Skip to main content
NCBI home page
The Journals of Gerontology Series B: Psychological Sciences and Social Sciences logoLink to The Journals of Gerontology Series B: Psychological Sciences and Social Sciences
. 2025 Feb 25;80(6):gbaf040. doi: 10.1093/geronb/gbaf040

Who’s On Your Team? Classifying Dementia Caregiving Networks and Associations With the Well-being of Caregivers and Care Recipients With Dementia

Amanda N Leggett 1,✉, Srabani Haldar 2, Sophia Tsuker 3, Wenhua Lai 4, Natasha Nemmers 5, HwaJung Choi 6, Vicki Freedman 7
Editor: Markus Schafer8
PMCID: PMC12059475  PMID: 39994746

Abstract

Objectives

While caregivers are typically enmeshed in broad networks of family and friends assisting with care, this network has been neglected in favor of examining a “primary” caregiver. This study examines types of family and unpaid friend networks for individuals with dementia and how one’s network type relates to the well-being of care recipients with dementia and their caregivers.

Methods

Data are drawn from the nationally representative 2017 National Health and Aging Trends Study and associated National Study of Caregiving. The sample includes 336 dementia care networks (network size mean = 2.9). We first identified network types using latent class analysis and then examined the extent to which network type is associated with the well-being of care recipients with dementia (sleep, depressive symptoms) and their caregivers (emotional difficulty, overload, social support from family and friends) using ANOVA and linear regressions adjusting for demographics.

Results

3 network types were identified: “Siloed”—small networks, limited task sharing (29.8% of networks); “Small but Mighty”—small networks, high task sharing (23.0% of networks); and “Complex”—large networks, diverse membership, members who share and specialize in task assistance (47.2%). Individuals with dementia with a “Siloed” network had significantly poorer sleep quality and caregivers in “Siloed” networks reported receiving less social support from family and friends than those in “Small but Mighty” and “Complex” networks.

Discussion

Caregiver networks that are less collaborative may need supports to reduce isolation among caregivers and improve health outcomes for individuals with dementia.

Keywords: Caregiver network, Caregiver well-being, Dementia caregiving, Kinship caregiving

Beyond the Primary

As age-related dementia progresses, assistance with self-care or household activities (e.g., shopping, paying bills, bathing, dressing), compensation for memory impairments, management of behavioral and psychological symptoms of dementia, and assistance with medical decision-making become crucial to the survival and well-being of the individual living with dementia. The majority of persons living with dementia receive such care from unpaid, family caregivers—such as spouses, adult children, close relatives, and/or friends—with about 8% of individuals with dementia who live in the community not receiving care from a family caregiver (though potentially receiving care from other sources) (Alzheimer’s Association, 2024; de Vugt et al., 2004; Kasper et al., 2015; U.S. Department of Health and Human Services, 2018). More than 11 million Americans serve in a dementia care role, providing a total of 18 billion hours of care annually (Alzheimer’s Association, 2024). Community-dwelling older adults in the United States needing assistance with functional and mobility tasks have an average of four caregivers, with larger networks found for care recipients with dementia, yet the dominant model in family caregiving research has been to study a single, primary caregiver (Freedman & Spillman, 2014).

Because caregiving research tends to focus on the role of a “primary” caregiver for a care recipient with dementia, or the caregiving dyad of caregiver/care recipient with dementia from the caregiver’s perspective, this may be presenting incomplete evidence on the caregiving context, without attention to the role of the broader network within which caregivers function. This may be particularly true for individuals with dementia who are (1) women, who may receive more support from friends and family due to premature mortality of a spouse and/or as a form of coping through formation of support networks in a way that differs from men (Andersson & Monin, 2018; Taylor & Stanton, 2007), and (2) from Black and Latinx families, who tend to provide more assistance to care recipients with greater functional and cognitive impairment than do White caregivers (Chadiha et al., 1995; Cohen et al., 2019; Fabius et al., 2020; Fredman et al., 1995). Thus, in understanding the caregiving context in diverse populations, a network focus is necessary.

Characterizing Networks

The hierarchical compensatory model (Barker, 2002; Cantor, 1979) suggests caregivers enter a care role as a function of the closeness of their relationship with the care recipient (e.g., a spouse more likely to provide care than a cousin), while the task-specific model of caregiver selection (Litwak, 1985) theorizes a caregiving network is formed based on a match between care recipient needs and availability of helpers (e.g., family, friends, paid helpers). Indeed, prior work has characterized networks by their membership (e.g., size of the network, availability of a daughter in the network) and collaborative behaviors (e.g., network members sharing household tasks) (Andersson & Monin, 2018; Barker, 2002; Fast et al., 2004; Jørgensen et al., 2019; Spillman et al., 2020). Limited work, however, has considered how these characteristics cluster together to better characterize these dementia caregiving networks.

Relevant to the hierarchical and compensatory model, most studies examining network profiles consider who is in a care network—that is, a kinship relationship to an older adult care recipient. Representative studies out of Amsterdam and Canada draw distinction between networks that include paid versus unpaid care partners, coresidential versus non-coresiding, and/or including kin versus nonkin (Jacobs et al., 2018; Keating & Dosman, 2009). A recent study using the National Health and Aging Trends Study (NHATS) expanded beyond kinship to also include whether assistive technologies (e.g., the internet, a walker) supported care (Lin, 2024). Lin found networks were more diverse (including nonfamily and assistive technologies) when health was declining, and a spouse or adult child was unavailable. These studies, however, are not specific to dementia caregivers and do not consider the functioning of the network such as whether network members collaborate in providing care or types of care provided.

One study of 66 dementia caregivers (Friedman & Kennedy, 2021) identified four profiles of social support networks where caregivers detailed all members of their own social network as well as the social network of the individual with dementia and the types of support (e.g., emotional, functional) provided to each member of the dyad: (1) big networks including many helpers for both the caregiver and care recipient with dementia with above average help specific to caregiving; (2) big networks predominantly full of helpers assisting the caregiver directly; (3) small networks that primarily supported care recipients with dementia (i.e., indirect support to the caregiver); and (4) small networks that were not providing much assistance to either the caregiver or care recipient with dementia. However, these networks were defined from the perspective of one caregiver about their received social support, not defined around who all assists the care recipient with dementia with direct caregiving tasks. A second study, drawing from the 2011 NHATS, focused on gaps in assistance between primary and secondary caregivers of older adults. This study identified four network types: (1) smaller networks with fewer total care hours received by the older adult but shared care responsibilities between primary and secondary caregivers, (2) larger networks with average total care hours received and shared care responsibilities, (3) smaller networks predominantly led by a primary caregiver, and (4) mid-size networks led by a primary caregiver with intensive total care hours received (Hu et al., 2022). Most relevant to the current study, Ali and colleagues (2022) examined care networks based on intensity of the care provided by the networks and the regularity of caregiving task provision. Five profiles were identified: (1) high caregiving intensity and assistance with all care task domains, (2) assisted with all activities excluding health care assistance, (3) networks predominantly assisting with household tasks, (4) assisted on a more irregular care schedule and helped with mobility and self-care tasks, and (5) a network that provides limited assistance primarily with transportation.

