Evaluating the Effects of Inpatient Palliative Care Consultations on Subsequent Hospice Use and Place of Death in Patients With Advanced GI Cancers

Palliative care consultation was associated with increased hospice use, decreased likelihood of dying in a hospital, and increased likelihood of dying at home.

Abstract

Purpose:

Research has shown that patients with terminal illnesses prefer to die at home, yet more than 60% of patients with cancer are hospitalized in the last month of life. Less than half of these patients receive hospice care at the end of life. Inpatient palliative care consultations may serve as a bridge from hospitalization to receiving patient's preferred end-of-life care. GI tumors include some the deadliest cancers, and patients with these cancers may serve to benefit from palliative care services.

Methods:

The objective of this study was to evaluate the role of palliative care versus usual care on postdischarge outcomes and hospice use for patients with advanced GI cancers. Two hundred one adults, 82 of whom received a palliative care consult, were followed for 6 months after hospital discharge. Propensity scores were used to match palliative care patients to usual care patients. Outcome measures included hospice use, place of death, subsequent emergency department visits, hospital readmission, and survival.

Results:

Fifty-nine patients died in the 6 months postdischarge. Receiving a palliative care consult increased the odds of home death (odds ratio [OR] = 2.9; 95% CI, 1.02 to 8.44; P = .046) and decreased the odds of hospital death (OR = 0.159; 95% CI, 0.05 to 0.52; P = .002). At 2 and 4 months, more patients in the palliative care group were receiving hospice services at death, compared with the usual care group (75% v 18%, P = .001% and 93% v 30%, P = .000 respectively). There were no differences with respect to emergency department visits, hospital readmission, and survival.

Conclusion:

Palliative care consultation was associated with increased hospice utilization, decreased likelihood of dying in a hospital, and increased likelihood of dying at home.

Introduction

It is largely recognized that care for patients with advanced cancer needs improvement. Though studies show that patients with advanced cancer prefer to be cared for and to die at home,¹ more than 60% of patients with advanced cancer are admitted to a hospital in the last month of their life.² Tumors of the GI tract include some of the most prevalent and deadliest cancers. Colon cancer remains the most common GI cancer and the second leading cause of cancer death, and hepatocellular carcinoma is the fastest growing cause of cancer-related death in the United States.^3,4 Others, including pancreatic, gastric, and biliary cancers, have 5-year survival rates of less than 30%.³

Although the benefits of hospital palliative care on measures such as symptom burden, quality of life, and patient and family satisfaction have been demonstrated^5,6; less is known about palliative care's effect on outcomes such as hospice use and place of death. The objective of our study was to evaluate the role of an inpatient palliative care consultation service on postdischarge outcomes and hospice use for patients with advanced GI cancers given the increase in prevalence and high mortality associated with these cancers. We hypothesized that those patients who received palliative care consultation while hospitalized would have higher rates of hospice referral, a higher likelihood of receiving hospice care at the time of death, and an increased likelihood of dying outside of a hospital. In addition, we hypothesized that palliative care consultations would lead to fewer postdischarge emergency department visits and hospital readmissions.

Methods

We conducted secondary analysis on data collected from a large, multicenter National Cancer Institute–sponsored study. The parent project was a prospective study comparing the effects of palliative care versus usual care on patient, family, and health care use outcomes in patients with advanced solid tumors. Patients admitted with advanced cancer were enrolled at five US hospitals from August 2007 to July 2010. Palliative care consultation at all sites followed a standardized approach to care adapted from the National Comprehensive Cancer Network (NCCN) Guidelines. Specifically, this included (1) formal assessment of pain and symptoms using standardized validated instruments; (2) recommendations, communicated to the primary physician, for management of pain and symptoms using protocols developed by the NCCN; (3) formal communication with patients and caregivers using consistent protocols for establishing goals of care, communicating bad news, and withholding or withdrawing treatments that do not meet goals; and (4) comprehensive care transition management based on agreed-on goals of care and available resources.

