Transition Into the Caregiver Role Among Older Adults: A Study of Social Participation and Social Support Based on the Canadian Longitudinal Study on Aging

Abstract

Objectives

Older adults (65 years and older) are increasingly providing care for their spouses, family members, and nonkin others (e.g., friends and neighbors). However, available knowledge of older caregivers is limited to spousal caregivers and their psychological outcomes. Other caregiver role types or social outcomes among older caregivers are less well studied. Thus, this study examines the social participation and social support among older caregivers by comparing 3 types of older caregivers, including spousal caregivers, nonspouse family caregivers, and nonkin caregivers.

Methods

Participants for this study were drawn from the Baseline and Follow-up 1 data from the Canadian Longitudinal Study on Aging. A total of 3,789 older adults became caregivers during the 2 data collection time points. Linear mixed models were applied to examine the change of social participation and social support among the three caregiver role types over the course of survey.

Results

The study finds that after transitioning into the caregiving role, spousal caregivers, and nonkin caregivers experienced a decline in social participation, and spousal caregivers also received less social support over time. When comparing the 3 caregiver role types, spousal caregivers reported the greatest decline in social participation and social support.

Discussion

This study adds to the relatively limited knowledge of older caregivers by presenting the changes in social participation and social support after transitioning into 3 types of caregiver roles. The results indicate the need to provide support for caregivers, particularly spousal and nonkin caregivers, to help them maintain social relationships and networks for participation and support.

Approximately one-quarter (1.5 million) of Canadians aged 65 years and older are caregivers (hereafter referred to as older caregivers) supporting their spouses/partners, other family members (e.g., parents, children, grandchildren, and siblings), and nonkin others (e.g., friends and neighbors) with chronic health conditions or age-related needs (Arriagada, 2020). In Canada, about one-third (34%) of older caregivers provided care primarily to their spouses, and the remaining two-thirds provided care to others (23% for parents or parents-in-law, 20% for nonkin others, 14% for relatives, and 9% for children) in 2018 (Arriagada, 2020). Also, older caregivers tend to provide the longest caregiving hours with personal care and emotional support compared to caregivers in other age groups, despite their smaller proportion among all caregivers (Sinha, 2015).

In Canada, the proportion of older caregivers among all caregivers has gradually increased from 12%in 2012 to 20% in 2018 (Arriagada, 2020; Sinha, 2015). A similar rise in the number of older caregivers also has been observed in other countries, such as the United States and the United Kingdom (Greenwood et al., 2019). Several demographic and socioeconomic trends underlie the fact that older adults shoulder more caregiving responsibility for family and friends. Higher female labor force participation and dual wage-earning families, and greater geographic dispersal of families limit the availability of care and support originating from younger generations (Li & Wister, 2021). Also, filial obligation may be weakened due to increasing rates of divorce and remarriage among younger generations in developed countries (Lowenstein & Daatland, 2006). As a result, older adults need to undertake more caring responsibility for their spouses, siblings, or friends, which is traditionally provided by or shared with their adult children or other family members. There may also be normative reasons (e.g., past family practices, gendered expectations) for why older caregivers, especially older women, feel that they should provide care to spouses, other kin, or friends (Moen, 2016).

The existent research on older caregivers is predominated by studies on spousal caregivers, whereas other caregiver types (e.g., older caregivers to relatives or nonkin others) are typically overlooked (Lapierre & Keating, 2013; Marino et al., 2020). The focus on spousal caregivers reflects the hierarchies of obligation framework, which provides a rationale for care patterns based on the relationship closeness (Penning, 1990). However, a lack of study on older caregivers to other family members or friends not only limits our understanding of their caregiving experience but also fails to recognize their contribution (Lapierre & Keating, 2013).

Furthermore, our knowledge of social well-being of older caregivers is relatively limited compared to caregiver burden and psychological well-being. Social participation and social support stemming from active social interactions and networks of relationships are central features of successful aging. However, both social participation and support are not well documented for older caregivers, considering that caregiving is negatively related to social participation and social support (Greenwood et al., 2019; Li & Lee, 2020). Also, social participation and social support have rarely been examined using longitudinal data that can specifically map trajectories over time to capture the fluidity of care, social connectedness, and aging (Hawkley et al., 2020). Previous caregiving studies have primarily relied on static cross-sectional data or convenience samples, which limits the inferences and generalization that can be drawn (Kaufman et al., 2019). Modeling transitions in caregiving status or progression over time are a preferable approach to specifying interactions and disentangling bidirectional effects of trajectories of caregiving on the health and social well-being outcomes among caregivers, which is particularly beneficial when there may be lag effects (Liu & Lou, 2019). Furthermore, it is critical to assess caregiving patterns and relationships at key transition points, especially the early period of the caregiving role. This is essential to understand the onset and temporality of the caregiving role for effective and timely program and policy development (Ducharme et al., 2011; Williams et al., 2021).

Given these knowledge gaps, this study aims to investigate the association between transitions into three caregiver role types (spousal caregiver, nonspouse family caregiver, nonkin caregiver) among older adults and two significant social outcomes (social participation and social support) using two waves of data drawn from the Canadian Longitudinal Study on Aging (CLSA).

