Abstract
Context
African Americans are less likely to receive hospice care and more likely to receive aggressive end-of-life care than Whites. Little is known about how palliative care consultation to discuss goals-of-care (hereafter “PCC”) is associated with hospice enrollment by race.
Objectives
To compare enrollment in hospice at discharge between propensity-matched cohorts of African Americans with and without PCC, and Whites with and without PCC.
Methods
Secondary analysis of a retrospective cohort study at a high-acuity hospital; employing stratified propensity-score matching for 35,154 African Americans and Whites age 18+ admitted for conditions other than childbirth or rehabilitation, who were not hospitalized at end of study, and didn’t die during index hospitalization (hospitalization during which first PCC occurred).
Results
Compared to African Americans without PCC, African Americans with PCC were 15 times more likely to be discharged to hospice from index hospitalization (2.4% vs. 36.5%, P < 0.0001). Compared to White patients without PCC, White patients with PCC were 14 times more likely to be discharged to hospice from index hospitalization (3.0% vs. 42.7%, P < 0.0001).
Conclusion
In propensity-matched cohorts of seriously ill patients, palliative care consultation to discuss goals-of-care was associated with significant increases in hospice enrollment at discharge among both African Americans and Whites. Research is needed to understand how PCC influences decision-making by race, how PCC is associated with post-discharge hospice outcomes such as disenrollment and hospice length of stay, and if PCC is associated with improving racial disparities in end-of-life care.
Keywords: Palliative care, hospice care, healthcare disparities, advance care planning, patient care planning, decision-making process, terminal care
INTRODUCTION
In the United States, African Americans are less likely than Whites to have advance care planning discussions (1) despite evidence that patient-provider conversations about goals-of-care are associated with preference-concordant end-of-life care and less aggressive treatments (2, 3). Aggressive end-of-life care is associated with family perceptions of lower quality care (4, 5), greater decision regret (6), and lower patient quality-of-life (7). Communication disparities and differences in discussion effectiveness may help explain why African Americans are more likely than Whites to receive aggressive end-of-life care and die in hospitals, and less likely to receive comfort-oriented hospice care (1, 8). Hospice offers holistic illness management and the possibility of higher quality-of-life and satisfaction near end-of-life (9–13), yet a significantly lower proportion of African Americans die on hospice compared to Whites (14). Evidence suggests lack of knowledge about hospice strongly contributes to African American underutilization of hospice (15, 16). Palliative care consultations to discuss goals-of-care (“PCC”) can improve communication about prognosis and care options such as hospice (17), with earlier consultations resulting in significantly longer hospice length of stay (18). However, despite the National Academy of Medicine’s recommendation to implement palliative care and reduce racial disparities in end-of-life care (3), little is known about PCC associations with hospice use by race. The purpose of this study was to compare enrollment in hospice at discharge between propensity-matched cohorts of seriously ill African Americans with and without PCC and seriously ill Whites with and without PCC to understand if PCC is associated with increased hospice use within racial groups.
METHODS
Setting
This secondary analysis of a retrospective cohort study employed medical records data from a large, urban, high-acuity academic medical system in the Northeast United States (19). The medical center serves a diverse area composed of 46% African Americans, 36% Whites, 9% Asians, and 6% Hispanics and receives referrals and transfer patients from suburban areas (20). Over half of families in the service area live in poverty (20). The center’s palliative care team is well-established and predominantly works as a consultation service (19). Physicians, advance practice nurses, registered nurses, social workers, a pharmacist, and a chaplain make up the team. Intensive care unit teams request one-third of all palliative care consultations (19). The parent study, which analyzed outcomes of 41,363 seriously ill hospitalized patients using propensity score matching, found PCC was associated with reduced future acute care costs and utilization resulting in savings of over $6000 per patient, but did not assess outcomes by race or analyze hospice enrollment (19). Although race was not a factor in whether or not a patient received PCC in the parent study (19), race may be a factor in cost and utilization outcomes associated with PCC, warranting its study. In our study, supplementary Medicaid data was pulled from electronic medical records.
Participants
Patients were included if they were 18 years or older; self-identified as African American or White; were admitted between July 1, 2014 and October 31, 2016 for conditions other than childbirth or rehabilitation; were not hospitalized at the end of the study period; and did not die during index hospitalization, the admission during which patients first received PCC. The hospital’s palliative care registry, which includes demographic and clinical information (e.g., the reason for consultation), was used to identify all patients who received a palliative care consultation during the study period. Patients who received a consult specifically to discuss goals-of-care were included. Patients who received a consult for reasons other than goals-of-care were excluded (e.g., exclusively pain management).
Ethics
This study was approved by the University of Pennsylvania institutional review board (August 21, 2018; authorized by 45 CFR 46.104, category #4) and followed strict procedures to ensure patient data privacy, security, and ethics.
Analysis
The independent variable is PCC as recorded in the palliative care registry according to patient self-identified racial group (African American, White). All hospital admissions during the study period were placed into one of four groups: (1) African Americans with PCC, (2) African Americans without PCC, (3) Whites with PCC, and (4) Whites without PCC. Based on data available in the parent study, the outcome is patient discharge to hospice from index hospitalization. The sample size (N = 35,154) met requirements for a statistical power of 0.91 and an alpha level of 0.01 to detect an effect size of 0.136. There was no missing data in our study, as the parent study had previously excluded 0.4% of patients in its sample due to missing data (representing <5% of patients).
