Change in Social Engagement among Incident Caregivers and Controls: Findings from the Caregiving Transitions Study

Abstract

Objectives:We aimed to compare incident family caregivers and matched controls on change in social network variables and satisfaction with social activities. Methods: A total of 479 participants from the Caregiving Transitions Study were included in the analysis, 244 (50.9%) of whom began substantial and sustained caregiving between baseline and follow-up interviews. We assessed associations between incident caregiving and change in social network size, change in monthly social contact, and satisfaction with social activities. Results: Incident caregivers reported lower satisfaction with social activities and greater decline in social network size than controls. Among only the caregivers, dementia caregivers reported greater decline in social network size while caregivers who were spouses, experienced higher strain, and provided more hours of care reported lower satisfaction with social activities. Discussion: Social engagement declined more among incident caregivers than controls. Future studies should aim to understand why subgroups of caregivers reported lower social engagement.

Keywords

incident caregiving social activities social network social contact

Introduction

Social engagement becomes increasingly important over the life course. As one ages, a series of transitions such as retirement, relocation, loss of loved ones, and the onset of chronic conditions may lead to the loss of existing ties to one’s social network or the creation of new ones (Cornwell et al., 2008, 2014). Taking on responsibilities to provide ongoing care for a family member or friend with a disability is another transition that is associated with aging and can affect one’s engagement in social activities. Several trends have been observed with respect to how social relationships tend to change in later life. Older adults who experience relocation tend to lose non-family social network members while strengthening ties to family members (Badawy et al., 2019). Others have shown that older adults sometimes experience growth in the size of their social networks due to the creation of weak ties to adapt to the loss of strong ones (Cornwell et al., 2014). Given the mixed evidence, it may require additional context about an individual older adult’s life to understand how social engagement changes in later life.

Carstensen’s socioemotional selectivity theory posits that older people are more selective in their social interactions, with a focus on close relationships rather than extensive casual relationships (Carstensen, 1991, 1992), which may explain the observed reduction in social engagement as one ages. For older adults, loneliness resulting from a lack of social engagement may lead to adverse physical and emotional outcomes such as higher depressive symptom scores, more functional limitations, and higher mortality risk (Cacioppo et al., 2010; Perissinotto et al., 2012), as well as declines in cognition, increased risk of Alzheimer’s disease, and other physical problems (Institute of Medicine, 2015). Evidence from these studies highlight the importance of maintaining social engagement in later life, which may be particularly important for older adults who are providing care for other older adults with chronic conditions since they have less time to participate in social activities (Clark & Bond, 2000).

Previous work has shown that social engagement is positively associated with various aspects of well-being in caregivers. In a study of female caregivers, more social ties independently predicted decreased risk of depressive symptoms (Cannuscio et al., 2004). A study using the National Health and Aging Trends Study found that higher caregiver social engagement was associated with better self-reported health for the care recipient (Kelley et al., 2017). Furthermore, social support and activity are significantly correlated with lower depression, higher life satisfaction, and self-rated health in studies of white and African American caregivers of persons with dementia (Clay et al., 2008; Haley et al., 1987). Another study found that caregivers’ social capital, which consists of social contacts that provide access to social, emotional, and practical support, is associated with lower caregiving-related burden (Papastavrou et al., 2015).

Prior studies have sought to understand differences in social engagement between caregivers and controls and demographic variables moderating this relationship. Older adults who are caregivers have less time to participate in social activities compared to those who are not (Clark & Bond, 2000), and this has important implications for the well-being of both caregivers and care recipients. Haley et al. (1995) found that both African American and white caregivers indicated lower levels of social activity and lower levels of satisfaction with the quality of their social support than noncaregivers. Another report found that satisfaction with leisure activities was lower among caregivers of persons recovering from stroke than noncaregivers, an effect that persisted over time (Haley et al., 2015). Studies on race differences on social engagement in caregivers have shown mixed results. Several studies have shown that social support networks for African American caregivers are more diverse in the types of relationships than those of white caregivers (Cox, 1993; Smerglia et al., 1998), while Burgio et al. (2003) found that African American caregivers reported greater declines in satisfaction with social support over time than white caregivers. In these studies, caregiving status of participants at baseline may confound outcomes since caregivers who enter a study may differ from noncaregiving controls on other important demographic or health-related variables. Furthermore, a review by Dilworth-Anderson et al. (2002) found that a majority of the studies on social engagement among caregivers use convenience or purposive sampling designs, which could yield nonrepresentative findings.

Certain subgroups of caregivers may be at higher risk of decreased social engagement than others. To our knowledge, studies have not assessed differences in aspects of social engagement between dementia caregivers and nondementia caregivers, spouse and nonspouse caregivers, or caregivers reporting varying levels of strain. Prior evidence has shown that dementia caregivers and spouse caregivers face higher burden and depression associated with the death of the care recipient than nondementia and nonspouse caregivers, respectively (Pinquart & Sörensen, 2003, 2011), and that spouse caregivers of persons with dementia are at higher risk of loneliness than nonspouse caregivers (Adams, 2008; Eloniemi-Sulkava et al., 2002). Furthermore, a report found that spouse caregivers had lower levels of social support and smaller social networks than adult child caregivers for persons with dementia (Rigby et al., 2019). Caregivers with higher strain reported lower health-related quality of life in both physical and mental domains (Litzelman et al., 2014) as well as fewer social contacts (Roth et al., 2009). Worse health-related outcomes in dementia caregivers, spouse caregivers, and caregivers experiencing high strain suggest that they may be at higher risk for decreased social engagement than nondementia caregivers, nonspouse caregivers, and caregivers reporting low or no strain, respectively.