Taken together, however, important gaps remain in existing studies. Researchers have largely not explored networks for individuals with dementia who require intensive and long-term care arrangements. While kinship composition has been explored, the connection of network membership with how network members collaborate to provide care is unexplored, and whether one’s network type predicts critical outcomes for both caregivers and individuals with dementia.

How Network Characteristics Relate to the Well-being of Care Recipients With Dementia and Their Caregivers

Existing theories of mutual influence such as the life course perspective of linked lives (Elder et al., 2003; Mejía & Gonzalez, 2017) and the transitive model (Ruiz et al., 2006) posit that spouses and partners have mutual influence on one another, and it therefore follows that caregiver traits may influence care recipient’ well-being outcomes and vice versa. Additionally, the Multilevel Conceptual Framework of Care Network Collaboration (Ellis et al., 2022) suggests that both individual caregiver dynamics and collective care network dynamics (e.g., network cooperation, doing one’s “fair share” of care) simultaneously influence the task engagement of the network (i.e., which tasks caregivers take on and how they are shared across the network), which may have direct influence on both care recipient and caregiver outcomes. However, limited work has considered how caregiver networks may influence outcomes for individuals with dementia or their caregivers.

Prior research has considered how specific aspects of one’s caregiving network can influence care outcomes. For example, larger networks with more family members and females represented and networks with closer geographic proximity of members to the care recipient were associated with greater hours of care received by a care recipient (Fast et al., 2004). However, a larger network did not relate to greater collaboration among network members generally (Ellis et al., 2022). Further, relative to receiving care from a spouse alone, Potter (2019) found that other care configurations (e.g., non-spouse caregiver, multiple caregivers not including a spouse) were associated with more unmet needs (e.g., not having needed assistance with bathing or dressing) for the care recipient. Larger caregiving networks have also been associated with more depressive and anxiety symptoms reported among care recipients, particularly at levels of higher morbidity (Andersson & Monin, 2018). Health services utilization is also related to caregiver network composition. For instance, if a care recipient has a spouse or partner in their network, it reduced their risk of skilled nursing admission, whereas having a caregiver who helped with tasks related to medical care (i.e., attending a physician visit or managing medications) was associated with increased risk of admission (Jørgensen et al., 2019).

In the current study, we consider whether one’s caregiving network type is associated with two important aspects of well-being for a care recipient with dementia: sleep quality and depressive and anxiety symptoms. As many as 60%–70% of individuals with dementia report sleep impairment and problems with sleep are associated with caregiver burden and risk for institutionalization, suggesting a link between the care role and the sleep quality of the care recipient with dementia (Kim et al., 2014; Ornstein & Gaugler, 2012; Wennberg et al., 2017). While prior research has shown that caregiving task management, and in particular medical care management (e.g., managing medications, wound care), is associated with caregivers own sleep disturbances, it is unknown whether this may carry into sleep disturbances of the care recipient with dementia (Polenick et al., 2018). In addition, while prior work has found network size to relate to an older adult’s mental health in the context of multimorbidity (Andersson & Monin, 2018), and the caregiver stress process (e.g., a caregiver’s own mental health, coping styles) relates to the well-being of a care recipient with dementia more broadly (Ejem et al., 2015; García-Alberca et al., 2013; Pristavec, 2019), we consider the network type as a predictor of the mental health of a care recipient with dementia. Based on prior research, it is hypothesized the larger, more diverse networks that are contributing to more task collaboration may be associated with worse well-being for the care recipient with dementia.

Limited work, on the other hand, has considered how the caregiving network one finds oneself in can influence a caregiver’s own well-being. A meta-analysis identified worse psychological, financial, and physical outcomes among spousal caregivers relative to adult child and children-in-law caregivers. Yet children-in-law reported worse relationship quality and less caregiving gains (Pinquart & Sörensen, 2011). A study of caregivers of adults aged 50 and older found that poorer quality care coordination was associated with greater levels of caregiver stress (Xu et al., 2021). Additionally, while associations between caregiver task provision and social support with burden and well-being are well established (e.g., social support buffering caregiver stress), moving beyond associations at the primary caregiver level to interrelations within the care network is a critical next step.

Hence, the current study builds on prior theory and research to (1) define profiles of dementia family caregiving networks based on their compositional makeup and caregiving collaboration (e.g., membership, task sharing) and (2) explore how these networks relate to key care outcomes for both family caregivers and care recipients with dementia. We utilize the 2017 wave of the nationally representative NHATS and affiliated National Study of Caregiving (NSOC) to build network typologies from perspective of the care recipient with dementia. Defining network typologies provides a person-centered approach to considering the variability in types of networks, offering a critical next step to understanding the function and influence of networks on care outcomes and how to intervene within the caregiving network.

Method

Data and Sample

Akin to sampling methods of other network analyses (e.g., Hu et al., 2022), our primary sample consists of community-dwelling individuals with probable dementia who have two or more family or unpaid nonfamily caregivers in their network. Data are drawn from the 2017 NHATS and linked 2017 NSOC. The 2017 wave was selected as the most recent, pre-COVID-19 dataset that could be linked with NSOC and which also provided the greatest sample size given our eligibility criterion. NHATS is a nationally representative sample of Medicare beneficiaries aged 65 years or older. We were specifically interested in care recipients with dementia and thus selected for NHATS participants classified as having probable dementia. In NHATS, probable dementia status is based upon (1) a self or proxy report of a diagnosis, (2) a score of 2 or higher on the AD8 Dementia Screening Interview, and (3) performance on other cognitive tests (scoring ≥1.5 SDs below the mean in two or more domains on tests of memory, executive function, and orientation) as detailed in Kasper and colleagues (2013). This measure has strong sensitivity and specificity relative to diagnosis by a consensus panel based on an in-home assessment and medical records. Additionally, in the NHATS interview, participants are asked if a family member or unpaid nonfamily member helped them with any self-care tasks, mobility, and/or household chores. NHATS participants then provide the names of all individuals who helped with these tasks. All listed helpers for each NHATS participant populate a file representing the full caregiving networks for each participant and what tasks each network member assisted.

In 2017, Wave 7 of the NHATS study, there were 6,312 participants, of whom 1,953 had at least one caregiver in the network file who met NSOC caregiver eligibility criteria (NSOC eligibility includes being a family member or unpaid nonfamily member who provides help with mobility, self-care, household, or other activities, e.g., transportation). These 1,953 participants had 4,206 NSOC-eligible caregivers. Of these, 1,572 participants were community-dwelling (with 3,421 caregivers), and of these, 440 had probable dementia (with 1,075 caregivers). Finally, our analytic sample included 336 community-dwelling participants with probable dementia who had at least two NSOC-eligible caregivers (i.e., constituting 336 unique networks) with a total of 971 caregivers in their networks.