At baseline, before propensity score matching, African Americans and Whites in PCC and Non-PCC groups differed significantly (Table 1). To account for these differences and reduce bias, we used a two-step matching process, per race, to make the PCC and Non-PCC groups similar in their mean propensity for having had PCC (21). First, we used logistic regression analysis to identify factors associated with the likelihood of a patient receiving PCC (Table 2). These variables included: age, gender, Medicaid status, primary diagnosis, All Patient Refined (APR)-DRG Severity of Illness (the extent of physiological decompensation) at time of index hospitalization discharge, APR-DRG Risk of Mortality (the likelihood of dying) at time of index hospitalization discharge, intensive care during index hospitalization, intensive care duration greater than six days during index hospitalization, oncology services during index hospitalization, acute care 30 days before index hospitalization, and acute care costs accumulated during index hospitalization (defined as the impact of $1,000 to represent acute care utilization and show the expected increase in getting PCC) (19).
Table 1.
Description of African American patients with and without palliative care consultation (PCC), and White patients with and without PCC before stratified propensity score matching
| Total N = 35,154 | African Americans Non-PCC | African Americans with PCC | P value | Whites Non-PCC | Whites with PCC | P value | ||
|---|---|---|---|---|---|---|---|---|
| N = 10,777 | N = 383 | N =23,180 | N = 814 | |||||
| Age (years) | 18–39 | 5,739 (16%) | 2385 (22.1%) | 19 (5.0%) | <.0001 | 3280 (14.2%) | 55 (6.8%) | <.0001 |
| 40–45 | 2,419 (7%) | 980 (9.1%) | 4 (1.0%) | 1405 (6.1%) | 30 (3.7%) | |||
| 46–50 | 2,642 (8%) | 953 (8.8%) | 21 (5.5%) | 1639 (7.1%) | 29 (3.6%) | |||
| 51–55 | 3,614 (10%) | 1215 (11.3%) | 45 (11.8%) | 2294 (9.9%) | 60 (7.4%) | |||
| 56–60 | 4,338 (12%) | 1330 (12.3%) | 55 (14.4%) | 2854 (12.3%) | 99 (12.2%) | |||
| 61–65 | 4,320 (12%) | 1165 (10.8%) | 53 (13.8%) | 2987 (12.9%) | 115 (14.1%) | |||
| 66–70 | 4,223 (12%) | 955 (8.9%) | 51 (13.3%) | 3093 (13.3%) | 124 (15.2%) | |||
| 71–75 | 3,199 (9%) | 668 (6.2%) | 41 (10.7%) | 2386 (10.3%) | 104 (12.8%) | |||
| >75 | 4,660 (13%) | 1126 (10.5%) | 94 (24.5%) | 3242 (14.0%) | 198 (24.3%) | |||
| Gender | Male | 17,286 (49%) | 4543 (42.2%) | 172 (44.9%) | .2836 | 12148 (52.4%) | 423 (52.0%) | 0.80 |
| Female | 17,868 (51%) | 6234 (57.8%) | 211 (55.1%) | 11032 (47.6%) | 391 (48.0%) | |||
| Medicaid | Yes | 4,819 (14%) | 3454 (32.0%) | 83 (21.7%) | <.0001 | 1243 (5.4%) | 39 (4.8%) | 0.48 |
| No | 30,335 (86%) | 7323 (68.0%) | 300 (78.3%) | 21937 (94.6%) | 775 (95.2%) | |||
| Primary Diagnosis | Cancer | 6,955 (19.8%) | 1063 (9.9%) | 103 (26.9%) | <.0001 | 5472 (23.6%) | 317 (38.9%) | <.0001 |
| Cardiovascular disorder / Heart Failure | 6,430 (18.3%) | 1774 (16.5%) | 76 (19.8%) | 4444 (19.2%) | 136 (16.7%) | |||
| Endocrine disorder | 1,942 (5.5%) | 902 (8.4%) | 8 (2.1%) | 1014 (4.4%) | 18 (2.2%) | |||
| GI disorder | 4,120 (11.7%) | 1147 (10.6%) | 29 (7.6%) | 2872 (12.4%) | 72 (8.9%) | |||
| Gynecologic or urologic disorder | 2,393 (6.8%) | 1086 (10.1%) | 17 (4.4%) | 1267 (5.5%) | 23 (2.8%) | |||
| Infectious disease and Sepsis | 2,950 (8.4%) | 1227 (11.4%) | 63 (16.5%) | 1565 (6.8%) | 95 (11.7%) | |||
| Neurologic disorder | 3,508 (10%) | 1204 (11.2%) | 31 (8.1%) | 2227 (9.6%) | 46 (5.7%) | |||
| Respiratory disorder | 1,394 (4.0%) | 493 (4.6%) | 23 (6.0%) | 819 (3.5%) | 59 (7.3%) | |||
| Other | 5,462 (15.5%) | 1881 (17.5%) | 33 (8.6%) | 3500 (15.1%) | 48 (5.9%) | |||
| APR-DRG Severity of Illness | Minor | 8,044 (22.9%) | 2343 (21.7%) | 6 (1.6%) | <.0001 | 5680 (24.5%) | 15 (1.8%) | <.0001 |
| Moderate | 13,876 (39.5%) | 4446 (41.3%) | 36 (9.4%) | 9316 (40.2%) | 78 (9.6%) | |||
| Major | 9,912 (28.2%) | 3140 (29.1%) | 167 (43.6%) | 6270 (27.1%) | 335 (41.2%) | |||
| Severe | 3,322 (9.4%) | 848 (7.9%) | 174 (45.4%) | 1914 (8.3%) | 386 (47.4%) | |||
| APR-DRG Risk of Mortality | Minor | 15,914 (45.3%) | 5142 (47.7%) | 11 (2.9%) | <.0001 | 10741 (46.3%) | 20 (2.5%) | <.0001 |
| Moderate | 10,112 (28.8%) | 3072 (28.5%) | 50 (13.1%) | 6854 (29.6%) | 136 (16.7%) | |||
| Major | 6,637 (18.9%) | 1940 (18.0%) | 179 (46.7%) | 4177 (18.0%) | 341 (41.9%) | |||
| Severe | 2,491 (7.1%) | 623 (5.8%) | 143 (37.3%) | 1408 (6.1%) | 317 (38.9%) | |||
| Acute care hospitalization 30 days prior to index hospitalization | Yes | 653 (1.9%) | 127 (1.2%) | 80 (20.9%) | <.0001 | 240 (1.0%) | 206 (25.3%) | <.0001 |