In this study, we aimed to understand how social engagement changes after beginning substantial and sustained caregiving, a significant life transition that can induce changes in routine activities and existing social networks. In the primary analysis, we assess whether there is greater longitudinal decline in self-reported social network size and social contact as well as lower self-reported satisfaction with the amount of time spent on social activities among incident caregivers, defined as those who transitioned into a caregiving role between baseline and follow-up study assessments, than noncaregiving controls. We also assessed interactions between incident caregiving and age, sex, or race on social engagement. Last, we aimed to identify subgroups of caregivers at the highest risk of social isolation by assessing how the relationship of the caregiver to the care recipient, dementia caregiving status, and caregiving strain are associated with incident caregivers’ social engagement.

Design and Methods

REGARDS

The Reasons for Geographic and Racial Differences in Stroke (REGARDS) study is a national longitudinal investigation of 30,329 adults in the United States, with enrollment taking place between 2003 and 2007. Participants were randomly sampled from a commercially available nationwide list and were over 45 years of age, white or African American, free of previous diagnoses of cancer requiring chemotherapy, and not living in or on the waiting list for a nursing home (NH). African Americans and residents of the Stroke Belt (the states of Alabama, Arkansas, Georgia, Louisiana, Mississippi, North Carolina, South Carolina, and Tennessee) were oversampled by design (Howard et al., 2005). Briefly, participants were sent a letter and contacted by telephone using a computer-assisted telephone interview (CATI); following verbal consent, trained interviewers obtained demographic information, medical history, and indices of quality of life. Demographic information included questions on caregiving status and family relationships. Three to 4 weeks after the CATI, a brief in-home physical exam including blood pressure measurements and blood samples was conducted and written informed consent was obtained. A second extensive CATI and in-home assessment were conducted a median of 9.4 years after the first one, between 2013 and 2016 (Howard et al., 2018).

Approximately 12 years after enrollment, REGARDS participants were queried by telephone to obtain updated information on caregiving status and family relationship variables. Each participant was asked, “Are you currently providing care on an ongoing basis to a family member, friend, or neighbor with a chronic illness or a disability? This would include any kind of regular help with basic activities such as dressing, bathing, grooming this person, managing bills, arranging for medical care, watching or supervising this person, or providing transportation.” Participants who answered “yes” were potentially eligible to be enrolled into the Caregiving Transitions Study (CTS) as incident caregivers, while those who answered “no” to this question at the REGARDS baseline and follow-up caregiver screening were potentially eligible to be enrolled into CTS as noncaregiving controls. Participants who answered “yes” to this updated caregiving screening question were later contacted by telephone to determine if they met the study’s inclusion and exclusion criteria for being a caregiver.

CTS

Detailed information on the design, sampling, and enrollment procedures of the CTS have been described elsewhere (Roth et al., 2020). Briefly, potentially eligible incident caregivers from REGARDS were contacted by telephone, and trained interviewers surveyed them on when they began providing care, whether they had been continuously providing care since that time, their relationship with the care recipient, whether their care recipient currently or previously lived in a NH or assisted living facility (ALF), and the number hours of care they provide on a typical week. Incident caregivers were only eligible if they transitioned into the caregiving role at least 6 months after the first REGARDS in-home assessment and at least 3 months before the second REGARDS in-home assessment, if their care recipient was over 18 years of age and did not reside in an NH, ALF, or other residential care setting, if they provided at least 5 hours of care per week, if they lived with or within 50 miles of the care recipient, and if they provided usable blood samples at both of the REGARDS in-home assessments. Of 546 potential incident caregivers who were contacted, 298 were determined to be eligible; 47 out of those 298 declined participation, leaving 251 participants to be enrolled as incident caregivers in CTS (Roth et al., 2020).

Incident caregivers enrolled in CTS were surveyed on additional questions about the care recipient including age, dementia status, whether they lived together, and other major health problems. Incident caregivers were also asked about the strain that they had experienced: “In general, how much of a mental or emotional strain is it on you to provide this care? Would you say no strain, some strain, or a lot of strain?”

After enrolling incident caregivers, lists of up to five potentially eligible noncaregiving control participants were generated for each incident caregiver from 10,254 eligible noncaregivers in REGARDS. The noncaregiving control must have matched the enrolled caregiver on up to seven demographic and health history factors (age ±5 years, sex, race, education level, marital status, self-rated health at baseline, and self-reported history of serious cardiovascular disease) and must have provided usable blood samples at both REGARDS in-home assessments. Additionally, potential noncaregiving controls who were matched to spouse caregivers had to be married, and potential controls matched to adult child caregivers had to have at least one living parent. Potential controls were randomly called and screened for eligibility and were excluded if they reported serving as a family caregiver at any point during their participation in REGARDS. Calls continued until one participant was deemed eligible and provided verbal consent to participate in the project, and a total of 251 noncaregiving controls were enrolled to match the 251 incident caregivers. REGARDS and CTS were approved by the institutional review boards (IRBs) of all participating institutions.

Among 502 participants enrolled in CTS, our analysis excluded those with missing data on social network score or monthly social contact score (n = 9). We also excluded those missing data on satisfaction with social activities (n = 14). Our final analytic sample included 479 CTS participants, 244 (50.9%) of whom were incident caregivers (Figure 1).

Figure 1.

Flowchart of inclusion criteria for participants.

Variables

Age was collected by self-report at the CTS interview. Sex, race, education level, income, and urban-rural status were collected by self-report at REGARDS baseline. Education was defined as a categorical variable with four categories: less than high school, high school graduate, some college, and college graduate and above. Annual income was defined as less than $20,000, $20,000–$34,000, $35,000–$74,000, and above $75,000. Urban–rural status was defined using the Rural/Urban Commuting Area classification at the census tract level and categorized as urban (≥75% urban), mixed (25%–75% urban), or rural (≤25% urban) (Howard et al., 2017). Among incident caregivers, dementia status of the care recipient was coded as a binary variable and hours of caregiving per week as a continuous variable. Relationship to the care recipient (spouse, adult child, other) and self-reported caregiving strain (“no strain,” “some strain,” and “a lot of strain”) were coded as categorical variables.