Theoretical Framework

This study is guided by an applied role strain theory, which outlines how the potential stress and conflict arising from performing multiple roles (Goode, 1960; Reid & Hardy, 1999). Role strain theory has been widely applied in studies related to life course transitions, and the consequences of managing multiple roles (e.g., familial, social, and productive roles), both of which are integral to caregiving patterns among older adults (Polenick & DePasquale, 2019). Role strain arises from multiple reasons, including the potential contradictions or ambivalence created by multiple and sometimes incompatible or ambiguous role expectations, and the conflict of allocation of finite time, place, or resources to manage different roles (Goode, 1960), a hallmark of caregiving (Wister et al., 2021). To manage multiple roles and potential role overload, people adopt the “inter-domain transition” strategy (Matthews et al., 2014) to shift resources from one domain, such as social life, to another, such as caregiving. Thus, it is reasonable to expect that older adults prioritize caregiving responsibilities and allocate their time and energy from social life to undertake caregiving tasks. As a result, older adults experience reduced social participation after becoming caregivers. Also, the linkage between multiple roles and social support is multiple facets. In caregiving situation, due to the role demand and exhaustion from performing caregiving tasks, it becomes hard for older caregivers to maintain their existing social network where considerable social support stems from. Older caregiver may also find it hard to learn new information or knowledge to seek support from formal resources.

Social Participation

Social participation is defined as an individual’s engagement in various activities which enable interaction with others in the community or society (Goll et al., 2015). Generally, it is challenging for older adults to care for others although maintaining active participation in social or leisure activities, due to their own health conditions (e.g., chronic illness) or age-related frailty (Carabante et al., 2018). With respect to caregiver role types, some studies have shown that spousal caregivers experience greater restrictions in social activities compared to other types of caregiver roles (Carabante et al., 2018; Ott et al., 2007). Spousal caregivers tend to spend most of their time at home with their care receivers, due to feelings of guilt about leaving them home alone or avoidance of embarrassing behaviors outside of home (Carabante et al., 2018). Other research does not support the disparity of social participation among different caregiver role types (Li & Wister, 2021), or finds no change in engagement in social activities among older spousal caregivers over time (Queen et al., 2019). George and Gwyther (1986) found that the impact of caregiving on social participation is related to the type of social activities. In their study, spousal caregivers reported the lowest level of club or organization attendance, whereas adult–child caregivers showed the most restriction in visiting family and friends, but no difference regarding church attendance or hobby-related activities. Taken together, the evidence pertaining to associations between types of caregiver roles and social participation remains equivocal.

Social Support

Social support refers to the emotional, instrumental, or informational resources people receive to cope with stress or other life events (Cohen, 2004). The research findings on social support across caregiver role types also vary. Based on a meta-analysis, Pinquart and Sörensen (2011) suggested that spousal caregivers reported less informal/unpaid support than adult–child caregivers. One potential reason is that spousal caregivers have a stronger sense of duty or pride, which prevents them from reaching out for help (Greenwood et al., 2019). However, another study indicated that spousal caregivers are more likely to receive informal/unpaid support than adult-child caregivers or other relative/friend caregivers (Rogero-García & Rosenberg, 2011). Other studies have examined different types of social support, and the findings are domain specific. For instance, Bierhals et al. (2019) revealed that spousal caregivers to older adults tend to receive more help from friends and have better personal relationships with others than nonspousal caregivers, but nonspousal caregivers received more financial support from others. In addition, Galvin et al. (2010) also found that spousal caregivers attend more support groups than nonspousal caregivers, but receive fewer other tangible services (e.g., meal services).

In light of these research gaps and inconsistent findings, the present research applies longitudinal data and examines the associations between the onset of different caregiver role types among older adults and their social participation and social support outcomes over time.

Thus, based on the existing literature and role strain theory, we propose to test Hypothesis 1: Older adults experience reduced levels of social participation and social support after becoming caregivers. In addition, as suggested by the existing evidence on the higher caregiving workload among spousal caregivers than other types of caregiver roles, we also propose to examine Hypothesis 2: Spousal caregivers will experience more decline in social participation and social support than nonspouse family caregivers and nonkin caregivers.

Method

Data and Sample

In this study, participants were selected from the baseline and follow-up 1 (FUP1) of the CLSA. The CLSA is a national-level population-based panel survey focusing on social, physiological, biological, and clinical health and well-being of Canadians aged 45–85 years old from 2011. Currently, two waves of CLSA data apart from approximately three years are available, including the Baseline (51,338 participants, 2011–2015) and FUP1 (44,817 participants, 2015–2018). The CLSA is comprised of two cohorts of participants, the Comprehensive cohort who were randomly selected among population residing within 25 km (or 50 km in a lower population density area) of the 11 data collection sites (Victoria, Vancouver, Surrey, Calgary, Winnipeg, Hamilton, Ottawa, Montréal, Sherbrooke, Halifax, and St. John’s) across Canada, and the Tracking cohort who were randomly selected from the ten provinces by the computer-assisted interview system. Besides the survey interview, the Comprehensive cohort participants also visit the data collection sites to undertake physical examinations and provide biological specimens. The CLSA initially contacted about 152,000 potential participants based on the provincial health registries and previous national health studies (i.e., the Canadian Community Health Survey). As a result, the CLSA received roughly 83,600 consent forms for further contact, which led to 51,338 participants successfully completing the CLSA Baseline questionnaire (response rate 61.5%). Detailed information about the CLSA has been published elsewhere (Kirkland et al., 2015; Raina et al., 2009, 2019). Researchers can access the de-identified data and information on weighting through the CLSA website (https://www.clsa-elcv.ca/).

The unit of analysis in this study contains older adults who transit into caregiver roles during the two CLSA data collection time points (i.e., noncaregiver at Baseline and caregiver at FUP1). At the Baseline, there were 21,491 older adults, of whom 13,001 did not identify them as caregivers. During the FUP1, among the 13,001 noncaregiver older adults from Baseline, 3,986 became caregivers by providing support to family members or friends during the past 12 months prior to the survey. A total of 3,789 participants out of the 3,986 were included as the final sample, excluding those whose main care receivers had passed away before the FUP1 survey.

