Use of Home-based Clinical Care and Long-Term Services and Supports Among Homebound Older Adults

Abstract

Objectives:

Describe use of home-based clinical care and home-based long-term services and supports (LTSS) using a nationally representative sample of homebound older Medicare beneficiaries.

Design:

Cross-sectional study

Setting and Participants:

Homebound, community-dwelling fee-for-service Medicare Beneficiaries participating in the 2015 National Health and Aging Trends Study (n=974)

Methods:

Use of home-based clinical care (i.e., home-based medical care, skilled home health services, other home-based care (e.g., podiatry) was identified using Medicare claims. Use of home-based LTSS (i.e., assistive devices, home modification, paid care, ≥40 hours/week of family caregiving, transportation assistance, senior housing, home-delivered meals) was identified via self or proxy report. Latent class analysis was used to characterize patterns of use of home-based clinical care and LTSS.

Results:

Approximately 30% of homebound participants received any home-based clinical care and about 80% received any home-based LTSS. Latent class analysis identified 3 distinct patterns of service use: 1) Class 1 High Clinical with LTSS (8.9%); 2) Class 2 Home Health Only with LTSS (44.5%), and: 3) Class 3 Low Care and Services (46.6% homebound). Class 1 received extensive home-based clinical care, but their use of LTSS did not meaningfully differ from Class 2. Class 3 received little home-based care of any kind.

Conclusions and Implications:

While home-based clinical care and LTSS utilization was common among the homebound, no single group received high levels of all care types. Many who likely need and could benefit from such services do not receive home-based support. Additional work focused on better understanding potential barriers to accessing these services and integrating home-based clinical care services with LTSS is needed.

Brief Summary:

While home-based clinical care and LTSS utilization was common among the homebound, no single group received high levels of all care types. Many who could likely benefit from such services do not receive home-based support.

Introduction

Home-based clinical services (e.g., home-based medical care, skilled home health) are a patient-centered, safe, and cost-effective way to meet the clinical care needs of older adults with complex medical needs.^1–3 Such clinical care typically exists alongside home-based long-term services and supports (LTSS) that help people meet functional needs in the home setting over time. Home-based LTSS include a wide range of services (e.g., home-delivered meals, home care workers) and supports (e.g., assistive devices, home modifications).⁴ While Medicaid is a key funder of home-based LTSS,⁵ home-based LTSS are also funded by non-Medicaid payers (e.g., individuals and families, long-term care insurance, community based organizations) and provided directly by unpaid family caregivers.⁶

Previous research has established that the homebound experience high levels of functional impairment, clinical comorbidity, and socioeconomic disadvantage⁷ and may benefit from care, services, and supports in the home. However, while utilization of discrete components of home-based clinical care and specific types of home-based LTSS have been documented,^1,8,9 the simultaneous use of home-based clinical and long-term care among the older adult population is unknown. Such an approach is necessary to characterize and adequately support the total care needs of the growing number of homebound older adults who wish to remain living at home and support the integration of the clinical and LTSS that help make this possible. In this study, we used a nationally representative sample of older, fee-for-service Medicare beneficiaries to describe the full patterns of home-based clinical care and home-based LTSS use among community-dwelling homebound older adults.

Methods:

Study Population

We drew cross-sectional data from the nationally-representative 2015 cohort of the National Health and Aging Trends Study (NHATS) linked to Medicare Fee-for-Service claims. The annual NHATS survey examines late-life disability and function among Medicare beneficiaries ≥65 years of age.¹⁰

Of the 8,087 living 2015 NHATS participants, we excluded those living nursing homes (n=403) or assisted livings (n=269) to construct a sample of community-dwelling Medicare beneficiaries.^7. To identify a broad range of individuals who may benefit from care, services, and supports in the home,⁸ we then identified homebound older adults based on self or proxy report that they rarely or never left home or left home only with assistance or difficulty (n=1716). Finally, we excluded those who did not have at least 12 months of fee-for-service Medicare claims data prior to the interview (n=742) for a final analytic sample of 974 individuals.