Data on incident caregivers’ social network and social contact were collected at the REGARDS baseline CATI and the CTS interview. The size of participants’ social network was measured using the sum of the responses from the first two questions of a 3-item questionnaire based on the Revised Lubben’s Social Network Scale (LSNS): “(1) How many close friends do you have? That is, people that you feel at ease with, can talk to about private matters, and can call for help? (2) How many relatives do you have that you feel close to?” Monthly social contact was measured using the third question on the Revised LSNS: “How many of these friends or relatives do you see at least once a month?” (Lubben, 2008). Numeric responses for each question were converted to a score on an ordinal scale according to LSNS: a response of 0, 1, or 2 was coded as 0, 1, or 2, respectively; a response of 3 or 4 was coded as 3, a response between 4 and 8 was coded as 4, and a response greater than 8 was coded as 5. Change in social network and change in social contact were measured by subtracting the REGARDS baseline score from the CTS interview score.

Data on satisfaction with social activities were derived from a broader measure of satisfaction with leisure activities collected at the CTS interview and was measured on an ordinal scale of 0 = “not at all satisfied,” 1 = “a little satisfied,” or 2 = “a lot satisfied.” This questionnaire measured the participant’s satisfaction with the amount of time spent over the past month in activities that the participant enjoys, in quiet time alone, attending church or going to other meetings of groups or organizations, taking part in hobbies or interests, going out for meals or other social activities, doing fun things with other people, or visiting with family and friends (Stevens et al., 2004). This scale was found to have high validity when compared against satisfaction with support, social networks, and the well-being subscore of the Centers for Epidemiological Studies Depression Scale (Stevens et al., 2004). Scores from all items on the questionnaire other than “in quiet time alone” and “taking part in hobbies and interests” were added to obtain a summary score from a five-item scale of activities that involve social interactions (range = 0–10; Cronbach’s alpha = .82).

Statistical Analysis

The independent variable is incident caregiving, defined as having become a caregiver between the baseline REGARDS assessment and the CTS assessment. Matched noncaregiving controls were not caregivers at either interview. We used three multiple linear regression models to separately assess the associations between incident caregiving and (1) the change in social network score from the baseline REGARDS assessment to the CTS assessment, (2) change in the number of social network members with whom the participant has monthly contact, and (3) satisfaction with social activities at the CTS assessment. All multiple linear regressions were adjusted for age, sex, race, education level, income categories, and urban–rural status. The REGARDS baseline assessment score was also included as a covariate for the analyses of change in social network score and change in social contact. We also assessed whether age, sex, or race modified the relationship between incident caregiving and social engagement by adding interaction terms in subsequent analytic models.

To understand the influence of care recipient’s dementia status, care recipient’s relationship to the caregiver, and caregiving strain on social engagement, we conducted separate analyses on only the incident caregivers. We ran three multiple linear regressions using the change in social network score, change in monthly social contact score, and satisfaction with social activities as the outcomes, respectively. Multiple linear regressions were adjusted for age, sex, race, education level, income categories, urban–rural status, relationship to care recipient, dementia status of care recipient, caregiving strain, and hours of caregiving per week.

Results

Table 1 shows demographic characteristics and descriptive means for social network score, monthly social contact score, and satisfaction with social activities. Due to the matching process implemented in CTS (Roth et al. 2020), caregivers and controls were closely comparable on demographic characteristics. Incident caregivers and controls at the time of the CTS interview were both around 72 years of age on average. In both groups, over one-third were African Americans, two-thirds were women, less than half had graduated from college, and nearly two-thirds lived in urban areas. Among caregivers, 47% had care recipients with dementia. Other care recipients had a stroke (n = 16), cancer (n = 6), heart disease (n = 15), other weakness or physical disability (n = 22), mental symptoms (n = 6), or another condition (n = 64). About half of the caregivers were spouses or cohabiting nonmarried significant others of care recipients. Caregivers provided 43.1 hours of care per week on average (SD: 29.3), and one-quarter (n = 63; 25.8%) reported experiencing a lot of strain.

Table 1.

Demographic Characteristics, Social Network Score at Baseline and Follow-up, and Social Activities at Follow-up of Caregivers and Controls.

Mean (SD) or N (%)	Caregivers (N = 244)	Controls (N = 235)
Age	71.8 (8.1)	71.7 (7.7)
Race (African American)	87 (35.7%)	86 (36.6%)
Sex (female)	159 (65.2%)	157 (66.8%)
Education level
Less than high school	9 (3.7%)	10 (4.3%)
High school graduate	56 (23.0%)	54 (23.0%)
Some college	75 (30.7%)	63 (26.8%)
College graduate/above	104 (42.6%)	108 (46.0%)
Income1
Less than $20k	19 (7.8%)	19 (8.1%)
$20k–$34k	47 (19.3%)	44 (18.7%)
$35k–$74k	103 (42.2%)	93 (39.6%)
$75k and above	55 (22.5%)	61 (26.0%)
Urban–rural status2
Urban	160 (65.6%)	157 (66.8%)
Mixed	28 (11.5%)	32 (13.6%)
Rural	33 (13.5%)	22 (9.4%)
Care recipient has dementia	115 (47.1%)	—
Relationship to care recipient		—
Spouse	125 (51.2%)
Adult child	62 (25.4%)
Other	57 (23.4%)
Caregiving strain		—
No strain	35 (14.3%)
Some strain	146 (59.8%)
A lot of strain	63 (25.8%)
Hours of caregiving per week	43.1 (29.3)	—
Total social network score3
Baseline	7.3 (2.0)	7.2 (1.8)
Follow-up	6.3 (2.2)	6.9 (2.1)
Change	−.95 (2.2)	−.31 (2.1)
Monthly social contact4
Baseline	3.7 (1.3)	3.7 (1.3)
Follow-up	3.2 (1.5)	3.5 (1.5)
Change	−.46 (1.5)	−.26 (1.6)
Satisfaction with social activities	6.4 (2.9)	8.3 (2.0)

Social network score categorized based on Lubben’s Social Network Scale: 1 = 1, 2 = 2, 3–4 = 3, 4–8 = 4, and >8 = 5.