Measurement

Dependent variables

The dependent variables are social participation and social support. Social participation was measured by eight questions inquiring about the frequency of participating in community-based activities (SPA), including family or friendship-based activities outside the household, church or religious activities, sports or physical activities, educational and cultural activities, service club or fraternal organization activities, neighborhood, community or professional association activities, volunteer or charity work, and any other recreational activities involving other people, including hobbies, gardening, and other games. A 5-point Likert scale was used to measure the frequency from Never (1) to At least once a day (5) (Cronbach’s alpha = 0.61 at Baseline, Cronbach’s alpha = 0.64 at FUP1). Social participation variable was summed up based on those eight items, and a higher number means greater social participation. This SPA scale was originally developed for the Canadian Community Health Survey and adopted by the CLSA. A variety of studies have been conducted based on the SPA scale and supported its validity and reliability by establishing the relationship between social participation and depressive symptoms, loneliness, health care utilization, etc. (e.g., Chamberlain et al., 2022; Wang et al., 2021).

Social support was measured using the Medical Outcome Study (MOS) Social Support Survey (Sherbourne & Stewart, 1991). The MOS consists of 19 items capturing four aspects of social support from both informal and formal sources, including emotional/information support (i.e., empathetic understanding, and the encouragement of expressions of feelings; the offering of advice, information and so on), affectionate support (i.e., expression of love and affection), tangible support (i.e., material aid or behavioral assistance), and positive social interaction (i.e., the availability of others to do enjoyable and relaxing things; Sherbourne & Stewart, 1991). Each item was measured based on a 5-point Likert scale from None of the time (1) to All of the time (5), and the social support variable was calculated based on these 19 items (Cronbach’s alpha = 0.89 at Baseline, Cronbach’s alpha = 0.89 at FUP1). A higher score on the MOS scale indicates a higher level of social support.

Independent variable

The independent variable is types of caregivers, including spousal caregivers, nonspouse family caregivers, and nonkin caregivers. The caregiver status was captured by the question “During the past 12 months, have you provided any of the following types of assistance to another person because of a health condition or limitation?” Any participant providing at least one type of support (personal care, medical care, transportation, meal preparation, etc.) to another person was identified as a caregiver. This reference question was asked at both Baseline and FUP1, which indicated the participants’ caregiving status at two data collection time points. Participants who provided support to another person needed to answer one more question regarding their relations with the main care receiver they dedicated the most time and resources to assist: “What is the relationship between you and this person? Is s/he your...?” For this study, the relations were further categorized into three groups, including spousal caregivers (husband/wife, and common-law partner), nonspouse family caregivers (parent, child, sibling, grandchild, in-laws, other family members), and nonkin caregivers (friends, neighbor, or others). Thus, the group of caregivers included in this study are unpaid and informal caregivers.

Covariates

The data analysis also included several key demographic and socioeconomic social determinants of health as covariates, as well as functional and health-related factors essential to social participation and social support among both family caregivers and the general aging population (Mikkonen & Raphael, 2010; Schulz et al., 2016). The demographic and socioeconomic factors were all extracted from the Baseline, including age (65–74 years old and 75 years and older), gender (male and female), marital status (married/common-law and not married), living arrangement (living with someone and living alone), education level (less than postsecondary education and postsecondary education), employment status (retired/unemployed and employed), personal annual income (less than $20,000, $20,000–$49,999, and $50,000 or more), and country of birth (Canada and foreign born).

Another three functions and health-related factors from both Baseline and FUP1 were also adjusted in the data analysis, including Older Americans’ Resources and Services Multidimensional Functional Assessment Questionnaire (basic activities of daily living [ADLs] and instrumental activities of daily living [IADLs; Fillenbaum & Smyer, 1981], and chronic illness). The ADLs and IADLs both contain seven items related to daily activities, such as dress and walk for ADLs, and use telephone and grocery shopping for IADLs. All the items were measured at three levels, from completely unable (0) to completely able (2). The ADLs and IADLs variables were summed based on each set of items, and higher scores indicate better functional capability. The chronic illness was represented by the number of 27 types of chronic diseases collected in the CLSA, such as Alzheimer’s disease, back problems, ulcers, urinary incontinence, etc. The number of reported chronic conditions was counted, and a higher number means severer chronic illness.

In addition to the participants’ information, survey wave (Baseline and FUP1) and participant cohorts (Comprehensive and Tracking) were included as survey-related factors in to adjust the effect of time and sampling strategies.

Data Analytic Procedure

The SPSS 26 software package was used for data analysis. Descriptive statistics of all studied variables were generated for the total sample and the three types of caregivers (Table 1). Group comparisons based on three types of caregivers were conducted (Table 1). The between- and within-group differences based on three types of caregivers from Baseline to FUP1 were further examined (Table 1). To analyze the association between different types of caregivers and social participation and social support, longitudinal multivariate data analyses were conducted using the linear mixed models (LMM; Brown & Prescott, 2015). LMM is an extension of simple linear models to allow both fixed and random effects. LMM has been developed for analyzing panel data because it accounts for the nesting of repeated time points within individuals; the result is that it models the variation over time within individuals as well as the variation across individuals. During the modeling, survey wave was included in the data analysis to examine the change of outcomes from Baseline to FUP1. Interaction terms between all studied variables and survey wave were added in the analysis to examine the association between the independent variables and the change of outcome variables over the survey.

Table 1.

