Missing the Mark: High Rates of Absent and Untimely Access to Specialty Palliative Care in Patients with Peri-Hospital Mortality

Abstract

Background:

Despite mounting evidence that specialty palliative care (PC) improves patients' symptoms, quality of life, and goal concordant care, these services are likely underutilized.

Objective:

To determine the rate of missed and delayed opportunities for specialty PC in patients with peri-hospital death.

Design:

A retrospective, cross-sectional analysis, using electronic medical records of a state-wide healthcare system in Colorado, was performed. Included were adults who died during admission or within seven days of discharge from January 2015 to October 2018 at an academic medical center and had prior encounters within the affiliated state-wide healthcare system in the last year of life. Excluded were patients with sudden or obstetrics-related deaths. Referral orders from the electronic medical record identified specialty PC consultation. Data from the Colorado Department of Public Health and Environment linked with the medical record determined time from first PC consultation to death.

Results:

The sample included 2088 decedent patients, with most deaths (81%) occurring in the hospital. Only 33% of patients had PC consultation, which was higher for patients with cancer (42%) than for those without cancer (26%). Of patients with specialty PC consultation, the median time from first referral to death was eight days (interquartile range: 3.25–25 days).

Conclusions:

Patients with peri-hospital death have low rates of specialty PC consultation, which, when present, often occurs close to death. This suggests there is a high rate of missed opportunities for specialty PC in this population.

Introduction

The burden of serious illness is tremendous. Conditions such as heart failure, cancer, and dementia account for more than 1.7 million deaths in the United States annually.¹ Morbidity is equally high for these patients who suffer myriad and often debilitating physical and psychological symptoms.² Beyond patients, serious illness weighs on caregivers. Half of informal caregivers experience a decline in their own health.³ All these issues will escalate as the population ages and therapeutic options increase, thereby expanding the incidence and burden of chronic, progressive serious illnesses.

Palliative care (PC) is ideally positioned to address these challenges, as it is a specialty that is aimed at the amelioration of physical, psychological, social, and spiritual suffering at all stages of serious illness. We define serious illness as conditions with a high risk of mortality that negatively impact quality of life and well-being of patients and/or caregivers.⁴ PC may include hospice as the final phase of care, prioritizing comfort in the last months of life, yet it also encompasses a broader purview by integrating symptom alleviation with life-prolonging therapies at any point from diagnosis. Specialty PC involves consultation by inpatient and/or outpatient teams separate from hospice services. In contrast, primary PC, delivered by clinicians without specialty PC training, provides a basic level of PC such as initial goals-of-care discussions as well as management of uncomplicated symptoms and patient or family distress.⁵ Any remaining unmet needs warrant specialty referral.

Growing evidence demonstrates that specialty PC, especially when initiated early, improves person-centered outcomes, including patients' symptoms, caregiver burnout, and alignment of care with patients' goals.^6–8 In particular, quality-of-life benefits of PC extend over three months.^6–8 Specialty PC has also been associated with decreased utilization of potentially burdensome interventions at the end of life.^6,7,9 Unfortunately, there is a shortage of specialty-trained clinicians,¹⁰ which has prompted re-emphasis on primary PC. However, primary PC is currently constricted by insufficient training and time.^9,11–15

In particular, hospital death and hospice referral within seven days of death are common but are generally undesired end-of-life outcomes.¹⁶ Around 30% of patients still die in the hospital,¹⁷ even though it is the least preferred place to die¹⁸ and is associated with more unmet end-of-life needs.^19,20 Likewise, 28% of hospice referrals occur within seven days of death.²¹ However, both the National Quality Forum and the American Society of Clinical Oncology identify such short duration of hospice as an indicator of suboptimal end-of-life care given worse symptom control and quality-of-life scores when patients receive seven or fewer days of hospice.^22–24

Although specialty PC is believed to be underutilized,^2,9,14,15 there is a paucity of data on missed opportunities for specialty PC referral. To the authors' knowledge, in the United States, only one study of the Veteran's Affairs health system has investigated the rate of specialty PC referral in patients with hospital deaths.²⁵

The principal aim of this study was to determine the proportion of patients who died during or within seven days of hospitalization without specialty PC consultation in the last 12 months before death (rate of missed specialty PC). Secondary aims were to (1) determine the proportion of patients who received specialty PC consultation and who died within 90 days of referral (rate of delayed specialty PC), (2) identify the number and type of medical encounters in the past 12 months of life for patients who did not receive specialty PC (missed opportunities for specialty PC), and (3) determine patient characteristics associated with missed PC consultation.