Participants missing data for income: n = 20 for caregivers and n = 18 for controls.

Participants missing data for urban–rural status: n = 23 for caregivers and n = 24 for controls.

Total number of friends and relatives that participant feels close to.

Number of friends or relatives seen at least once a month.

From the REGARDS baseline visit to the CTS interview, the social network scores and monthly social contact scores for both incident caregivers and controls declined. Social network scores for incident caregivers decreased by approximately one unit, while among controls the score decreased by one-third of a unit; monthly social contact scores decreased by approximately half of a unit among incident caregivers and by one-quarter of a unit among controls (Figure 2). At the CTS interview, the mean satisfaction with social activities was 1.9 points lower for incident caregivers (mean = 6.4, SD = 2.9) than for controls (mean = 8.3, SD = 2.0).

Figure 2.

Unadjusted change (mean, SD) in social network and contact scores from REGARDS baseline visit to Caregiving Transitions Study baseline visit (approximately 12 years).

After adjusting for covariates, incident caregivers had significantly greater decline in social network score (−.62; SE = .18; p value < .01) but only marginally greater decline in monthly social contact (−.24; SE = .13; p value = .06) than noncaregiving controls. Incident caregivers had significantly lower satisfaction with social activities (−1.95; SE = .23; p value < .01) than controls (Table 2). This association was also observed for the full leisure satisfaction scale that included these five social activities and two nonsocial activities (satisfaction with time spent “in quiet time alone” and “taking part in hobbies and interests”) as well as for a subscale consisting of only the two nonsocial activities. None of the interaction tests were statistically significant (all p values > .10); therefore, age, sex, and race were all not found to modify the relationship between incident caregiving and social engagement.

Table 2.

Multivariable Regression of Change in Social Network, Change in Monthly Social Contact, and Satisfaction with Social Activities (N = 479).

	Total change in social network score1		Change in monthly social contact score2		Satisfaction with social activities
	Estimate (SE)	p value	Estimate (SE)	p value	Estimate (SE)	p value
Age	.005 (.01)	.67	.0007 (.009)	.93	.03 (.02)	.11
Sex (female)	.17 (.20)	.40	.05 (.14)	.73	.21 (.26)	.41
Race (African American)	−.03 (.20)	.87	−.17 (.14)	.23	.27 (.26)	.30
Education level
Less than high school	Ref	Ref	Ref	Ref	Ref	Ref
High school graduate	.24 (.49)	.62	−.05 (.35)	.85	.10 (.63)	.87
Some college	.19 (.48)	.70	−.11 (.35)	.67	−.21 (.63)	.74
College graduate and above	−.10 (.48)	.84	−.38 (.34)	.27	.08 (.62)	.90
Income
Less than $20k	Ref	Ref	Ref	Ref	Ref	Ref
$20k–$34k	.24 (.38)	.52	.09 (.27)	.73	−.19 (.49)	.70
$35k–$74k	.48 (.36)	.18	.24 (.25)	.35	−.008 (.46)	.99
$75k and above	.30 (.38)	.42	.08 (.27)	.76	−.48 (.49)	.33
Urban–rural status
Urban	Ref	Ref	Ref	Ref	Ref	Ref
Mixed	.39 (.28)	.17	.01 (.20)	.94	−.71 (.36)	.05
Rural	.03 (.29)	.93	−.08 (.21)	.69	−.24 (.38)	.53
Baseline total social network score	−.50 (.05)	<.01	—	—	—	—
Baseline monthly social contact score	—	—	−.61 (.05)	<.01	—	—
Caregiving status
(ref: control)	−.62 (.18)	<.01	−.24 (.13)	.06	−1.95 (.23)	<.01

Total number of friends and relatives that participant feels close to.

Number of friends or relatives seen at least once a month.

Among incident caregivers, those caring for care recipients with dementia had less decline in social network score compared to those caring for care recipients without dementia (.63; SE = .28; p value = .02), and every additional hour of care per week was associated with greater decline in social network score (−.01; SE = .005; p value = .04), after adjusting for covariates. Caregivers in the “other” category who were neither spouses nor adult children had significantly higher satisfaction with social activities than spouse caregivers (1.37; SE = .46; p value < .01). Caregivers reporting high strain had significantly lower satisfaction with social activities than those reporting no strain (−1.39; SE = .62; p value = .02) and more hours of care was associated with lower satisfaction with social activities (−.02; SE = .006; p value < .01). Older age was associated with higher satisfaction with social activities among incident caregivers (.06; SE = .03; p value = .04) but sex, race, education, and income were not. Participants who lived in mixed (25%–75% urban) census tracts had lower satisfaction with social activities than participants who lived in urban (≥75% urban) census tracts (−1.64; SE = .53; p value < .01) (Table 3).

Table 3.

Multivariable Regression of Change in Social Network, Change in Monthly Social Contact, and Satisfaction with Social Activities only among Caregivers (N = 244).

