30.3% in 2017 (p = .010). The share of Jewish patients decreased over the period, and the share of Arab patients increased (p = .001).

No differences were found in the percentage of mortality during the weekday. However, the highest mortality rate (46.3%) was observed in the morning shift, followed by evening (37.4%) and night (16.3%). Over time, the proportion of patients dying during heavy-workload shifts in- creased (p = .001).

Most patients were classified as high triage priority level (P1) and therefore immediately entered the shock room (resuscitation bay); a smaller fraction of patients were admitted first to the urgent area during the study period (p = .001) (Table 1).

Table 1

Descriptive statistics for the study sample.a

2015

(n = 226)

2016

(n = 220)

2017

(n = 228)

p-value

Cause of death n (%) DOA

41 (18.1)

57 (25.9)

69 (30.3)

0.010

AND

137 (60.6)

109 (49.5)

106 (46.5)

0.007

LET

48 (21.2)

54 (24.5)

53 (23.2)

0.705

Patient characteristics Age (mean +- SD)

74.92 +- 15.38

72.19 +- 16.89

73.59 +- 16.23

0.207

Gender

% Male

52.7

49.1

54.4

0.521

Ethnicity (%)

Jewish

82.7

72.3

78.5

0.001

Arabs

14.2

19.5

20.2

Others

3.1

8.2

1.3

Team characteristics n (%) Physician ethnicity

Jewish

86 (38.1)

59 (26.8)

68 (29.8)

0.132

Arabs

86 (38.1)

101 (45.9)

99 (43.4)

Others

54 (23.9)

60 (27.3)

61 (26.8)

Physician position

Intern

58 (25.7)

89 (40.5)

96 (42.1)

0.000

Senior

Nurse ethnicity

168 (74.3)

131 (59.5)

132 (57.9)

Jewish

94 (41.6)

80 (36.4)

106 (46.5)

0.000

Arabs

132 (58.4)

126 (57.3)

121 (53.1)

Others

14 (6.4)

1 (0.4)

ED setting n (%)

Shift

Morning

106 (46.9)

92 (41.8)

114 (50.0)

0.344

Evening

81 (35.8)

94 (42.7)

77 (33.8)

Night

39 (17.3)

34 (15.5)

37 (16.2)

Day

Sunday

34 (15.0)

34 (15.5)

33 (14.5)

0.999

Middle of week

120 (53.1)

117 (53.2)

122 (53.5)

Weekend

72 (31.9)

69 (31.4)

73 (32.0)

Workload

If >=250 patients

127 (56.2)

140 (63.6)

167 (73.2)

0.001

Triage classification

1

133 (58.8)

131 (59.5)

124 (54.4)

0.471

2

18 (8.0)

18 (8.2)

26 (11.4)

3

64 (28.3)

55 (25.0)

55 (24.1)

4

9 (4.0)

12 (5.5)

18 (7.9)

5

2 (0.9)

4 (1.8)

5 (2.2)

Area admission Shock room

101 (44.7)

109 (49.5)

117 (51.3)

0.001

Emergent area

96 (42.4)

98 (44.5)

105 (46.1)

Urgent area

29 (12.8)

13 (5.9)

6 (2.6)

a DOA = Dead on arrival; AND = Allow natural death; LET = Life-extending treatment.

The leading causes of death were stage 4 metastatic cancer, called “terminal oncology” (20.3%), sepsis (13.9%), and multi-organ failure (12.9%) (Table 2).

A positive association was found between age and the AND approach (t(483) = 2.864, p b .001), while the average age for natural death was higher than for resuscitation (AND: mean = 76.54 +- 14.20, LET: mean = 70.34 +- 17.57; p b .001). A correlation was found between gender and type of therapeutic approach: men tended to receive higher rates of CPR (p b .001; 56.1% vs. 38.7%) and women tended to experience more natural death (p b .001; 61.3% vs. 43.9%). Jewish patients experienced more natural death than Arab patients did (54.3% vs. 44.3%; p = .035).

Physicians and nurses were mostly Arab (42.4% and 56.2%, respec- tively). The number of interns who participated in mortality events in- creased over the period (p b .001).

No correlations were found between staff shift, day, and dying pro- cess. We found a positive association between Jewish nurses and the AND approach (?2 = 16.71, p = .001).