We additionally consider whether caregiver network types are associated with well-being outcomes for caregivers. Well-being outcomes are collected in NSOC, a linked study of family and unpaid nonfamily caregivers for NHATS participants. NSOC samples from the previously described care network file for eligible caregivers. If more than five caregivers are listed per NHATS participant, five were randomly selected for inclusion in NSOC such that each NHATS participant eligible for NSOC could have up to five caregivers represented. Of the 971 caregivers reflected in the care networks in the analytic sample for this analysis, 557 were NSOC participants.

Measures

Network components

Based on prior literature, we included several variables representing the composition and diversity of caregiver networks (Spillman et al., 2020). First, network size constituted the number of NSOC-eligible caregivers an NHATS participant had in their network. For inclusion in the latent class analysis (LCA), this measure was dichotomized as two caregivers or three or more caregivers based on the mean network size. We also include availability of a spouse (yes/no), daughter (yes/no), non-immediate family member (any family other than a spouse, adult child, daughter-in-law or son-in-law; yes/no), or nonfamily member (e.g., friend, neighbor; yes/no) in the network.

Finally, we consider variables detailing the task sharing and participation within the network. First, based on prior research by Spillman and colleagues (2020), we define the caregivers as being either “generalists” or “specialists.” Caregivers are defined as “specialists” if they assist with only one activity domain, whereas “generalists” are those assisting with multiple task domains. Four care task domains were defined in Spillman and colleagues’ prior work: household activities (four activities; e.g., cooking meals), mobility and self-care activities (seven activities; e.g., dressing, helping in and out of a bed or chair), medical activities (two activities; e.g., managing medications, communicating with physicians), and transportation (two activities; e.g., helping someone get around outside the home) (Spillman et al., 2020). Therefore, a “specialist” caregiver would help with one of these domains, whereas “generalists” would help with two or more of these domains. We create indicators of the extent to which there is task sharing, that is overlap in the network members who assist with a given domain of care.

Outcomes

Sleep quality and psychological well-being of the care recipient with dementia

Well-being of care recipients with dementia is assessed by psychological well-being and sleep impairment. Psychological well-being was assessed with the Patient Health Questionnaire-4 (PHQ-4), a combination of the PHQ-2 (e.g., had little interest or pleasure in doing things) and GAD-2 (e.g., felt nervous, anxious, or on edge) which is on a scale of 1 (not at all) to 4 (nearly every day) where a higher mean score indicates more depressive and anxiety symptoms (α = .76) (Kroenke et al., 2009; Löwe et al., 2010). Sleep impairment in the last month was measured with a three-item scale ranging from 1 (never) to 5 (every night) (e.g., difficulty falling asleep, taking medications to help with sleep) where a higher mean score indicates worse sleep (α = .65).

Caregiver’s stress and well-being

Four caregiver outcomes are considered, capturing both positive and negative aspects of the caregiving stress process. Emotional difficulty of care is created by combining two items. First, caregivers are asked if providing assistance to the NHATS participant has been emotionally difficult (yes/no). If yes, caregivers rate the level of difficulty on a 5-point scale from 1 (a little difficult) to 5 (very difficult). Respondents who said “no” to the first question are scored as 0 so the scale ranges from 0 not at all difficult to 5 very difficult. Caregiver overload is a mean score of four items asking the caregiver how much they “haven’t had time for oneself, have gotten a routine going but the NHATS participant’s needs have changed, have had more things to do than one can handle, and have been exhausted when they have gone to bed at night.” Items are on a Likert scale from 1 (not so much) to 3 (very much) (α = .77). Limited caregiving gains is measured with a mean of four items on whether caregiving “helps you deal better with difficult situations, gives confidence in abilities, helps you feel satisfied that your care recipient is well cared for, and helps bring you closer to the care recipient” on a Likert scale of 1(very much) to 3 (not so much) (α = .72) such that a higher score indicates fewer caregiving gains. Lack of social support from family and friends included a sum of three dichotomous (yes/no) questions regarding whether the caregiver had friends and family to help with daily activities (e.g., running errands), help in care provision for the care recipient, and to talk to about important things in life (recoded where a higher score is indicative of an absence of support). For ease of interpretability, all outcomes were coded such that a higher score indicates a more negative outcome (e.g., lack of social support, more emotional difficulty).

Demographic Characteristics and Medical Conditions

Models adjust for characteristics of the care recipients with dementia including: gender (female = 1), race/ethnicity (non-Hispanic White, non-Hispanic Black, Other [including Hispanic] = reference), and count of current chronic medical conditions excluding dementia (sum of nine conditions, e.g., cancer, diabetes). Models with caregivers’ well-being outcomes adjust for their age, gender, and race.

Analysis

First, a LCA was run to identify profiles of caregiver network clusters. LCA is a person or network-oriented latent variable approach that looks for subtypes of “networks” sharing certain patterns of characteristics. Two sets of parameters were estimated: (1) class membership/posterior probabilities (the proportion of the population that belongs to a class; i.e., the proportion of care recipients with dementia having a particular caregiver network type) and (2) item-response probabilities conditional on class membership (probability of a class endorsing a particular response to a specified item) (Collins & Lanza, 2010; Goodman, 1974). From these two sets of parameters, the posterior probability of class membership for an individual can be estimated. Models were run iteratively by number of classes and goodness of fit criterion (Akaike’s Information Criterion [AIC]; the Bayesian Information Criterion [BIC]; lower values reflecting more optimal model fit and parsimony) and were compared, along with theoretical and empirical background on caregiving networks, to determine the best solution (Akaike, 1974; Schwarz, 1978). We then utilized the maximum probability assignment rule to assign care recipients with dementia to their best-fitting network type based on their posterior probabilities (Collins & Lanza, 2010). Unadjusted comparisons across the network types were then made on the demographic characteristics, medical conditions, and outcome variables using chi-square tests and ANOVAs.

Finally, network types determined through LCA analysis were dummy coded as network membership versus not, holding one network type as the reference group, and included as key predictors of both the well-being of care recipients with dementia and their caregivers. Two linear regression models were run adjusting for the previously described care recipient with dementia characteristics with network type predicting sleep impairment and psychological well-being (PHQ-4) of the care recipient with dementia. Next four linear regression models were run adjusting for caregiver characteristics with network type predicting the caregiver’s emotional difficulty of care, overload, caregiver gains, and social support. Models are weighted (using the final analytic weight) to adjust for differential probabilities of sample selection, and statistical tests are adjusted to address the study’s complex survey design (Kasper & Freedman, 2021), and all models utilize listwise deletion for missing values. Standardized coefficients are provided for individual predictors, and an F test is provided for overall goodness of fit of the model. All analyses were run using SAS 9.4, copyright SAS Institute, Inc.

Results

On average, care recipients with dementia were 83 years of age, had 0.45 chronic medical conditions, 60% were female, 53.6% were White, 15.2% were Black, and 27.8% were of Hispanic or from another racial/ethnic background. Weighted descriptive statistics for full sample characteristics, as well as characteristics by network type, can be found in Table 1 (with caregiver characteristics in Supplementary Table 1).

Table 1.

