Work Impact and Emotional Stress Among Informal Caregivers for Older Adults

Abstract

Objectives:

With the growing aging population and reliance on informal caregivers in the United States, many individuals will take on the role of caregiver as an adult. We examined whether informal caregivers experience work interference or a change in work status (i.e., retiring/quitting) due to caregiving. We also explored whether experiencing work interference or a change in work status was associated with greater emotional stress.

Method:

This secondary analysis is drawn from the Fifth National Survey of Older Americans Act (OAA) program participants, which included 1,793 family caregivers. The present analysis is on caregivers of working age (18–64 years) providing care to another adult, which included 922 caregivers. Ordinal logit models were used to assess associations between experiencing work interference or a change in work status and emotional stress. Study weights were applied for all analyses.

Results:

At the time of the survey, more than half (52.9%) of caregivers were employed full- or part-time. Among nonworking caregivers (i.e., not working or retired) at the time of the survey, 39.8% responded that they had quit or retired early due to caregiving demands. Among employed caregivers, 52.4% reported that informal caregiving had interfered with their employment. Importantly, those respondents who reported work interference or a change in work status were more likely to report higher levels of emotional stress associated with caregiving demands.

Discussion:

These findings suggest the need to further explore work among informal caregivers and associations with emotional stress, as well as consider work-based policy approaches, organizational and/or societal, to support informal caregivers.

A high proportion of Americans provide informal care to an older adult, and this type of care can be particularly intensive in terms of time commitment and tasks performed (Kim & Schulz, 2008; National Alliance for Caregiving, 2009). The demand for informal assistance for older adults will only increase over the next several decades as the older adult population in the United States grows (Institute of Medicine, 2008; Redfoot, Feinberg, & Houser, 2013).

It is well recognized that providing informal care may not be without consequence to the caregiver as certain caregivers are at elevated risk for burden and distress (i.e., anxiety and depression) (Bevans & Sternberg, 2012; Etters, Goodall, & Harrison, 2008; Given & Northouse, 2011; Kim & Schulz, 2008). Caregiving-related characteristics and caregiver characteristics have been shown to be associated with a caregiver’s emotional response to caregiving (Sherwood et al., 2008). Indeed, experiencing greater objective burden or care intensity (i.e., high hours of care and assistance with a higher number of functional deficits) as a caregiver is associated with higher levels of anxiety, depressive symptoms, and emotional stress (Cannuscio et al., 2002; Grunfeld et al., 2004; Nijboer, Triemstra, Tempelaar, Sanderman, & van den Bos, 1999; Palos et al., 2011; Sherwood et al., 2008). Further, caregiver characteristics, such as non-Hispanic white race, female gender, and residence in the same household as the care recipient are also shown to be associated with greater emotional stress (Dunkin & Anderson-Hanley, 1998; Etters et al., 2008; National Alliance for Caregiving & AARP, 2004; Yee & Schulz, 2000).

Though comprehensive, it is not to say that the caregiving body of literature is exhaustive in understanding caregiver emotional response. For example, the association between paid work and emotional response has not yet been adequately explored within the informal caregiving context, despite being named in theory as a factor that might contribute to a caregiver’s emotional response to caregiving (Sherwood et al., 2008). Consideration of paid work (herein referred to as “work”) and caregiver distress is increasingly relevant given that caregivers are often employed at the start of caregiving and might find great meaning or financial stability through employment. Indeed, nearly 6 in 10 informal caregivers in the United States in 2009 were employed while providing care and nearly half were employed full-time (National Alliance for Caregiving, 2009). The ways in which caregiving can affect work vary, including modifying one’s work schedule while still fulfilling time requirements, reducing hours of work, or leaving the workforce altogether (National Alliance for Caregiving, 2009; National Alliance for Caregiving & AARP, 2004; Van Houtven, Coe, & Skira, 2013). A report based on a U.S. nationally representative sample of working-age adults showed that women providing informal care to another adult decreased their hours worked per week by 9% (Van Houtven et al., 2013). These findings are especially important given that the majority of informal caregivers are middle-aged women (Redfoot et al., 2013). Although a lower proportion of men provide informal care than women, an increasing number of men are providing such care and, thus, are also susceptible to adverse effects on work. Men providing informal care to another adult were also shown to have a mean employment reduction rate of 3.9% (Van Houtven et al., 2013).

