Nearly 4 in 10 extended families of older adults in the United States include an older relative with dementia

Abstract

INTRODUCTION

The growing number of older adults with dementia could have implications for their family members, many of whom will be called upon to provide care.

METHODS

Leveraging the familial design of the 2021 Panel Study of Income Dynamics, we estimate dementia prevalence among older adults, their households, immediate families, and extended families.

RESULTS

About 21% of adults ages 65 and older have dementia. About 26% of both households and immediate families with an adult age 65 and older include an individual with dementia. This figure rises to 37% among extended families of older adults. Among those with older adults, less‐educated households and families have higher dementia rates than do more‐educated ones; extended families with racial/ethnic minorities have higher dementia rates than do their non‐Hispanic White counterparts.

DISCUSSION

Nearly four in 10 extended families of older adults include someone with dementia, potentially placing family members at risk of becoming caregivers.

Highlights

• We provide the first national estimates of dementia in extended families.

• About 26% of immediate families with an older adult include someone with dementia.

• Nearly four in 10 extended families of older adults include someone with dementia.

• Findings have implications for targeting care‐related supports to families.

1. BACKGROUND

The aging of the U.S. population will bring with it a rise in the number of adults with dementia, from six million in 2020 to nearly 12 million by 2040. ¹ Alzheimer's disease and related dementias (hereafter dementia) are debilitating conditions that impair cognitive function, ultimately leading to a loss of independence in everyday activities. Family members are most likely to partner with and care for older adults with dementia when they need assistance; indeed, most care to older adults with dementia, by some estimates 80%, is provided by family and friends. ² These family care partners and caregivers (hereafter caregivers) spend more time providing assistance and experience worse emotional health than other types of family caregivers. ³ Caregiving also has implications for long‐term financial well‐being and for physical and mental health. ⁴ Even family members who do not serve as caregivers may experience consequences from having someone with dementia in the family. For instance, if a parent develops dementia and requires long‐term services and supports, providing those services and supports might entail a reduction in resources available to all adult children, even those who are not caregivers. ⁵ Despite its potential impact, there is limited information on the prevalence of dementia among household and family units and no estimates of exposure to someone with dementia within extended family networks.

Identifying household and family members of older adults with dementia and therefore “at risk” of being a caregiver now or in the future is a crucial first step for developing appropriate and targeted support mechanisms for caregivers, a major thematic area highlighted at the 2023 Summit on Care, Services, and Supports for Persons Living with Dementia and Their Care Partners/Caregivers. It is now well established that older adults with care needs receive assistance from multiple caregivers ⁶ ; this is particularly true for care provided to older adults with dementia. ⁷ Even family members who are not currently providing care may be called upon to serve as caregivers in the future or to provide respite care. In addition, even well before family members begin serving as caregivers to a relative, they may already be preparing to provide care ⁸ , ⁹ (e.g., moving closer to parents with dementia, shifting work hours). The first step to developing interventions for addressing caregiver needs is to identify the pool at risk of needing such services.

Studies such as those of the Health and Retirement Study (HRS) and the National Health and Aging Trends Study (NHATS) are well suited for studying family caregiving for dementia. HRS has rich information on couples along with a roster of children that can be used to explore dementia exposure and care within a household and immediate family. Similarly, NHATS interviews an older adult, obtains a roster of adult children, and interviews up to five family caregivers, which can be used to explore dementia exposure within a household and immediate family as well as care networks. These studies, however, do not provide information on potential caregivers exposed to someone with dementia in the extended family. We leverage the unique familial design of the Panel Study of Income Dynamics (PSID) to examine dementia prevalence at the individual, household, immediate family, and extended family levels. Because dementia prevalence is higher among individuals with lower educational attainment and those who are non‐White, ¹⁰ we also explore differences in dementia prevalence at the household and family level among those with and without college degrees and for units with at least one person who is a racial/ethnic minority.

RESEARCH IN CONTEXT

1. Systematic review: The authors reviewed the literature using traditional sources and did not find national estimates of the likelihood of having someone with dementia in one's household, immediate family, or extended family. These individuals are at risk of being a care partner to a family member with dementia and make up the pool of individuals who may benefit from targeted interventions and supports.

2. Interpretation: Findings demonstrate that having a family member with dementia is common. While only 21% of individuals age 65 and older have dementia, almost 37% of extended families of older adults include someone age 65 and older with dementia.

