A Survey of Hospice and Palliative Care Clinicians' Experiences and Attitudes Regarding the Use of Palliative Sedation

Abstract

Background:

A variety of terms and attitudes surround palliative sedation (PS) with little research devoted to hospice and palliative care (HPC) clinicians' perceptions and experiences with PS. These factors may contribute to the wide variability in the reported prevalence of PS.

Objective:

This study was designed to better identify hospice and palliative care (HPC) clinician attitudes toward, and clinical experiences with palliative sedation (PS).

Methods:

A 32-question survey was distributed to members of the American Academy of Hospice and Palliative Medicine (n = 4678). The questions explored the language clinicians use for PS, and their experiences with PS.

Results:

Nine hundred thirty-six (20% response rate) responded to the survey. About 83.21% preferred the terminology of PS compared with other terms. A majority felt that PS is a bioethically appropriate treatment for refractory physical and nonphysical symptoms in dying patients. Most felt PS was not an appropriate term in clinical scenarios when sedation occurred as an unintended side effect from standard treatments. Hospice clinicians use PS more consistently and with less distress than nonhospice clinician respondents. Benzodiazepines (63.1%) and barbiturates (18.9%) are most commonly prescribed for PS.

Conclusion:

PS is the preferred term among HPC clinicians for the proportionate use of pharmacotherapies to intentionally lower awareness for refractory symptoms in dying patients. PS is a bioethically appropriate treatment for refractory symptoms in dying patients. However, there is a lack of clear agreement about what is included in PS and how the practice of PS should be best delivered in different clinical scenarios. Future efforts to investigate PS should focus on describing the clinical scenarios in which PS is utilized and on the level of intended sedation necessary, in an effort to better unify the practice of PS.

Introduction

Palliative Sedation (PS) is a well-established medical procedure that has professional support from major medical organizations.^1,2 The presence of three cardinal elements make PS legally and ethically distinct from more controversial practices like physician-assisted death and euthanasia; it is used (1) to relieve intractable symptoms; (2) in terminal illness, (3) by applying proportionate doses of pharmacotherapies to intentionally lower awareness.^1,3

While several national and international medical organizations have frameworks to guide the practice of PS, there remains neither a standardized protocol nor agreed-upon language to define which medical circumstances constitute PS.^2,4,5 In fact, multiple other terms and phrases have been utilized in the medical literature to describe the practice—controlled sedation for refractory symptoms,⁶ terminal sedation,⁷ end-of-life sedation,⁸ continuous deep sedation,⁹ PS to unconsciousness,¹⁰ and sedation for intractable distress.⁷ These factors may contribute to the wide variability in the published literature regarding the frequency in which PS is employed and its effect on survival.^11,12

While much of the discourse in the medical literature surrounding PS has pertained to ethics,⁶ only one other study to our knowledge specifically surveyed the attitudes and experiences of hospice and palliative care (HPC) clinicians regarding PS.¹³ The survey by Lux et al. focused on the preferred medication choice to achieve PS and found that most of the respondents were hospice clinicians.¹³ Our principal aim was to elicit a preferred terminology for the definition of PS by HPC clinicians by surveying hospice and palliative medicine (HPM) clinicians. Secondary aims were to better define clinical scenarios that constitute PS, to elicit attitudes regarding the use of PS in clinical practice, and to evaluate how clinician experiences correlate with those attitudes.

Methods

We conducted an electronic survey of demographics, clinician attitudes, and experiences with PS among HPC clinicians. Participation was voluntary, and no remuneration was offered. The study was reviewed and approved by the Medical College of Wisconsin (MCW) Institutional Review Board (PRO00023847). The survey concept was approved by the American Academy of Hospice and Palliative Medicine (AAHPM) Research Committee, but members did not have influence over the design of the survey itself.

Participants

We invited all members of the AAHPM with available electronic mail addresses to participate in the study (N = 4678).

Study procedures

The electronic survey was conducted over six weeks using an initial invitation letter and two follow-up letters sent by electronic mail.

