Living in the Community With Dementia: Who Receives Paid Care?

Abstract

OBJECTIVES:

Paid caregivers (e.g. home health aides, personal care attendants) provide hands-on care that helps individuals with dementia live in the community. This study 1) characterizes paid caregiving among community-dwelling individuals with dementia, and 2) identifies factors associated with receipt of paid care.

DESIGN:

Cross-sectional analysis

SETTING:

2015 National Health and Aging Trends Study (NHATS), a nationally representative study of Medicare recipients age 65 and older

PARTICIPANTS:

Community-dwelling individuals with dementia (n=899)

MEASUREMENTS:

Paid and family caregiving support was determined by participant or proxy report of help received with functional tasks. Multivariable logistic regression was used to examine factors associated with receipt of paid care. NHATS population sampling weights were used to produce national paid caregiving prevalence estimates.

RESULTS:

Only 25.5% of community-dwelling individuals with dementia received paid care and 10.8% received ≥ 20 hours of paid care per week. For those who received it, paid care accounted for about half of the 83 total caregiving hours (paid and family) that they received each week. Among the subgroup of individuals with advanced dementia (those with impairment in dressing, bathing, toileting, and managing medications and finances), nearly half (48.3%) received paid care. Multivariable analysis adjusting for sociodemographic, family caregiving support, functional, and clinical characteristics found that the odds of receiving paid care were higher among men (OR 1.91, 1.24–2.95), the unmarried (OR 2.20, 1.31–3.70), those with Medicaid (OR 2.16, 1.27–3.66), and those requiring more help with ADLs (OR 1.32, 1.18–1.48) and IADLs (1.29, 1.14–1.46.)

CONCLUSIONS:

New ways of making paid caregiving more accessible throughout the income spectrum are required to support family caregivers and respect the preferences of individuals with dementia to remain living in the community.

INTRODUCTION

The nearly 6 million individuals living with dementia in the United States rely on an estimated 16 million family caregivers (e.g., spouses, children, friends), totaling over 18 billion hours of care per year with an estimated cost of over $221 billion.¹ This current reliance on family caregivers is unsustainable; by 2050 the number of individuals living with dementia is expected to triple to 14 million, yet the pool of younger people to provide care is shrinking.^1,2 Newer care models that shift the locus of long-term care from institutions to the home and community in order to provide dementia care better aligned with personal preference^3–5 may further exacerbate demand on family caregivers.

While paid caregivers including home health aides, personal care attendants, and other direct care workers may be able to provide additional care (e.g. bathing, shopping, companionship) that keeps those with dementia safe in the community,⁶ most existing studies of paid caregivers focuses on workforce concerns including job satisfaction and retention^7,8 and fail to consider the impact of paid caregivers on those they care for. Receipt of care from the estimated 2 million paid caregivers providing home care⁹ is likely driven not only by functional need but also by socioeconomic characteristics and the caregiving context (e.g., living arrangements, number and geographic proximity of children).^10,11 Income and Medicaid status likely play important roles in the receipt of paid care; in the United States, paid caregiving in the community is predominantly paid for by individuals and families themselves or Medicaid, whose eligibility and covered services vary significantly by state.^12,13

In this study, we used data from the National Health and Aging Trends Study, a nationally-representative cohort of community-dwelling older adults to: 1) characterize paid caregiving among community-dwelling individuals with dementia and 2) identify factors associated with receipt of paid care. Such information is essential to guide evidence-based policies to support long-term care of individuals with dementia living in the community.

METHODS

Study Population

Our data were drawn from the nationally representative 2015 National Health and Aging Trends Study (NHATS) (n=8,334), a longitudinal study of late-life disability and function among Medicare beneficiaries age 65 and older that began in 2011.¹⁴ NHATS conducts annual in-person interviews and cognitive and physical performance tests with participants or proxies. Among those 2015 NHATS participants who lived in the community (n=7,499), we focused on those with dementia (n=899).