Measures:

Home-based care, services, and supports use:

We defined home-based care, services, and supports broadly in an effort to capture the wide range of both episodic and ongoing care, services, and supports that may support homebound older adults in the home (Figure 1).¹¹ These care, services, and supports are not mutually exclusive and can occur in combination (e.g., an individual may receive personal care as part of a skilled home health episode).

Figure 1:

Receipt of Home-based Care, Services, And Supports Among the Homebound

Following previous work,^1,12 we defined home-based clinical care using Medicare claims from the 12 months prior to the NHATS interview: 1) “home-based medical care” defined as receipt of a visit with a service location of home or domiciliary from physicians, physician assistants, or nurse practitioners as determined by the Healthcare Common Procedure Coding System, 2) “skilled home health care” defined as an home health episode as determined by the Medicare Home Health File, and 3) “other home-based clinical care” defined as other provider (e.g., podiatrist) claims with a service location of home or domiciliary.

We defined LTSS using NHATS survey responses that corresponded to existing LTSS definitions^4–6 and identified seven LTSS variables: 1) assistive device defined as reported use of a walker, wheelchair, or scooter in the last month; 2) home modification defined as report that the respondent added at least one elevator, stair lift, grab bars, shoer seat, or raised toilet seat to their home; 3) paid caregiver defined as paid helper assistance with a functional task, (4) transportation assistance defined as getting to places outside the home in the last month using a van or shuttle service for seniors or people with disability; 5) senior housing defined consistent with the previous literature and based on the NHATS facility questionnaire;¹³ 6) home-delivered meals defined as hot meals in the last month from Meals on Wheels or other food assistance program; and 7) high family caregiving defined as ≥40 hours per week of total caregiving support among all non-paid helpers.¹⁴

Other variables:

All other variables were drawn from NHATS survey responses and were based on self or proxy report unless otherwise noted.

Sociodemographic and contextual variables included gender, age, race and Hispanic ethnicity, income,¹⁵ receipt of Medicaid, marital status, education, number of family caregivers, total family caregiving hours, living alone, social isolation,¹⁶ and living in a metropolitan area.

Clinical and functional variables included number of chronic conditions, dementia status,¹⁷ self-reported health, mean number of activities of daily living (eating, getting out of bed, showering, toileting, dressing, getting around inside, getting around outside) in which difficulty or impairment was reported, mean number of instrumental activities of daily living (laundry, shopping, meal preparation, medication management, bills and banking, addressing money matters, and going to the doctor) in which difficulty or impairment was reported, and depression (≥ 3 on the PHQ-2).

Analysis:

After describing use of home-based clinical care and home-based LTSS, we used latent class analysis to characterize groups of individuals based on home-based care, service, and support use. We started with a one-class model and then added classes comparing models to determine the optimal number of latent classes. We examined model fit based on indices (Log-likelihood, Akaike Information Criterion, and Bayesian Information Criterion), model interpretability, and clinical judgment.¹⁸ (Supplemental Table 1). After selecting the optimal 3-class model (BIC=8148.1), we assigned each participant to a specific class based on their posterior latent class probabilities. Within each class, we identified the proportion of members using each care, service, or support and used Chi-square and Student’s t-test to examine overall and pair-wise differences in characteristics among classes. To adhere to CMS reporting guidelines that prohibit reporting of small values that may compromise participant anonymity, we report these proportions and prevalence of characteristics as low (0–25%), medium (25–50%), and high (50–100%).

Survey weights were used throughout. The study was approved by the Icahn School of Medicine at Mount Sinai IRB.

Results

Homebound Medicare beneficiaries received a variety of home-based clinical care and home-based LTSS (Figure 1). About 30% received any form of home-based clinical care; use of home health services was most common (27.1%). The vast majority receive some type of LTSS (80%); use of assistive devices was most common (nearly 50%) while receipt of home-delivered meals was least common (<5%).

Those who received home-based clinical care were more likely to receive home-based LTSS as compared to those who did not. However, use of the four most common types of home-based LTSS (i.e., assistive devices, home modifications, paid caregiving, high family caregiving) was similar regardless of type of home-based clinical care used. For example, 71.0% of those who received skilled home health care alone (21.2% of the sample) used assistive devices; 71.7% of those who also used other forms of home-based clinical care (9.3% of the sample) used assistive devices.

