Abstract
As more people live and die in the community despite complex health needs and functional impairment, the need for hospice increases. We found high and increasing penetration of hospice in community-based residential settings, compared with hospice use in private residences and nursing homes.
Where people live as they age has implications for health care delivery, costs, and policy. Consistent with care preferences and evidence of cost savings, both long-term care and end-of-life care are shifting from institutional care facilities to the home setting.1,2 There are an estimated 28,900 residential care communities in the US, including assisted living facilities, retirement communities, personal care homes, and senior citizen housing.3 As more people live and die in the community despite complex health needs and functional impairment, the need for hospice increases. Using nationally representative data, we found that 9.8 percent of Medicare beneficiaries spend the end of their life in community-based residential settings (data not shown) and that these people had significantly higher rates of hospice use over the study period, compared with those in private residences and nursing homes (51.2 percent versus 37.3 percent and 43.9 percent, respectively; p < 0.001) (exhibit 1).
Exhibit 1. Proportion of Medicare decedents who used hospice, by residential setting, 2002–17.
SOURCE Authors’ analysis of data for Medicare beneficiaries who died during 2002–17 (excluding data for 2014) from the Medicare Current Beneficiary Survey. NOTES N = 8,394. Community-based residential settings include assisted living facilities, continuing care communities, retirement communities, and senior citizen housing. The error bars indicate 95% confidence intervals.
Because older adults consistently prefer to live and receive care at home, the use of various alternatives to institutional long-term care has the potential to improve the quality of life and promote function and independence for people as they age. State and federal policies have encouraged the use of alternatives to nursing homes, including adult foster care, board and care facilities, and assisted living facilities. Efforts by state Medicaid programs to allocate spending for long-term services and support in home and community-based settings rather than for nursing home care has been a key driver of the growth in community-based residential care models.4,5 However, evaluations of community-based residences have often focused on specific settings (for example, assisted living facilities only)6–10 or on individual states, and none has focused on the population at the end of life. Thus, the extent to which hospice supports patients and families at the end of life across diverse types of community-based residential care models is unknown.
Study Data And Methods
We conducted a pooled cross-sectional study using data from the Medicare Current Beneficiary Survey (MCBS) for the period 2002–17. These data exclude survey results from 2014, which were not released by the Centers for Medicare and Medicaid Services. The survey sample is representative of the Medicare population in total and within age groups. A unique aspect of the MCBS is its residential timeline, which provides a continuous measure of participants’ residence throughout their four-year participation in the study.
The MCBS is conducted in person. Measures are self-reported for respondents in the community and proxy reported for respondents in facilities. For our measure of dementia, we used an inclusive case definition developed for use with the survey data.11 All measured variables are as of a person’s last MCBS interview date before death or the after-death proxy interview.
We defined residential setting at the end of life as where a person (or their proxy) reported residing at the time of death—a private residence, community-based residential setting, or nursing home. Consistent with prior work that used the MCBS,4 we categorized respondents as living in community-based residences if the place in which they lived provided long-term supportive services in a home-like setting. They included assisted living facilities, retirement communities, continuing care facilities, senior citizen housing, rest homes, adult/group homes, and personal care homes. Respondents in a nursing home, rehabilitation facility, hospital, mental health facility, or inpatient hospice were categorized as living in a nursing home. All other respondents were classified as living in a private residence, a category that included one-family dwellings, duplexes, apartments or condominiums, and mobile homes. Our focus on residential setting at the end of life is distinct from the body of work that examined the site of death.2,12–14 Our goal was to examine data on older adults who were living in a diverse and emerging group of community settings to better understand the services and support that these adults received as they aged.
To estimate associations between demographic, clinical, and socioeconomic characteristics by residential setting (that is, private residence, community-based residential setting, or nursing home), we conducted multinomial logistic regression. We used multivariable logistic regression to estimate associations between residential setting and use of hospice, controlling for age, sex, race/ethnicity, educational attainment, Medicaid eligibility, serious illness (cancer, heart disease, dementia, stroke, lung disease, or diabetes), impairment in activities of daily living (ADLs), region, and survey year. All analyses incorporated MCBS survey weighting and design variables.