	Total change in social network score1		Change in monthly social contact score2		Satisfaction with social activities
	Estimate (SE)	p value	Estimate (SE)	p value	Estimate (SE)	p value
Age	−.02 (.02)	.33	−.01 (.01)	.42	.06 (.03)	.04
Sex (female)	.22 (.30)	.46	.15 (.21)	.46	.55 (.39)	.16
Race (African American)	.27 (.30)	.38	−.15 (.21)	.47	.42 (.40)	.30
Education level
Less than high school	Ref	Ref	Ref	Ref	Ref	Ref
High school graduate	.27 (.75)	.71	.28 (.52)	.59	.90 (.98)	.36
Some college	.16 (.74)	.83	.17 (.51)	.74	−.24 (.97)	.80
College graduate and above	−.04 (.74)	.96	−.13 (.51)	.81	.03 (.98)	.98
Income
Less than $20k	Ref	Ref	Ref	Ref	Ref	Ref
$20k–$34k	−.17 (.55)	.76	.09 (.38)	.81	−.06 (.73)	.94
$35k–$74k	.33 (0.53)	.54	.25 (.37)	.50	−.34 (.70)	.63
$75k and above	.22 (.58)	.70	.14 (.40)	.73	−.80 (.76)	.29
Urban-rural status
Urban	Ref	Ref	Ref	Ref	Ref	Ref
Mixed	−.08 (.44)	.86	−.52 (.30)	.09	−1.64 (.57)	<.01
Rural	.05 (.40)	.90	−.16 (.28)	.57	−.72 (.53)	.17
Baseline total social network score	−.52 (.07)	<.01	—	—	—	—
Baseline monthly social contact score	—	—	−.58 (.07)	<.01	—	—
Care recipient has dementia	.63 (.28)	.02	.32 (.19)	.10	−.59 (.37)	.11
Relationship to care recipient
Spouse	Ref	Ref	Ref	Ref	Ref	Ref
Adult child	−.58 (.37)	.12	−.30 (.26)	.25	.77 (.49)	.12
Other	.42 (.35)	.23	.31 (.24)	.20	1.37 (.46)	<.01
Caregiving strain
No strain	Ref	Ref	Ref	Ref	Ref	Ref
Some strain	.23 (.40)	.56	.17 (.27)	.53	−.18 (.52)	.73
A lot of strain	−.17 (.47)	.71	−.12 (.32)	.71	−1.39 (.62)	.02
Hours of care per week	−.01 (.005)	.04	−.002 (.003)	.46	−.02 (.006)	<.01

Total number of friends and relatives that participant feels close to.

Number of friends or relatives seen at least once a month.

Discussion and Implications

In this analysis of social engagement among incident caregivers and matched controls, we found that incident caregivers had greater decline in total social network score, marginally greater decline in monthly social contact score, and lower satisfaction with social activities than noncaregiving controls. Among incident caregivers, no differences in either the change in social network or social contact were observed by relationship to the care recipient, while dementia caregivers had less decline in social network size compared to nondementia caregivers after covariate adjustment. Spouse caregivers, high-strain caregivers, and caregivers providing more hours of care per day had lower satisfaction with social activities than caregivers who were neither spouses nor adult children, caregivers who experienced no strain, and caregivers providing fewer hours of care per day, respectively, after covariate adjustment.

Social engagement for older adults has become a particularly timely issue during the COVID-19 pandemic. The lack of respite options for caregivers of older adults who can no longer stay at long-term care facilities may lead to increased stress and fewer opportunities to engage in social or leisure activities. This may exacerbate trends from studies conducted prior to the pandemic showing that caregivers experience lower satisfaction with leisure activities than noncaregiving controls (Haley et al., 2015). Our study corroborated these findings in a subsample of REGARDS participants with exclusion criteria that ensured all caregivers provided a minimum of 5 hours of care per week, showing that satisfaction with social activities was lower among incident caregivers than controls. Incident caregivers likely had less time to participate in social activities due to caregiving responsibilities, which may explain why our finding of lower satisfaction with social activities among caregivers also held for the two nonsocial activities on the full leisure satisfaction scale.

Although our findings showed that caregivers experienced greater decline in total social network size and marginally greater decline in monthly social contact than controls, this may not necessarily translate to worse health outcomes for caregivers. Studies have shown that the quality of relationships is more predictive of outcomes such as mortality, physical and mental health than the quantity of relationships (Antonucci et al., 1997; 2013). Marginally significant results for monthly social contact suggest that caregivers may have remained in frequent contact (i.e., at least monthly) with a similar number of people despite the decline in social network size, prioritizing the quality rather than the quantity of relationships in their social network. More evidence is needed on whether social engagement is associated with changes in caregivers’ health over time.

Another study from REGARDS found that caregivers who experience high strain have fewer social contacts than noncaregivers (Roth et al., 2009). Other analyses of CTS data have shown that dementia caregivers provide, on average, more hours of care than nondementia caregivers (Sheehan et al., 2020). After adjusting for this greater level of care in our study, dementia caregivers did not differ from nondementia caregivers on satisfaction with social activities, and dementia caregivers actually reported less decline in social network size compared to nondementia caregivers. Regardless of the type of caregiving, we found that caregivers who provided more hours of care reported lower levels of satisfaction with leisure activities, greater decline in social network size but no difference in the frequency of social contact. This is likely because these caregivers have less time to spend on social activities with most members of their social networks but maintained social contacts related to caregiving support. Furthermore, our findings and those of Sheehan et al. (2020) from the CTS suggest that previous findings of dementia caregivers’ worse subjective caregiving experiences than those of nondementia caregivers (Ory et al., 1999) may be explained by the greater quantity of care (e.g., hours, type of problems) that dementia caregivers typically provide.