| No | 34,501 (98.1%) | 10650 (98.8%) | 303 (79.1%) | 22940 (999.0%) | 608 (74.7%) | |||
| ICU care during index admission | Yes | 11,448 (32.6%) | 2792 (25.9%) | 197 (51.4%) | <.0001 | 8025 (34.6%) | 434 (53.3%) | <.0001 |
| No | 23,706 (67.4%) | 7985 (74.1%) | 186 (48.6%) | 15155 (65.4%) | 380 (46.7%) | |||
| ICU care > 6 days during index admission | Yes | 2,637 (7.5%) | 624 (5.8%) | 104 (27.2%) | <.0001 | 1683 (7.3%) | 226 (27.8%) | <.0001 |
| No | 32,517 (92.5%) | 10153 (94.2%) | 279 (72.8%) | 21497 (92.7%) | 588 (72.2%) | |||
| Visited by Oncology service during index admission (1st or 2nd service) | Yes | 2,984 (8.5%) | 456 (4.2%) | 77 (20.1%) | <.0001 | 2188 (9.4%) | 263 (32.3%) | <.0001 |
| No | 32,170 (91.5%) | 10321 (95.8%) | 306 (79.9%) | 20992 (90.6%) | 551 (67.7%) | |||
| DNR documented during index admission | Yes | 1,425 (4.1%) | 242 (2.2%) | 172 (44.9%) | <.0001 | 594 (2.6%) | 417 (51.2%) | <.0001 |
| No | 33,729 (95.9%) | 10535 (97.8%) | 211 (55.1%) | 22586 (97.4%) | 397 (48.8%) | |||
| Mean number of days hospitalized during index admission (SD) | 5.96 (7.63) | 17.05 (19.5) | <.0001 | 6.29 (7.67) | 16.53 (19.06) | <.0001 | ||
| Median number of days hospitalized during index admission (IQR) | 4.0 (2.0–7.0) | 10.0 (6.0–20.0) | <.0001 | 4.0 (2.0–7.0) | 10.0 (6.0–20.0) | <.0001 | ||
| Mean number of ICU days during index admission (SD) | 1.18 (4.26) | 6.03 (12.70) | <.0001 | 1.45 (4.00) | 6.36 (13.78) | <.0001 | ||
| Median number of ICU days during index admission (IQR) | 0 (0–0) | 1.0 (0–6.0) | <.0001 | 0 (0–1.0) | 1.0 (0–7.0) | <.0001 | ||
| Mean direct acute care costs during index admission (SD) | $15665 ($22667) | $35982 ($49024) | <.0001 | $19583 ($25209) | $40126 ($59607) | <.0001 | ||
| Median direct acute care costs during index admission (IQR) | $9,723 ($5,898–$15,641) | $18,578 ($10,811–$38,119) | <0001 | $12,114 ($7,639–$21,373) | $20,017 ($10,455–$41,445) | <.0001 | ||
| Changed goals-of-care during PCC index admission (% yes) | n/a | 60.1% | - | n/a | 67.7% | 0.01 | ||
Significance tests for the percentage variables involved a Chi-squared test; significance tests for parametric continuous variables involved a t test; significance tests for non-parametric continuous variables involved a Kruskal-Wallis non-parametric test of ranks.
Table 2.
Logistic Regression Analysis of Likelihood of Receiving Palliative Care Consult (PCC)
| African American patients | White patients | ||||||
|---|---|---|---|---|---|---|---|
| Parameter | Odds Ratio | 95% Confidence Interval for Odds Ratio | Odds Ratio | 95% Confidence Interval for Odds Ratio | |||
| Age (years) | 18–39 | 0.201 | (0.112,0.359) | 0.511 | (0.356,0.733) | ||
| 40–55 | 0.392 | (0.266,0.576) | 0.604 | (0.461,0.79) | |||
| 56–65 | 0.591 | (0.416,0.839) | 0.789 | (0.626,0.994) | |||
| 66–75 | 0.511 | (0.364,0.717) | 0.691 | (0.553,0.864) | |||
| >75 | - | - | - | - | |||
| Gender | Male | 0.896 | (0.71,1.131) | 0.732 | (0.624,0.858) | ||
| Female | - | - | - | - | |||
| Medicaid* | Yes | 1.256 | (0.911,1.733) | 1.325 | (0.912,1.925) | ||
| No | - | - | - | - | |||
| Primary Diagnosis | Cancer | - | - | - | - | ||
| Cardiovascular disorder and Heart Failure | 0.32 | (0.22,0.466) | 0.365 | (0.284,0.469) | |||
| Endocrine disorder | 0.226 | (0.105,0.487) | 0.649 | (0.383,1.098) | |||
| GI disorder | 0.35 | (0.213,0.573) | 0.638 | (0.472,0.863) | |||
| Gynecologic or urologic disorder | 0.292 | (0.162,0.525) | 0.461 | (0.284,0.748) | |||
| Infectious disease and Sepsis | 0.33 | (0.223,0.488) | 0.607 | (0.461,0.8) | |||
| Neurologic disorder | 0.244 | (0.151,0.396) | 0.463 | (0.324,0.663) | |||
| Respiratory disorder | 0.469 | (0.275,0.798) | 1.105 | (0.786,1.553) | |||
| Other | 0.311 | (0.196,0.493) | 0.448 | (0.316,0.635) | |||
| APR-DRG Risk of Mortality | Minor | 0.021 | (0.011,0.041) | 0.016 | (0.01,0.026) | ||
| Moderate | 0.096 | (0.065,0.143) | 0.11 | (0.086,0.142) | |||
| Major | 0.47 | (0.35,0.633) | 0.464 | (0.38,0.566) | |||
| Severe | - | - | - | - | |||
| ICU during index admission | Yes | 1.317 | (1.018,1.704) | 1.347 | (1.138,1.595) | ||
| No | - | - | - | - | |||
| ICU > 6 days during index admission | Yes | 1.589 | (1.118,2.258) | 1.698 | (1.362,2.118) | ||
| No | - | - | - | - | |||
| Seen by Oncology in index admission | Yes | 1.764 | (1.251,2.487) | 1.703 | (1.394,2.08) | ||
| No | - | - | - | - | |||
| Admitted to hospital 30 days prior | Yes | 17.643 | (12.064,25.801) | 23.996 | (18.703,30.786) | ||
| No | - | - | - | - | |||
| Parameter (Continuous Variable) | Estimate | SE | P-Value | Estimate | SE | P-Value | |
| Direct Cost of index admission (impact per $1,000) * | 2.92*10−3 | 1.52*10−3 | .0535 | 1.72*10−3 | 9.55*10−4 | 0.072 | |