Materials and Methods

Study design

This study is a retrospective analysis of a dataset created by an enterprise that extracts administrative and clinical information from an electronic medical record of a state-wide healthcare system and links this to data from several outside sources. The healthcare system includes one academic medical center and nine community hospitals as well as their associated outpatient specialty and primary care clinics. The academic medical center and three of the community hospitals have multidisciplinary, inpatient PC teams, including social workers, chaplains and physicians, and/or advanced practices providers. The academic medical center provides inpatient and outpatient PC services. The inpatient team maintains Joint Commission Advanced Palliative Care Certification (one of only 81 programs awarded this honor to date).²⁶ Joint Commission certification, among other metrics, requires the availability of PC services 24 hours a day, as well as collection and use of standardized performance measures to improve care. This clinical PC service is mature, having worked within the academic center for more than 15 years. All hospitals in the system have access to community-based PC.

Patients were included if they (1) died during admission or within seven days of discharge from January 2015 to October 2018 at the hospital of the academic medical center, (2) were ≥19 years old at time of death, and (3) had opportunities for specialty PC consultation in the 12 months before death (defined as at least one hospitalization in addition to the terminal hospitalization, two emergency department visits, or two face-to-face outpatient encounters at any facility within the state-wide healthcare system). A single emergency department or outpatient encounter in one year may be insufficient contact to demonstrate PC needs in many patients, especially those with nonmalignant disease. For this reason, two of either type of visits were required for inclusion. By including patients seen before terminal hospitalization within the system, the risk of receiving PC outside the system was minimized.

Death was identified from the state's Department of Public Health and Environment. Patients were excluded if they had International Statistical Classification of Diseases and Related Health Problems, Ninth Revision (ICD9) or International Statistical Classification of Diseases and Related Health Problems, Tenth Revision (ICD10) codes for obstetric-related problems (O00-O99, O9A, Z03, Z31-34, Z36-7, Z93A, V22-24, V28, 630-679) documented in the electronic medical record during the last 12 months of life. Patients were also excluded if ICD9 or ICD10 codes that are indicative of sudden death (as compiled from several prior studies^1,12,27,28) were documented during their terminal hospitalization. Sample size was determined by the number of patients meeting inclusion criteria during the study period. The study was reviewed by the institution's Institutional Review Board and deemed not to be human subject research.

Outcomes

The primary outcome was the rate of specialty PC referral in the last 12 months of life. Specialty PC was defined as the presence of an order in the electronic medical record for inpatient PC consultation or a referral for outpatient PC or home-based PC. Secondary outcomes were the number of days between PC consultation and death, and the number, type, and location of clinical encounters (outpatient appointments, emergency department visits, and hospitalizations) in the last 12 months of life. The former was calculated while comparing data from the medical record with that of the state's Department of Public Health and Environment, and the latter was obtained from the medical record.

Patient factors

Data on patient factors were derived from the electronic medical record, including: age at time of death, sex, ethnicity, race, insurance class at time of death, presence of advance directive (defined as presence of an uploaded document into the category of “Advance Directive(s),” “Do Not Resuscitate Order,” or “Medical Durable Power of Attorney”), intensive care unit (ICU) stay during terminal admission, hospital death, and hospice referral. The presence of medical conditions (cancer, end-stage renal disease, chronic kidney disease, chronic pulmonary disease, cirrhosis, congestive heart failure, stroke, dementia, depression, chronic pain, and drug abuse) was identified by using previously validated ICD9 and ICD10 codes.²⁹ Frailty was identified for patients with dementia, admission from skilled nursing facility, or frailty-associated ICD9 and ICD10 codes. Variables were only included in the dataset if they could be abstracted from the medical record, so no participants had missing data.