Demographic and Socioeconomic Background of Participants based on Caregiver Role Types (N = 3,789)

	All participants, %/Mean (SD)	Spousal caregivers, %/Mean (SD) (a)	Nonspouse family caregivers, %/Mean (SD) (b)	Nonkin caregivers, %/Mean (SD) (c)	χ2(df)/F-test
Age groups (Baseline)					21.55 (2) ***
65–74 years old	66.95	64.50	71.12	65.81
75 years and older	33.05	35.50	28.88	34.19
Gender (Baseline)					17.82 (2) ***
Male	48.15	53.34	43.43	47.84
Female	51.85	46.66	56.57	52.16
Marital status (Baseline)					384.93 (2) ***
Not married	28.64	4.32	30.79	43.52
Married/common law	71.36	95.68	69.21	56.48
Living arrangement (Baseline)					326.86 (2) ***
Living with someone	75.23	97.94	74.22	60.68
Living alone	24.77	2.06	25.78	39.32
Education (Baseline)					4.43 (2)
Without postsecondary education	53.33	49.08	52.18	56.96
Postsecondary diploma/degrees	46.67	50.92	47.82	43.04
Employment status (Baseline)					1.09 (2)
Retired/unemployed	90.66	89.80	91.55	90.64
Employed	9.34	10.20	8.45	9.36
Personal annual income (Baseline)					15.06 (4) **
Less than $20,000	22.93	24.03	29.31	17.92
$20,000–$49,999	46.26	43.08	41.51	51.58
$50,000 and more	24.39	26.62	22.21	24.36
Not stated	6.41	6.28	6.96	6.14
Country of birth (Baseline)					33.42 (2) ***
Foreign born	18.80	19.62	11.06	23.43
Canada	81.20	80.38	88.94	76.57
ADLs
Baseline	13.87 (0.52)	13.85 (0.55)	13.88 (0.53)	13.88 (0.49)	0.55
FUP1	13.78 (0.70)	13.82 (0.57)	13.76 (0.74)	13.76 (0.75)	1.36
IADLs
Baseline	13.89 (0.47)	13.86 (0.51)	13.91 (0.41)	13.89 (0.48)	1.29
FUP1	13.78 (0.69)	13.75 (0.69)	13.77 (0.70)	13.81 (0.68)	0.59
Chronic conditions
Baseline	3.66 (2.20)	3.89 (2.11)	3.52 (2.15)	3.61 (2.28)	2.77
FUP1	4.43 (2.44)	4.57 (2.34)	4.31 (2.33)	4.43 (2.58)	1.35
SPA
Baseline	19.37 (4.98)	18.75 (5.15)	18.99 (4.91)	20.05 (4.82)	14.56***, a, b < c
FUP1	19.01 (4.95)	18.24 (4.87)	18.88 (4.76)	19.62 (5.05)	20.60***, a < b < c
t-Test	2.18*	4.78***	−0.66	2.51*
MOS
Baseline	82.61 (17.60)	85.96 (14.59)	83.60 (17.73)	79.62 (18.88)	19.81***, a > b > c
FUP1	82.01 (17.81)	82.53 (16.84)	82.92 (18.29)	80.97 (18.10)	1.70
t-Test	1.52	5.31***	0.62	−0.86
Caregiver type (FUP1)		—	—	—	—
Spousal caregiver	28.66
Nonspouse family caregiver	28.60
Nonkin caregiver	42.75

Notes: N = 3,789; CLAS = Canadian Longitudinal Study on Aging; MOS = Medical Outcome Study Social Support Survey; SD = standard deviation; SPA = social participation activities.

^* p < .05.

^** p < .01.

^*** p < .001.

Two sets of LMM were performed for social participation and social support (Table 2, Figures 1 and 2). In Model 1, survey wave, demographic and socioeconomic factors, function, and health-related factors, and all the interactions with survey wave were included. The caregiver type variable and its interaction with the survey wave were added to Model 2. The demographic and social–economic variables, function and health-related variables, caregiver type variables, and the survey wave were included as fixed effects. A random intercept was included to model the variation in the dependent variable outcomes across participants. Likelihood ratio tests based on the Akaike Information Criterion was performed to compare the model fit. Also, the same analyses were conducted for the four domains of MOS social support scale, and results are illustrated in Supplementary Tables 2 and 3.

Table 2.

Linear Mixed Models for Social Participation and Social Support Among Older Adults Transitioning Into Caregiver Roles