Inclusion criteria for the parent study included age 18 and older and a diagnosis of metastatic or locally advanced, inoperable tumor. Exclusion criteria included cognitive impairment, admission for routine chemotherapy, death or discharge within 48 hours of admission, and any previous consultation with a palliative care specialist. Subjects were followed prospectively. Medical record abstraction was performed retrospectively to determine which patients received palliative care and which did not. Patients and designated proxies were followed up via telephonic interview by trained study investigators at 2, 4, and 6 months after discharge.

Subjects from the parent study with a GI cancer diagnosis were included in these analyses. Diagnoses included stage IV colorectal cancer; stages III-IV gastric, pancreatic, or liver cancers; and stages II-IV gallbladder or biliary cancers. These cutoffs for inclusion were chosen on the basis of overall poor prognosis associated with these various stages of disease and the potential to benefit from palliative care. There was a total of 427 patients with a GI cancer diagnosis, 82 of whom received palliative care consultation during hospitalization. Palliative care patients were matched to usual care patients by propensity score (PS). Various PS matching strategies were tested in order to create balanced palliative care and usual care groups. Matching strategies tested included radius, kernel, Mahalanobis, nearest-neighbor, and caliper. Ultimately, the caliper strategy with 1:many with replacement matching gave the largest sample size with the smallest mean bias of 5.7%. The caliper used was 0.2 standard deviations of the logit of the treated subject's PS. Eighty-two palliative care patients were matched to 119 usual care subjects.

The primary outcome measures of the study were referral to hospice (home or inpatient), hospice use, and place of death. Secondary outcome measures included postdischarge emergency department visits, hospital readmission, and survival. Continuous data were compared by using Student's t test, and qualitative data were compared by using the χ² test. All analyses were performed using SPSS version 20 and STATA version 10. Institutional review board approval was obtained through the Icahn School of Medicine at Mount Sinai and all collaborating sites.

Results

A total of 201 unique subjects were included in the analysis; 82 palliative care subjects were matched to 119 usual care subjects. The demographics and clinical characteristics of the two samples were similar (Table 1); mean age was approximately 60 years, approximately 65% of subjects were white, and a little less than half were female. The majority of patients in both groups presented with primary colorectal cancer, followed by pancreatic cancer. During the study period, 8% of patients died during their initial hospital stay, 40% died in the 6-month follow-up period, and 23% were lost to follow-up. Of the patients discharged alive from the hospital, 8% of patients in the usual care group were discharged to hospice (home or inpatient) versus 38% in the palliative care group (P < .001). At the 2-month follow-up mark, there were no significant differences between the palliative and usual care groups in terms of survival and those living in hospice (Table 2). However, significantly more patients in the palliative care group died in hospice care (75% v 18%, P = .001). At the 4- and 6-month follow-up intervals, there were no significant differences in survival and those living with hospice care in the palliative and usual care groups. At 4 months, significantly more patients died while receiving hospice care in the palliative care group (30% v 93%; P = .001). The same was observed at 6 months; however, this difference was not significant given the small sample of patients who died between months 4 and 6 (n = 8).

Table 1.

Patient Baseline Characteristics

Characteristic	Usual Care (n = 119)		Palliative Care (n = 82)		P
	No.	%	No.	%
Mean age, years	61		62		.68
Female		47		45	.82
Race
Black		26		29	.6
White		68		67	.87
Religion
Catholic		15		12	.65
Protestant		15		9	.21
Jewish		2		1	.55
ADL complete assist		7		11	.43
Mean Elixhauser comorbidity	4.16		4.33		.60
Insurance
Medicare		42		43	.93
Medicaid		18		18	1.00
Uninsured		4		9	.20
Cancer site					.45
Colorectal		46		35
Pancreas		20		29
Liver		9		10
Other		26		26

NOTE. Data reflect weighted values.

Abbreviation: ADL, activities of daily living.

Table 2.

Patient Outcomes

Outcome	Usual Care (%)		Palcare (%)		P
	No.	%	No.	%
Alive at discharge		95		89	.155
Discharge to hospice		8		38	.000
Living with hospice 2 mo		17		21	.605
Hospice at death at 2 mo		18		75	.001
Hospice at death 4 mo		30		93	.001
Living with hospice 4 mo		9		25	.096
Hospice at death 6 mo		67		80	.673
Living with hospice 6 mo		9		13	.682
Any hospice at death		27		85	.000
ED visit any time after discharge		60		59	.984
Mean No. of ED visits	2.06		1.86		.699
Hospitalized after discharge		68		67	.960
Mean No. of hospitalizations	1.09		1.02		.755

NOTE. Data reflect weighted values.