Survey development

We reviewed the PS literature to identify current practices as well as gaps in our knowledge. We utilized SurveyMonkey technology to create the survey. It involved 32 questions—7 of which involved participant demographics, 1 involved terminology preferences, and the remaining 24 explored participants' attitudes and experiences with PS. Of the latter 24 questions, 10 were yes/no questions, 8 were multiple choice questions, and 6 used Likert scales. The terminology-preference question asked respondents to select the best terminology to describe the “the intentional lowering of awareness toward, and including, unconsciousness for patients with severe and refractory symptoms in the setting of a terminal illness”¹ from a multiple choice list. The six possible responses reflected the five most commonly utilized terms from our literature review and an option to choose “other.” To minimize bias, none of the possible multiple choice responses was included in the invitation letter, the consent to participate, or the terminology preference question. Furthermore, this preferred terminology question was the first of the 32 survey questions and had to be answered before the remaining questions. All participating authors reviewed the survey content for validity as well as two external committees: (1) a panel of HPC clinicians at the MCW, and (2) a four-person committee of MCW experts in biostatistics, psychiatry, bioethics, and palliative care.

Statistical analysis

We utilized descriptive biostatistics analysis to examine for correlations in how respondents answered different questions. Pearson's chi-square tests were used to test for difference in categorical independent variables. Fisher's exact tests were used to test for differences in categorical independent variables with small sample sizes. Kruskal–Wallis test was used to test for difference between three or more dependent variable groups in ordered categorical independent variables. Wilcoxon rank-sum test, also known as Mann–Whitney U test, was used to test for difference between two dependent variable groups in ordered categorical independent variables. Friedman's test was used to test for differences between ordered categorical dependent variables. McNemar's exact test was used to test for differences of a matched set of variables. Cochran's Q test was used to test for differences of two or more matched sets of variables. Statistical analysis was performed using SAS 9.3 (The SAS Institute, Cary, NC). To assess for nonresponder bias, we performed a Pearson chi-square to test for categorical independent values between survey responders and known demographic values of 2014 AAHPM members. We also performed a Jonckheere–Terpstra wave analysis to compare responses to specific survey questions to examine for potential bias between those who responded after the first, second, and third mailings.

Results

Nine hundred fifty-five persons opened the survey and 936 responded to at least one question, yielding a response rate of 20.4%. See Table 1 for baseline characteristics.

Table 1.

Baseline Characteristics of Study Participants

Question	% of respondents (n)	Total No. of responses
How many years of experience in hospice or palliative care?
No experience	0.8 (7)	926
<5	25.0 (231)
5–9	27.5 (255)
10–15	22.0 (204)
>15	24.7 (229)
Do you practice in the United States?
Yes	94.6 (879)	929
No	5.4 (50)	929
Do you practice in a U.S. state, where physician-assisted death is decriminalized?
Yes	4.67 (39)	836
No	95.3 (797)	836
What is your primary work environment?
Hospice	39.5 (336)	850
Long-term care facility	1.9 (16)
Inpatient community hospital	21.2 (180)
Academic care center	32.9 (280)
Palliative care outpatient clinic	4.6 (38)
What predominantly comprises your patient population?
Adults	74.9 (674)	900
Pediatrics	5.1 (46)
Both	20.0 (180)
Have you completed a HPM Fellowship?
Yes	30.5 (275)	901
No	67.6 (609)
Currently completing	1.9 (17)
Are you board certified in HPM?
Yes	79.2 (715)	903
No	20.8 (188)	903
Do you practice at a religiously affiliated healthcare institution?
Yes	20.2 (160)	791
No	78.6 (622)
Unsure	1.1 (9)

HPM, hospice and palliative medicine.

Respondents' experiences with PS

Most respondents had been asked about PS by a patient or family member (74.2% and 80.9%, respectively). Over half of respondents had prescribed PS (52.3%) for a nondecisional patient at the request of a surrogate. Nearly two thirds of respondents (66.0%) had prescribed PS in the last year and only 14.9% of respondents had never prescribed PS. Neither completion of a HPM fellowship (p 0.549) nor board certification in HPM (p 0.326) was significantly associated with the timing of the last PS prescription. Benzodiazepines were the most commonly preferred PS medication (63.1% of respondents), followed by barbiturates (18.9%) and opioids (9.4%) (Table 2).

Table 2.

Clinicians' Experiences with Palliative Sedation

Question	% of respondents (n)	Total n
Has a patient asked you about PS?
Yes	74.2 (637)	858
No	25.8 (221)	858
Has family of a patient asked you about PS?
Yes	80.9 (694)	858
No	19.1 (164)	858
What is your preferred medication class to achieve PS?
Benzodiazepines	63.1 (476)	755
Barbiturates	18.9 (143)
Opioids	9.4 (71)
Propofol	4.9 (37)
Antipsychotics	2.7 (20)
Dexmedetomidine	1.1 (8)
Have you prescribed PS for a nondecisional patient?
Yes	52.3 (448)	857
No	47.7 (409)	857
When is the last time you prescribed PS?
Within the last week	7.3 (57)	786
Within the last month	17.8 (140)
Within the last six months	25.1 (197)
Within the last year	15.9 (125)
Within the last five years	19.1 (150)
Never	14.9 (117)

PS, palliative sedation.