Statistical analyses

Given the degree of right skew in the time between first PC referral and death, medians and interquartile ranges (IQRs) rather than means and standard deviations are reported. To determine which demographic and clinical variables were associated with the presence of specialty PC consultation, chi-squared analysis was performed. Variables associated with specialty PC consultation (p-value <0.25) were fitted into the multivariate logistic regression model.³⁰ Covariates included in the logistic regression model are gender, age, race, ethnicity, insurance status, comorbidity, admission to ICU during terminal hospitalization, visit with primary care, death in hospital, and presence of advance directive in chart. Analysis was performed by using R i386 3.5.1 software.

Results

Our study identified 2928 patients who died during admission or within seven days of discharge from the academic medical center. Of these patients, 39 had insufficient contact within the state-wide healthcare system before their terminal admission so they were excluded. In addition, 801 patients were excluded for obstetric-related or sudden death. This resulted in the inclusion of 2088 patients, of whom 1688 (81%) patients died in the hospital. Patients were 56% male, 66% white, 15% Hispanic, and 11% black. The average age was 64 years old. There was a high level of medical illness, as 45% had cancer, 73% had organ failure (defined as having heart failure, cirrhosis, chronic kidney disease, or chronic lung disease), and 61% had frailty (defined as having dementia, being admitted from a nursing facility, or having ICD9 or ICD10 codes indicative of frailty).

Regarding the primary and secondary outcomes (Tables 1 and 2), only 33% of patients received PC consultation in the 12 months before death and this was higher for patients with cancer (42%) than for those without cancer (26%). The median time from initial PC consultation to death was eight days (IQR: 3.25–25 days). Of patients with specialty PC referral, 47% had consultation within just seven days of death and only 10% had consultation >90 days before death (Fig. 1). During the last 12 months of life, patients without PC had fewer healthcare contacts than those with PC with an average of 1.8 versus 2.4 hospitalizations, 1.5 versus 2.1 emergency department visits, and 9.4 versus 18.2 clinic appointments respectively.

FIG. 1.

Histogram of patients per duration of specialty palliative care.

Table 1.

Percentage of Patients with Specialty Palliative Care by Patient Variables

Variables	Overall population	Received specialty PC
Variables	% (n)	% (n)
Total	100.0 (2088)	33.4 (698)
Gender
Male	56.4 (1177)	32.7 (385)
Female	43.6 (911)	34.4 (313)
Age
<65 years old	48.9 (1022)	36.4 (372)
65–84 years old	41.7 (870)	30.9 (269)
>84 years old	9.4 (196)	29.1 (57)
Ethnicity
Hispanic	14.8 (308)	31.2 (96)
Non-Hispanic	80.7 (1685)	35.0 (589)
Other	4.5 (95)	13.7 (13)
Race
Native	1.1 (24)	20.8 (5)
Asian	2.5 (53)	34.0 (18)
Black	10.6 (222)	29.3 (65)
Multiple	1.5 (31)	41.9 (13)
Other	18.6 (388)	27.1 (105)
White	65.6 (1370)	35.9 (492)
Insurance
Medicaid	7.5 (157)	28.7 (45)
Private	11.4 (238)	37.8 (90)
Indigent	0.4 (8)	12.5 (1)
Medicare	30.9 (645)	28.5 (184)
Self-pay	47.1 (983)	36.6 (360)
Other	2.7 (57)	31.6 (18)
Condition
Cancer	45.2 (943)	42.4 (400)
No cancer	54.8 (1145)	26.0 (298)
Chronic pulmonary disease	40.1 (838)	31.7 (266)
Cirrhosis	16.9 (352)	30.7 (108)
Congestive heart failure	40.7 (850)	33.3 (283)
Chronic kidney disease	36.6 (764)	35.1 (268)
End-stage renal disease	8.2 (172)	34.9 (60)
Dementia	8.5 (178)	33.1 (59)
Stroke	25.9 (540)	38.5 (208)
Frailty	60.6 (1266)	40.5 (513)
Depression	30.6 (639)	41.3 (264)
Chronic pain	48.9 (1021)	39.7 (405)
Drug abuse	16.4 (343)	46.6 (160)
Other
Seen by primary care	16.8 (350)	43.4 (152)
Died in hospital	80.8 (1688)	30.2 (510)
Advance directives in chart	93.7 (1956)	34.7 (679)
In ICU during terminal hospitalization	58.4 (1219)	22.3 (272)

ICU, intensive care unit; n, number of patients; PC, palliative care; %, percentage of patients.