Table 2

Causes of death during the study period.b

Frequency

%

DOA

152

22.6

Oncology terminally ill

137

20.3

Sepsis

94

13.9

MOF

87

12.9

Respiratory failure

72

10.7

cardiac failure

55

8.2

Multiple trauma

25

3.7

Brain bleeding

12

1.8

Aortic dissection

7

1

GI bleeding

7

1

Abdominal peritonitis/perforation

6

0.9

Acute mesenteric ischemia/event

5

0.7

Hypothermia

3

0.4

ARF

3

0.4

Hypoglycemia

2

0.3

DIC

2

0.3

Drug overdose

2

0.3

STEMI

1

0.1

Massive PE

1

0.1

Massive burn

1

0.1

Total

674

100

b DOA = Dead on arrival; MOF = Multi-organ failure; GI = Gastrointestinal; ARF = Acute renal failure; DIC = disseminated intravascular coagulation; STEMI = ST elevation myocardial infarction; PE = Pulmonary embolism.

Three logistic regression models were run to predict significant fac- tors that influence the therapeutic approach in EOL patients (Table 3). Throughout the first two years, age was a protective factor for the LET approach (in 2015: 95% CI 0.954-0.992, OR 0.97; in 2016: 95% CI

0.960-0.995; OR 0.98). In 2017, male patients were more likely to re- ceive LET than females (95% CI 1.704-5.246; OR 2.99), and Jewish phy- sicians predicted more use of the AND approach (95% CI 0.289-0.940; OR 0.52). In 2016 and 2017, Jewish ethnicity for nurses was found to be predictive of AND in comparison to Arab nurses (in 2016: 95% CI 0.324-0.993, OR 0.57; in 2017: 95% CI 0.314-0.921, OR 0.54).

Nevertheless, in 2015, workload was a significant factor predicting LET (95% CI 1.005-6.503, OR 2.56). The higher the patient’s level of ur- gency, the greater his or her chance of receiving LET during 2016 and 2017 (95% CI 1.491-5.016, OR 2.74; 95% CI 0.108-5.830, OR 3.25;

respectively).

ROC analysis was used to determine the additive effects of each fac- tor on LET (Fig. 1). The results indicated that during the study period, the contribution of patient characteristics was most crucial to LET (C- statistics: 2015: 0.62; 2016: 0.64; 2017: 0.67). After the ED-setting char- acteristics were added to the model, the accumulated contribution in- creased slightly (C-statistics: 2015: 0.59; 2016: 0.63; 2017: 0.68). However, entering the team characteristics decreased the impact (C- statistics: 2015: 0.52; 2016: 0.57; 2017: 0.62).

Limitation

Our study has several limitations. First, neither personal approach nor religion of ED staff have been elucidated. This may have an impact on the therapeutic approaches chosen.

Second, the study did not include and investigate the EOL patients who received LET or AND and survived the ED visit. Therefore, we pres- ent an incomplete portrait of ED therapeutic approaches in EOL patients. Third, DOA patients who received LET by EMS were admitted to the ED in a regular manner and documented as usual in the medical records. Maybe there is a need to regard these patients as a separate sub-popu- lation. In addition, we related to all cases together. It is possible that ex- ternal injuries (e.g., trauma) and internal injuries (e.g., respiratory failure) should be divided into two separate groups, and that age groups should be considered separately.

Table 3

Logistic regression and receiver operating characteristic models for predicting life-extending treatment.

2015

2016

2017

Characteristics

OR

95% CI

Low

High

Sig.

Area

OR

95% CI

Low

High

Sig.

Area

OR

95% CI

Low

High

Sig.

Area

Patient

Age

0.97

0.954

0.992

0.005

0.980

0.9600

0.9950

0.011

0.970

0.9680

1.005

0.140

Male

1.49

0.851

2.567

0.165

1.44

0.8330

2.478

0.193

2.99

1.704

5.246

0.000

Jewish

0.93

0.443

1.960

0.852

0.740

0.4040

1.452

0.414

1.21

0.5980

2.438

0.600

C-statistics Team

Jewish physician

0.860

0.550

0.4970

0.699

1.502

0.002

0.605

0.624

0.880

0.542

0.4780

0.690

1.600

0.002

0.664

0.636

0.520

0.606

0.2890

0.746

0.9400

0.000

0.030

0.675

Senior physician

0.99

0.535

1.823

0.969

0.90

0.5220

1.557

0.710

1.65

0.9560

2.850

0.072

Jewish nurse

1.16

0.6750

1.994

0.591

0.570

0.3240

0.9930

0.047

0.540

0.3140

0.9210

0.024

C-statistics ED

Morning

0.98

0.446

0.570

0.601

1.697

0.549

0.952

0.524

1.14

0.491

0.643

0.644

2.004

0.042

0.661

0.568

0.44

0.547

0.253

0.692

0.772

0.002

0.004

0.619

Weekday

0.57

0.213

1.528

0.264

1.61

0.422

6.159

0.485

0.96

0.412

2.224

0.918

Workload

2.56

1.005

6.503

0.049

0.83

0.230

3.019

0.782

0.75

0.305

1.845

0.532

High-severity P scale

1.05

0.586

1.863

0.882

2.74

1.491

5.016

0.001

3.25

0.108

5.830

0.000

Admission to shock room

0.90

0.516

1.575

0.716

1.44

0.832

2.489

0.193

0.990

0.570

1.727

0.977

C-statistics

0.510

0.661

0.030

0.586

0.560

0.707

0.000

0.634

0.620

0.758

0.000

0.689

CI = Confidence interval; OR = Odds ratio; Sig. = Significance.