	SPA		MOS
	Model 1 estimate/95% CI	Model 2 estimate/95% CI	Model 1 estimate/95% CI	Model 2 estimate/95% CI
Survey wave (Baseline) FUP1	3.30 [−2.63, 9.23]	3.05 [−2.87, 8.97]	−15.17 [−35.89, 5.56]	−16.29 [−36.92, 4.34]
Cohorts (tracking) Comprehensive	1.05*** [0.61, 1.49]	1.07*** [0.64, 1.51]	−2.97*** [−4.49, −1.44]	−2.99*** [−4.51, −1.47]
Cohorts * Survey wave (tracking) Comprehensive	0.70*** [0.38, 1.01]	0.69*** [0.38, 1.00]	−0.54 [−1.66, 0.57]	−0.62 [−1.73, 0.49]
Age groups (75 years and older) 65–74 years old	−0.35 [−0.82, 0.12]	−0.37 [−0.84, 0.10]	0.37 [−1.26, 1.99]	0.13 [−1.50, 1.76]
Age groups × survey wave (75 years and older) 65–74 years old	0.60*** [0.26, 0.94]	0.54** [0.20, 0.88]	0.63 [−0.57, 1.84]	0.43 [−0.77, 1.64]
Gender (female) Male	−1.24*** [−1.73, −0.76]	−1.26*** [−1.74, −0.78]	−4.20*** [−5.87, −2.54]	−4.30*** [−5.97, −2.64]
Gender × survey wave (female) Male	−0.13 [−0.48, 0.22]	−0.14 [−0.49, 0.21]	0.06 [−1.17, 1.29]	−0.01 [−1.24, 1.22]
Marital status (married/common law) Not married	−0.94* [−1.76, −0.12]	−1.23** [−2.07, −0.40]	−8.19*** [−11.05, −5.33]	−8.77*** [−11.70, −5.84]
Marital status × survey wave (married/common law) Not married	0.80* [0.19, 1.41]	0.61 [−0.02, 1.24]	2.04 [−0.12, 4.20]	−0.98 [−1.22, 3.18]
Living arrangement (Living alone) Living with someone	−0.33 [−1.20, 0.55]	−0.12 [−0.99, 0.75]	5.33*** [2.28, 8.38]	5.21*** [2.16, 8.27]
Living arrangement × survey wave (living alone) Living with someone	0.73* [0.07, 1.38]	0.75* [0.09, 1.40]	0.78 [−1.54, 3.11]	1.15 [−1.17, 3.47]
Education (postsecondary diploma/degrees) Without postsecondary education	−0.84*** [−1.28, −0.40]	−0.88*** [−1.32, -0.44]	−1.12 [−2.65, 0.41]	−1.04 [−2.57, 0.49]
Education × survey wave (postsecondary diploma/degrees) Without postsecondary education	0.20 [−0.12, 0.52]	0.20 [−0.12, 0.52]	−0.01 [−1.13, 1.12]	−0.08 [−1.20, 1.04]
Employment status (employed) Retired/unemployed	1.04** [0.31, 1.77]	1.01** [0.28, 1.73]	2.60* [0.07, 5.12]	2.58* [0.06, 5.09]
Employment status × survey wave (employed) Retired/unemployed	−0.21 [−0.74, 0.33]	−0.22 [−0.76, 0.31]	−0.10 [−2.00, 1.79]	−0.22 [−2.10, 1.67]
Personal annual income ($50,000 and more) Less than $20,000 $20,000–$49,999	−2.11*** [−2.80, −1.42] −0.71* [−1.27, −0.15]	−2.08*** [−2.76, −1.40] −0.75** [−1.30, −0.19]	−2.11 [−4.49, 0.27] −1.53 [−3.47, 0.40]	−2.16 [−4.54, 0.21] −1.52 [−3.45, 0.40]
Personal annual income * survey wave ($50,000 and more) Less than $20,000 $20,000–$49,999	−0.03 [−0.53, 0.46] −0.34 [−0.74, 0.07]	−0.06 [−0.56, 0.44] −0.36 [−0.77, 0.04]	−0.20 [−1.95, 1.55] 0.42 [−1.01, 1.85]	−0.30 [−2.04, 1.44] 0.26 [−1.16, 1.69]
Country of birth (Canada) Foreign born	−0.23 [−0.81, 0.34]	−0.37 [−0.94, 0.20]	0.39 [−1.59, 2.37]	0.57 [−1.42, 2.56]
Country of birth × survey wave (Canada) Foreign born	−0.60** [−1.03, −0.17]	−0.59** [−1.01, −0.16]	−0.43 [−1.83, 1.07]	−0.57 [−2.08, 0.93]
ADLs	0.30* [0.01, 0.60]	0.30 [−0.01, 0.60]	1.83*** [0.80, 2.87]	1.93*** [0.90, 2.96]
ADLs × survey wave	−0.22 [−0.57, 0.14]	−0.20 [−0.55, 0.15]	−0.69 [−1.92, 0.53]	−0.73 [−1.94, 0.49]
IADLs	0.46* [0.08, 0.84]	0.48* [0.11, 0.86]	−1.24 [−2.58, 0.10]	−1.25 [−2.58, 0.09]
IADLs × survey wave	−0.10 [−0.49, 0.29]	−0.12 [−0.51, 0.27]	1.73* [0.36, 3.11]	1.72* [0.35, 3.09]
Chronic conditions	−0.08 [−0.16, 0.01]	−0.08* [−0.16, −0.01]	−0.18 [−0.46, 0.11]	−0.18 [−0.46, 0.10]
Chronic conditions × survey wave	−0.07* [−0.15, −0.01]	−0.06 [−0.14, 0.01]	−0.28* [−0.54, −0.03]	−0.24 [−0.50, 0.01]
Caregiving type (spousal caregiver) Nonspouse family caregiver Nonkin caregiver		0.11 [−0.46, 0.69] 1.36*** [0.80, 1.92]		2.18* [0.19, 4.18] 0.92 [−1.01, 2.85]
Caregiving type × survey wave (spousal caregiver) Nonspouse family caregiver Nonkin caregiver		0.60** [0.20, 1.01] 0.50* [0.10, 0.90]		2.36** [0.94, 3.78] 3.59*** [2.20, 4.99]
AIC	24,000.38	23,954.59	33,787.21	33,743.32

Notes: Reference group is listed in (--); N = 3,789. ADLs = activities of daily living; AIC = Akaike Information Criterion; CI = confidence interval; CLAS = Canadian Longitudinal Study on Aging; IADLs = instrumental activities of daily living; MOS = Medical Outcome Study Social Support Survey; SPA = social participation activities.

^* p < .05.

^** p < .01.

^*** p < .001.

Figure 1.

Social participation over time.

Figure 2.

Social support over time.

Among the studied variables, personal income has a significant number of missing observations (6%), which were included as a separate category and retained in the analysis as “not stated.” All other variables have minimum missing values. LMM has the function to handle the missing data for different waves on the outcome variables via restricted maximum-likelihood estimation; and listwise deletion was used for independent variables (e.g., demographic factors) with missing cases. In addition, as requested by CLSA, the inflation weight (v1.2) was applied to generate descriptive statistics, and the analytic weight (v1.2) was applied for bivariate and multivariate analyses.