Table 2.

Time from First Palliative Care Consultation to Death by Patient Variable

Variable	Duration of PC (days)		% of patients with PC duration
Variable	Median	IQR	≤7 days	≤90 days
Total	8.0	3.25–25	46.8	89.5
Gender
Male	8.0	4–22	46.5	91.2
Female	8.0	3–28	47.3	87.5
Age
<65 years old	9.0	4–25.25	43.5	87.9
65–84 years old	8.0	3–27	46.8	91.4
>84 years old	5.0	3–9	68.4	91.2
Ethnicity
Hispanic	7.0	3–16	53.1	92.7
Non-Hispanic	9.0	4–27	45.0	88.8
Other	3.0	3–5	84.6	100.0
Race
Native	2.0	2–15	60.0	100.0
Asian	12.0	6–137	33.3	66.7
Black	16.0	7–47	27.7	80.0
Multiple	27.0	5–57	38.5	84.6
Other	5.0	3–13	59.0	96.2
White	8.0	3–24	47.4	90.2
Insurance
Medicaid	17.0	6–52	33.3	77.8
Private	8.5	4–19	46.7	94.4
Indigent	8.0	8–8	0.0	100.0
Medicare	8.0	3–23.5	48.9	88.6
Self-pay	8.0	3–24	48.3	90.6
Other	14.0	5–35.75	33.3	83.3
Conditions
Cancer	9.0	4–33	41.3	87.3
No cancer	6.5	3–16	54.4	92.6
Chronic pulmonary disease	9.0	4–32	45.5	86.8
Cirrhosis	7.5	3–22.25	50.0	90.7
Congestive heart failure	8.0	4–26	46.3	89.0
Chronic kidney disease	9.5	4–32	42.5	88.1
End-stage renal disease	13.0	4–49	36.7	85.0
Dementia	9.0	4–40	42.4	83.1
Stroke	6.0	3–16	54.8	88.9
Frailty	11.0	4–34	40.4	86.7
Depression	13.0	5–45.5	34.5	81.8
Chronic pain	11.0	4–44	38.0	84.2
Drug abuse	11.0	5–47.25	40.0	82.5
Other
Seen by primary care	9.5	4.75–32	52.2	89.2
Died in hospital	7.0	3–19.75	39.5	86.8
Advance directives in chart	8.0	4–26	46.2	89.2
In ICU during terminal hospitalization	6.0	2–21.75	52.9	89.0

IQR, interquartile range.

In addition, of patients receiving specialty PC, the majority (91%) received inpatient consultation whereas only 20% and 4% had outpatient and home PC referrals, respectively. In comparison to patients without cancer, those with cancer had a higher degree of outpatient referrals (28% vs. 9%) and a lower rate of inpatient consultation (89% vs. 95%). Although 94% had some form of an advance directive document uploaded into the chart, only 10% received hospice care.

We used results from the unadjusted analysis to fit the logistic regression model (Table 3), which found that age <65 years old, age 65–84 years old, self-pay insurance, cancer, stroke, frailty, drug abuse, and presence of some type of advance directive document were positively associated with specialty PC referral and, therefore, negatively associated with missed opportunities for specialty PC referral. Conversely, other ethnicity, chronic pulmonary disease, and admission to ICU during terminal hospitalization were negatively associated with specialty PC referral and, thus, positively associated with missed opportunities for specialty PC referral.

Table 3.