Weighted Characteristics for Three Caregiver Network Clusters

Sample characteristics Total Siloed (Cluster 1; 29.8%) Small but Mighty (Cluster 2; 23.0%) Complex (Cluster 3; 47.2%) Χ 2 /F
%/M SE Range %/M SE Range %/M SE Range %/M SE Range
Characteristics of NHATS care recipients with dementia (N = 336)
 Female 60.1 2.9 — 52.8 4.3 — 59.6 4.8 — 65.1 3.3 — 1.2
 Race
  White 53.6 3.7 — 56.2 4.7 — 54.8 4.7 — 51.1 4.5 — 1.5
  Black 15.2 2.2 — 9.7 2.0 — 11.1 2.4 — 21.1 3.1 —
  Othera 27.8 4.7 — 34.1 5.4 — 34.2 5.6 — 27.8 4.7 —
 Age 83.2 0.5 68–106 82.8 0.9 68–99 82.9 0.7 69–98 83.7 0.6 68–106 0.5
 Medical conditions 3.2 0.1 0–8 3.5 0.1 1–8 3.0 0.2 0–8 3.1 0.1 0–7 2.9*
 Psychological well-beingb 1.9 0.1 1–4 2.0 0.1 1–4 1.8 0.1 1–4 1.9 0.1 1–4 0.5
 Sleep quality 2.5 0.1 1–5 2.9 0.1 1–5 2.3 0.1 1–5 2.3 0.1 1–5 8.6***
NSOC caregiver outcomes (N = 557)
Emotional difficulty 1.5 0.1 0–5 1.5 0.1 0–5 2.0 0.3 0–5 1.3 0.1 0–5 5.9***
Caregiving overload 1.5 0.0 1–3 1.5 0.1 1–3 1.6 0.0 1–3 1.5 0.0 1–3 1.6
Caregiving gains 1.4 0.0 1–3 1.5 0.0 1–3 1.4 0.1 1–2.75 1.4 0.0 1–3 0.7
Family and friends support 1.2 0.0 1–2 1.3 0.0 1–2 1.1 0.0 1–2 1.2 0.0 1–2 12.0***
Open in a new tab

Notes: NHATS = National Health and Aging Trends Study; NSOC = National Study of Caregiving; SE = standard error. Chi-square and ANOVA tests were run for mean comparisons by network type.

aOther race includes Hispanic ethnicity and any non-White and non-Black racial group. bPsychological well-being is assessed with the Patient Health Questionnaire-4 (PHQ-4).

* p < .05. ***p < .001.

A series of latent class models (2 through 5 class models) were compared utilizing the AIC, the BIC, and for overall interpretability of the results. While the four-class solution showed the best statistical fit, network clusters were most distinct and more interpretable at the three-class solution (AIC and BIC values are found in Supplementary Table 2). Average posterior probabilities for class assignment were high (Siloed networks: M = 0.99, SD = 0.08, range = 0.51–1.00; Small but Mighty networks: M = 0.97, SD = 0.05, range = 0.82–1.00; Complex networks: M = 0.94, SD = 0.10, range = 0.53–1.00), suggesting that response patterns clearly aligned within a particular class. Table 2 presents the membership class probabilities for each class (GAMMA, γ) and the probability that members of each class endorsed each item response category (RHO, ρ). Figure 1 represents the average item response probability for each component of the caregiving networks (e.g., network size, availability of spouse, mobility task sharing).

Table 2.

Item-Response Probabilities and Distribution of Latent Classes

Parameter Variable Siloed (Cluster 1; 29.8%) Small but Mighty (Cluster 2; 23.0%) Complex (Cluster 3; 47.2%)
GAMMA .30 .23 .47
RHO Three or more helpers in network .10 .00 1.00
RHO Spouse in the network .36 .17 .27
RHO Daughter in the network .60 .72 .71
RHO Non-immediate family in the network .29 .29 .55
RHO Nonfamily in the network .21 .11 .19
RHO Specialist in the network .97 .00 .82
RHO Network shares medical tasks .06 .57 .55
RHO Network shares household tasks .39 .74 .82
RHO Network shares mobility tasks .14 .59 .67
RHO Network shares transportation tasks .37 .46 .59
Open in a new tab

Notes: GAMMA parameters are class membership probabilities—i.e., the proportion of care recipients with dementia having a particular caregiver network type; and RHO parameters are item response probabilities conditional on class membership—i.e., the probability of a class endorsing a particular response to a specified item.

Figure 1.

Alt text: Figure 1 displays a bar chart of the predicted probabilities of component characteristics for three care networks fully detailed in Table 2.

Open in a new tab

Predicted probabilities of care network characteristics across 3 network types.

Siloed Networks

Siloed networks made up 29.8% of network types. Siloed networks were smaller networks (M = 2.0 members; ρ = 0.1 for three or more helpers in network) with slightly greater likelihood of having a spouse (ρ = 0.4) and slightly lower likelihood of having an adult daughter (ρ = 0.6) in the network than the other network types. Siloed networks stand out for having the highest likelihood of having a specialist in the network (ρ = 1.0) and the lowest likelihood of engaging in task sharing in any of the four task domains (ρ = 0.1 for medical tasks, 0.4 for household tasks, 0.1 for mobility tasks, and 0.4 for transportation tasks), hence the “siloed” name.

Small but Mighty Networks

The least common network type (23%) was the Small but Mighty network. Small but Mighty networks contained only two caregivers (M = 2.0 members; ρ = 0.0 for three or more helpers in network). Small but Mighty networks were least likely of the network types to have a spouse (ρ = 0.2) and nonfamily member in the network (ρ = 0.1), and also had a low likelihood of having a non-immediate family member (ρ = 0.3). Despite the small size of this network, they were least likely to have a specialist in the network (ρ = 0.0) yet had nearly as significant levels of task sharing as the Complex network (ρ = 0.6 for medical tasks, 0.7 for household tasks, 0.6 for mobility tasks, and 0.5 for transportation tasks). Therefore, while a small network in size, they are “mighty” in their collaborative task sharing as they were least likely to specialize in a single caregiving domain and likely to collaborate across multiple task domains.

Complex Networks

The most common network type was “Complex” (47.2%). Complex networks were larger (M = 3.8 members; ρ = 1.0 for three or more helpers in network) and were most likely to include non-immediate family members (ρ = 0.6). They were equivalently likely as the Siloed network to include nonfamily (ρ = 0.2) and as the Small but Mighty network to include an adult daughter caregiver (ρ = 0.7). They were also most likely to be involved in task sharing across the four domains (ρ = 0.6 for medical tasks, 0.8 for household tasks, 0.7 for mobility tasks, 0.6 for transportation tasks). Despite significant task sharing, they were also likely to include a specialist caregiver in the network (ρ = 0.8). Hence, “complexity” is drawn from diverse network membership and extensive collaboration in caregiving tasks.

Next, mean difference tests were run to compare care characteristics by network type and linear regression models were run to analyze network classification as a predictor of well-being outcomes for both care recipients with dementia and their caregivers. In adjusted linear regression models, the Siloed network was utilized as the reference group. Siloed networks were selected as the reference group as they were most distinct in their individualistic, less collaborative nature from the other network types.