As proposed by Hobfoll (2002), caregiving-related changes in work may negatively impact resource reserve, which can lead to greater stress. In addition, studies of working parents demonstrate that conflict in work–family balance can have an adverse effect on the parent’s psychological health (Canivet et al., 2010; Carr, 2002; Shimazu, Bakker, Demerouti, & Peeters, 2010), possibly due to a mismatch with expected gender or sociocultural roles (Carr, 2002). Similarly, among employed women who were informal caregivers, having work interfere with caregiving duties or vice versa (caregiving was interfering with their work duties) was associated with levels of stress due to the work–family imbalance (Gordon & Rouse, 2013).

Guided by prior work exploring the emotional response of informal caregivers (Sherwood et al., 2008) as well as prior literature from work–family conflict (Carr, 2002; Frone, Russell, & Cooper, 1992; Gordon & Rouse, 2013; Hobfoll, 2002), we sought to characterize work interference or a change in work status among informal caregivers. Specifically, among nonworking caregivers, what proportion reported that they had quit or retired due to their caregiving responsibilities? And among employed caregivers, what proportion had experienced work interference due to caregiving? Next, we explored associations among work interference and quitting or retiring early with emotional stress. Based on prior findings, it was hypothesized that individuals who experienced work interference or a change in work status due to caregiving (i.e., quitting or retiring early) would report higher levels of emotional stress compared with informal caregivers who did not report such experiences.

Method

Data Source and Sample

Data for this secondary analysis are drawn from the 2009 Fifth National Survey of Older Americans Act (OAA) program participants and specifically the caregiver module that included 1,793 informal caregivers. These caregivers were invited to participate because they had utilized support services through OAA state and community programs on aging under the National Family Caregiver Support Program (NFCSP) (2000). The survey involved a two-stage sample design that involved selecting a sample of Area Agencies on Aging (AAA) followed by a sample of clients for each of the six services (e.g., home-delivered meals, transportation, the Family Caregiver Support program) within the sampled AAA. Data were weighted specifically adjusting for nonresponse, followed by trimming of extreme weights. A full description of the sampling technique and weighting is publically available (see http://www.agid.acl.gov/DataFiles/NPS/Files.aspx?year=2009&serviceid=1). The present analysis focused on adult caregivers of traditional working age (i.e., age 18–64 years), which included 922 caregivers. All care recipients of these working-age caregivers were 60 years of age or older with the majority (79%) being 75 years of age or older.

Measures

This research sought to characterize work interference or a change in work status among informal caregivers and explore associations with emotional stress.

Emotional stress

The primary outcome of interest was perceived emotional stress of caregiving reported by the caregiver. To measure emotional stress of caregiving, the caregivers were asked: “Where 1 is ‘not at all stressful’ and 5 is ‘very stressful,’ how emotionally stressful would you say that caring for the care recipient is for you?” (response option of 1, 2, 3, 4, or 5.)

Caregiving and work

The primary predictor variables included quitting or retiring early because of caregiving (for nonworking caregivers) or caregiving interfering with work (for employed caregivers). Quitting or retiring early was identified using the following item: Caregivers who reported not working at the time of the survey were asked, “Did your caregiving responsibilities cause you to quit working or retire early?” (Yes or no.) Caregiving interfering with work was assessed as follows: Caregivers who reported that they were employed at the time of the survey were asked, “Has providing care for the care recipient interfered with your job?” (Yes or no.)

Potential covariates

Potential covariates were based on prior literature and included caregiver demographic variables (i.e., gender, race/ethnicity, age, income, education level), caregiving/caregiver characteristics (i.e., duration of care, proximity to care recipient, providing care for other individuals, relationship to the care recipient), and care recipient functioning variables (i.e., number of activities of daily living (ADLs), hours of care required). These measures are described below.