Chi-Squared Analysis and Logistic Regression of Association with Palliative Care Consultation by Patient Variable

Variable	Proportion with specialty PC		Unadjusted		Logistic
Variable	With variable, % (n)	Without variable, % (n)	X²	p	Regression, OR (CI)
Gender
Male	32.7 (385)	34.4 (313)	0.6	0.5	1.0 (0.8–1.3)
Female	34.4 (313)	32.7 (385)	0.6	0.5	Reference
Age
<65 year old	36.4 (372)	30.6 (326)	7.7	<0.01	2.2 (1.5–3.3)
65–84 year old	30.9 (269)	35.2 (429)	4.0	0.04	1.6 (1.1–2.4)
>84 year old	29.1 (57)	33.9 (641)	1.6	0.2	Reference
Ethnicity
Hispanic	31.2 (96)	33.8 (602)	0.7	0.4	0.8 (0.6–1.2)
Non-Hispanic	35.0 (589)	27.0 (109)	8.8	<0.01	Reference
Other	13.7 (13)	34.4 (685)	16.5	<0.01	0.5 (0.2–0.9)
Race
Native	20.8 (5)	33.6 (693)	1.2	0.3	0.7 (0.2–2.0)
Asian	34.0 (18)	33.4 (680)	<0.01	1.0	1.1 (0.6–2.0)
Black	29.3 (65)	33.9 (633)	1.7	0.2	0.7 (0.5–1.0)
Multiple	41.9 (13)	33.3 (685)	0.7	0.4	1.7 (0.8–3.7)
Other	27.1 (105)	34.9 (593)	8.3	<0.01	1.1 (0.8–1.6)
White	35.9 (492)	28.7 (206)	10.7	<0.01	Reference
Insurance
Medicaid	28.7 (45)	33.8 (653)	1.5	0.2	1.0 (0.6–1.5)
Private	37.8 (90)	32.9 (608)	2.1	0.1	1.4 (0.9–2.0)
Indigent	12.5 (1)	33.5 (697)	0.8	0.4	0.3 (0.02–2.1)
Medicare	28.5 (184)	35.6 (514)	9.8	<0.01	—
Self-pay	36.6 (360)	30.6 (338)	8.2	<0.01	1.5 (1.1–1.9)
Other	31.6 (18)	33.5 (680)	0.02	0.9	1.1 (0.6–2.1)
Conditions
Cancer	42.4 (400)	26.0 (298)	61.7	<0.01	1.5 (1.2–1.8)
Chronic pulmonary disease	31.7 (266)	34.6 (432)	1.7	0.2	0.8 (0.6–0.9)
Cirrhosis	30.7 (108)	34.0 (590)	1.3	0.3	—
Congestive heart failure	33.3 (283)	33.5 (415)	<0.01	1.0	—
Chronic kidney disease	35.1 (268)	32.5 (430)	1.4	0.2	1.2 (1.0–1.5)
End-stage renal disease	34.9 (60)	33.3 (638)	0.1	0.7	—
Dementia	33.1 (59)	33.5 (639)	<0.01	1.0	—
Stroke	38.5 (208)	31.7 (490)	8.2	<0.01	1.6 (1.3–2.1)
Frailty	40.5 (513)	22.5 (185)	71.9	<0.01	1.9 (1.5–2.4)
Depression	41.3 (264)	30.0 (434)	25.2	<0.01	1.1 (0.9–1.4)
Chronic pain	39.7 (405)	27.5 (293)	34.4	<0.01	1.1 (0.9–1.4)
Drug abuse	46.6 (160)	30.8 (538)	31.5	<0.01	1.5 (1.2–2.0)
Other
Seen by primary care	43.4 (152)	31.4 (546)	18.4	<0.01	1.3 (1.0–1.7)
Died in hospital	30.2 (510)	47.0 (188)	40.2	<0.01	1.1 (0.8–1.4)
Advance directives in chart	34.7 (679)	14.4 (19)	22.0	<0.01	2.1 (1.2–3.6)
In ICU during terminal hospitalization	22.3 (272)	49.0 (426)	161.4	<0.01	0.3 (0.3–0.4)

CI, confidence interval; OR, odds ratio; X², chi squared.