  1. Discussion

We found that 47.8% of patients in this study received LET. Likewise, previous studies found that the proportion of patients undergoing in- hospital CPR before death increased over time and was higher for mi- nority patients [2,4]. This finding correlates with evidence that some pa- tients at EOL experience a dying process that does not always comply with the basic thought of a “good death.” [2,6,7] 52% experienced natu- ral death, although this percentage decreased during the study period. Reduction in the AND approach despite an increase in the number of palliative-care patients may indicate a widening gap between the ED

team and patient perceptions, or lack of information upon arrival.

The main accepted purpose of the ED team is to treat undifferenti- ated patients, to assess the patient quickly, and to resuscitate and stabi- lize the patient [2,14]. In the case of patients at EOL, these principles

cannot be implemented permanently because these patients are treated dissimilarly [15-17]. Interactions between EOL models and ED team have been investigated by Chan [15], even though the ED might not be the most appropriate setting for giving EOL care. Additionally, LET for these patients may not be the best or preferred option [18]. Elderly patients were found to have greater odds for AND, which is reasonable given patient multi-morbidity and poor prognosis [4,5].

We found a substantial number of DOA patients (24.8%). For these patients, an out-of-hospital resuscitation was begun by EMS, in most cases without taking the patient’s condition into account. It seems that EMS is not prepared for palliative care. Previous studies point out that EMS teams frequently did not avoid resuscitating EOL patients, particu- larly if there was no documentation of precise instructions [19,20].

Arab patients received LET at higher rates, despite their prognosis, than Jewish patients (p = .001). These results might imply Racial bias.

Fig. 1. Receiver operating characteristic analysis.

Previous studies showed that it is possible that the quality of care be- fore, during, and after cardiac arrest is lower for minority patients [4,21,22]. In our study this is a surprising finding, because most of the team members were Arab. Therefore, it is not certain whether there is a concern of malpractice or whether there is discrimination in treat- ment. Maybe the difference in results is due to cultural variance [23].

The ED is a tumultuous area, especially in cases of overcrowding. Lo- gistic regression analysis, in 2015, indicated that as ED workload in- creased, more LETs were performed (OR = 1.67, p = .003 and OR = 1.42, p b .001, respectively). This finding may indicate that in workload scenarios at the ED, teams must focus on doing rather than on being. There is significant agreement across research papers and guidelines concerning the core elements of palliative care [1,24]. Beck and her col- leagues concluded [25], in a study focused on nurse assistants’ experi- ence of palliative care in municipal residential care settings, that more focus is needed on the trajectory of older peoples’ dying, and that there is also a need for engaged care leaders (doing vs. being).

Our results indicate that the contribution of patient characteristics was most crucial to the selection of therapeutic approach. These results correlate with previous studies finding an association between patient characteristics and the EOL dying process, such as age and ethnicity [2,4]. Surprisingly, ED-setting characteristics decrease the variance ex- planation for therapeutic approach. From what we know, there is a lack of evidence in the literature regarding this finding in an ED setting. On the contrary, team characteristics increase the proportion of out- come explained. This finding corresponds with prior studies showing an association between team characteristics and EOL dying process, such as physician and nurse ethnicity [12,13].

  1. Conclusion

The type of therapeutic approach used for EOL patients in the ED de- pends on variables in all three treatment layers, especially patient char- acteristics. There is a need for a unique national protocol that contains precise instructions for EOL patients in various situations, both outside and inside the hospital.

Author contributions

All authors interpreted the data and edited and approved the final article. SM and PH drafted and conceived the study. SM, SR and DA de- signed the intervention. SM, PH, ZL and DA analyzed the data, designed the study and performed data collection. SM, PH and SR take responsi- bility for the paper as a whole.