Results

As shown in Table 1, among the 3,789 participants, the majority were aged between 65 and 74 years old (67%), female (52%), married or living with a common-law partner (71%), living with someone (75%), without a postsecondary diploma/degree (53%), retired (91%), receiving personal income between $20,000 and $49,999 (46%), and born in Canada (81%) at the Baseline. All participants had high levels of functional capability in ADLs and IADLs at both Baseline and FUP1. Also, participants reported an average of 3.7 types of chronic diseases at Baseline, and 4.4 types at the FUP1 out of 27.

Table 1 also presents the social participation and social support levels at both Baseline and FUP1 for all participants and each type of caregiver role. From Baseline to FUP1, 28.7% of participants become spousal caregivers, 28.6% become nonspouse family caregivers, and 42.8% become nonkin caregivers. Among all participants, a significant decrease from Baseline to FUP1 was found in SPA score (t = 2.18, p < .05) but not MOS score. When it comes to each type of caregiver role, both spousal caregivers (t = 4.78, p < .001) and nonkin caregivers (t = 2.51, p < .05) reported significantly lower SPA scores at FUP1 than Baseline. And only spousal caregivers (t = 5.31, p < .001) reported lower MOS scores at FUP1 compared to Baseline. Also, nonkin caregivers reported the highest SPA score at both Baseline (F = 14.56, p < .001) and FUP1 (F = 20.60, p < .001), but the lowest MOS score at Baseline (F = 19.81, p < .001). Conversely, spousal caregivers reported the lowest SPA scores at both Baseline and FUP1, but the highest MOS score at Baseline. The spousal caregivers, nonspouse family caregivers, and nonkin caregivers also differed in most demographic and socioeconomic factors (age group, gender, marital status, living arrangement, personal income, and country of birth), but none of the functional and health-related factors (see Table 1 for detailed information).

The results of LMM for social participation and social support are listed in Table 2, and only results from Model 2 with all studied variables are presented later. When all the covariates and independent variables were adjusted, survey wave was not significantly associated with SPA or MOS scores. Nonkin caregivers reported a higher SPA score than spousal caregivers at Baseline (estimate = 1.36, 95% CI: [0.80,1.92]). Also, the interactive effect between caregiver type and survey wave was significant, and both nonspouse family caregivers (estimate = 0.60, 95% CI: [0.20, 1.01]) and nonkin caregivers (estimate = 0.50, 95% CI: [0.10, 0.90]) had a greater increase in SPA scores than spousal caregivers from Baseline to FUP1. In addition, nonspouse family caregivers reported a higher MOS score than spousal caregivers at Baseline (estimate = 2.18, 95% CI: [0.19, 4.18]). The significant interactive effect between caregiving type and survey wave indicated that nonspouse family caregivers (estimate = 2.36, 95% CI: 0.94, 3.78) and nonkin caregivers (estimate = 3.59, 95% CI: [2.20, 4.99]) reported a greater increase in MOS scores from Baseline to FUP1. Refer to Figures 1 and 2 for the plotted changes of social participation and social support among three types of caregivers from Baseline to FUP1. Therefore, the results did not support Hypothesis 1, and Hypothesis 2 was fully supported.

Most of the covariates were associated with social participation and/or social support concurrently or longitudinally. Participants aged 65–74 years old reported a higher SPA score than those aged 75 years and older at Baseline (estimate = 0.54, 95% CI: [0.20, 0.88]). Male participants reported lower scores of SPA (estimate = −1.26, 95% CI: [−1.74, −0.78]) and MOS (estimate = −4.30, 95% CI: [−5.97, −2.64]) than females at Baseline. Similarly, participants who were not married or partnered (estimate = −1.23, 95% CI: [−2.07, −0.40] for SPA, and estimate = −8.77, 95% CI: [−11.70, −5.84] for MOS) reported lower scores. Participants living with someone reported higher SPA over time (estimate = 0.75, 95% CI: [0.09, 1.40]) and MOS scores at Baseline (estimate = 5.21, 95% CI: [2.16, 8.27]). A lower level of education was negatively associated with the SPA score at Baseline (estimate = −0.88, 95% CI: [−1.32, −0.44]). Participants who retired reported higher SPA (estimate = 1.01, 95% CI: [0.28, 1.73]) and MOS (estimate = 2.58, 95% CI: [0.06, 5.09]) scores at Baseline than those who were employed. Personal income was associated with SPA score at Baseline, and participants with lower income reported lower SPA scores than those with an annual income of $50,000 and more (estimate = −2.08, 95% CI: [−2.76, −1.40] for less than $20,000, and estimate = −0.75, 95% CI: [−1.30, −0.19] for $20,000–$49,999). Foreign born participants reported lower SPA scores than Canadian born ones over time (estimate = −0.59, 95% CI: [−1.01, −0.16]). ADLs were positively associated with MOS score at Baseline (estimate = 1.93, 95% CI: [0.90, 2.96]). IADLs were positively related to SPA score at Baseline (estimate = 0.48, 95% CI: [0.11, 0.86]), and MOS score over time (estimate = 1.72, 95% CI: [0.35, 3.09]). The number of chronic conditions was negatively correlated with SPA score at Baseline (estimate = −0.08, 95% CI: [−0.16, −0.01]).

Discussion

This study adds to the limited knowledge of social participation and social support among older caregivers by comparing transitions into one of three types of caregiver roles (spousal caregivers, nonspouse family caregivers, and nonkin caregivers) using two waves of data from the CLSA. The two main findings include (a) after transitioning into caregiver role, spousal caregivers, and nonkin caregivers experienced declines in social participation, and spousal caregivers also received less social support; and (b) spousal caregivers reported a greater decline in social participation and social support than nonspouse family caregivers and nonkin caregivers after becoming caregivers.