Study Results
Our sample consisted of 8,394 Medicare decedents with a mean age of 80.1 years at death, who represented 28.4 million Medicare decedents. Of the decedents in our sample, 52.9 percent were female, 79.6 percent were non-Hispanic white, and 12.5 percent had a college education. Decedents had a mean 1.9 of the six measured serious illnesses. Specifically, 37.0 percent had cancer, 39.1 percent had heart disease, and 41.3 percent had dementia. More than half of the sample (62.4 percent) received help with three or more ADLs.
Characteristics Of Beneficiaries By Residential Setting
At the end of life, 63.6 percent of the sample resided in private residences, 9.8 percent in community-based residential settings, and 26.6 percent in nursing homes. Those in community-based residential settings reported living in a wide range of housing types, including assisted living facilities (30.4 percent), retirement communities (23.2 percent), continuing care facilities (16.4 percent), and senior citizen housing (12.3 percent). The proportion of people in community-based residential settings increased from 7.7 percent in 2002 to 11.0 percent in 2017 (test for trend by year: p = 0.127).
Demographic and socioeconomic characteristics differed across residential settings. Compared to Medicare beneficiaries in private residences, those in community-based residential settings at the end of life were more likely to be female, non-Hispanic white, not married, and college educated, and they had similar rates of Medicaid enrollment (exhibit 2). Clinical and functional characteristics also differed across residential settings. Compared to beneficiaries in private residences, those in community-based residential settings at the end of life were more likely to have dementia, less likely to have cancer, and more likely to receive help with three or more ADLs (exhibit 3). In analyses that adjusted for all measured characteristics, we found that the greatest differential effects across residential settings were by race/ethnicity, marital status, Medicaid coverage, dementia status, and receiving help with three or more ADLs (data not shown).
Exhibit 2. Demographic and socioeconomic characteristics of Medicare decedents, by residential setting, 2002–17.
SOURCE Authors’ analysis of data for Medicare beneficiaries who died in 2002–17 (excluding data for 2014) from the Medicare Current Beneficiary Survey. NOTES The sample size is in the notes to exhibit 1. The error bars indicate 95% confidence intervals. Community-based residential settings are defined in the notes to exhibit 1. “Medicaid” means enrolled in Medicaid.
Exhibit 3. Clinical and functional characteristics of Medicare decedents, by residential setting, 2002–17.
SOURCE Authors’ analysis of data for Medicare beneficiaries who died in 2002–17 (excluding data for 2014) from the Medicare Current Beneficiary Survey. NOTES The sample size is in the notes to exhibit 1. The error bars indicate 95% confidence intervals. Community-based residential settings are defined in the notes to exhibit 1. ADL is activity of daily living.
Hospice Use By Residential Setting And Year
Hospice use for Medicare beneficiaries in community-based residential settings at the end of life was significantly (p < 0.001) higher (51.2 percent) than for those in private residences (37.3 percent) and those in nursing homes (43.9 percent) (exhibit 1). There was variation in hospice use by type of community-based residential setting, with the highest hospice use in assisted living facilities (61.1 percent) and the lowest use in senior citizen housing (32.7 percent).
Longitudinal trends in the proportions of Medicare beneficiaries by residential setting and hospice use indicated significantly increasing rates of hospice use over time for each residential setting type (p < 0.05 for the associations between year and hospice use in both unadjusted and multivariable models for each of the three residential settings), with the greatest increases for Medicare beneficiaries in private residences (exhibit 4). In models that accounted for demographic, clinical, and socioeconomic differences among Medicare beneficiaries across settings, people in community-based residential settings had significantly higher rates of hospice use compared to those in private residences (45.9 percent versus 37.0 percent) (exhibit 5).