Measures

Dementia Status and Severity

Dementia status was determined using criteria for probable dementia established by NHATS¹⁵ which incorporated self-report of dementia, proxy responses to the Alzheimer’s Disease (AD)-8 screening tool, and a cognitive interview that assessed memory, orientation, and function. Consistent with a Functional Assessment Staging Tool (FAST)¹⁶ score of 6–7, we categorized individuals as having advanced dementia who had impairment (as defined below) with dressing, bathing, toileting, and managing medications and finances.

Receipt of Care

If NHATS participants or proxies note that help is needed to perform a particular functional task (eating, getting out of bed, showering, toileting, dressing, laundry, shopping, meal preparation, medication management, getting around outside, bills and banking, addressing money matters, and going to the doctor) in the last month, the helper is identified. The following information about the helper is subsequently obtained: 1) relationship to study participant, 2) hours of help provided, 3) whether care was paid or unpaid, 4) if paid, sources of payment (self or family, a government program e.g., Medicaid or Medicare, insurance, other). We calculated total weekly hours of care provided from paid and family caregivers. We determined the proportion of participants receiving any paid care and those receiving high-intensity paid support, previously defined as 20 or more hours of paid caregiving per week.¹⁷

Additional measures

We assessed additional sociodemographic, caregiving, functional, and clinical variables (Table 1). Imputed values for missing income values are provided by NHATS.¹⁸ Given Medicaid was missing for >10%, we performed multiple imputation for Medicaid status; predictors of Medicaid status were income, education, race, owning one’s own home, participation in Meals on Wheels, and receipt of food stamps or heating/utility assistance. Chronic conditions were assessed by self-report. In order to capture a comprehensive picture of functional abilities, we defined impairment in a given activity of daily living (ADL) or instrumental activity of daily living (IADL) based on report that: (1) activity performed with some or a lot of difficulty, (2) activity never or rarely done alone, (3) activity completed with the help of others, (4) activity not performed because of lack of available assistance, or (5) adverse consequence experienced because no one available to help with the activity.

Table 1:

Characteristics of Those Who Do and Do Not Receive Paid Care Among Individuals with Dementia (n=899) and Advanced Dementia¹ (n=177)

	Dementia			Advanced Dementia2
	Full Sample	Paid Caregiver	No Paid Caregiver	Full Sample	Paid Caregiver	No Paid Caregiver
Sample N	899	240	659	177	92	85
National Estimate	3,361,954	858,947	2,503,007	585,846	282,688	303,158
Any paid care, %	25.5			48.3
≥20 hours a week paid caregiving, %	10.8	42.3		29.1	56.2
Sociodemographic Characteristics
Age, mean	81.8	82.8	81.6	83.1	84.2	81.9
Male, %	44.4	45.9	43.8	33.1	31.4	34.8
White, Non-Hispanic, %	68.5	65.2	69.6	57.2	65.2	48.6
≥ High school, %	60.5	62.9	59.6	59.2	68.9	48.6
Income3a, 3b
Quartile 1 (low)	25.0	28.3	23.8	24.5	22.4	26.7
Quartile 2	25.0	20.2	26.6	24.9	21.4	28.7
Quartile 3	25.0	24.3	25.2	25.4	29.9	20.5
Quartile 4 (high)	25.0	27.2	24.3	25.2	26.3	24.0
Income, median	21,809	21,591	21,392	21,629	22,784	19,847
Medicaid, %	28.3	39.4**	24.4**	39.6	39.6	39.5
Caregiving Context
Unmarried, %	60.9	69.1**	58.2**	61.2	66.2	56.0
Number of living children, mean	3.0	2.8	3.1	3.3	2.7**	3.9**
Lives alone, %	31.4	38.2	29.1	17.0	22.4	11.2
Assisted living, %	19.6	20.0	19.4	10.6	11.3	9.8
Region
Northeast, %	20.2	25.8	18.2	16.7	22.2	10.8
Midwest, %	16.1	13.5	16.9	14.3	15.8	12.6
South, %	38.7	38.5	38.8	48.8	46.8	50.9
West, %	25.1	22.2	26.0	20.2	15.2**	25.7**
Metro area, %	84.0	86.7	83.1	85.1	86.1	84.0
Functional and Clinical Characteristics
ADLs impaired,4 mean	2.3	3.7**	1.9**	5.5	5.5	5.4
IADLs impaired,5 mean	2.5	3.5**	2.2**	4.7	4.6	4.8
Count medical conditions, mean	4.5	4.8	4.4	5.2	5.0	5.5
Hospital stay in the last year, %	34.2	41.9*	31.5*	50.3	46.6	54.3
Died in the next 12 months, %	11.3	15.3	10.0	24.8	26.0	23.5

¹Tests for difference are between those who received paid care and those who did not within each group.