Spousal caregivers are most likely to experience reduced social participation and social support compared to nonspouse family caregivers and nonkin caregivers, which is consistent with some studies (Ott et al., 2007; Pinquart & Sörensen, 2011; Tanji et al., 2005). There are several interpretations of this finding. First, it is well documented that spousal caregivers generally provide the most intensive caregiving compared to other caregivers, leading to limited time or energy for out-of-home social activities (Carabante et al., 2018; Li & Lee, 2020; Li et al., 2021). Thus, spousal caregivers may find it challenging to balance caregiving roles and other out-of-home roles (e.g., volunteer, sports club member, etc.), and transfer most of their resources to supporting the care recipient. Also, spousal caregivers usually assume the primary caregiver role, although other caregivers normally provide secondary assistance (Lima et al., 2008). Therefore, spousal caregivers may find it hard to disengage from their caregiving responsibilities, even temporarily for respite services, due to their affection and obligation (Choi et al., 2007). Also, older spousal caregivers tend to experience worse mental health issues (depression, anxiety, etc.), and poorer health behaviors than other caregivers (Li & Lee, 2020), which may compromise social participation. In addition, besides caring for their spouses or partners, spousal caregivers also need to perform the housework or other family-related responsibilities previously performed by care receivers.

Older adults also reported the most decline in social support after becoming spousal caregivers than nonspouse family caregivers and nonkin caregivers. Some researchers have pointed out that older spousal caregivers may refuse to admit that they need support due to their preference to maintain autonomy and independence in caregiving (Dam et al., 2017). Older spousal caregivers also insist on undertaking the caregiving responsibilities by themselves to avoid being a burden to their children, or to protect the privacy of the care receiver, particularly when intimate care is needed (Dam et al., 2017). This situation is also reflected by the fact that most spousal caregivers are solo caregivers (55%; Ornstein et al., 2019). Also, spouses and partners are the main sources of support for each other. Once one spouse or partner becomes the care receiver due to chronic illness, accident, or other age-related frailty, the other one (spousal caregiver) loses the most significant source of support. This can also explain the fact that in our study, spousal caregivers report the highest level of social support before becoming caregivers, although the social support level declines rapidly after care onset for their spouses.

Another notable finding in our study is that nonkin caregivers also experienced a significant decline in social participation after shouldering caregiving responsibilities. Nonkin support has been understood as occasional, but more recent research has recognized that nonkin caregiving can be long-term (e.g., over 1 year) or even permanent, including providing support in daily life, intimate care, and end of life care among the aging population (Burns et al., 2011; Pleschberger & Wosko, 2017). When someone needs continuous care from a nonkin friend or neighbor, they tend to live alone (single or widowed), have no child around, or even have no active social network to provide the necessary support (Li & Lee, 2020). Therefore, nonkin caregivers may need to provide considerable help and care, particularly when the health of care receivers declines (Pleschberger & Wosko, 2017). Also, nonkin caregivers tend to provide instrumental support, including transportation and grocery shopping (Himes & Reidy, 2000). Those instrumental activities usually entail a light workload but are time consuming, which may affect the opportunity to attend or participate in other social activities. In addition, due to the transition into caregiver role, nonkin caregivers may notice the nature of interactions with their care receivers changes from social gathering or friendship activities into a kind of supportive behaviors. This shifting nature of interaction may also explain the declined social participation from the perspective of nonkin caregivers.

In our study, most of the social–demographic, function, and health-related factors are associated with social participation and social support. New older caregivers who are younger (65–74 years old), female, married or partnered, living with someone, with postsecondary education, retired, and receiving higher levels of income reported higher levels of social participation and social support. These findings are consistent with previous literature emphasizing the salience of socioeconomic status, and accessibility to social resources in successful aging and maintaining good health and well-being among caregivers (Mikkonen & Raphael, 2010; Pinquart & Sörensen, 2011). Finally, older caregivers with better functional capability in ADLs and IADLs, and fewer chronic diseases also reported higher levels of social participation and social support. This finding resonates with previous caregiving studies (Li et al., 2021; Schulz et al., 2016), and further emphasizes the significance of promoting the health and well-being of older caregivers to enable them to better manage both caregiving roles and other out-of-home roles.

Limitations

This study has several limitations worthy of clarification. First, the CLSA does not collect information related to prior caregiving history. Fast et al. (2020) have pointed out that from middle-age, individuals may become caregivers several times, such as caring for aging parents during the middle-age and supporting spouse or siblings entering old-age. Therefore, it is possible that those older adults with prior caregiving experience have higher resilience in managing caregiving responsibilities and maintaining active social participation and networking. Additionally, the proportion of nonkin caregivers was higher than spousal or nonspouse family caregivers in our study compared to the broader literature (Wolff et al., 2018). However, this finding may reflect the trend that increasing care is demanded and provided among nonkin networks due to sociodemographic trends related to smaller family sizes and greater mobility of the younger generation (Broese van Groenou & De Boer; 2016; Lapierre & Keating, 2013; Pleschberger et al., 2019). Another explanation is related to the way of identifying caregivers in different surveys. Based on the CLSA, informal caregivers are defined based on whether they provide any type of support to another person. Some other surveys or studies may have more restrictive criteria, such as a cutoff on the caring hours per week to define caregiving role, for instance more than one hour per week as observed in the Canada General Social Survey (Statistics Canada, 2013). As a result, some surveys may rule out individuals, usually nonkin others (e.g., friends, neighbors), who provide occasional support to others, such as pet feeding, snow shoveling, and other tasks supporting independent living.