Table 1 describes service use patterns for each of the three classes. Class 1 (“High Clinical with LTSS”) was the smallest group (n=120, 8.9% of the sample) and was characterized by highest receipt of home-based clinical care. Class 2 (“Home Health Only with LTSS,” n=442, 44.5% of the sample) members received some home health services but few other forms of home-based clinical care. Members of both Class 1 and Class 2 received similar amounts of home-based LTSS_. Finally, Class 3 (“Low Care and Services,” n=412, 46.6% of the sample) members received few home-based care, supports, or services.

Table 1.

Home-Based Care, Services, and Supports Received by Persons by Class^*

	Class 1 (N=120)	Class 2 (N=n=442)	Class 3 (N=412)
	“High Clinical with LTSS”	“Home Health Only with LTSS”	“Low Care or Services”
Home-Based Clinical Care
Skilled Home Health	High†	Medium †	Low
Other Home-Based Care	High†	Low†	Low
Home-Based Medical Care	Medium†	Low†	Low
Home-Based Long Term Services and Supports
Assisted Device	High	High	Low
Home Modification	Medium†	High†	Medium
Paid Caregiver	Medium	Medium	Low
High Family Caregiving‡	Medium	Medium	Low
Transportation Services	Low	Low	Low
Senior Housing	Low	Low	Low
Home-Delivered Meals	Low	Low	Low

^*Proportion of individuals using care, service, or support: Low= 0–25%, Medium= 25–50%, High= 50–100%

^†p<0.05 comparison of group 1 and 2

^‡defined as ≥40 hours of support from non-paid helpers

We then examined differences in characteristics among members of each of the classes identified (Table 2). Consistent with their pattern of home-based care, services, and supports use, members of Class 3 “Low Care or Services” were younger, had fewer chronic conditions, and were less functionally impaired than those in either Class 1 or Class 2. As compared to the “Home Health Only with LTSS” class, members of the “High Clinical with LTSS” class were older (mean age 84.0 vs 78.8, p<0.01), more likely to have dementia (37.0% vs 25.4%, p=0.03), and more likely to live alone (47.2% vs 26.4%, p<0.01). However, there were no other differences in sociodemographic, contextual, clinical, or functional, characteristics between the two classes.

Table 2.

Sample Characteristics by Class^*

		Class 1 (N=120)	Class 2 (N=442)	Class 3 (N=412)
	Overall (N=974)	“High Clinical with LTSS”	“Home Health Only with LTSS”	“Low Care or Services”	p-value comparing all 3 groups
Sociodemographic and Contextual Variables
Female, %	65.4%	High	High	High	0.86
Age, mean (SD)	78.1 (9.6)	84.0 (11.9) †	78.8 (9.7) †	76.3 (8.5)	<0.01
Race, %					0.14
White Non-His	74.3%	High	High	High
Black Non-His	11.2%	Low	Low	Low
Hispanic/Other	14.5%	Low	Low	Low
Income, mean (SD)	43,869 (676,139)	36,801 (52,628)	46,139 (72,965)	43,053 (62,389)	0.42
Medicaid, %	19.8%	Low	Low	Low	<0.01
Married, %	42.5%	Medium	Medium	Medium	0.09
High school, %	72.9%	High	High	High	0.32
Family caregivers (mean, SD)	2.1 (1.4)	2.3 (1.6)	2.3 (1.5)	1.8 (1.2)	<0.01
Family care hours (mean, SD)	32.1 (57.3)	49.1 (98.0)	45.1 (66.0)	16.5 (31.1)	<0.01
Lives alone, %	30.7%	Medium†	Medium†	Medium	<0.01
Socially isolated, %	39.7%	Medium	Medium	Medium	<0.01
Metropolitan Area, %	74.8%	High	High	High	<0.01
Clinical and Functional Variables
Chronic conditions, mean (SD)	4.28 (2.24)	4.9 (2.8)	4.5 (2.4)	4.0 (1.9)	<0.01
Dementia, %	21.0%	Medium†	Medium†	Low	<0.01
Self-reported health fair/poor, %	52.6%	Medium	High	Medium	<0.01
ADLs impaired, mean (SD)	1.4 (2.1)	2.6 (2.8)	2.2 (2.4)	0.5 (1.1)	<0.01
IADLs impaired, mean (SD)	1.7 (2.0)	2.4 (2.4)	2.40 (2.11)	0.80 (1.41)	<0.01
Depression, %	28.5%	Medium	Medium	Low	<0.18