Few studies have assessed differences in the change in social network or the level of satisfaction with social activities by relationship to the care recipient. In a study by Fredman et al. (2006), both spouse caregivers and nonspouse caregivers were found to spend less time on leisure exercise compared to married noncaregivers, and the effect was nonsignificantly larger in spouse caregivers. This is consistent with our findings showing that spouse caregivers experience lower satisfaction with social activities than caregivers who were neither spouses nor adult children, a group that likely experiences the least burden and caregiving exposure. A study using the Berkman–Syme Social Network Index found that female caregivers with high spousal care time commitment and few social ties were more likely to report depressive symptoms than those with no spousal care responsibilities and many social ties; the effect was smaller for women who cared for their parents (Cannuscio et al., 2004). In our study, spouse caregivers did not experience greater decline in social network score or monthly social contact score than caregivers who were neither spouses nor adult children. Further study is needed to determine whether there are any interactions between social network score and relationship to care recipient on incident caregivers’ health outcomes.

In our study, neither sex nor race moderated the effect of incident caregiving on the change in social network score or satisfaction with social activities. Previous work with convenience samples has found that African American caregivers have more diverse types of relationship in their social networks (Ajrouch et al., 2001; Cox 1993; Smerglia et al., 1998). This discrepancy may be attributed to the different types of study samples; CTS used a population-based sample with controls matched on demographic characteristics, whereas most previous studies have used convenience samples that may have recruited minority caregivers who were healthier at baseline. In our study, older age was associated with higher satisfaction with social activities among incident caregivers. This is consistent with findings in a study with both caregivers and noncaregivers showing that older age is associated with psychological aspects of quality of life, life satisfaction, and satisfaction with leisure activities (Haley et al., 2015). These findings suggest that older caregivers have better appraisals of caregiving experience and higher measures of well-being, independent of other demographic variables.

We found that caregivers residing in mixed areas, defined as areas that are 25%–75% urban, reported significantly lower satisfaction with social activities than caregivers in urban areas. This effect was marginally significant in the entire sample. It is unclear what might explain these findings, since not a lot of research has been done in caregivers residing in mixed areas. As such, this group may be especially understudied in addition to having lower measures of social well-being. Since the ability to virtually connect with others may be less accessible to certain subpopulations due to existing geographic disparities in broadband access and digital literacy in the United States (Pew Research Center, 2013), more research is needed to understand how the relationship between geographic region and caregivers’ social well-being changes post-pandemic.

Our study has several strengths. Prior to CTS, few studies have examined outcomes in older adults who became caregivers while participating in a longitudinal population-based study. CTS enrolled caregivers who were actively providing substantial and sustained care as well as controls who reported no history of providing care, individually matched on multiple demographic and health-related variables. Prior evidence on the negative impact of family caregiving primarily comes from cross-sectional studies of convenience samples, whereas our analyses aim to leverage the population-based sample of CTS to provide a more rigorous understanding of social engagement among incident caregivers and controls. The Revised LSNS distinguished the total number of members in participants’ social networks from the members with whom the participant had monthly contact, an important distinction in assessing the quality of relationships. This allowed us to understand the extent to which participants felt close to members of their social network in addition to the actual contact made with those members.

Our study also has some limitations. Social network variables did not provide any information on the type of contact that participants had with members. The contacts could have been related to support for caregiving task or for leisure, which may have different implications for various domains of the caregivers’ well-being. This may present a barrier to translating findings into strategies to improve caregivers’ social support. Furthermore, we did not have longitudinal data for satisfaction with social activities that would allow us to assess how incident caregivers’ satisfaction with time spent on activities and interests changed over time compared to controls. As such, it was unclear whether incident caregivers had lower satisfaction with social activities at baseline or had experienced a decline in satisfaction with social activities after becoming a caregiver.

Our findings highlight the public health importance of social engagement among caregivers, particularly subgroups such as caregivers with higher exposure, including dementia caregivers and caregivers providing more hours of care per week. Though it should be noted that our study included only white and African American participants and our findings may not be generalizable to other racial or ethnic groups, there are several broad implications for practice and policy. As part of caregiver education, clinicians should regularly check in with caregivers, whether during online or in-person medical visits, emphasize the importance of maintaining close social relationships, and connect them to support groups or communities when needed (Dang et al., 2020). Policymakers should provide increased funding for caregiver respite programs to allow caregivers to take more time to attend to their own needs, such as by expanding the number of employers covered under the Families First Coronavirus Response Act and providing paid leave for more caregivers. Previous research has shown that an intervention for spouse caregivers of persons with dementia that increased other family members’ engagement with caregiving also increased caregivers’ satisfaction with their social support network, which then mediated the positive impact of the intervention on caregivers’ depressive symptoms (Roth et al., 2005) and on delayed NH placement for the care recipients (Mittelman et al., 2006). Our findings also provide support for interventions involving social engagement, a critical next step for addressing social isolation in older adults (Institute of Medicine, 2015). Future studies should aim to understand the purpose of different types of social connections maintained by caregivers and the types of social activities that caregivers engage in, which would allow communities to provide resources to facilitate the maintenance of those connections and involvement in those activities, respectively. In the context of COVID-19 and physical distancing measures, researchers should aim to understand whether and how caregivers are maintaining social connections virtually. An understanding of these aspects of social engagement is imperative to supporting family caregivers, especially as more are needed to care for the aging population in the coming years.

Footnotes

Acknowledgments

The authors thank the other investigators, the staff, and the participants of the REGARDS study for their valuable contributions. A full list of participating REGARDS investigators and institutions can be found at .