Respondents' attitudes regarding PS

In response to the preferred-terminology question, 83.2% selected palliative sedation. About 8.8% preferred sedation for intractable distress in the dying; 5.2% preferred terminal sedation; 1.7% preferred continuous deep sedation; and 1.1% preferred end-of-life sedation. There was no significant difference in how clinicians who completed a HPM fellowship or clinicians who predominantly practice in a hospice answered this question (p 0.417 and 0.668, respectively). However, there was a significant difference in how HPM board-certified clinicians answered this question (p 0.002) with a trend toward higher preference for the term palliative sedation. Wave analysis demonstrated a significant difference in responses between early and late responders (p < 0.001) with early responders preferring the term palliative sedation. Figure 1 reports how frequently respondents felt various clinical scenarios represented by PS.

FIG. 1.

Which of the following scenarios would you consider to be the appropriate use of the medical phrase “Palliative Sedation”? (Choose all that apply). Total number of responses: 805. GTT, continuous intravenous infusion; SI, sedation intended; SU, sedation unintended; VW, ventilator withdrawal.

A large majority of respondents felt that refractory pain, dyspnea, delirium, and nausea were appropriate indications for PS (96.3%, 91.9%, 82.1%, and 68.6%, respectively). More than half of respondents felt clinicians should be allowed to prescribe PS for nonphysical symptoms (53.3%; Fig. 2), including anxiety (58.9%), existential distress (42.1%), and depression (22.2%). There was no significant difference in how non-U.S. clinicians or clinicians working in religiously affiliated institutions answered this question (p 0.270, and 0.289, respectively).

FIG. 2.

Clinicians' attitude regarding access to PS (a) p-Value 0.270 Wilcoxon rank-Sum test for place of practice in the United States; (b) p-Value 0.001 Kruskal–Wallis test for importance of access to PS. APPs, advance practice providers; HPC, hospice and palliative care; PS, palliative sedation.

Most respondents felt there is a clear bioethical distinction between PS and physician-assisted death (94.3%) and euthanasia (94.7%). Practicing at a religiously affiliated healthcare institution did not significantly alter responses to these questions (p 0.787 and 0.148, respectively).

Most respondents felt that access to PS was important (95.7%) with 41.5% classifying it as extremely important. There was a significant correlation in how respondents classified the importance of access to PS and the appropriateness for experienced advanced practice providers (APPs) to prescribe PS (p 0.001): respondents who described access as “very” or “extremely important” were more likely to view the prescription of PS by experienced APPs as appropriate (Fig. 2). The largest number of respondents (46.1%) felt that PS is appropriate when the estimated life expectancy was less than two weeks (Fig. 3). Interestingly, the second largest cohort (25.8%) indicated that there should not be prognostic limitations when prescribing PS. Wave analysis of prognosis and timing of PS showed no statistically significant difference (p 0.335) in responses.

FIG. 3.

“What should be the estimated life expectancy for a patient receiving palliative sedation?”

Most respondents reported infrequent distress when prescribing PS. Using a distress scale, the most common response was feeling “occasionally distressed” (31.7%). Nearly two-thirds of respondents reported feeling “never distressed,” “rarely distressed,” or “occasionally distressed,” whereas one-third of respondents described their distress level as either “often distressed,” “frequently distressed,” or “always distressed.” There was no significant difference in how HPM board-certified or HPM fellowship-trained clinicians answered this question (p 0.218, p 0.461, respectively). Clinicians who predominantly work in a hospice environment were significantly less distressed when prescribing PS than clinicians working in nonhospice settings (p 0.042).

There was a significant difference in respondents' confidence with assessing whether PS dosing is proportionate, whether the underlying illness is terminal, and whether a symptom is refractory or intractable (p < 0.001). There was a trend toward respondents feeling less confident about assessing whether a symptom is intractable compared with determining proportionate dosing or when an illness is terminal (Fig. 4). The subset of AAHPM respondents, board certified or fellowship trained in HPM, were significantly more confident at assessing all three of these cardinal components.

FIG. 4.