3. Future directions: Further study is needed to understand the extent to which living with or having an immediate or extended family member with dementia increases the risk of being a caregiver and resulting effects on health and well‐being.

2. METHOD

2.1. Data

We used the 2021 wave of PSID, ¹¹ the longest running nationally representative sample of U.S. households, first begun in 1968. These data are uniquely suited for analyzing dementia prevalence across families, as descendants of the original 1968 families become respondents themselves once they move to independent households. Two strengths of PSID are its oversample of Black families (about one‐third) and families below the poverty line (about 15%). We focused on adults aged 65 and older and their respective households, immediate families (i.e., spouse and children), and extended families going back to 1968, as depicted in Figure 1.

FIGURE 1.

Data structure and embeddedness of older adults in households, immediate families, and extended families, Panel Study of Income Dynamics (PSID).

2.2. Analyses

To measure dementia status, we used the validated “Eight‐item Informant Interview to Differentiate Aging and Dementia” (AD8), administered for each adult age 65 and older. ¹² The AD8 asks about changes in eight cognitively relevant domains because of a thinking or memory problem and may be self‐ or proxy‐reported. This screener has high sensitivity and specificity for distinguishing people with clinical dementia and those without dementia, ¹³ including in the African American community. ¹⁴ Individuals with a score of two or more screen in for dementia. We calculated the number of adults with dementia within households, immediate families, and extended family units. We defined households as relatives in the physical dwelling unit where the adult with dementia resides. Immediate families are defined as a current spouse or long‐term cohabiting partner and living children (biological, adoptive, step), who are present in PSID 2021. For extended families, we leverage the study design of PSID “clan” and include all family members, including extended kin, who link back to a common 1968 family (including, for example, siblings, cousins, nieces, nephews, and grandchildren). Our analytic sample consisted of 2603 individuals who completed the AD8 in 2021.

We begin our analyses by describing the sample of individuals ages 65 and older. We then show the percentage of households, immediate families, and extended families with an individual screening in with dementia, overall and among units with at least one adult age 65 and older. Finally, we show predicted percentages with dementia by race‐ethnicity and education of the unit (household, immediate family, extended family). In a sensitivity analysis, we also control for the number of older adults and number of total members identified in each unit (i.e., for household, we adjust for household size; for immediate family, immediate family size, and so forth). We classify a unit according to whether any member of that unit has at least a college education (vs all do not) and is part of a racial/ethnic minority group (vs all are not). Missing data on race‐ethnicity and education are imputed either with information from other vertical kin ties in the family unit or, when that is unavailable, modal categories. Data are weighted using individual cross‐sectional weights and adjusting for survey design and clustering of individuals within family/household units. Statistics are weighted and account for PSID's complex survey design. Additional information about the data, sample, and imputations are available in the online Supplemental Appendix.

3. RESULTS

Adults age 65 and older in this study with dementia differ from those without dementia in several ways: They are older, slightly more likely to be female, more likely to be a member of a racial/ethnic minority group, less likely to have a bachelor's degree, and slightly less likely to be a primary respondent to the PSID interview or spouse/partner (Table 1). They are also more likely to have a proxy reporting on their behalf.

TABLE 1.

Demographic characteristics of households, immediate families, and extended families of U.S. adults age 65 and older with and without dementia, 2021.

	All individuals age 65 and older (N = 2603)	Individuals age 65 and older without dementia (N = 2096)	Individuals age 65 and older with dementia (N = 507)
Demographic characteristics
Age (mean, SD)	73.5 (5.6)	72.8 (5.2)	76.2 (6.5)
Female (%)	55.6	55.4	56.5
Race‐ethnicity a (%)
White, non‐Hispanic	74.9	76.4	69.2
Black, non‐Hispanic	7.3	6.9	8.8
Hispanic	8.7	8.0	11.0
Other, non‐Hispanic	9.1	8.7	11.0
Bachelor's degree or higher a (%)	29.1	31.4	20.7
Survey respondent type (%)
Reference person	64.9	65.4	63.3
Spouse/cohabiting partner	32.3	33.2	29.2
Other family unit member (OFUM) or proxy	2.7	1.4	7.4
Unit size (mean, SD)
Household	2.3 (1.0)	2.3 (1.0)	2.2 (1.1)
Immediate family	3.8 (1.3)	3.9 (1.3)	3.8 (1.3)
Extended family	14.0 (11.4)	14.1 (11.3)	14.0 (11.7)

Notes: Individuals include reference persons, spouses/partners, and other family unit members (OFUMs) present in PSID household at the time of interview. Unit size includes adults age 65 and older. Households include any co‐residing family member. Immediate families include partner (legal spouse or long‐term cohabiting partner) and children (biological, adopted, step).