Exhibit 4. Trends in proportions of Medicare decedents, by residential setting and hospice use, 2001–17.
SOURCE Authors’ analysis of data for Medicare beneficiaries who died in 2002–17 (excluding data for 2014) from the Medicare Current Beneficiary Survey. NOTES The sample size is in the notes to exhibit 1. Community-based residential settings are defined in the notes to exhibit 1. For the lines showing hospice use, p value of test for trend in hospice use over time for each of the three settings is less than 0.05.
Exhibit 5.
Adjusted and unadjusted proportion of Medicare decedents who used hospice, by residential setting, 2002–17
| Unadjusted proportion | Difference in unadjusted proportion | Adjusted proportion | Difference in adjusted proportion | |||||
|---|---|---|---|---|---|---|---|---|
| Residential settings | % | 95% CI | % | 95% CI | % | 95%CI | % | 95% CI |
| Private residence | 37.3 | (35.0, 39.6) | —a | —a | 37.0 | (34.8, 39.1) | —a | —a |
| Community-based residential setting | 51.2 | (47.3, 55.1) | 13.9 | (9.8, 18.0)**** | 45.9 | (42.1, 49.6) | 8.9 | (4.5, 13.4)**** |
| Nursing home | 43.9 | (40.8, 47.0) | 6.6 | (3.8, 9.4)**** | 40.3 | (37.2, 43.4) | 3.3 | (0.1, 6.5)** |
SOURCE Authors’ analysis of data for Medicare beneficiaries who died in 2002–17 (excluding data for 2014) from the Medicare Current Beneficiary Survey. NOTES Adjusted proportions account for differences from the reference group in demographic, clinical, functional, and socioeconomic characteristics. Community-based residential settings are defined in the notes to exhibit 1. CI is confidence interval.
Reference group.
p < 0:05
p < 0:001
Discussion
In this first national study of decedents categorized by residential setting, we found that those who lived in community-based residential settings at the end of life were generally non-Hispanic white, college educated, and not married. From a clinical perspective, their defining characteristics were having a diagnosis of dementia and receiving help with three or more ADLs. These results provide a more nuanced understanding of the factors associated with where people live at the end of life. Specifically, the desire to age in place may introduce complex decision-making issues, particularly for those who may have mild dementia, increasing functional impairment, and no spouse to provide care. Furthermore, the high proportion of people in community-based residential settings who had serious illness underscores the potential need for palliative care programs to serve this population before they become eligible for hospice.
Our finding of high rates of hospice penetration in community-based residential settings is not surprising, given the prevalence of people with serious illness and functional impairment and the often limited medical and nursing support in these settings. Increasing rates of hospice use in these settings, coupled with the fact that a principle of hospice care is to support people who wish to die at home, are consistent with recent reports of increasing rates of death in a “home” setting and declining rates of hospital death.14 Though evidence through 2012 from the Office of Inspector General of the Department of Health and Human Services highlights concerns regarding the potential overuse of hospice in this setting,15 high rates of hospice enrollment may be appropriate given the care needs of residents. In addition, the built-in supportive services in community-based residential settings may reduce disparities in hospice access for people living alone in the community.16
Older adults are not only living but also dying in community-based residential settings with high and increasing support from hospices. Future work to characterize the specific services that these settings offer will enable a greater understanding of how hospice can be appropriately used to improve end-of-life care in the community.
Acknowledgments
Melissa Aldridge has received funding from the National Institute of Nursing Research (Grant No. R01 NR018462-01) and from the Claude D. Pepper Older Americans Independence Center, Icahn School of Medicine at Mount Sinai.
Contributor Information
Melissa D. Aldridge, Department of Geriatrics and Palliative Medicine at the Icahn School of Medicine at Mount Sinai, in New York City, and the James J. Peters Bronx Veterans Affairs Medical Center, in the Bronx, New York..
Katherine A. Ornstein, Department of Geriatrics and Palliative Medicine and the Institute for Translational Epidemiology, Icahn School of Medicine at Mount Sinai..