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

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 a cooperative agreement [U01 NS041588] co-funded by the National Institute of Neurological Disorders and Stroke (NINDS) and the National Institute on Aging (NIA), National Institutes of Health, Department of Health and Human Services and by an investigator-initiated grant [RF1 AG050609] from the NIA. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NINDS or NIA. Representatives of the NINDS were involved in the review of the manuscript but were not directly involved in the collection, management, analysis, or interpretation of the data.

ORCID iDs

Chelsea Liu

Chanee D. Fabius

References

Adams

K. B.

(2008). Specific effects of caring for a spouse with dementia: Differences in depressive symptoms between caregiver and noncaregiver spouses. International Psychogeriatrics, 20(3), 508–520.

Ajrouch

K. J.

Antonucci

T. C.

Janevic

M. R.

(2001). Social networks among blacks and whites: the interaction between race and age. The journals of gerontology. Series B, Psychological sciences and social sciences, 56(2), S112-S118.

Antonucci

T. C.

Ajrouch

K. J.

Birditt

K. S.

(2013). The convoy model: Explaining social relations from a multidisciplinary perspective. The Gerontologist, 54(1), 82–92. doi: 10.1093/geront/gnt118

Antonucci

T. C.

Fuhrer

Dartigues

J. F.

(1997). Social relations and depressive symptomatology in a sample of community-dwelling French older adults. Psychology and aging, 12(1), 189-195.

Badawy

P. J.

Schafer

M. H.

Sun

(2019). Relocation and Network Turnover in Later Life: How Distance Moved and Functional Health Are Linked to a Changing Social Convoy. Research on Aging, 41(1), 54-84. doi: 10.1177/0164027518774805

Burgio

Stevens

Guy

Roth

D. L.

Haley

W. E.

(2003). Impact of two psychosocial interventions on White and African American family caregivers of individuals with dementia. The Gerontologist, 43(4), 568-579.

Cacioppo

J. T.

Hawkley

L. C.

Thisted

R. A.

(2010). Perceived social isolation makes me sad: 5-year cross-lagged analyses of loneliness and depressive symptomatology in the Chicago health, aging, and social relations study. Psychology and Aging, 25, 453–463.

Cannuscio

C. C.

Colditz

G. A.

Rimm

E. B.

Berkman

L. F.

Jones

C. P.

Kawachi

(2004). Employment status, social ties, and caregivers’ mental health. Social science & medicine (1982), 58(7), 1247-1256.

Carstensen

L.L.

(1991). Socioemotional selectivity theory: Social activity in life-span context. Annual Review of Gerontology and Geriatrics, 11, 195-217.

10.

Carstensen

L. L.

(1992). Social and emotional patterns in adulthood: Support for socioemotional selectivity theory. Psychology and Aging, 7, 331-338.

11.

Clark

M. S.

Bond

M. J.

(2000). The effect on lifestyle activities of caring for a person with dementia. Psychology, Health & Medicine, 5, 13–27. doi:10.1080/135485000105972

12.

Clay

O. J.

Roth

D. L.

Wadley

V. G.

Haley

W. E.

(2008). Changes in social support and their impact on psychosocial outcome over a 5-year period for African American and White dementia caregivers. International journal of geriatric psychiatry, 23(8), 857-862. doi:10.1002/gps.1996

13.

Cornwell

Laumann

E. O.

Schumm

L. P.

, (2008). The social connectedness of older adults: a national profile*. American Sociological Review, 73(2), 185–203. doi:10.1177/000312240807300201

14.

Cornwell

Schumm

L. P.

Laumann

E. O.

Kim

Y. J.

(2014). Assessment of social network change in a national longitudinal survey. The Journals of Gerontology. Series B, Psychological Sciences and Social Sciences, 69(Suppl 2), S75–S82. doi:10.1093/geronb/gbu037

15.

Cox

(1993). Service needs and interests: A comparison of African American and White caregivers seeking Alzheimer’s assistance. American Journal of Alzheimer’s Care and Related Disorders & Research, 8(3), 33-40.

16.

Dang

Penney

L. S.

Trivedi

Noel

P. H.

Pugh

M. J.

Finley

Pugh

J. A.

Van Houtven

C. H.

Leykum

(2020). Caring for caregivers during COVID-19. Journal of the American Geriatrics Society. Advance online publication. doi:10.1111/jgs.16726

17.

Dilworth-Anderson

Williams

I. C.

Gibson

B. E.

(2002). Issues of race, ethnicity, and culture in caregiving research: A 20-year review (1980-2000). The Gerontologist, 42(2), 237-272.

18.

Eloniemi-Sulkava

Rahkonen

Suihkonen

Halonen

Hentinen

Sulkava

(2002). Emotional reactions and life changes of caregivers of demented patients when home caregiving ends. Aging & Mental Health, 6(4), 343-349.

19.

Fredman

Bertrand

R. M.

Martire

L. M.

Hochberg

Harris

E. L.

(2006). Leisure-time exercise and overall physical activity in older women caregivers and noncaregivers from the caregiver-SOF study. Preventive Medicine, 43(3), 226-229.

20.

Haley

W. E.

Levine

E. G.

Brown

S. L.

Bartolucci

A. A.

(1987). Stress, appraisal, coping, and social support as predictors of adaptational outcome among dementia caregivers. Psychology and Aging, 2, 323-330.

21.

Haley

W. E.

Roth

D. L.

Hovater

Clay

O. J.

(2015). Long-term impact of stroke on family caregiver well-being: A population-based case-control study. Neurology, 84(13), 1323-1329. doi:10.1212/WNL.0000000000001418

22.