Demographic and Care Context Differences in the Care Recipient With Dementia by Care Network

Care networks did not differ by the care recipient’s age, gender, or race. However, the mean number of chronic medical conditions differed significantly by network type (Table 1), with Siloed networks reporting care for individuals with the most disease burden (i.e., multimorbidity; M = 3.50) and Small but Mighty networks caring for the least amount of disease burden (M = 3.01) (F(2,333) = 3.10, p < .05). Full results on chi-square and ANOVA tests of mean differences in the demographic and health characteristics of the care recipient with dementia are given in Table 1.

Psychological Well-being and Sleep Quality Differences in the Care Recipient With Dementia by Care Network

Depressive and anxiety symptoms did not significantly differ by one’s network type (Table 1). However, care recipients with dementia differed significantly in their sleep impairment by their caregiver network type (F(2,333) = 8.61, p < .001); care recipients with Siloed networks reported significantly worse sleep (M = 2.86) than both those with Small but Mighty and Complex networks (M = 2.30). This finding held in adjusted linear regressions (Table 3) with individuals with Small but Mighty (β = −.20, SE = 0.23, p < .05) and Complex (β = −.22, SE = 0.19, p < .05) network types having significantly better sleep quality than care recipients with Siloed networks.

Table 3.

The Linear Regression Model of Caregiver Network Type on Sleep Quality and Psychological Well-being of Care Recipients with Dementia

Predictors Psychological well-being Sleep quality
β (SE) β (SE)
Small but Mighty −.03 (0.16) −.19 (0.22)*
Complex .02 (0.17) −.20 (0.20)*
Siloed (reference) — —
Age −.05 (0.01) −.07 (0.01)
Female −.07 (0.13) −.02 (0.15)
Male (reference) — —
Race
 White −.05 (0.17) −.09 (0.17)
 Black −.01 (0.18) −.03 (0.19)
 Hispanic and other (reference) — —
Medical conditions .20 (0.04)*** .25 (0.04)*
Open in a new tab

Notes: SE = standard error. Psychological well-being is assessed with the Patient Health Questionnaire-4 (PHQ-4) scale.

* p < .05. ***p < .001.

Stress and Well-being Differences in the Caregiver by Care Network

Caregivers differed significantly by network type on their reported emotional difficulty of caregiving (F(2,537) = 5.88, p < .01). Caregivers in Small but Mighty networks had the greatest emotional difficulty (M = 2.02), whereas caregivers in Complex networks had the least emotional difficulty (M = 1.33). Caregivers also differed by their network type in the social support they received from family and friends (F(2,537) = 11.95, p < .001). Small but Mighty caregivers reported receiving more social support (M = 1.14), whereas Siloed network caregivers reported receiving less support (M = 1.32). This association held in an adjusted linear regression such that caregivers in both Small but Mighty networks (β = −.22, SE = 0.05, p < .05) and Complex networks (β = −.19, SE = 0.05, p < .05) reported receiving significantly more social support than caregivers in Siloed networks (Table 4).

Table 4.

The Association of Caregiving Network Type With Caregiver’ Stress and Well-being

Predictors Emotional difficulty Caregiving overload Caregiving gains Family and friends’ support
β (SE) β (SE) β (SE) β (SE)
Small but Mighty .09 (0.31) .07 (0.09) −.01 (0.09) −.23 (0.06)**
Complex .00 (0.25) −.06 (0.08) .01 (0.07) −.19 (0.06)*
Siloed (reference) — — — —
Age −.04 (0.01) −.00 (0.00) .05 (0.00) .22 (0.00)**
Female .12 (0.24) .10 (0.07) −.08 (0.06) .08 (0.03)
Race
 White −.05 (0.58) −.32 (0.12)** .08 (0.22) −.08 (0.08)
 Black −.15 (0.61) −.28 (0.13)** −.14 (0.23) −.05 (0.08)
 Hispanic and other race (reference) — — — —
Open in a new tab

Notes: SE = standard error.

* p < .05. **p < .01.

Discussion

Using LCA with a national sample of Medicare beneficiaries with dementia and their family caregivers, we find that dementia caregiving networks vary in their makeup and in the way they divide and delegate caregiving responsibilities. Makeup varies from primarily daughter-led networks to more diverse networks including non-immediate family and friends, whereas division of caregiving tasks vary from members specializing in one domain of caregiving task to more extensive sharing and collaborating across domains of caregiving tasks.

Profiling Networks—Distinctions From Prior Literature

Prior analyses that have attempted to identify classes or clusters of caregivers have emphasized distinct aspects of the care network, primarily looking at network size and membership, with some examining the intensity of care provision, looking more broadly at a caregiver’s social support, or how much the primary caregiver specifically collaborates with secondary network members (Ali et al., 2022; Friedman & Kennedy, 2021; Hu et al., 2022). Akin to our classifications, these studies have found network clusters that vary in size, caregiving intensity, and membership. Our model builds on these prior studies and, uniquely, defines clusters specifically from the perspective of the care recipient with dementia including all individuals who assisted the individual with household, self-care, mobility, and transportation tasks. We do not attempt to identify a “primary caregiver” responsible for the majority of care given prior research by Hu and colleagues (2022) that found 40% of caregiving networks with two or more caregivers did not include a “primary caregiver.” Unique to prior research, we consider the extent to which network members specialize or collaborate to provide caregiving tasks and the types of tasks network members collaborate on, going beyond network size and membership. This perspective allows us to better describe the functioning of the network with respect to care provision dynamics.

In line with these differences in constructs included in network models, we identified both similarities and distinctions in our network profiles from the prior literature. Similar to Lin (2024) and Keating and Dosman (2009), we identified networks that were more immediate family (i.e., spouse, adult daughter) centric and small (i.e., Siloed and Small but Mighty), whereas one profile that was larger and more diverse in membership (i.e., the Complex network which was more likely to include non-immediate family members). We did not examine intensity of care hours provided, as did Friedman and Kennedy (2021) and Hu and colleagues (2022); however, similar to Ali and colleagues (2022), we examined types of task provision. Ali and colleagues found, for example, one network profile that primarily assisted with transportation, whereas another network profile predominantly aided in household tasks, which may reflect more the type of assistance a care recipient needs than the care network’s response. Hence, moving beyond Ali’s examination of what tasks a network assisted with, we considered whether multiple network members were sharing in the assistance with particular task domains, and whether network members assisted broadly with task domains or specialized in one type of care. While this may also reflect care needs, it also provides nuance in how network members collaborated, or independently went about meeting these needs. Further, given our sample was individuals classified in NHATS as having probable dementia, we presume our sample had more significant care needs than the population of older adults more broadly, considered by most existing care network studies. In contrast to Ellis and colleagues (2022) who found that larger networks did not necessarily increase collaboration of members, we found that Complex networks were most engaged in task sharing. Yet, as the Siloed and Small but Mighty networks had only two network members on average, the capacity for sharing was greater among our Complex networks (i.e., greater availability of members with which to help with tasks). We also distinguish between the Siloed network, which was small but had a member(s) who specialized in only one type of caregiving task with the Small but Mighty network who universally assisted with multiple task domains and shared in these caregiving tasks.