Source: Authors’ analysis of data from 2021 Panel Study of Income Dynamics (PSID).

Abbreviation: SD, standard deviation

^a Missing data on race‐ethnicity and education are imputed either with information from other vertical kin ties in the family unit or, when that information is unavailable, modal categories. Data are weighted using individual cross‐section weights and adjusting for survey design and clustering of individuals within family units.

Figure 2 shows the percentage of the population age 65 or older with dementia in 2021, and the percentage among households, immediate families, and extended families with at least one adult age 65 or older in the household or family unit. Whereas only 21% of individuals age 65 and older have dementia, about 26% of households and immediate families with someone age 65 and older include at least one person with dementia. This figure rises to almost 37% among extended families with someone age 65 and older.

FIGURE 2.

Percentage with dementia in 2021 among individuals, households, immediate families, and extended families of older adults in the United States. Source: Authors’ analysis of data from 2021 Panel Study of Income Dynamics (PSID). Notes: Data are weighted and adjusted for sample design by unit of analysis (i.e., individuals, households, immediate families, extended families). N = 2603 individuals, 1924 households, 2021 immediate families, 1387 extended families.

Table 2 shows the weighted estimates of older adults with dementia at the individual, household, immediate family, and extended family levels, first overall and then by education and race‐ethnicity of the unit. Across households, almost 7% include an older adult who screens in for dementia. As noted earlier, this rises to 26.3% when limiting analyses to only those households with an older adult age 65 and older. We see fairly similar percentages among immediate families (6.4% overall, 25.6% among those with someone age 65 or older). A much higher percentage of extended families have a relative with dementia: 26.1% overall and 36.6% among those with at least one relative age 65 or older.

TABLE 2.

Percentage of adults with dementia among U.S. households, immediate families, and extended families overall and by race‐ethnicity and education of the unit, 2021.

	All households (N = 8833) %	Households with at least one adult age ≥65 (N = 1924), %	All immediate families (N = 12,872) %	Immediate families with at least one adult age ≥65 (N = 2021), %	All extended families (N = 2359), %	Extended families with at least one adult age 65+ (N = 1387), %
At least one adult age ≥65 with dementia	6.9	26.3	6.4	25.6	26.1	36.6
By education
No one in unit has college degree	8.2	29.7	5.8	29.9	19.9	38.0
At least one person in unit has college degree	5.3*	22.0*	7.1	22.8*	27.9*	36.4
By race‐ethnicity
At least one person in the unit is a racial/ethnic minority	5.8	28.1	5.0	27.2	26.7	39.8
All unit members are White, non‐Hispanic	7.7*	25.2	7.6*	24.8	25.1	31.4 *

Notes: Data are weighted and adjusted for sample design by respective unit of analysis (i.e., households, immediate families, extended families).

Source: Authors’ analysis of data from the 2021 Panel Study of Income Dynamics.

^* p < .05 significant difference from pairwise comparison.

The second and third panels of Table 2 show the same results, but by education (no one with a college degree vs at least one member with a college degree) and race‐ethnicity (at least one racial/ethnic minority member vs none) of the household and family units. Dementia prevalence is higher in household and family units without college‐educated adults than in household and family units with at least one member with a college degree. About 30% of less‐educated households and immediate families that include someone age 65 and older also include someone with dementia, whereas 22% of households and 23% of immediate families in which at least one person has a college degree have an older adult with dementia. Households and immediate families with at least one adult age 65 or older who have at least one racial/ethnic minority member in the unit have a dementia prevalence similar to those with only White non‐Hispanic members; however, extended families with at least one member who is a racial/ethnic minority have a greater percentage with dementia (39.8%) relative to units with all White, non‐Hispanic members (31.4%).

If we control for the number of older adults and number of members in each unit (see Appendix Table S1), findings shift in two key ways. First, educational differences among all immediate families and extended families with at least one adult age 65 or older strengthen and become statistically significant. Second, racial/ethnic differences among all immediate families and extended families with at least one adult age 65 or older weaken and are no longer statistically significant.