Abbreviation: ED, emergency department.

Cumulatively, during the 6-month follow-up period, significantly more patients in the palliative care group died while receiving hospice care (85% v 27%; P < .001). Information regarding place of death was gathered for 91% of the patients who died during the 6-month follow-up period. Of the patients in the palliative care group who died, 58% died at home versus 31% in the usual care group (P = .040). Forty-two percent of usual care patients died during a subsequent hospital admission versus 15% in the palliative care group (P = .020). In sum, palliative care consultation was associated with significantly increased odds of dying at home (odds ratio = 2.9; 95% CI, 1.02 to 8.44; P = .046) and significantly decreased odds of dying in a hospital (odds ratio = 0.0159; 95% CI, 0.05 to 0.52; P = .002). Finally, we observed no significant differences between the two groups with respect to postdischarge emergency department visits and hospital readmission (Table 2).

Discussion

We compared postdischarge outcomes in patients with advanced GI cancers who received palliative care consults versus those who received usual care while hospitalized. We found that significantly more patients in the treated group were discharged to hospice care. Although there were no significant differences in survival between the two groups, of the patients who did die in the 6 months after discharge, those who had received palliative care consultation in the hospital had significantly higher rates of dying under hospice care. Furthermore, receiving a palliative care consult was associated with three-fold increased odds of dying at home and an 85% reduction in odds of dying in a hospital. No significant differences were found with regards to subsequent emergency department visits and hospital readmission rates.

While many studies to date have shown the effects of early integration of palliative care for oncology patients on quality of life and survival,⁷ the goal of this study was to address some of the deficits in end-of-life care for patients with advanced GI cancers with regard to hospice use and place of death. Hospice still remains an underused resource despite its demonstrated ability to improve quality of care and patient satisfaction and reduce costs.^8,9 Our study addresses this gap in care by linking inpatient palliative care consultations to hospice use, especially at the time of death, when its resources are so crucial not only to patients, but also to their family members and caregivers. By initiating goals of care discussions and transitions in care, inpatient palliative care consultations enable patients and families to discuss a plan and preferences for future care. Although research has shown that people prefer to die at home, this is often not the reality for oncology patients, who frequently die in a hospital setting, as was the case with the usual care patients in our study.¹⁰ Conversely, the patients in our study who received palliative care consultation were three times more likely to die at home. Of importance, there was no significant difference in survival between the two groups.

The first limitation of our study was the small sample size of 201 subjects, with 25% lost to follow-up. This reflects the challenge of conducting studies with terminally ill patients and their caregivers, who are struggling through a physically and emotionally difficult time in their lives. Nonetheless, a sensitivity analysis showed that even if all of the usual care patients lost to follow-up had died in hospice and all of the palliative care patients lost to follow-up had died without hospice, the difference between to the two groups would still be statistically significant (P ≤ .001). In terms of our population, we were limited by the fact that less than 1% of subjects in both groups were Hispanic. Lastly, our study does not account for variability in exposure to palliative care. Specifically, we did not account for time that the palliative team monitored the patient when measuring the effect of palliative care consultation on outcomes.

Unplanned hospitalizations for patients with advanced cancer predict survival of less than 6 months¹¹ and serve as an appropriate time to initiate advance care planning and goals of care discussions. Inpatient palliative care consultations, with a focus on symptom management, goals of care discussions, and care transitions, can help to initiate these talks and bridge patients from hospitalization to appropriate end-of-life care. While this study suggests that initiating inpatient palliative care consultation for hospitalized patients with advanced GI cancers leads to increased hospice use before death and increased likelihood of death in the home, further investigation is necessary to explore this relationship with other cancer types.

Authors' Disclosures of Potential Conflicts of Interest

The authors indicated no potential conflicts of interest.

Author Contributions

Conception and design: All authors

Collection and assembly of data: All authors

Data analysis and interpretation: All authors

Manuscript writing: All authors

Final approval of manuscript: All authors