References

  1. Grudzen CR, Richardson LD, Hopper SS, Ortiz JM, Whang C, Morrison RS. Does palli- ative care have a future in the emergency department? Discussions with attending emergency physicians. J Pain Symptom Manage 2012;43:1-9.
  2. Forero R, McDonnell G, Gallego B, et al. A literature review on care at the end-of-life in the emergency department. Emerg Med Int 2012;2012:486516.
  3. Holt GE, Sarmento B, Kett D, Goodman KW. An unconscious patient with a DNR tat- too. N Engl J Med 2017;377:2192-3.
  4. Ehlenbach WJ, Barnato AE, Curtis JR, et al. epidemiologic study of in-hospital cardio- pulmonary resuscitation in the elderly. N Engl J Med 2009;361:22-31.
  5. Etkind SN, Bone AE, Gomes B, et al. How many people will need palliative care in 2040? Past trends, future projections and implications for services. BMC Med 2017;15:102.
  6. Wright AA, Keating NL, Balboni TA, Matulonis UA, Block SD, Prigerson HG. Place of death: correlations with quality of life of patients with cancer and predictors of be- reaved caregivers’ mental health. J Clin Oncol 2010;28:4457-64.
  7. Emanuel EJ, Emanuel LL. The promise of a good death. Lancet 1998;351 [SII21-SII9].
  8. Glare P, Virik K, Jones M, et al. A systematic review of physicians’ Survival predictions in EOL ill cancer patients. BMJ 2003;327:195-8.
  9. Kardamanidis K, Lim K, Da Cunha C, Taylor LK, Jorm LR. hospital costs of older peo- ple in New South Wales in the last year of life. Med J Aust 2007;187:383-6.
  10. Pauly BM, Varcoe C, Storch J. Framing the issues: moral distress in health care. HEC Forum 2012;24:1-11.
  11. Hamel MB, Lynn J, Teno JM, et al. Age-related differences in care preferences, treat- ment decisions, and clinical outcomes of seriously ill hospitalized adults: lessons from SUPPORT. J Am Geriatr Soc 2000;48 [S176-S82].
  12. Barnato AE, Anthony DL, Skinner J, Gallagher PM, Fisher ES. Racial and ethnic differ- ences in preferences for end-of-life treatment. J Gen Intern Med 2009;24:695-701.
  13. Mebane EW, Oman RF, Kroonen LT, Goldstein MK. The influence of physician race, age, and gender on physician attitudes toward advance care directives and prefer- ences for end-of-life decision-making. J Am Geriatr Soc 1999;47:579-91.
  14. Le Conte P, Riochet D, Batard E, et al. Death in emergency departments: a multicen- ter cross-sectional survey with analysis of withholding and withdrawing life sup- port. Intensive Care Med 2010;36:765-72.
  15. Chan GK. End-of-life models and emergency department care. Acad Emerg Med 2004;11:79-86.
  16. Chan GK. Understanding end-of-life caring practices in the emergency department: developing Merleau-Ponty’s notions of intentional arc and maximum grip through praxis and phronesis. Nurs Philos 2005;6:19-32.
  17. Chan GK. End-of-life and palliative care in the emergency department: a call for re- search, education, policy and improved practice in this frontier area. J Emerg Nurs 2006;32:101-3.
  18. Travis SS, Bernard M, Dixon S, McAuley WJ, Loving G, McClanahan L. Obstacles to palliation and end-of-life care in a long-term care facility. Gerontologist 2002;42: 342-9.
  19. Feder S, Matheny RL, Loveless RS, Rea TD. Withholding resuscitation: a new ap- proach to prehospital end-of-life decisions. Ann Intern Med 2006;144:634-40.
  20. Stone SC, Abbott J, McClung CD, Colwell CB, Eckstein M, Lowenstein SR. Paramedic knowledge, attitudes, and training in end-of-life care. Prehosp Disaster Med 2009; 24:529-34.
  21. Borum ML, Lynn J, Zhong Z. Study to Understand P, Preferences for O, Risks of T. The effects of Patient race on outcomes in seriously ill patients in SUPPORT: an overview of economic impact, medical intervention, and end-of-life decisions. J Am Geriatr Soc 2000;48 [S194-S8].
  22. Olomu AB, Watson RE, Siddiqi A-e-A, et al. Changes in rates of beta-blocker use in community hospital patients with acute myocardial infarction. J Gen Intern Med 2004;19:999-1004.
  23. Cox CL, Cole E, Reynolds T, Wandrag M. Implications of cultural diversity in do not attempt resuscitation (DNAR) decision-making. J Multicult Nurs Health 2006;12:20.
  24. Ferrell BR. Overview of the domains of variables relevant to end-of-life care. J Palliat Med 2005;8 [S22-S9].
  25. Beck I, Tornquist A, Brostrom L, Edberg A-K. Having to focus on doing rather than being–nurse assistants’ experience of palliative care in municipal residential care settings. Int J Nurs Stud 2012;49:455-64.