Associations of Network Profiles With Care Recipient and Caregiver Characteristics

Although prior research has found Black and Latinx families have larger networks, are more likely to live with children, provide greater caregiving intensity, and assist with a wider variety of activities (Choi et al., 2020; Cohen et al., 2019; Fabius et al., 2020; Reyes et al., 2020), our network groupings did not vary by the race of the care recipient with dementia. However, Siloed networks were significantly more likely to be providing care for individuals with greater prevalence of multimorbidity, which may suggest these networks are assisting with greater medical management or care coordination on top of that being provided for dementia care. This was surprising given lower task sharing among the Siloed network members, and particularly for medical care tasks. This may reflect a single caregiver taking on significant medical care burden. It may be posited that more complex care contexts (i.e., multimorbidity) elicit the use of professional care services such as in-home nursing or respite care, thus reducing the need for collaborative family or other unpaid care networks. However, while we could not examine paid community services such as respite care or support groups with the present data, a follow-up comparison found that Complex networks were more likely to include a caregiver that was paid (46.1% among Complex networks vs. 36.7% among Siloed networks). Future work should consider the intersections and collaborations of paid care providers within family care networks.

Surprisingly, caregivers’ emotional strain, overload, and gains did not vary by network type. For example, our findings did not replicate those of Andersson and Monin (2018) who observed greater depressive and anxiety symptoms among individuals with larger network sizes. These discrepancies with the prior literature may have emerged because the network clusters that we identified included a myriad of attributes aside from network size (e.g., care task sharing, specialized vs. generalized care roles). Individual associations between care intensity and task provision and caregiver burden are well established (Pinquart & Sörensen, 2003; Riffin et al., 2019), and thus it may be that psychosocial well-being is more a factor of the caregiver’s own stress process of caregiving regardless of the network in which they find themselves. Task sharing, for example, may be necessary given the care needs of the care recipient with dementia, yet may not in and of itself affect the overall intensity of the individual caregiver’s care provision or the caregiver’s appraisal of the stressor (i.e., Lazarus and Folkman’s Transactional Model of Stress, Pearlin’s Stress Process Model) (Lazarus & Folkman, 1984; Pearlin et al., 1990). Caregiver’ mental health may depend more on quality of task sharing as prior work has found discordant quality care coordination to be associated with caregiver burden (Xu et al., 2021). This topic is an area ripe for further investigation.

Finally, our analysis did point to the Siloed care network in general relating to worse well-being including poorer care recipient with dementia sleep quality and less social support received by caregivers from their friends and family. Future research into the mechanisms behind what may be described as poorer psychosocial functioning of this network type is needed.

Limitations

The use of both NHATS and NSOC offers the benefits of nationally representative data and combined reports of the care needs and care network of an individual with dementia, as well as self-reports from multiple caregivers within the network about their caregiving experience. However, nonresponse in NSOC prevented us from including all network members, in the caregiver outcomes analysis. Paid nonfamily helpers were also not considered as part of the network. Additionally, as we specifically chose to focus on collaborative networks, individuals with only one caregiver, which may reflect particularly isolated networks, were excluded from this analysis. Comparisons of solo caregivers and caregiving networks including more than one caregiver warrant further exploration. Finally, as this was a cross-sectional study, we cannot establish causal associations between caregiver networks and health outcomes or discern how networks may change in composition and diversity over time. It may be that a care recipient with dementia and his or her caregiver(s)’s health and well-being influence how a network is formed. In addition to longitudinal analyses, future work should also explore whether network structure varies by race/ethnicity and/or gender of the care recipient with dementia.

In conclusion, caregiving networks for individuals are varied yet often unexplored in caregiving research. Some networks function by significant task sharing of two immediate family members whereas others divide and conquer. These networks vary in the support they receive from others and the sleep quality of individuals with dementia varies by their network type. Future work should consider how networks adapt and change over time in correspondence with changing care needs of the care recipient with dementia. It is critical for the caregiving field to broaden its perspective beyond the primary caregiver to acknowledge more diverse network types. Intervening at the network level may have important implications for quality of care provided and long-term viability of aging in place for community-dwelling individuals living with dementia.

Supplementary Material

gbaf040_suppl_Supplementary_Materials
gbaf040_suppl_supplementary_materials.docx (16.3KB, docx)

Contributor Information

Amanda N Leggett, Institute of Gerontology, Wayne State University, Detroit, Michigan, USA.

Srabani Haldar, Institute of Gerontology, Wayne State University, Detroit, Michigan, USA.

Sophia Tsuker, Institute of Gerontology, Wayne State University, Detroit, Michigan, USA.

Wenhua Lai, Institute of Gerontology, Wayne State University, Detroit, Michigan, USA.

Natasha Nemmers, Institute of Gerontology, Wayne State University, Detroit, Michigan, USA.

HwaJung Choi, Department of Internal Medicine, University of Michigan, Ann Arbor, Michigan, USA.

Vicki Freedman, Institute of Social Research, University of Michigan, Ann Arbor, Michigan, USA.

Markus Schafer, (Social Sciences Section).

Funding

This work is funded by the National Institute on Aging (NIA; R01AG079097, PI: A.N. Leggett). A.N. Leggett is also funded by the NIA (NIAP30AB072931).

Conflict of Interest

None.

Data Availability

The National Health and Aging Trends Study and National Study of Caregiving are produced and distributed by www.nhats.org with funding from the National Institute on Aging (U01AG032947, R01AG054004). This study was not preregistered.

Author Contributions

A.N. Leggett conceptualized the study and hypotheses, supervised the data analysis, and wrote the paper. S. Haldar managed the data, conducted the data analysis, and assisted in interpretation of findings. S. Tsuker helped with literature review and revising the manuscript. W. Lai helped with interpretation of findings and revising the manuscript. N. Nemmers helped with interpretation of findings and revising the manuscript. H. Choi helped to plan the study and revise the manuscript. V. Freedman helped to plan the study and revise the manuscript.