Abbreviation: SE, standard error
For every increase of $1000 in direct costs, the estimate shows the expected increase in getting a palliative care consult.
The APR-DRG variables have four subcategories (minor, moderate, major, extreme) that are assigned systematically at discharge based on diagnoses and procedures coded for billing during hospitalization (19, 22). These variables depend on the patient’s underlying acute or chronic condition and comorbidity interactions (19, 22). The presence and interaction of multiple serious diseases characterize high Severity of Illness and high Risk of Mortality (23). As a secondary analysis, our study was limited to using these available measures to indicate illness severity and comorbidity, however a 2018 study found APR-DRG Risk of Mortality and Severity of Illness are better predictors of in-hospital mortality than the Charlson Comorbidity Index (24), making them dependable proxies for illness severity and enabling effective severity adjustment. Because we could not account for socioeconomic variables such as education and health literacy that were unavailable in the dataset, we used Medicaid insurance status as a proxy for socioeconomic status (25–27).
The logistic regression model of African American PCC and Non-PCC patients had a C statistic of 0.907, indicating adequate model fit. The model of White PCC and Non-PCC patients had a C statistic of 0.901, also indicating adequate model fit.
Preliminary propensity models showed that many variables were significant in predicting which patients received PCC. To maximize the quality of matching and prevent substantial data loss, we followed the subclassification or stratification approach suggested by Rosenbaum and Rubin (28) and used stratified propensity score matching, an established method in health services research (19, 29–31). Based on existing methodological discourse, no propensity score matching method is a priori superior to others (32), however this method has been shown to reduce 90–95% of bias (28, 33). When compared to traditional propensity score matching, stratified matching has been found to produce similar results (31, 34).
After developing logistic regression models by race, we created propensity scores for each patient using output from the logistic regression models, by race. To account for possible inherent differences and nonrandom assignment of variables in the PCC and Non-PCC groups, we then created stratified propensity score models, by race, by ranking and stratifying the scores into six mutually exclusive propensity tiers, from highest to lowest, to identify which patients were most likely to have received PCC (Table 3). All patients were assigned to a tier. For each race’s model, patients in the PCC and Non-PCC groups were similar in terms of their propensity scores for having had PCC (Table 3), enabling rigorous comparison of PCC patients to Non-PCC patients within tiers, within races (21, 35). With this method, PCC and Non-PCC groups do not need to match on each individual variable. Instead, confounding variables are accounted for through logistic regression to predict PCC. To reduce bias, the method’s goal is for top tiers to have similar mean propensities for PCC. We then used a Cochran-Mantel-Haenszel test to analyze discharge to hospice for patients in the two highest propensity tiers (blended), by race, because these patients were most likely to have received PCC (Table 4). The lowest propensity tiers had too few PCC cases, so we did not analyze them. Statistical significance was taken at the 0.05 level.
Table 3.
Propensity Scores of Patients Most Likely to Receive Palliative Care Consultation1 (PCC), by Race
| African American patients | |||||||||
| Patients with Palliative Care Consult,1 n = 383 | Patients without Palliative Care Consult,1 n = 10,777 | ||||||||
| Tier | Propensity Score Mean | n | Min, | Max | Propensity Score Mean | n | Min, | Max | |
| Highest | 0.303 | 178 | 0.202 | 0.481 | 0.276 | 1,667 | 0.202 | 0.498 | |
| 2nd Highest | 0.134 | 129 | 0.078 | 0.201 | 0.114 | 1,717 | 0.073 | 0.202 | |
| White patients | |||||||||
| Patients with Palliative Care Consult,1 n = 814 | Patients without Palliative Care Consult,1 n = 23,180 | ||||||||
| Tier | Propensity Score Mean | n | Min, | Max | Propensity Score Mean | n | Min, | Max | |
| Highest | 0.331 | 350 | 0.212 | 0.498 | 0.298 | 3,604 | 0.212 | 0.498 | |
| 2nd Highest | 0.153 | 257 | 0.097 | 0.212 | 0.141 | 3,698 | 0.096 | 0.212 | |
Palliative care consult to discuss goals-of-care
Table 4.