Caregivers reported the following demographic information: education (less than high school diploma, high school diploma or GED, some college/business/technical school, bachelor’s degree, some graduate or postgraduate degree); ethnicity/race (white, African American, Asian, American Indian, Native Hawaiian Pacific Islander, Other race); gender; household income ($10,000 or less, $10,001–$20,000, $20,001–$30,000, $30,001–$50,000, above $50,000); age (18–34, 35–59, and 60–64 years); and marital status (married or divorced/separated/not married). Caregivers were also asked: “What is your relationship to the care recipient?” Responses included: husband, wife, son, son-in-law, daughter, daughter-in-law, mother, brother, sister, granddaughter, grandson, niece, nephew, friend/neighbor/another person, other relative. For this analysis, relationship was categorized according to: (1) Spouse; (2) adult child or son/daughter in-law; or (3) other (relative or friend).

We assessed the following caregiving characteristics: To measure duration of care, caregivers were asked, “For how long have you been providing care to the care recipient?” Response options included: (1) 6 months or less; (2) more than 6 months but less than 1 year; (3) at least 1 year but less than 2 years; (4) 2–5 years; (5) 5–10 years; or (6) 11 or more years. Due to modest sample sizes across the various categories, the responses were collapsed into: (1) less than 2 years; (2) 2–5 years; (3) 5–10 years; or (4) 11 or more years. Proximity to care recipient was measured using the following question: “How far away do you live from the care recipient?” (In-home; 1 hour or less away; more than 1 hour away). Because of a small sample size for “more than 1 hour away,” we collapsed this variable into a binary outcome (in-home or out of home). Caregivers were also asked about the type of care provided or the ADL limitations of the care recipient. ADL limitations were categorized into having three or more ADL limitations (yes or no). Because emotional well-being may also be negatively affected by the number of hours of care, caregivers were asked “In your judgment, how many hours per day of help, care, or supervision does the care recipient need?” Responses were categorized as (1) 0–4 hours; (2) 5–12 hours; (3) 13–23 hours; or (4) 24 hours. Finally, the caregivers were asked about other care responsibilities by responding to the following question, “How many persons in total are you caring for, not counting the care recipient?” Responses were categorized in the original data set as: 1 other person; 2 other people; or 3–5 other people (although the survey did not detail or describe who they were caring for or the type of care provided, such as informal care or child care.)

Analysis

Descriptive statistics were used to characterize work interference or a change in work status among working-age informal caregivers for frail older adults. We conducted chi-square analyses to assess factors (demographics and caregiving characteristics) associated with work interference or a change in work status. Next, to examine whether work interference or a change in work status is associated with higher levels of emotional stress, ordinal logistic regression analyses were performed separately for nonworking caregivers and employed caregivers. Because we are using an ordinal logit modeling approach, parameters are deemed statistically significant using t tests. Positive parameters are associated with a positive effect on the variation in stress levels (dependent variable), whereas negative parameters are associated with an inverse relationship on the variation in stress levels. All analyses in this study applied sample weights to account for the survey design. We estimated standard errors via Taylor linearization to account for sample weights (Lumley, 2011). All analyses were performed using Stata (version 12.0).

Results

Caregivers in this sample were predominantly women (80.1%) and the adult child or son/daughter in-law of the care recipient (78.3%). Additional sample characteristics are presented in Table 1.

Table 1.