Discussion

Absent specialty PC as missed care

The majority of patients with peri-hospital deaths in this sample (67%) did not receive specialty PC. This is particularly true for patients without malignancy (74%) as compared with those with cancer (58%). This proportion is higher than a comparable study of nearly 58,000 U.S. veterans who died in hospitals or nursing homes: This study found that 42% of patients did not receive specialty PC.²⁵ Although 20% of these veterans died in nursing homes, this is also a population at high risk for unmet end-of-life needs. Although the ideal rate of specialty PC consultation per death is unknown, a referral rate of 58% in patients with high risk for unmet end-of-life needs as seen within the Veteran's Affair healthcare system can serve as a minimal benchmark.

The discrepancy in rates of specialty PC referral in our cohort and the Veteran's Affair population likely represents a missed opportunity for care, because peri-hospital death is associated with poor outcomes. First, patients and their families generally prefer to avoid dying in the hospital.^1,11,18 In addition, in comparison to home or hospice death, hospital death is associated with more unmet end-of-life needs with worse symptom control, caregiver support, and communication regarding patients' illnesses.^19,20 One study of hospital decedents showed that up to 29% of symptoms received insufficient management, up to 57% of their caregivers' needs were inadequately addressed, and up to 33% of these patients received confusing or contradictory information.³¹ Importantly, more than half of complaints to hospitals pertain to the dying process.¹⁹

Similarly, dying within seven days of admission to hospice is a risk for poor end-of-life care.^22–24 Patients with fewer than seven days of hospice care are more likely to have undertreated pain, worse quality-of-life scores, and discordance between their actual and preferred place of death.^22,23 Moreover, caregivers of patients with such short duration of hospice care, compared with those with longer periods of care, have higher rates of depression up to six months after patients' deaths.²⁴ Because of this and related data, both the National Quality Forum and the American Society of Clinical Oncology identify admission to hospice within seven days of death as an indicator of poor end-of-life care.²²

Patients who did not receive specialty PC had numerous missed opportunities for referral to either outpatient or community-based PC, with an average of nine face-to-face clinical visits and two admissions during the last 12 months of life. The frequency of outpatient visits suggests that the clinic is an opportune setting for interventions to increase PC consultation. Likewise, hospitalized patients and their families may be particularly receptive to PC, as exacerbations of their illnesses can prompt reflection and highlight unmet needs.

Late specialty PC as delayed care

When consulted, specialty PC in our cohort occurred exceedingly close to death. The median time from referral to death was eight days. In previous studies, the number of days from first contact with specialty PC and death is variable, with the median ranging from 10 to 70 days.^12,14,32,33 The short time from specialty PC to death implies a delay in care. Although the precise timing of PC is still debated,³⁴ the body of literature supports positive effects of PC extending to at least three months.^6–8 In our cohort, 90% of patients had PC consults less than 90 days before death, suggesting that they did not fully benefit from PC.

Missed and delayed PC as a potential medical error

As we have stated, peri-hospital death is often not goal concordant care and confers a high risk of unmet end-of-life care needs. However, these patients had numerous opportunities in their last 12 months of life for specialty PC referral, which improves concordance of place of death with patient preference and reduces end-of-life hospitalizations.^2,9 In other words, peri-hospital death is a poor but potentially preventable end-of-life outcome. Predicting the precise timing of death, even with the best tools, is inherently limited. However, specialty PC can occur at any stage of serious illness, which makes receiving consultation before the last weeks of life much more feasible. Therefore, we argue that the majority of cases of absent and delayed specialty PC in patients with peri-hospital deaths may constitute a medical error. Further investigations of the quality of end-of-life care management in patients with peri-hospital mortality is necessary to confirm and clarify this claim.

Clinician bias, rather than patient and family resistance, is the main source of absent PC consultation.¹⁴ First, clinicians overestimate prognosis (optimism bias) by up to 530%.³⁵ In addition, there is often a misconception that PC is indicated only when death is imminent (commission bias) and PC is sometimes seen as a personal failure.^14,36 The high degree of untimely PC consultation in this study, despite access to highly resourced PC teams, indicates the impact of these biases. As such, a high index of suspicion is often required to recognize the need for specialty PC and there are many tools and triggers to identify patients who may benefit from PC.^37,38

Limitations

The retrospective review of data from the medical record carries several limitations, including the inaccuracies of charted information. Although this health care system encompasses a wide network of practices, some patients may have received uncaptured out-of-network PC services. In addition, we do not know the quality of end-of-life care for these patients and their families, so it is possible that all end-of-life needs were met by primary PC and specialty care was not warranted.