^*p<0.05,

^**p<0.01

²Advanced dementia was defined as those requiring assistance with dressing, bathing, toileting, and managing medications and finances

^3aIncome quartiles for dementia: quartile 1 <$12,248; quartile 2 $12,248-$21,809; quartile 3 $21,809-$42,704; quartile 4 >$42,704

^3bQuartiles for advanced dementia: quartile 1 <$12,058; quartile 2 $12,058-$21,629; quartile 3 $21,629–37,467; quartile 4 >$37,467

⁴Six ADLs evaluated: eating, getting out of bed, showering, toileting, dressing, getting around inside

⁵Five IADLs evaluated: laundry, shopping, meal preparation, medication management, bills and banking

Analysis

In order to produce national prevalence estimates of individuals with dementia and receipt of paid caregiving, we applied NHATS population sampling weights to adjust for differential selection probabilities and nonresponse based on participant characteristics.¹⁹ We used Chi-square and Student’s t-test to examine differences between who did and did not receive paid care. We then tested predictors of paid caregiving using a multivariable logistic regression model that included both variables significantly associated with receipt of paid care in univariate analysis and variables hypothesized to impact receipt of paid care: age, sex, race and ethnicity, marital status, education, Medicaid, income quartile, ADL impairment, IADL impairment, comorbidities, and hospitalization in the last year. All analyses were completed using Stata 15 (College Station, TX).

RESULTS

Of the 7,499 community-dwelling participants in NHATS 899 (7.9% of the weighted sample) had dementia, representing nearly 3.4 million individuals with dementia living in the community. Only 25.5% received paid care and 10.8% received ≥ 20 hours of paid care per week. Over 2.5 million did not receive any paid care. Even among the 17.4% of those with dementia who had advanced dementia, only 48.3% received paid care and 29.1% received ≥20 hours of paid care per week (Table 1.)

Those with dementia who received paid care received about twice as many total hours of care per week (paid and family) compared to those who did not receive paid care (83.1 vs. 41.1, p<0.01). Among those with advanced dementia, there was no statistically significant difference in total care hours between those who received paid care and those who did not (117.9 vs 106.3, p=0.4). Regardless of dementia severity, paid caregiving when present accounted about 40% of total care hours (Figure 1).

Figure 1: Hours of Caregiving Support Among Individuals with Dementia and the Subgroup with Advanced Dementia¹.

Orange: Paid Caregiving

Blue: Unpaid Caregiving

¹Advanced dementia was defined as those requiring assistance with dressing, bathing, toileting, and managing medications and finances.

Those with dementia who received paid care were sicker and more functionally impaired than those who did not receive paid care. They had higher mean levels of ADL and IADL impairment (3.7 vs. 1.9, p<0.01 and 3.5 vs 2.2, p<0.01 respectively) and were more likely to have been hospitalized in the last year (41.9% vs 31.5%, p=0.03). They were more likely to be unmarried (69.1% vs 58.2%, p<0.01). Those who received paid care were significantly more likely to have Medicaid (39.4% vs 24.4%, p<0.01). Among individuals with advanced dementia who, by definition, had profound functional impairment we also found that individuals who received paid care were likely to have fewer children (Table 1).

Paid caregiving was most often paid for by individuals and families (45.2%) and government payers (i.e. Medicaid or Medicare) (30.7%) and less frequently by insurance (15.5%). Payment source varied with income. Among those in the lowest income quartile, government payers accounted for almost half (42.9%) of paid caregiving while among those in the highest income quartile, self-pay covered the majority (50.9%) of care. Though not statistically significant, receipt of any paid care was highest among those in the lowest and highest income quartiles (27.7% and 28.9% respectively) as compared to the second and third quartiles (20.1% and 24.9% respectively) (Figure 2).