Discharge to Hospice After Propensity Score Matching in Patients Most Likely to Receive Palliative Care Consultation1 (PCC), by Race
| African American patients | Propensity Group | PCC | Non-PCC | P value b |
| Highest Tier (n)a | n = 178 | n = 1,667 | ||
| Second Highest Tier (n)a | n = 129 | n = 1,717 | ||
| Top Two Tiers Blended (n)a | n = 307 | n = 3,384 | ||
| Discharge to hospice from index admission (% yes) | Highest propensity group | 30.3% | 2.6% | <0.0001 |
| Second highest propensity group | 45.0% | 2.3% | <0.0001 | |
| Top Two Tiers Blended | 36.5% | 2.4% | <0.0001 | |
| White patients | Propensity Group | PCC | Non-PCC | P value b |
| Highest Tier (n)a | n = 350 | n = 3,604 | ||
| Second Highest Tier (n)a | n = 257 | n = 3,698 | ||
| Top Two Tiers Blended (n)a | n = 607 | n = 7,302 | ||
| Discharge to hospice from index admission (% yes) | Highest propensity group | 41.1% | 3.3% | <0.0001 |
| Second highest propensity group | 44.8% | 2.8% | <0.0001 | |
| Top Two Tiers Blended | 42.7% | 3.0% | <0.0001 | |
Sample sizes are consistent within each column (across all outcome variables). There were six mutually exclusive tiers, but only the top two were analyzed to show comparison among patients most likely to have received consult.
Significance tests involved a Cochran-Mantel-Haenszel test.
Palliative care consult to discuss goals-of-care
RESULTS
Our sample included 11,158 African American patients (PCC = 383; No PCC = 10,777) and 23,994 White patients (PCC = 814; No PCC = 23,180) (N = 35,154). As shown in Table 1, both groups differed significantly at baseline before matching across every variable except gender and use of Medicaid. At baseline, Whites with and without PCC did not differ in having Medicaid (P=0.48), however African Americans with and without PCC differed (PCC = 21.7% with Medicaid; No PCC = 32% with Medicaid; P=<0.0001). After matching, patients across racial groups did not differ in their propensity for having had PCC (Table 3).
Compared to African American patients without PCC, African American patients with PCC were 15 times more likely to be discharged to hospice from index hospitalization (top two propensity tiers blended, 2.4% vs. 36.5%, P < 0.0001) (Table 4). Compared to White patients without PCC, White patients with PCC were 14 times more likely to be discharged to hospice from index hospitalization (top two propensity tiers blended, 3.0% vs. 42.7%, P < 0.0001).
DISCUSSION
This is the first known study to assess hospice use associated with PCC by race. Our study demonstrated a large and significant impact of PCC on hospice enrollment in propensity-matched cohorts of seriously ill White patients and seriously ill African American patients.
Dramatic increases in patient discharge to hospice among both racial groups is important given the well-documented benefits of hospice care (36, 37). In a recent propensity-score matched study of heart failure patients with past history of hospitalization, for example, authors found that in the six months following a second hospitalization discharge, patients who enrolled in hospice had longer median survival and lower acute care use, including fewer emergency department visits, hospital days, and ICU stays; and were less likely to die in the hospital—all evidence that suggests fewer disruptive transitions in care and higher quality of life near death (38). In addition to evidence of hospice patients living longer than similarly ill patients who do not receive hospice (39, 40), hospice has been found to be associated with patients having fewer unmet needs, higher quality of care, better quality of dying, improved emotional support for patient or family, and better overall satisfaction among families (3, 41–43). Hospice also benefits health systems and payers by reducing unnecessary acute care use and related costs (2, 44). Palliative care consultations to discuss goals-of-care are similarly associated with cost reductions (2, 19, 45), likely partly due in part to increased use of hospice.
Even though our study found PCC was associated with hospice enrollment increases in both racial groups, absolute enrollment still differed among races. More research is needed to understand if PCC helps reduce racial disparities in end-of-life care after discharge to hospice (3), including disparities in hospital readmission, disenrollment from hospice, and the number of care transitions hospice patients experience (46–48). A recent study showed about half of informal caregivers of hospice patients had unmet information needs related to hospice care (49) and that inadequate understanding of hospice among family caregivers is a cause of hospitalization after hospice enrollment (50). Research is also needed to understand how PCC is associated with future hospice enrollment, duration of enrollment, future acute care use, and place of death—and if racial differences exist.
In our study, hospice increases were greater among African Americans, a population that has historically demonstrated lower use of hospice care (51). Increases in hospice use among African Americans may help reduce racial differences in end-of-life pain management, quality of care, perceptions and concerns about care quality, and other outcomes (52). Although distrust is often cited as a reason for African Americans choosing life-sustaining treatments over comfort care, a recent systematic review found trust in healthcare providers and advance care planning outcomes did not differ between African Americans and Whites engaged in advance care planning (53). Coupled with our results, this finding suggests improved interpersonal communication between clinicians and patients, which occurs during palliative care consultations (54), may be driving increased hospice use across races and should be investigated in studies aiming to improve racial disparities in hospice outcomes (46, 55).