Our findings may overestimate the rates of specialty PC in other health care systems that may have fewer specialty PC resources than this academic medical center with a Joint Commission certified inpatient PC team with a low rate of missed consults, 2.4% during the study period (Dr. Jeanie Youngwerth, Director of Clinical Palliative Care, e-mail communication, February 2020). Finally, when to initiate specialty PC services is still debated. Defining delayed PC as consultation within three months of death carries support from several high-quality trials of early PC.^6–8 In addition, many guidelines and programs use the “surprise question” to trigger PC before the last six to twelve months of life.³⁹

Next steps

This study prompts further investigation of missed and delayed specialty PC on a population level in the United States. The proportion of patients with peri-hospital death with absent or delayed PC consultation can serve as an outcome for research on identifying patients with high likelihood of needing specialty PC.

Conclusion

Patients who die in the hospital or shortly after discharge have a high degree of missed opportunities for specialty PC. This may represent a medical error that warrants greater attention, especially given the general preference to die at home.

Footnotes

Acknowledgments

The results of this study were supported by the Health Data Compass Data Warehouse project (), and the data analysis owes much to the guidance of Tracy Schifeling, PhD.

Funding Information

No funding was received for this article.

Author Disclosure Statement

No competing financial interests exist.

References

Heron

: Deaths: Leading causes for 2016. Natl Vital Stat Rep, 2018; 67:1–77.

Kelley

, Morrison

: Palliative care for the seriously ill. N Engl J Med, 2015; 373:747–755.

Collins

, Swartz

: Caregiver care. Am Fam Physician, 2011; 83:1309–1317.

Kelley

, Bollens-Lund

: Identifying the population with serious illness: The “denominator” challenge. J Palliat Med, 2018; 21(S2):S7–S16.

Quill

, Abernethy

: Generalist plus specialist palliative care—Creating a more sustainable model. N Engl J Med, 2013; 368:1173–1175.

Kavalieratos

, Corbelli

, Zhang

, et al.: Association between palliative care and patient and caregiver outcomes: A systematic review and meta-analysis. JAMA, 2016; 316:2104–2114.

Davis

, Temel

, Balboni

, Glare

: A review of the trials which examine early integration of outpatient and home palliative care for patients with serious illnesses. Ann Palliat Med, 2015; 4:99–121.

Haun

, Estel

, Rücker

, et al.: Early palliative care for adults with advanced cancer. Cochrane Database Syst Rev, 2017; 6:CD011129.

Hearn

, Higginson

: Do specialist palliative care teams improve outcomes for cancer patients? A systematic literature review. Palliat Med, 1998; 12:317–332.

10.

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.

11.

French

, McCauley

, Aragon

, et al.: End-of-life medical spending in last twelve months of life is lower than previously reported. Health Aff Proj Hope, 2017; 36:1211–1217.

12.

Seow

, O'Leary

, Perez

, Tanuseputro

: Access to palliative care by disease trajectory: A population-based cohort of Ontario decedents. BMJ Open, 2018; 8:e021147.

13.

Fassbender

, Fainsinger

, Carson

, Finegan

: Cost trajectories at the end of life: The Canadian experience. J Pain Symptom Manage, 2009; 38:75–80.

14.

Beernaert

, Deliens

, Pardon

, et al.: What are physicians' reasons for not referring people with life-limiting illnesses to specialist palliative care services? A nationwide survey. PLoS One, 2015; 10:e0137251.

15.

Higginson

, Finlay

, Goodwin

, et al.: Do hospital-based palliative teams improve care for patients or families at the end of life?. J Pain Symptom Manage, 2002; 23:96–106.

16.

Earle

, Landrum

, Souza

, et al.: Aggressiveness of cancer care near the end of life: Is it a quality-of-care issue?. J Clin Oncol, 2008; 26:3860–3866.

17.

Cross

, Warraich

: Changes in the place of death in the United States. N Engl J Med, 2019; 381:2369–2370.

18.