4. DISCUSSION

Although the impact of dementia on health, wealth, and well‐being extends beyond individuals, studies typically examine dementia prevalence among individuals or couples. We find that having someone with dementia in the household or in one's immediate or extended family is common. Among households and immediate families with at least one older adult, 26% include an older adult with dementia. Among extended families, the figure is higher – amounting to 37%. Estimates are even higher among household and family units without college graduates. The estimate of 37% among extended families with an older adult is comparable to that of a 2011 Associated Press poll that found that about 31% of respondents reported knowing someone with dementia in their families. ¹⁵

There are many ways having an older relative with dementia can influence broader family units. The most obvious mechanism is through caregiving. While spouses are generally the first to provide care in later life, adult children, often daughters, are more likely to be caregivers for the subset of older adults living with dementia. ⁴ Moreover, a growing body of work suggests that caregiver networks assisting older adults with dementia are more likely than networks assisting other older adults to be larger and include extended family members. ² , ⁷

While some of the extended family members included in our calculations may be less likely than others to serve as caregivers, having a relative with dementia may nonetheless affect them. For example, it could affect decisions about purchasing long‐term care insurance, allocating resources to paid long‐term services and supports, or planning for the future possibility of dementia. Attitudes toward dementia testing and openness to treatment may also be influenced by having a relative with dementia. ¹⁶

This work also has implications for health professionals as having a relative with dementia may be a risk factor for being a caregiver. Asking if people are living with or have a relative outside the household with dementia at primary care visits could be one way to identify individuals at risk of being caregivers and associated health risks. In addition, with new efforts being made to integrate care for patients with support for their caregivers, ¹⁷ it will become increasingly important for medical professionals to get a better sense of the full network of potential caregivers, even those who are not coresiding or immediate kin.

As the only nationally representative dataset with information on dementia status across households, immediate families, and extended families, these data are uniquely suited for this investigation. Still, several limitations bear noting. Because of the household design, few older adults living in institutional settings, such as nursing homes, are captured. In addition, not all types of family members are followed over time, and while the design effectively captures vertical ties (e.g., grandchild, child, parent, grandparent), other family relationships may be less accurately represented. This may result in underestimates of the size of family networks. Our focus on proportions instead of counts avoids some potential underreporting bias, but we may still underrepresent some families’ exposure to dementia more than others (e.g., immigrant sample). There are several other limitations related to the design of the PSID. Race‐ethnicity is only available for reference persons and spouses/partners and needed to be imputed for other household members. Because of this we used a simplified dichotomized measure of race. In addition, due to PSID following rules, only one member of married couples will have upward generational ties in the data. Since this feature is random, we do not expect it to create bias in the estimates, but it would lead to an underestimate of family size. In addition, geographic proximity of extended family members could be related to the impact of dementia exposure, but that is beyond the scope of this study.

This work represents only a first step toward showing dementia exposure in families using PSID. Our measure of families is simplified. If one family has multiple children, they are treated as a single family with exposure to someone with dementia. A more complex analysis could take account of exposure to individuals and could take advantage of the longitudinal nature of the data and look at years exposed as well. Families could alternatively be explored based on whether they consist of vertical or horizontal ties. These are potential areas for future work.

This study confirms the broad reach of dementia across households and families in the U.S. As many as 37% of extended families with at least one adult age 65 and older – nearly four in 10 – have a relative with dementia in their extended family network. As the population ages, it will become increasingly important to consider the potential effects of dementia on these broader family units. A family network perspective is important for quantifying the broader costs of dementia on families, as well as for capturing the full effects of policies providing support to persons with dementia and their care partners. Future work could take advantage of the unique design of PSID and its linkages between adults with dementia and broader family units to explore whether and how being part of a household, immediate family, and extended family of an older adult with dementia may affect the health and well‐being of individuals, families, and society.

CONFLICT OF INTEREST STATEMENT

The authors have no conflicts of interest to report. Author disclosures are available in the Supporting Information.

CONSENT STATEMENT

All human subjects provided informed consent to be included in PSID.

Supporting information

Supporting Information

Supporting Information

Friedman EM, Patterson SE, Freedman VA. Nearly 4 in 10 extended families of older adults in the United States include an older relative with dementia. Alzheimer's Dement. 2025;21:e70451. 10.1002/alz.70451