^*Proportions of individuals with characteristic: Low= 0–25%, Medium= 25–50%, High= 50–100%

^†p<0.05 comparison of group 1 and 2

Discussion

Use of home-based LTSS was moderate among those receiving home-based clinical care regardless of type of home-based clinical care received, and a sizable subgroup of homebound older adults used minimal home-based supports of any kind. While latent class analysis differentiated patterns of home-based clinical care use, no single group appeared to receive all types of home-based clinical care and LTSS. These results highlight the heterogeneity of how home-based care, services, and supports are used and identify the need for more work to understand barriers and facilitators of access to home-based care among the homebound.

While our findings describe receipt of rather than need for home-based care, services, and supports, we see that many types of home-based services were used at low frequency. This confirms previous work indicating that forms of home-based clinical care are underutilized¹⁹ and suggests that several types of home-based LTSS (e.g., home-delivered meals) may be underutilized as well. Indeed, a post-hoc analysis of all homebound NHATS participants found receipt of multiple forms of home-based LTSS was rare and most individuals received either 1 or 2 services or supports (Supplemental Figure 1). Given that this is a homebound population, such findings likely reflect not a lack of need, but rather a lack acceptability, availability, and/or affordability of home-based LTSS.^20–22 Affordability of LTSS is a particular concern, especially given supports and services are covered by a patchwork of payers with substantial geographic variation coverage.²³ Future work should examine how factors including care preferences and access to care shape receipt of services and supports in the home.²⁴

A strength of our analysis is that the range of LTSS described are agnostic to payer or sources. Not only are home-based LTSS funded by a wide variety of sources, individual homebound older adults may simultaneously draw support from multiple payment sources (e.g., Medicaid-funded paid care, Medicare-funded wheelchair, home-delivered meals sponsored by a community-based organization) to remain living at home. Models of care that integrate home-based clinical services and home-based LTSS have the potential to improve outcomes without increasing costs,²⁵ yet evidence to guide the integration is limited and frequently focused on specific populations (e.g., those with both Medicare and Medicaid).^26,27 Future work should consider the wide range of ways people access home-based LTSS and prompt CMS to expand efforts to improve clinical and long-term care integration for the larger population.

This study has several limitations. First, our analysis was limited to those receiving fee-for-service Medicare and our results may not be generalizable to those not covered by Medicare or those covered by Medicare Advantage. In particular, those enrolled in Medicare Advantage plans that cover LTSS (e.g., home meal delivery) may have difference patterns of care, services, and supports in the home. Second, our measures of LTSS were limited to existing NHATS survey items and additional forms of LTSS (e.g., case management) were not included. Third, the size of the NHATS sample limited our ability to evaluate regional variation in LTSS, which is particularly relevant for Medicaid-funded supports and services. Finally, CMS reporting guidelines prohibit reporting of small values that may compromise participant anonymity. As a result, more nuanced comparisons between groups (e.g., percentages in each group receiving each type of home-based LTSS, more nuanced LTSS definitions) were not possible.

Conclusion and Implications

This study provides an innovative, person-centered examination of the array of home-based care, services and supports that enable homebound older adults to remain at home. Additional work focused on robustly characterizing and assessing the adequacy of home-based LTSS within the context of other forms of home-based care will help support efforts to improve access to the care, services, and supports that keep the homebound safe at home.

Supplementary Material

1

Supplemental Figure 1: Patterns of Overlapping Use of 7 Home-Based Long-Term Services and Supports Among Homebound Older Adults

Supplemental Table 1: Model Fit Indices for Models with 2, 3 and 4 Groups