Haley

W. E.

West

C. A.

Wadley

V. G.

Ford

G. R.

White

F. A.

Barrett

J. J.

Harrell

L. E.

Roth

D. L.

(1995). Psychological, social, and health impact of caregiving: A comparison of black and white dementia family caregivers and noncaregivers. Psychology and Aging, 10, 540-552.

23.

Howard

Cushman

Moy

C. S.

Oparil

Muntner

Lackland

D. T.

Manly

J. J.

Flaherty

M. L.

Judd

S. E.

Wadley

V. G.

Long

D. L.

Howard

V. J.

(2018). Association of clinical and social factors with excess hypertension risk in black compared with White US adults. JAMA, 320(13), 1338-1348. doi:10.1001/jama.2018.13467

24.

Howard

V. J.

Cushman

Pulley

Gomez

C. R.

R. C.

Prineas

R. J.

Graham

Moy

C. S.

Howard

(2005). The reasons for geographic and racial differences in stroke study: objectives and design. Neuroepidemiology, 25(3), 135-143.

25.

Howard

Kleindorfer

D. O.

Cushman

Long

D. L.

Jasne

Judd

S. E.

Higginbotham

J. C.

Howard

V. J.

(2017). Contributors to the excess stroke mortality in rural areas in the United States. Stroke; a Journal of Cerebral Circulation, 48(7), 1773-1778. doi:10.1161/STROKEAHA.117.017089

26.

Institute of Medicine (2015). Cognitive aging: Progress in understanding and opportunities for action. The National Academies Press.

27.

Kelley

D. E.

Lewis

M. A.

Southwell

B. G.

(2017). Perceived support from a caregiver’s social ties predicts subsequent care-recipient health. Preventive medicine reports, 8, 108–111. doi:10.1016/j.pmedr.2017.08.001

28.

Litzelman

Skinner

H. G.

Gangnon

R. E.

Nieto

F. J.

Malecki

Witt

W. P.

(2014). Role of global stress in the health-related quality of life of caregivers: evidence from the survey of the health of Wisconsin. Quality of life research: an international journal of quality of life aspects of treatment, care and rehabilitation, 23(5), 1569-1578. doi: 10.1007/s11136-013-0598-z

29.

Lubben

(2008). Assessing social networks among elderly populations. Family & Community Health: The Journal of Health Promotion & Maintenance, 11, 42-52.

30.

Mittelman

M. S.

Haley

W. E.

Clay

O. J.

Roth

D. L.

(2006). Improving caregiver well-being delays nursing home placement of patients with Alzheimer disease. Neurology, 67(9), 1592–1599, doi:10.1212/01.wnl.0000242727.81172.91

31.

Ory

M. G.

Hoffman

R. R.

Yee

J. L.

Tennstedt

Schulz

(1999). Prevalence and impact of caregiving: A detailed comparison between dementia and nondementia caregivers. The Gerontologist, 39(2), 177–186, doi:10.1093/geront/39.2.177

32.

Papastavrou

Andreou

Middleton

Tsangari

Papacostas

(2015). Dementia caregiver burden association with community participation aspect of social capital. Journal of advanced nursing, 71(12), 2898-2910. doi:10.1111/jan.12762.

33.

Perissinotto

C. M.

Stijacic Cenzer

Covinsky

K. E.

(2012). Loneliness in older persons: A predictor of functional decline and death. Archives of Internal Medicine, 172, 1078-1083.

34.

Pew Research Center (2013). Home Broadband 2013 [Report]. https://www.pewresearch.org/internet/2013/08/26/home-broadband-2013/

35.

Pinquart

Sörensen

(2003). Differences between caregivers and noncaregivers in psychological health and physical health: A meta-analysis. Psychology and aging, 18(2), 250-267.

36.

Pinquart

Sörensen

(2011). Spouses, adult children, and children-in-law as caregivers of older adults: A meta-analytic comparison. Psychology and aging, 26(1), 1-14.

37.

Rigby

Ashwill

R. T.

Johnson

D. K.

Galvin

J. E.

(2019). Differences in the experience of caregiving between spouse and adult child caregivers in dementia with lewy bodies. Innovation in aging, 3(3), igz027. doi:10.1093/geroni/igz027

38.

Roth

D. L.

Haley

W. E.

Rhodes

J. D.

Sheehan

O. C.

Huang

Blinka

M. D.

(2020). Transitions to family caregiving: Enrolling incident caregivers and matched non-caregiving controls from a population-based study. Aging Clinical and Experimental Research. Epub ahead of print.

39.

Roth

D. L.

Mittelman

M. S.

Clay

O. J.

Madan

Haley

W. E.

(2005). Changes in social support as mediators of the impact of a psychosocial intervention for spouse caregivers of persons with Alzheimer’s disease. Psychology and aging, 20(4), 634.

40.

Roth

D. L.

Perkins

Wadley

V. G.

Temple

E. M.

Haley

W. E.

(2009). Family caregiving and emotional strain: Associations with quality of life in a large national sample of middle-aged and older adults. Quality of Life Research: an International Journal of Quality of Life Aspects of Treatment, Care and Rehabilitation, 53, 1117-1125.

41.

Sheehan

O. C.

Haley

W. E.

Howard

V. J.

Huang

Rhodes

J. D.

Roth

D. L.

(2020). Stress, burden, and well-being in dementia and nondementia caregivers: Insights from the caregiving transitions study. Manuscript under review.

42.

Smerglia

V. L.

Deimling

G. T.

Barresi

C. M.

(1998). Black/White family comparisons in helping and decision-making networks of impaired elderly. Family Relations, 37, 305-309.

43.

Stevens

A. B

Coon

Wisniewski

Vance

Arguelles

Belle

Mendelsohn

Ory

Haley

(2004). Measurement of leisure time satisfaction in family caregivers. Aging & mental health, 8(5), 450-459.