References

  1. Akaike, H. (1974). New look at statistical model identification. IEEE Transactions on Automatic Control, 19(6), 716–723. https://doi.org/ 10.1109/tac.1974.1100705 [DOI] [Google Scholar]
  2. Ali, T., McAvay, G. J., Monin, J. K., & Gill, T. M. (2022). Patterns of caregiving among older adults with and without dementia: A latent class analysis. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 77(Suppl. 1), S74–S85. https://doi.org/ 10.1093/geronb/gbab237 [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Alzheimer’s Association (2024). Alzheimer’s disease facts and figures. Alzheimer’s and Dementia, 20(5), 1–146. https://doi.org/ 10.1002/alz.13809 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Andersson, M. A., & Monin, J. K. (2018). Informal care networks in the context of multimorbidity: Size, composition, and associations with recipient psychological well-being. Journal of Aging and Health, 30(4), 641–664. https://doi.org/ 10.1177/0898264316687623 [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Barker, J. C. (2002). Neighbors, friends, and other nonkin caregivers of community-living dependent elders. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 57(3), S158–167. https://doi.org/ 10.1093/geronb/57.3.s158 [DOI] [PubMed] [Google Scholar]
  6. Cantor, M. H. (1979). Neighbors and friends: An overlooked resource in the informal support system. Research on Aging, 1, 434–463. https://doi.org/ 10.1177/016402757914002 [DOI] [Google Scholar]
  7. Chadiha, L. A., Proctor, E. K., Morrow-Howell, N., Darkwa, O. K., & Dore, P. (1995). Post-hospital home care for African-American and white elderly. The Gerontologist, 35(2), 233–239. https://doi.org/ 10.1093/geront/35.2.233 [DOI] [PubMed] [Google Scholar]
  8. Choi, H., Schoeni, R. F., Wiemers, E. E., Hotz, V. J., & Seltzer, J. A. (2020). Spatial distance between parents and adult children in the United States. Journal of Marriage and the Family, 82(2), 822–840. https://doi.org/ 10.1111/jomf.12606 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Cohen, S. A., Sabik, N. J., Cook, S. K., Azzoli, A. B., & Mendez-Luck, C. A. (2019). Differences within differences: Gender inequalities in caregiving intensity vary by race and ethnicity in informal caregivers. Journal of Cross-Cultural Gerontology, 34(3), 245–263. https://doi.org/ 10.1007/s10823-019-09381-9 [DOI] [PubMed] [Google Scholar]
  10. Collins, L. M., & Lanza, S. T. (2010). Latent class analysis and latent transition analysis with applications in the social, behavioral, and health sciences. John Wiley & Sons, Inc. [Google Scholar]
  11. de Vugt, M. E., Stevens, F., Aalten, P., Lousberg, R., Jaspers, N., Winkens, I., Jolles, J., & Verhey, F. R. J. (2004). Do caregiver management strategies influence patient behaviour in dementia? International Journal of Geriatric Psychiatry, 19(1), 85–92. https://doi.org/ 10.1002/gps.1044 [DOI] [PubMed] [Google Scholar]
  12. Ejem, D. B., Drentea, P., & Clay, O. J. (2015). The effects of caregiver emotional stress on the depressive symptomatology of the care recipient. Aging and Mental Health, 19(1), 55–62. https://doi.org/ 10.1080/13607863.2014.915919 [DOI] [PubMed] [Google Scholar]
  13. Elder, G. H., Johnson, M. K., & Crosnoe, R. (2003). The emergence and development of life course theory. In Mortimer J. T. & Shanahan M. J. (Eds.), Handbooks of sociology and social research.(pp. 3–19). Springer. https://doi.org/ 10.1007/978-0-306-48247-2_1 [DOI] [Google Scholar]
  14. Ellis, K. R., Koumoutzis, A., Lewis, J. P., Lin, Z., Zhou, Y., Chopik, W. J., & Gonzalez, R. (2022). Conceptualizing and operationalizing collaboration among multiple caregivers of older adults. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 78(Suppl. 1), S27–S37. https://doi.org/ 10.1093/geronb/gbac139 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Fabius, C. D., Wolff, J. L., & Kasper, J. D. (2020). Race differences in characteristics and experiences of black and white caregivers of older Americans. The Gerontologist, 60(7), 1244–1253. https://doi.org/ 10.1093/geront/gnaa042 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Fast, J., Keating, N., Otfinowski, P., & Derksen, L. (2004). Characteristics of family/friend care networks of frail seniors. Canadian Journal of Aging, 23(15), 5–19. https://doi.org/ 10.1353/cja.2004.0003 [DOI] [PubMed] [Google Scholar]
  17. Fredman, L., Daly, M. P., & Lazur, A. M. (1995). Burden among white and black caregivers to elderly adults. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 50(2), S110–118. https://doi.org/ 10.1093/geronb/50b.2.s110 [DOI] [PubMed] [Google Scholar]
  18. Freedman, V. A., & Spillman, B. C. (2014). Disability and care needs among older Americans. The Milbank Quarterly, 92(3), 509–541. https://doi.org/ 10.1111/1468-0009.12076 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Friedman, E. M., & Kennedy, D. (2021). Typologies of dementia caregiver support networks: A pilot study. The Gerontologist, 61, 1221. https://doi.org/ 10.1093/geront/gnab013 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. García-Alberca, J. M., Cruz, B., Lara, J. P., Garrido, V., Lara, A., Gris, E., & Gonzalez-Herero, V. (2013). The experience of caregiving: The influence of coping strategies on behavioral and psychological symptoms in patients with Alzheimer’s disease. Aging and Mental Health, 17(5), 615–622. https://doi.org/ 10.1080/13607863.2013.765833 [DOI] [PubMed] [Google Scholar]
  21. Goodman, L. (1974). Exploratory latent structure analysis using both identifiable and unidentifiable models. Biometrika, 61(2), 215–231. https://doi.org/ 10.1093/biomet/61.2.215 [DOI] [Google Scholar]
  22. Hu, M., Freedman, V. A., Patterson, S. E., & Lewis, N. (2022). Shared care networks assisting older adults: New insights from the National Health and Aging Trends study. The Gerontologist, 63(5), 840–850. https://doi.org/ 10.1093/geront/gnac155 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Jacobs, M. T., Broese van Groenou, M. I., Aartsen, M. J., & Deeg, D. J. H. (2018). Diversity in older adults’ care networks: The added value of individual beliefs and social network proximity. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 73(2), 326–336. https://doi.org/ 10.1093/geronb/gbw012 [DOI] [PubMed] [Google Scholar]
  24. Jørgensen, T. S. H., Allore, H., MacNeil Vroomen, J. L., Wyk, B. V., & Agogo, G. O. (2019). Sociodemographic factors and characteristics of caregivers as determinants of skilled nursing facility admissions when modeled jointly with functional limitations. Journal of the American Medical Directors Association, 20(12), 1599–1604.e4. https://doi.org/ 10.1016/j.jamda.2019.01.154 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Kasper, J. D., & Freedman, V. A. (2021). National Health and Aging Trends Study user guide: Rounds 1-10 final release. www.NHATS.org [Google Scholar]
  26. Kasper, J. D., Freedman, V. A., & Spillman, B. C. (2013). Classification of persons by dementia status in the National Health and Aging Trends Study. Technical Paper #5. www.NHATS.org [Google Scholar]