Clinician Confidence with Assessment of PS Components.

Respondents' experiences regarding PS policies

The majority of respondents were either unsure if their healthcare institution had guidelines or policies for prescribing PS, or reported that their institution did not have one (56.58%). Most respondents would feel either moderately, quite, or extremely supported by their healthcare institution if they were to prescribe PS (85.2%; Table 3). There was no statistical correlation between having established PS guidelines or policies and clinician distress with prescribing PS (p 0.958).

Table 3.

Clinicians' Experiences with Palliative Sedation Policies

Question	% of respondents (n)	Total No. of respondents
Does your institution have established guidelines or polices for prescribing PS?
Yes	43.4 (343)	790
No	44.9 (355)
Unsure	11.7 (92)
Which of the following should be embedded into a PS protocol for best practice? (Choose all that apply)
DNR	86.2 (677)		785
DNI	78.0 (612)
Cessation of artificial nutrition	60.9 (470)
Cessation of artificial hydration	52.5 (412)
Bioethics consult	20.1 (158)
Institution's risk management team	10.8 (85)
Chaplaincy consult	55.0 (432)
All of the above	15.2 (119)
None of the above	1.7 (13)
If you were to prescribe PS, how supported do you think you would be by your healthcare institution?
Not at all supported	3.2 (25)	789
Slightly supported	8.2 (65)
Moderately supported	20.7 (163)
Quite supported	37.8 (298)
Extremely supported	26.7 (211)
Not applicable	3.4 (27)

DNR, do not resuscitate; DNI, do not intubate.

Discussion

Our survey findings suggest that palliative sedation is the preferred terminology among a large majority of HPC clinicians compared with other terminology used in medical literature. Furthermore, most respondents felt it was appropriate to apply the term palliative sedation in hypothetical clinical situations in which a sedative was used with the intent to sedate a patient for refractory symptoms. In clinical situations in which sedation occurred as an unintended side effect from pharmacotherapies, the majority of clinicians surveyed did not feel palliative sedation was an applicable term. In other words, therapeutic intentions matter with regard to how our respondents applied the term palliative sedation. This should be considered when scrutinizing other published research investigating the prevalence of PS and its effect on survival. While a systemic review reported that PS likely had no overall effect on survival, 7 of the 11 studies reviewed were retrospective in nature, rendering it difficult to impossible for researchers to determine the clinical intentions.¹¹ Even published prospective studies, which have concluded that survival is not effected by PS, did not clearly describe how they determined the therapeutic intent of the treating clinician in cases deemed to be PS.^12,14

While these results suggest there is modest agreement among HPC clinicians regarding the preferred terminology and the appropriate application of the term palliative sedation, there are divergent views regarding the circumstances that warrant PS. For example, less than half of respondents felt it was appropriate to apply the term palliative sedation in a hypothetical clinical situation akin to respite sedation (short-term course of sedation followed by a sedation holiday after 48 hours). Also, nearly a quarter of respondents felt a continuous opioid drip to manage dyspnea during ventilator discontinuation constituted PS, even though opioids are not recommended for PS.^15–18 Therefore, there is no clear agreement among our respondents regarding the range of practices and therapies delivered under the rubric of palliative sedation. The observed discordance in how respondents view PS (including the terminology, the timing, the indications, and medications) may be a symptom of the disparate guidance from the professional organizations.²

Similar to other published research, our respondents described a clear bioethical distinction between PS and physician-assisted death and euthanasia.¹⁹ The majority of participants did not feel distressed when prescribing PS. Yet, when we broke the bioethical rationale for the appropriate use of PS into its three cardinal elements, clinicians expressed significantly less confidence assessing when symptoms are refractory than they did determining that the sedation is proportionate and that the illness is terminal. This discrepancy suggests that the determination of “refractory symptoms” may be more difficult than assessing that the sedation is proportionate. One possible solution would be to incorporate an itemized checklist of treatments for relevant symptoms into a PS guideline or protocol to better characterize when symptoms are sufficiently refractory. Additionally, best practice should include conferring with other providers if uncertainty remains before initiation of PS.