Higginson

, Daveson

, Morrison

, et al.: Social and clinical determinants of preferences and their achievement at the end of life: Prospective cohort study of older adults receiving palliative care in three countries. BMC Geriatr, 2017; 17:271.

19.

Al-Qurainy

, Collis

, Feuer

: Dying in an acute hospital setting: The challenges and solutions. Int J Clin Pract, 2009; 63:508–515.

20.

Hales

, Chiu

, Husain

, et al.: The quality of dying and death in cancer and its relationship to palliative care and place of death. J Pain Symptom Manage, 2014; 48:839–851.

21.

NHPCO Facts and Figures: Hospice Care in America. NHaPC: https://www.nhpco.org/sites/default/files/public/Statistics_Research/2017_Facts_Figures.pdf. 2017. (Last accessed November 13, 2018).

22.

Kumar

, Wright

, Hatfield

, et al.: Family perspectives on hospice care experiences of patients with cancer. J Clin Oncol, 2017; 35:432–439.

23.

Wright

, Zhang

, Ray

, et al.: Associations between end-of-life discussions, patient mental health, medical care near death, and caregiver bereavement adjustment. JAMA, 2008; 300:1665–1673.

24.

Bradley

, Prigerson

, Carlson

MDA

, et al.: Depression among surviving caregivers: Does length of hospice enrollment matter?. Am J Psychiatry, 2004; 161:2257–2262.

25.

Wachterman

, Pilver

, Smith

, et al.: Quality of end-of-life care provided to patients with different serious illnesses. JAMA Intern Med, 2016; 176:1095–1102.

26.

Quality Check: https://www.qualitycheck.org/search/?keyword=palliative%20care#keyword=palliative%20care&advancedcertification=Advanced%20Palliative%20Care. 2018. (Last accessed January 26, 2020).

27.

Ahmad

, Spencer

, Warner

, Sutton

Provisional drug overdose death counts: https://www.cdc.gov/nchs/nvss/vsrr/drugoverdose-data.htm. (Last accessed October 29, 2018).

28.

Jones

, Mack

, Paulozzi

: Pharmaceutical overdose deaths, United States, 2010. JAMA, 2013; 309:657–659.

29.

Tonelli

, Wiebe

, Fortin

, et al.: Methods for identifying 30 chronic conditions: Application to administrative data. BMC Med Inform Decis Mak, 2015; 15:31.

30.

Mosmer

, Lemeshow

: Applied Logistic Regression. New York, NY: Wiley, 1989.

31.

Carey

, Dose

, Humeniuk

, et al.: The experience of hospital death: Assessing the quality of care at an academic medical center. Am J Hosp Palliat Care, 2018; 35:189–197.

32.

Stubbs

, Assareh

, Achat

, Jalaludin

: Inpatient palliative care of people dying in New South Wales hospitals or soon after discharge. Intern Med J, 2019; 49:232–239.

33.

Harrison

, Cavers

, Campbell

, Murray

: Are UK primary care teams formally identifying patients for palliative care before they die?. Br J Gen Pract, 2012; 62:e344–e352.

34.

Courtright

, Cassel

, Halpern

: A research agenda for high-value palliative care. Ann Intern Med, 2018; 168:71–72.

35.

Christakis

, Lamont

: Extent and determinants of error in physicians' prognoses in terminally ill patients: Prospective cohort study. West J Med, 2000; 172:310–313.

36.

Callaway

: Timing is everything: When to consult palliative care. J Am Acad Nurse Pract, 2012; 24:633–639.

37.

Youngwerth

, Min

, Statland

, et al.: Caring about prognosis: A validation study of the caring criteria to identify hospitalized patients at high risk for death at 1 year. J Hosp Med, 2013; 8:696–701.

38.

Nelson

, Curtis

, Mulkerin

, et al.: Choosing and using screening criteria for palliative care consultation in the ICU: A report from the Improving Palliative Care in the ICU (IPAL-ICU) Advisory Board. Crit Care Med, 2013; 41:2318–2327.

39.

White

, Kupeli

, Vickerstaff

, Stone

: How accurate is the “Surprise Question” at identifying patients at the end of life? A systematic review and meta-analysis. BMC Med, 2017; 15:139.