Our study has several limitations. First, as is the case with all propensity-score studies, our models cannot account for unmeasured or unknown confounding variables (19, 56). We were unable to control for income, insurance, education, health literacy, religiosity, specific comorbidities, or duration of survival because these variables were not available in the dataset. Socioeconomic variables, such as income and education, which have been found to help explain differences in advance care planning in African Americans (53), were not available in the dataset and may have contributed. Compared to Whites, African Americans endure significant systematic disadvantages resulting in disproportionate and poor economic resources to higher unemployment and lower incomes, which impact health, health literacy, and access to quality care (53, 57). Less-educated patients in general have been found to have lower levels of health literacy (53, 58), which can influence how a patient understands, recalls, and applies information exchanged during healthcare encounters (58, 59). To account for this limitation, we used Medicaid status as a proxy. Still, unmeasured variables may be distributed differently in patients who received PCC and those who did not (19). Although our study controlled for severity of illness and risk of mortality using accepted variables (24), there are limitations to using secondary measures. As such, severity adjustment may be imperfect, possibly contributing to residual confounding.
Second, our analysis did not include hospice use among patients who enrolled in hospice after discharge. Third, PCC and Non-PCC patients were compared based on discharge from index admission, not last discharge during the study period. Propensity score matching should reduce any potential bias due to the distinction, but as a secondary analysis it is not possible to know if differences exist. Fourth, our study was conducted in a system with racially diverse patients and an experienced palliative care team. The size of the effect may vary in other settings. Finally, our study only analyzed African American and White patients. Evidence suggests Hispanic patients and Asian American patients also use hospice less than White patients, although enrollment rates are increasing in these groups (60). Future research on PCC and hospice enrollment among all racial/ethnic groups is needed (46).
Our findings provide evidence that PCC is associated with increased hospice use across races and that goals-of-care communication in PCC, rather than race alone, help determine end-of-life outcomes (6, 61). Clinicians should engage all patients in culturally-sensitive PCC discussions to clarify care preferences, enable informed decision-making, and improve understanding of how hospice might align with personal goals. Providing care consistent with patient values and goals enables the delivery of high-quality care (3, 62). Policies and practices that support goals-of-care discussions are needed to better meet the needs of vulnerable patients across racial groups.
Disclosures and Acknowledgments
The authors acknowledge the following financial support for the research, authorship, and/or publication of this article: Dr. Starr received support from the Ruth L. Kirschstein National Research Service Award training program in Individualized Care for At Risk Older Adults at the University of Pennsylvania, National Institute of Nursing Research [T32NR009356]; the Rita and Alex Hillman Foundation’s Hillman Scholars Program; and Jonas Philanthropies’ Jonas Nurse Leaders Scholars program. Dr. Ulrich is currently supported by a National Cancer Institute/National Institutes of Health Award [R01CA196131]. Dr. Meghani is currently supported by a National Institute of Health/National Institute of Nursing Research Award [R01NR017853]. The authors declare no conflicts of interest with respect to the research, authorship and/or publication of this article.
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Contributor Information
Lauren T. Starr, NewCourtland Center for Transitions and Health, University of Pennsylvania School of Nursing; Center for Bioethics, University of Pennsylvania School of Medicine..
Connie M. Ulrich, NewCourtland Center for Transitions and Health, University of Pennsylvania School of Nursing; University of Pennsylvania Perelman School of Medicine..
Paul Junker, Program for Clinical Effectiveness and Quality Improvement, University of Pennsylvania Health System..
Scott M. Appel, Center for Clinical Epidemiology and Biostatistics, University of Pennsylvania Health System.
Nina R. O’Connor, University of Pennsylvania Perelman School of Medicine.
Salimah H. Meghani, NewCourtland Center for Transitions and Health, University of Pennsylvania School of Nursing..
References
- 1.Clark M, Person S, Gosline A, Gawande A, Block S. Racial and Ethnic Differences in Advance Care Planning: Results of a Statewide Population-Based Survey. J Palliat Med 2018;0. [DOI] [PubMed] [Google Scholar]
- 2.Starr LT, Ulrich CM, Corey KL, Meghani SH. Associations Among End-of-Life Discussions, Health-Care Utilization, and Costs in Persons With Advanced Cancer: A Systematic Review. Am J Hosp Palliat Care 2019;36:913–926. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Institute of Medicine. Dying in America: improving quality and honoring individual preferences near the end of life. Mil Med 2015;180:365–7. [DOI] [PubMed] [Google Scholar]
- 4.Wachterman MW, Pilver C, Smith D, et al. Quality of End-of-Life Care Provided to Patients With Different Serious Illnesses. JAMA Intern Med 2016;176:1095–102. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Gieniusz M, Nunes R, Saha V, et al. Earlier Goals of Care Discussions in Hospitalized Terminally Ill Patients and the Quality of End-of-Life Care: A Retrospective Study. Am J Hosp Palliat Care 2018;35:21–27. [DOI] [PubMed] [Google Scholar]
- 6.Smith-Howell ER, Hickman SE, Meghani SH, Perkins SM, Rawl SM. End-of-Life Decision Making and Communication of Bereaved Family Members of African Americans with Serious Illness. Journal of Palliative Medicine 2016;19:174–182. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Lopez-Acevedo M, Havrilesky LJ, Broadwater G, et al. Timing of end-of-life care discussion with performance on end-of-life quality indicators in ovarian cancer. Gynecol Oncol 2013;130:156–61. [DOI] [PubMed] [Google Scholar]
- 8.Hanchate A, Kronman A, Young-Xu Y, Ash A, Emanuel E. Racial and ethnic differences in end-of-life costs: why do minoriteis cost more than whites? Arch Intern Med 2009;169:493–501. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Wright AA, Keating NL, Ayanian JZ, et al. Family Perspectives on Aggressive Cancer Care Near the End of Life. JAMA 2016;315:284–92. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Zhang B, Wright AA, Huskamp HA, et al. Health care costs in the last week of life: associations with end-of-life conversations. Arch Intern Med 2009;169:480–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Bernacki RE, Block SD, American College of Physicians High Value Care Task F. Communication about serious illness care goals: a review and synthesis of best practices. JAMA Intern Med 2014;174:1994–2003. [DOI] [PubMed] [Google Scholar]
- 12.Temel JS, Greer JA, Muzikansky A, et al. Early palliative care for patients with metastatic non-small-cell lung cancer. N Engl J Med 2010;363:733–42. [DOI] [PubMed] [Google Scholar]
- 13.Casarett D, Karlawish J, Morales K, et al. Improving the use of hospice services in nursing homes: a randomized controlled trial. JAMA 2005;294:211–7. [DOI] [PubMed] [Google Scholar]
- 14.Center for Disease Control. Estimates. March 2017. Available from: http://wonder.cdc.gov/. Accessed February 13, 2020.