Although we did not specifically ask our respondents to comment on their most common indication for PS, it is our hypothesis that physical symptoms (pain, dyspnea, delirium, or nausea) are a more common indication than symptoms signifying psychological distress (anxiety, existential distress, or depression). This trend could be explained by provider bias that ascribes different levels of suffering originating from the body as opposed to suffering of the mind. Alternatively, providers may feel less confident in treating emotional suffering, or have more difficulty in recognizing when the symptom is refractory. We were surprised that more respondents felt that existential distress was an appropriate indication for PS in comparison with depression. Perhaps this finding reflects the ethical and clinical challenges HPC clinicians face when determining decision-making capacity among depressed patients, in that the depression is clouding the patient's judgment and ability to consent to the PS. Conversely, clinicians may feel there are fewer allopathic treatments for existential suffering, making PS a relatively more attractive option.

Patients and families have asked most of our respondents about PS, and similar to other research, a majority of our respondents have prescribed PS in the last year.¹³ PS is likely more common in hospice settings because hospice clinicians were more likely to have prescribed PS within the last year (p < 0.0001). Nearly half of the providers felt that a prognosis of less than two weeks was the most appropriate estimated life expectancy for a patient to receive PS. We found this response intriguing considering the coincidence that a significant proportion of hospice patients die within 14 days (49.5% in 2012).²⁰ In practical terms, it can be clinically challenging to identify when a patient has less than two weeks to live, which may delay consideration of PS as an appropriate option.

Although access to PS was important to our respondents, the majority did not feel that experienced APPs should be able to prescribe PS. Likewise, a sizeable proportion of respondents (31.2%) indicated that non-HPC clinicians should not be allowed to prescribe PS. However, nearly 82% of the surveyed AAHPM members were physicians, which may have resulted in a sampling bias. Nevertheless, these findings may be worthy of further dialogue regarding the future role of non-HPC clinicians and APPs as PS prescribers considering the current and projected shortfall estimates of HPC physicians.²¹

Guidelines, policies, or explicitly written procedures, endorsed by a healthcare institution regarding the appropriate use of PS, have the potential to standardize practice and offer institutional support when PS is appropriately utilized. For institutions interested in creating their own PS policy or guideline, it may be worth noting that a significant majority of our respondents felt that “do-not-resuscitate” and “do-not-intubate” orders were appropriate elements of such a guideline. There was less agreement whether chaplaincy consultation and cessation of artificial hydration and nutrition were appropriate elements of such a policy or guideline. A significant majority did not feel risk management or bioethics consultations were necessary, consistent with the literature.¹³ However, the presence of institutional PS guidelines did not significantly correlate with clinician distress levels.

Study limitations

There may be an inherent selection bias among our responders compared with the general pool of HPC clinicians. Although AAHPM is the largest HPC organization, it is not known if AAHPM is a valid reflection of all HPC clinicians in the United States. Second, our response rate of AAHPM members was relatively low. A larger proportion of our respondents were hospice clinicians (39.5%) and board certified in HPM (79.2%) as compared with AAHPM members in general (28.0% and 75.0%, respectively).^22,23 Wave analysis showed a significant difference in early survey responders' preference of the term palliative sedation compared with late responders. However, other subanalyses did not show a responder bias when factoring in the differences between our survey population and that of the general AAHPM population. We regret that we did not inquire about respondents' clinical role (physician, APP, social worker, chaplain, and so on).

Although we attempted to minimize lead bias through question ordering and careful wording of our letter requesting study participation, it is likely that bias still occurred especially with the finding that palliative sedation is the preferred terminology. Specifically, AAHPM published a PS Position Statement shortly before our survey was dispersed. Hence, some of our respondents may have been compelled to choose palliative sedation not because of individual preference, but because it was perceived to be the more correct answer. Finally, despite the use of two editorial boards to phrase survey questions in a manner that would minimize response bias, it is likely this still occurred, especially since the implicit nature of multiple choice questions can inherently bias participants' responses.

Conclusion

Palliative sedation is the preferred term for the proportionate and intentional use of pharmacotherapies to lower awareness for refractory symptoms in terminally ill patients. A majority of HPC clinicians, our respondents, did not feel palliative sedation was an applicable term when sedation occurs as an unintended side effect from commonly utilized pharmacotherapies in terminally ill patients. However, there was less agreement among respondents regarding some of the more controversial practices regarding PS such as respite sedation and PS for psychological suffering.

Footnotes

Acknowledgments

The authors acknowledge Aniko Szabo, Anjishnu Banerjee, and Qun Xiang for biostatistical assistance; Katherine Ast and Jamal Turner at AAHPM for assistance in dispersing the survey; and David E. Weissman, Thomas Heinrich, and Nancy Havas for editorial assistance with the survey design.

Author Disclosure Statement

No competing financial interests exist.