Lihua Li, Department of Population Health, Icahn School of Medicine at Mount Sinai..
Notes
- 1.Gillsjö C, Schwartz-Barcott D, von Post I. Home: the place the older adult cannot imagine living without. BMC Geriatr. 2011;11:10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Aldridge MD, Bradley EH. Epidemiology and patterns of care at the end of life: rising complexity, shifts in care patterns and sites of death. Health Aff (Millwood). 2017;36(7): 1175–83. [DOI] [PubMed] [Google Scholar]
- 3.Sengupta M, Rome V, Harris-Kojetin L, Caffrey C. Long-term care providers and services users in the United States—residential care component: National Study of Long-Term Care Providers, 2015–2016. Hyattsville (MD): National Center for Health Statistics; 2017. [PubMed] [Google Scholar]
- 4.Degenholtz HB, Park M, Kang Y, Nadash P. Variations among Medicare beneficiaries living in different settings: demographics, health status, and service use. Res Aging. 2016;38(5):602–16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Ng T, Harrington C, Musumeci M, Reaves EL. Medicaid home and community-based service programs: 2010 data update. San Francisco (CA): Henry J. Kaiser Family Foundation; 2014. [Google Scholar]
- 6.Dougherty M, Harris PS, Teno J, Corcoran AM, Douglas C, Nelson J, et al. Hospice care in assisted living facilities versus at home: results of a multisite cohort study. J Am Geriatr Soc. 2015;63(6):1153–7. [DOI] [PubMed] [Google Scholar]
- 7.Sloane PD, Zimmerman S, Hanson L, Mitchell CM, Riedel-Leo C, Custis-Buie V. End-of-life care in assisted living and related residential care settings: comparison with nursing homes. J Am Geriatr Soc. 2003; 51(11):1587–94. [DOI] [PubMed] [Google Scholar]
- 8.Ball MM, Kemp CL, Hollingsworth C, Perkins MM. “This is our last stop”: negotiating end-of-life transitions in assisted living. J Aging Stud. 2014;30:1–13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Unroe KT, Bernard B, Stump TE, Tu W, Callahan CM. Variation in hospice services by location of care: nursing home versus assisted living facility versus home. J Am Geriatr Soc. 2017;65(7):1490–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Unroe KT, Stump TE, Effler S, Tu W, Callahan CM. Quality of hospice care at home versus in an assisted living facility or nursing home. J Am Geriatr Soc. 2018;66(4):687–92. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Yang Z, Zhang K, Lin PJ, Clevenger C, Atherly A. A longitudinal analysis of the lifetime cost of dementia. Health Serv Res. 2012;47(4): 1660–78. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Aldridge MD, Epstein AJ, Brody AA, Lee EJ, Cherlin E, Bradley EH. The impact of reported hospice preferred practices on hospital utilization at the end of life. Med Care. 2016; 54(7):657–63. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Teno JM, Gozalo P, Trivedi AN, Bunker J, Lima J, Ogarek J, et al. Site of death, place of care, and health care transitions among US Medicare beneficiaries, 2000–2015. JAMA. 2018;320(3):264–71. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Cross SH, Warraich HJ. Changes in the place of death in the United States. N Engl J Med. 2019;381(24): 2369–70. [DOI] [PubMed] [Google Scholar]
- 15.Department of Health and Human Services, Office of Inspector General. Medicare hospices have financial incentives to provide care in assisted living facilities [Internet]. Washington (DC): HHS; 2015. January [cited 2020 Apr 8]. Available from: https://oig.hhs.gov/oei/reports/oei-02-14-00070.pdf [Google Scholar]
- 16.Aldridge Carlson MD, Barry CL, Cherlin EJ, McCorkle R, Bradley EH. Hospices’ enrollment policies may contribute to underuse of hospice care in the United States. Health Aff (Millwood). 2012;31(12):2690–8. [DOI] [PMC free article] [PubMed] [Google Scholar]