- 15.Townsend A, March AL, Kimball J. Can Faith and Hospice Coexist: Is the African American Church the Key to Increased Hospice Utilization for African Americans? J Transcult Nurs 2017;28:32–39. [DOI] [PubMed] [Google Scholar]
- 16.Eneanya ND, Hailpern SM, O’Hare AM, et al. Trends in Receipt of Intensive Procedures at the End of Life Among Patients Treated With Maintenance Dialysis. Am J Kidney Dis 2017;69:60–68. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Gramling R, Norton SA, Ladwig S, et al. Direct observation of prognosis communication in palliative care: a descriptive study. J Pain Symptom Manage 2013;45:202–12. [DOI] [PubMed] [Google Scholar]
- 18.Robbins SG, Hackstadt AJ, Martin S, Shinall MC, Jr., Implications of Palliative Care Consultation Timing among a Cohort of Hospice Decedents. J Palliat Med 2019. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.O’Connor NR, Junker P, Appel SM, et al. Palliative Care Consultation for Goals of Care and Future Acute Care Costs: A Propensity-Matched Study. Am J Hosp Palliat Care 2017:1049909117743475. [DOI] [PubMed] [Google Scholar]
- 20.Public Health Management Corporation. University of Pennsylvania Health System 2016 Community Needs Assessment 2016. Available from: https://www.pennmedicine.org/~/media/documents%20and%20audio/annual%20reports/community/community_health_needs_assessment_uphs_chna_2016_1.ashx?la=en.2018. Accessed February 1, 2020.
- 21.Dehejia R WS. Propensity Score-Matching Methods for Nonexperimental Causal Studies. Review of Economics and Statistics 2002;84:151–161. [Google Scholar]
- 22.Baram D, Daroowalla F, Garcia R, et al. Use of the All Patient Refined-Diagnosis Related Group (APR-DRG) Risk of Mortality Score as a Severity Adjustor in the Medical ICU. Clin Med Circ Respirat Pulm Med 2008;2:19–25. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Averill R, Goldfield N, Hughes J, et al. All Patient Refined Diagnosis Related Groups (APR-DRGs) Methodology Overview. In: Wallingford, CT: 3M Health Solutions, 2003. [Google Scholar]
- 24.McCormick PJ, Lin HM, Deiner SG, Levin MA. Validation of the All Patient Refined Diagnosis Related Group (APR-DRG) Risk of Mortality and Severity of Illness Modifiers as a Measure of Perioperative Risk. J Med Syst 2018;42:81. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Foraker RE, Rose KM, Whitsel EA, et al. Neighborhood socioeconomic status, Medicaid coverage and medical management of myocardial infarction: atherosclerosis risk in communities (ARIC) community surveillance. BMC Public Health 2010;10:632. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Shen JJ, Wan TT, Perlin JB. An exploration of the complex relationship of socioecologic factors in the treatment and outcomes of acute myocardial infarction in disadvantaged populations. Health Serv Res 2001;36:711–32. [PMC free article] [PubMed] [Google Scholar]
- 27.Ayanian JZ, Kohler BA, Abe T, Epstein AM. The relation between health insurance coverage and clinical outcomes among women with breast cancer. N Engl J Med 1993;329:326–31. [DOI] [PubMed] [Google Scholar]
- 28.Rosenbaum PR, Rubin DB. Reducing bias in observational studies using subclassification on the propensity score. Journal of the American Statistical Association 1984;79:516–524. [Google Scholar]
- 29.Shah BR, Laupacis A, Hux JE, Austin PC. Propensity score methods gave similar results to traditional regression modeling in observational studies: a systematic review. J Clin Epidemiol 2005;58:550–9. [DOI] [PubMed] [Google Scholar]
- 30.Weitzen S, Lapane KL, Toledano AY, Hume AL, Mor V. Principles for modeling propensity scores in medical research: a systematic literature review. Pharmacoepidemiol Drug Saf 2004;13:841–53. [DOI] [PubMed] [Google Scholar]
- 31.Ross EG, Shah N, Leeper N. Statin Intensity or Achieved LDL? Practice-based Evidence for the Evaluation of New Cholesterol Treatment Guidelines. PLoS One 2016;11:e0154952. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Baser O Too much ado about propensity score models? Comparing methods of propensity score matching. Value Health 2006;9:377–85. [DOI] [PubMed] [Google Scholar]
- 33.Dehejia R, Wahba S. Causal effects in nonexperimental studies: re-evaluation of the evaluation of training programs. J Am Stat Assoc 1999;94:1043–62. [Google Scholar]
- 34.Khoza S, Barner JC, Richards K. Comparison of Multivariable-Adjusted Logistic Regression with Propensity Score-Matched, Propensity Score-Stratified, and Propensity Score-Adjusted Logistic Regression Models. In: ISPOR 16th Annual International Meeting Research Abstracts, Value in Heatlh, 2011:A92–A93. [Google Scholar]
- 35.Austin PC. An Introduction to Propensity Score Methods for Reducing the Effects of Confounding in Observational Studies. Multivariate Behav Res 2011;46:399–424. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Candy B, Holman A, Leurent B, Davis S, Jones L. Hospice care delivered at home, in nursing homes and in dedicated hospice facilities: A systematic review of quantitative and qualitative evidence. Int J Nurs Stud 2011;48:121–33. [DOI] [PubMed] [Google Scholar]