References

American Academy of Hospice and Palliative Medicine: Statement on Palliative Sedation. 2014. http://aahpm.org/positions/palliative-sedation (last accessed January 15, 2017).

Gurschick

, Mayer

, Hanson

: Palliative sedation: An analysis of international guidelines and position statements. Am J Hosp Palliat Care, 2015; 32:660–671.

Olsen

, Swetz

, Mueller

: Ethical decision making with end-of-life care: Palliative sedation and withholding or withdrawing life-sustaining treatments. Mayo Clin Proc, 2010; 85:949–954.

Schildmann

, Schildmann

, Kiesewetter

: Medication and monitoring in palliative sedation therapy: A systematic review and quality assessment of published guidelines. J Pain Symptom Manage, 2015; 49:734–746.

Editorial Board Palliative Care: Practice Guidelines. Palliative Sedation. Utrecht, NL: Association of Comprehensive Cancer Centers (ACCC), 2006, p. 33.

Salacz

, Weissman

: Controlled sedation for refactory symptoms. J Palliat Med, 2005; 8:136–138.

Chater

, Viola

, Paterson

, et al.: Sedation for intractable distress in dying—A survey of experts. Palliat Med, 1998; 12:255–269.

Papavasiliou

, Payne

, Brearley

: Current debates on end-of-life sedation: An international expert elicitation. Support Care Cancer, 2014; 22:2141–2149.

Rady

, Verheijde

: Uniformly defining continuous deep sedation. Lancet Oncol, 2016; 16:89.

10.

Reid

, Demme

, Quill

: When there are no good choices: Illuminating the borderland between proportionate palliative sedation and palliative sedation to unconsciousness. Pain Manag, 2011; 1:31–40.

11.

Maeda

, Morita

, Yamaguchi

, et al.: Effect of continuous deep sedation on survival in patients with advanced cancer (J-Proval): A propensity score-weighted analysis of a prospective cohort study. Lancet Oncol, 2016; 17:115–122.

12.

Maltoni

, Scarpi

, Rosati

, et al.: Palliative sedation in end-of-lifecare and survival: A systematic review. J Clin Oncol, 2012; 30:1378–1383.

13.

Lux

, McCrate Protus

, Kimbrel

, et al.: A survey of hospice and palliative care physicians regarding palliative sedation practices. Am J Hosp Palliat Med, 2017; 34:217–222.

14.

Maltoni

, Pittureri

, et al.: Palliative sedation therapy does not hasten death: Results from a prospective multicenter study. Ann Oncol, 2009; 20:1163–1169.

15.

Cherny

, Radbruch

: European Association for Palliative Care (EAPC) recommended framework for the use of sedation in palliative care. Palliat Med, 2009; 23:581–593.

16.

Reuzel

RPB

, Hasselaar

, et al.: Inappropriateness of using opioids for end stage palliative sedation: A Dutch study. Palliat Med, 2008; 22:641–646.

17.

Cherny

, Portenoy

: Sedation in the management of refractory symptoms: Guidelines for evaluation and treatment. J Palliat Care, 1994; 10:31–38.

18.

Hanks

, Cherny

, et al.: Oxford Textbook of Palliative Medicine, 4th ed. Oxford University Press, New York, 2010, p. 1566.

19.

Anneser

, Jox

, Thurn

, et al.: Physician-assisted suicide, euthanasia, and palliative sedation: Attitudes and knowledge of medical students. GMS J Med Educ, 2016; 33: Doc11. doi: 10.3205/zma001010. eCollection 2016.

20.

National Hospice and Palliative Care Organization. (2014). NHPCO's Facts and Figures–Hospice Care in America. www.nhpco.org/sites/default/files/public/Statistics_Research/2014_Facts_Figures.pdf (last accessed January 15, 2017).

21.

Lupu

: American Academy of Hospice and Palliative Medicine Workforce Task Force. Estimate of current hospice and palliative medicine physician workforce shortage. J Pain Symptom Manage, 2010; 40:899–911.

22.

American Academy of Hospice and Palliative Medicine. (2014). AAHPM by the Numbers. Our progress in 2014. http://aahpm.org/uploads/AAHPM14_AnnualReport_Web_pages.pdf (last accessed January 15, 2017).

23.

Kamal

, Bull

, et al.: Characterizing the hospice and palliative care workforce in the U.S.: Clinician demographics and professional responsibilities. J Pain Symptom Manage, 2016; 51:597–603.