  27. Kasper, J. D., Freedman, V. A., Spillman, B. C., & Wolff, J. L. (2015). The disproportionate impact of dementia on family and unpaid caregiving to older adults. Health Affairs, 34(10), 1642–1649. https://doi.org/ 10.1377/hlthaff.2015.0536 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Keating, N., & Dosman, D. (2009). Social capital and the care networks of frail seniors. Canadian Review of Sociology, 46(4), 301–318. https://doi.org/ 10.1111/j.1755-618x.2009.01216.x [DOI] [PubMed] [Google Scholar]
  29. Kim, S.-S., Oh Kyeung, M., & Richards, K. (2014). Sleep disturbance, nocturnal agitation behaviors, and medical comorbidity in older adults with dementia: Relationship to reported caregiver burden. Research in Gerontological Nursing, 7(5), 206–214. https://doi.org/ 10.3928/19404921-20140512-01 [DOI] [PubMed] [Google Scholar]
  30. Kroenke, K., Spitzer, R. L., Williams, J. B. W., & Löwe, B. (2009). An ultra-brief screening scale for anxiety and depression: The PHQ-4. Psychosomatics, 50(6), 613–621. https://doi.org/ 10.1016/S0033-3182(09)70864-3 [DOI] [PubMed] [Google Scholar]
  31. Lazarus, R. S., & Folkman, S. (1984). Stress, appraisal, and coping. Springer Publishing Company, Inc. [Google Scholar]
  32. Lin, Z. (2024). Diversity and dynamics in care networks of older Americans. Socius, 10, 23780231231223906. https://doi.org/ 10.1177/23780231231223906 [DOI] [Google Scholar]
  33. Litwak, E. (1985). Helping the elderly: The complementary roles of informal networks and formal systems. The Guilford Press. [Google Scholar]
  34. Löwe, B., Wahl, I., Rose, M., Spitzer, C., Glaesmer, H., Wingenfeld, K., Schneider, A., & Brähler, E. (2010). A 4-item measure of depression and anxiety: Validation and standardization of the Patient Health Questionnaire-4 (PHQ-4) in the general population. Journal of Affective Disorders, 122(1), 86–95. https://doi.org/ 10.1016/j.jad.2009.06.019 [DOI] [PubMed] [Google Scholar]
  35. Mejía, S. T., & Gonzalez, R. (2017). Couples’ shared beliefs about aging and implications for future functional limitations. The Gerontologist, 57(suppl_2), S149–S159. https://doi.org/ 10.1093/geront/gnx071 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Ornstein, K., & Gaugler, J. E. (2012). The problem with “problem behaviors”: A systematic review of the association between individual patient behavioral and psychological symptoms and caregiver depression and burden within the dementia patient-caregiver dyad. International Psychogeriatrics, 24(10), 1536–1552. https://doi.org/ 10.1017/S1041610212000737 [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Pearlin, L. I., Mullan, J. T., Semple, S. J., & Skaff, M. M. (1990). Caregiving and the stress process: An overview of concepts and their measures. The Gerontologist, 30(5), 583–594. https://doi.org/ 10.1093/geront/30.5.583 [DOI] [PubMed] [Google Scholar]
  38. Pinquart, M., & Sörensen, S. (2003). Associations of stressors and uplifts of caregiving with caregiver burden and depressive mood: A meta-analysis. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 58(2), P112–128. https://doi.org/ 10.1093/geronb/58.2.p112 [DOI] [PubMed] [Google Scholar]
  39. Pinquart, M., & Sörensen, S. (2011). Spouses, adult children, and children-in-law as caregivers of older adults: A meta-analytic comparison. Psychology and Aging, 26(1), 1–14. https://doi.org/ 10.1037/a0021863 [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Polenick, C. A., Leggett, A. N., Maust, D. T., & Kales, H. C. (2018). Medical care tasks among spousal dementia caregivers: Links to care-related sleep disturbances. The American Journal of Geriatric Psychiatry, 26(5), 589–597. https://doi.org/ 10.1016/j.jagp.2018.01.206 [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Potter, A. J. (2019). Care configurations and unmet care needs in older men and women. Journal of Applied Gerontology, 38(10), 1351–1370. https://doi.org/ 10.1177/0733464817733239 [DOI] [PubMed] [Google Scholar]
  42. Pristavec, T. (2019). The caregiving dyad: Do caregivers’ appraisals of caregiving matter for care recipients’ health? Archives of Gerontology and Geriatrics, 82, 50–60. https://doi.org/ 10.1016/j.archger.2019.01.020 [DOI] [PubMed] [Google Scholar]
  43. Reyes, A., Schoeni, R. F., & Choi, H. (2020). Race/ethnic differences in spatial distance between adult children and their mothers. Journal of Marriage and Family, 82(2), 810–821. https://doi.org/ 10.1111/jomf.12614 [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Riffin, C., Van Ness, P. H., Wolff, J. L., & Fried, T. (2019). Multifactorial examination of caregiver burden in a national sample of family and unpaid caregivers. Journal of the American Geriatrics Society, 67(2), 277–283. https://doi.org/ 10.1111/jgs.15664 [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Ruiz, J. M., Matthews, K. A., Scheier, M. F., & Schulz, R. (2006). Does who you marry matter for your health? Influence of patients’ and spouses’ personality on their partners’ psychological well-being following coronary artery bypass surgery. Journal of Personality and Social Psychology, 91(2), 255–267. https://doi.org/ 10.1037/0022-3514.91.2.255 [DOI] [PubMed] [Google Scholar]
  46. Schwarz, G. (1978). Estimating the dimensions of a model. The Annals of Statistics, 6, 461–464. https://doi.org// 10.1214/aos/1176344136 [DOI] [Google Scholar]
  47. Spillman, B. C., Freedman, V. A., Kasper, J. D., & Wolff, J. L. (2020). Change overtime in caregiving networks for older adults with and without dementia. The Journals of Gerontology, 75(7), 1563–1572. https://doi.org/ 10.1093/geronb/gbz065 [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Taylor, S. E., & Stanton, A. L. (2007). Coping resources, coping processes, and mental health. Annual Review of Psychology, 3, 377–401. https://doi.org/ 10.1146/annurev.clinpsy.3.022806.091520 [DOI] [PubMed] [Google Scholar]
  49. U.S. Department of Health and Human Services. (2018). National plan to address Alzheimer’s disease: 2018 update. https://aspe.hhs.gov/sites/default/files/private/pdf/259581/NatPlan2018.pdf [Google Scholar]
  50. Wennberg, A. M. V., Wu, M. N., Rosenberg, P. B., & Spira, A. P. (2017). Sleep disturbance, cognitive decline, and dementia: A review. Seminars in Neurology, 37(4), 395–406. https://doi.org/ 10.1055/s-0037-1604351 [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Xu, J., Liu, P. J., & Beach, S. (2021). Multiple caregivers, many minds: Family discord and caregiver outcomes. The Gerontologist, 61(5), 661–669. https://doi.org/ 10.1093/geront/gnaa067 [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

gbaf040_suppl_Supplementary_Materials
gbaf040_suppl_supplementary_materials.docx (16.3KB, docx)

Data Availability Statement

The National Health and Aging Trends Study and National Study of Caregiving are produced and distributed by www.nhats.org with funding from the National Institute on Aging (U01AG032947, R01AG054004). This study was not preregistered.

Articles from The Journals of Gerontology Series B: Psychological Sciences and Social Sciences are provided here courtesy of Oxford University Press

RESOURCES