- 37.Hughes NM, Noyes J, Eckley L, Pritchard T. What do patients and family-caregivers value from hospice care? A systematic mixed studies review. BMC Palliat Care 2019;18:18. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Gelfman LP, Barron Y, Moore S, et al. Predictors of Hospice Enrollment for Patients With Advanced Heart Failure and Effects on Health Care Use. JACC Heart Fail 2018;6:780–789. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Connor SR, Pyenson B, Fitch K, Spence C, Iwasaki K. Comparing hospice and nonhospice patient survival among patients who die within a three-year window. J Pain Symptom Manage 2007;33:238–46. [DOI] [PubMed] [Google Scholar]
- 40.Saito AM, Landrum MB, Neville BA, et al. Hospice care and survival among elderly patients with lung cancer. J Palliat Med 2011;14:929–39. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Teno JM, Clarridge BR, Casey V, et al. Family perspectives on end-of-life care at the last place of care. JAMA 2004;291:88–93. [DOI] [PubMed] [Google Scholar]
- 42.Teno JM, Gozalo PL, Lee IC, et al. Does hospice improve quality of care for persons dying from dementia? J Am Geriatr Soc 2011;59:1531–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Baer WM, Hanson LC. Families’ perception of the added value of hospice in the nursing home. J Am Geriatr Soc 2000;48:879–82. [DOI] [PubMed] [Google Scholar]
- 44.Hughes MC, Vernon E. Closing the Gap in Hospice Utilization for the Minority Medicare Population. Gerontol Geriatr Med 2019;5:2333721419855667. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45.Starr LT, Ulrich CM, Appel SM, et al. Goals-of-Care Consultations Are Associated with Lower Costs and Less Acute Care Use among Propensity-Matched Cohorts of African Americans and Whites with Serious Illness. J Palliat Med 2020. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46.Rizzuto J, Aldridge MD. Racial Disparities in Hospice Outcomes: A Race or Hospice-Level Effect? J Am Geriatr Soc 2018;66:407–413. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47.Wang SY, Hsu SH, Aldridge MD, Cherlin E, Bradley E. Racial Differences in Health Care Transitions and Hospice Use at the End of Life. J Palliat Med 2019;22:619–627. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48.Wang SY, Aldridge MD, Gross CP, et al. Transitions Between Healthcare Settings of Hospice Enrollees at the End of Life. J Am Geriatr Soc 2016;64:314–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49.Shalev A, Phongtankuel V, Reid MC, et al. Home Hospice Caregivers’ Perceived Information Needs. Am J Hosp Palliat Care 2019;36:302–307. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 50.Phongtankuel V, Scherban BA, Reid MC, et al. Why Do Home Hospice Patients Return to the Hospital? A Study of Hospice Provider Perspectives. J Palliat Med 2016;19:51–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 51.Cohen LL. Racial/ethnic disparities in hospice care: a systematic review. J Palliat Med 2008;11:763–8. [DOI] [PubMed] [Google Scholar]
- 52.Orlovic M, Smith K, Mossialos E. Racial and ethnic differences in end-of-life care in the United States: Evidence from the Health and Retirement Study (HRS). SSM Popul Health 2019;7:100331. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 53.Laury ER, MacKenzie-Greenle M, Meghani S. Advance Care Planning Outcomes in African Americans: An Empirical Look at the Trust Variable. J Palliat Med 2019;22:442–451. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 54.Seccareccia D, Wentlandt K, Kevork N, et al. Communication and Quality of Care on Palliative Care Units: A Qualitative Study. J Palliat Med 2015;18:758–64. [DOI] [PubMed] [Google Scholar]
- 55.Mayeda DP, Ward KT. Methods for overcoming barriers in palliative care for ethnic/racial minorities: a systematic review. Palliat Support Care 2019:1–10. [DOI] [PubMed] [Google Scholar]
- 56.Baek S, Park SH, Won E, Park YR, Kim HJ. Propensity score matching: a conceptual review for radiology researchers. Korean J Radiol 2015;16:286–96. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 57.Sudano JJ, Baker DW. Explaining US racial/ethnic disparities in health declines and mortality in late middle age: the roles of socioeconomic status, health behaviors, and health insurance. Soc Sci Med 2006;62:909–22. [DOI] [PubMed] [Google Scholar]
- 58.Baker DW. The meaning and the measure of health literacy. J Gen Intern Med 2006;21:878–83. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 59.Mantwill S, Monestel-Umana S, Schulz PJ. The Relationship between Health Literacy and Health Disparities: A Systematic Review. PLoS ONE 2015;10:e0145455. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 60.National Hospice and Palliative Care Organization. NHPCO Facts and Figures 2018 Edition, Revision 7–2-2019. In: Virginia: NHPCO, 2018. [Google Scholar]
- 61.Pullis B Perceptions of Hospice Care Among African Americans. Journal of Hospice & Palliative Nursing 2011;13:281–287. [Google Scholar]
- 62.Bernacki R, Hutchings M, Vick J, et al. Development of the Serious Illness Care Program: a randomised controlled trial of a palliative care communication intervention. BMJ Open 2015;5:e009032. [DOI] [PMC free article] [PubMed] [Google Scholar]