Figure 2: Amount of Paid Caregiving and Source of Payment for Paid Caregiving by Income Quartile.

Gray: No Paid Caregiving

Orange: Self or Family Funding

Yellow: Multiple Funding Source

Light Blue: Other or Unknown Funding

Green: Insurance Funding

Dark Blue: Government Funding¹

¹Includes those who reported funding from Medicare, Medicaid, State Program, other government payer, and unknown government payer.

In an adjusted model we found that the odds of receiving paid care were higher among men (OR 1.91, 1.24–2.95), the unmarried (OR 2.20, 1.31–3.70), those with Medicaid (OR 2.16, 1.27–3.66), and those requiring more help with ADLs (OR 1.32, 1.18–1.48) and IADLs (1.29, 1.14–1.46) (Supplementary Table S1).

DISCUSSION

Using a national survey of Medicare recipients, we found that only one in four of community-dwelling individuals with dementia received paid care. Among those with advanced dementia, only half received paid care. Even when paid caregiving was present, family caregivers continued to provide a substantial number of care hours (mean= 52 hours weekly.) Current care arrangements may be inadequate for many patients living in the community with dementia; adverse consequences of unmet care needs among those with dementia living in the community are very common,^20–22 suggesting that those with dementia may resist needed care and/or that family caregivers are not always available, able, or willing to provide all necessary assistance.

It is important to note that even when paid care was present paid caregivers provided less than half of total care hours, confirming the important collaboration between paid and family caregivers.²⁷ Future work should examine the range of paid and family caregiving arrangements and how they change over time to meet the basic needs of individuals with dementia.^28–30

Our findings highlight the fact that receipt of paid care is associated not only with functional need but also financial resources and Medicaid. Lower rates of paid caregiving among middle income individuals suggest that the middle class who neither qualify for Medicaid-funded home care nor have the means to pay for significant paid caregiving out-of-pocket face a unique challenge: they must either rely solely upon family caregivers to meet their needs or pay out-of-pocket for paid caregivers until their wealth is exhausted and they become Medicaid eligible. Additional factors that we were unable to assess in the current study (e.g., fear of loss of independence or introducing unfamiliar caregivers, expectation that care come from family alone)^23,24 may also impact receipt of paid care. Future work should explicitly examine these factors.

To date, there is no clear consensus or healthcare policy that addresses the long-term care needs of those with functional and cognitive impairment.^25,26 Expansion of programs that provide direct payment to family members for caregiving roles is needed to further support family caregivers as they provide the lion’s share of care in the community. Yet new ways of making paid caregiving more accessible are required if individual’s preferences to remain living in the community are to be met. In particular, expansion of formal policies that improve access to paid caregiving throughout the income spectrum through Medicaid expansion, increased Medicaid spending on home-and-community-based services, or tax credits or tax deductions for the money spent on paid caregiving is needed. Simultaneously, partnering with organizations committed to advancing the paid caregiver workforce through improved paid caregiver training will help ensure paid caregivers provide high-quality care.

Our study had several limitations. NHATS only assessed paid and family caregivers who helped with a discrete list of tasks and may have misclassified caregivers who provide support in other meaningful ways. Caregiver involvement relies on participant or proxy report and may be subject to recall bias or incomplete reporting. In addition, NHATS asks participants and proxies about help with tasks over the last month, which may include reports of help required only in the short-term. Finally, receipt of paid care among individuals with dementia may be associated with factors that were not assessed including caregiver characteristics, behavioral symptoms associated with dementia, and availability and use of community supports.

CONCLUSION

The growing number of initiatives to care for individuals with dementia outside of the nursing home setting will translate into more hands-on care delivered in the community. Yet despite a shrinking pool of family caregivers, a minority of individuals with dementia living in the community receive paid care. This study underscores the ways in which financial and social factors impact how needed care – both paid and family – is provided to older adults living with dementia. Policies to improve access to paid caregiving can support family caregivers and make it possible for individuals with dementia to live safely in the community.

Supplementary Material

Supp TableS1

Supplementary Table S1: Factors Associated with Receipt of Paid Care Among Those with Dementia (n=899)