In addition, some caregiving contextual factors are not included in data analyses due to the design of CLSA survey or the data requirement of LMM. The CLSA did not collect information related to the care receivers’ health and care needs, as well as the availability of other caregivers. Both factors are relevant to caregivers’ health and well-being, and well documented in other studies (e.g., Lin et al., 2012). In addition, caregiving hours per week or the caregiver–receiver living arrangement was not included in data analyses, although relevant information is available in the CLSA (see Supplementary Table 1 for further information). The decision to not include caregiving hour and caregiver–receiver living arrangement is made based on the requirement of LMM. To conduct LMM with two waves of data, valid data at Baseline or both Baseline and FUP1 are required, but caregiving hour is set at zero at Baseline and caregiver–receiver living arrangement is captured at FUP1. However, the caregiver role types can reflect these two factors because spousal caregivers are more likely to spend longer hours on caring and living with care receivers than nonspouse family caregivers and nonkin caregivers (Li & Lee, 2020). Another study focusing on continuous caregivers (based on the CLSA) has been conducted thorough examination of the relationship between caregiving hours, caregiver–receiver living arrangement, and social isolation (Li et al., 2021).

In addition, the CLSA did not collect information regarding the types of support caregivers provide to their main care receivers. Thus, it creates a challenge to examine the distribution of various types of caregivers (particularly nonkin caregivers) by different types and levels of support. Earlier studies have suggested that nonkin caregivers tend to engage in socializing and nonintimate instrumental tasks, although recent studies also revealed the trend of increasing engagement in intimate and end of life care among nonkin caregivers to older adults (Barker, 2002; Burns et al., 2011; Pleschberger & Wosko, 2017). Therefore, more comprehensive examination is needed to further identify the caring pattern among nonkin caregivers in future studies. An additional limitation is that the data were collected before the COVID-19 pandemic, thus the findings do not reflect the situation during the pandemic.

Implications

Our understanding of older caregivers regarding their health and well-being has lagged behind the growing prevalence of older caregivers. As older adults increasingly play a tremendous role in caring for others to supplement the public health care and social service system, caregiving needs to be considered as an essential component during the aging process. Indeed, caregiving consequences in the social domains may be as critical as the physical or mental health realms when considering the protective functions of active social interaction and social relationships for successful aging. The outbreak of the COVID-19 pandemic also emphasizes the salience of studying the social participation and support among older caregivers. Older caregivers face the double jeopardy of isolation and loneliness due to being disproportionately affected by the pandemic and providing care beyond their usual caregiving role (Cohen et al., 2020; Wister et al., 2022).

The findings also support the necessity to provide programs and services during the transitional stage to help new caregivers prepare for caregiving responsibilities. A previous intervention program discovered that new caregivers receiving the training program presented higher levels of self-efficacy and confidence, and better knowledge of available services and coping strategies (Ducharme et al., 2011). Yet, additional innovative programs and services to enable social participation and social support among older caregivers should be developed and evaluated, particularly targeting spousal and nonkin caregivers. Due to the influence of the pandemic, Internet or Internet-based support system may be widely needed, and those Internet-based programs have been approved as effective in helping caregivers maintain contact with family members and develop a new network.

Conclusion

This study demonstrates that older adults experience a greater decline in social participation and social support when becoming spousal caregivers compared to nonspousal family caregivers and nonkin. The findings of this study reinforce the need for the development, evaluation and implementation of innovative programs and services during the early transition stage for new caregivers, particularly spousal and nonkin caregivers, to better balance the caregiving role and other out-of-home roles. These initiatives will foster better quality caregiving experiences during a period in which caregiving is the backbone of the community care system.

Funding

Funding for the Canadian Longitudinal Study on Aging (CLSA) is provided by the Government of Canada through the Canadian Institutes of Health Research under grant reference: LSA 94473 and the Canada Foundation for Innovation, as well as the following provinces, Newfoundland, Nova Scotia, Quebec, Ontario, Manitoba, Alberta, and British Columbia. This research has been conducted using the CLSA data set Baseline Tracking v3.4, Baseline Comprehensive v4.0, Follow-up 1 Tracking v1.0, and Follow-up 1 Comprehensive v1.0, under Application Number 1909016. The CLSA is led by Drs. Parminder Raina, Christina Wolfson, and Susan Kirkland. The opinions expressed in this manuscript are the author’s own and do not reflect the views of the Canadian Longitudinal Study on Aging. The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by Social Sciences and Humanities Research Council of Canada, Insight Development Grants (File Number: 430-2021-00526).

Conflict of Interest

None declared.

Ethics Approval and Consent to Participate

This current project received ethics approval at two levels. Consent to participate was obtained for all participants under the CLSA harmonized multi-university ethics process approved by the Hamilton Integrated Research Ethics Board (HiREB), Hamilton Health Sciences/McMaster University. Written consent was obtained from all CLSA participants prior to enrollment. Individuals who were not deemed to be cognitively functional were excluded from the CLSA study. Simon Fraser University (SFU) was a participating institution in the CLSA data collection, and the SFU Office of Research Services Ethics Committee reviewed all consent material prior to data collection (SFU ORS #2010s0281).

Author Contributions

L. Li: main writer of manuscript. A. V. Wister, Y. Lee, and B. Mitchell: secondary authors made substantial contributions to the conception, design and editing of the manuscript. All authors have read and approved the final version of the manuscript and have agreed to be accountable for all parts.

Conflicts of Interest

The authors have declared that no competing interests exist. The opinions expressed in this manuscript are the authors' own and do not reflect the views of the Canadian Longitudinal Study on Aging.

Data Availability

Data are available from the Canadian Longitudinal Study on Aging (www.clsa-elcv.ca) for researchers who meet the criteria for access to de-identified CLSA data.