Descriptives of Caregivers Aged 18–64 Years (percentages are based on weighted samples)

Variable	%
Education
Less than high school	5.8
High school or GED	27.9
Some college/business/technical school	33.8
Bachelor’s degree	14.3
Some graduate or postgraduate degree	18.2
Caregiver’s race/ethnicity
White	78.9
Black/African American	15.4
Asian	2.9
American Indian/Alaskan Native	1.9
Other	1.7
Caregiver’s age (years)
18–34	2.5
35–59	66.5
60–64	31
Caregiver’s gender
Female	80.1
Male	19.9
Income
$10,000 or less	6
$10,001–20,000	15.4
$20,001–30,000	19.2
$30,001–50,000	22.4
More than $50,000	37
Marital status
Married	53.8
Not married (divorced, separated, or never married)	46.2
Care duration
Less than 2 years	20.6
2–5 years	41.7
5–10 years	27.3
11 or more years	10.4
Relationship to care recipient
Spouse	9.3
Adult child or son/daughter in-law	78.3
Other (relative or friend)	12.4
Proximity to care recipient
In-home	59
Out of home	41
Three or more ADLs for care recipient
Yes	69.9
No	30.1
Care hours needed
0–4 hours	29.8
5–12 hours	29.4
13–23 hours	10.2
24 hours	30.6
Provide care to others
1 other person	63.6
2 other people	19.9
3–5 other people	16.5
Employment status (at time of survey)
Employed (full- or part-time)	52.9
Retired/not working	47.1
Emotional stress of caregiving
Level 1 (not at all)	9.5
Level 2	18.9
Level 3	28.7
Level 4	21.7
Level 5 (very stressful)	21.2
Quit or retired due to caregiving among nonworking caregivers
Yes	39.8
No	60.2
Care interfered with work among employed caregivers
Yes	52.4
No	47.6

Note: ADL = activity of daily living.

Caregiving and Work

At the time of the survey, more than half (52.9%) of caregivers were employed full-or part-time. Among the nonworking caregivers, 39.8% responded that they had quit or retired early from work due to caregiving demands. In bivariate analyses, a higher proportion of caregivers who had quit or retired early because of caregiving were women, providing care to a spouse, providing three or more ADLs, residing in the care recipient’s home, and providing more hours of care compared with those who had not quit or retired early (Table 2). Among employed caregivers, 52.4% reported that caregiving had interfered with their employment. In bivariate analyses, caregivers who experienced work interference had higher levels of education and were more likely to be caring for a parent, providing three or more ADLs and providing more hours of care compared with employed caregivers who did not experience any work interference (Table 2).

Table 2.

Demographic and Caregiving-Related Variables and Work Change for Nonworking and Employed Caregivers

Variable	Nonworking		p Value	Employed caregivers		p Value
	Quit/retired early (%)			Work interference (%)
	Yes	No		Yes	No
Education			.582			.015
Less than high school	41.5	58.5		2.5	97.5
High school or GED	45.3	54.7		45.4	54.6
Some college/business/technical school	31.7	68.3		48.2	51.8
Bachelor’s degree	46.5	53.5		57.1	42.9
Some graduate or postgraduate degree	43.9	56.1		68.9	31.1
Caregiver’s race/ethnicity			.659
White	41.6	58.4		54	46	.323
Non-white	37.4	62.6		44.8	55.2
Caregiver’s age (years)			.924			.122
18–34	33	67		0	100
35–59	40	60		55	45
60–64	41	59		48	52
Caregiver’s gender			.020
Female	43.4	56.6		51.5	48.5	.137
Male	22.5	77.5		63.2	36.8
Income			.809			.100
$10,000 or less	42.5	57.5		37.5	62.5
$10,001–20,000	37.9	62.1		33.2	66.8
$20,001–30,000	36.5	63.5		36.7	63.3
$30,001–50,000	31.6	68.4		56.1	43.9
More than $50,000	43.7	56.3		59.7	40.3
Marital status			.337			.061
Married	42.7	57.3		58.6	41.4
Not married (divorced, separated, or never married)	36.9	63.1		43.1	56.9
Care duration			<.001			.345
Less than 2 years	15.6	84.4		43.6	56.4
2–5 years	49.7	50.3		52	48
5–10 years	54.9	45.1		62.1	37.9
11 or more years	24.3	75.7		42.3	57.7
Relationship to care recipient			<.001			<.001
Spouse	71	29		34.3	65.7
Adult child or son/daughter in-law	38.5	61.5		58.3	41.7
Other (relative or friend)	19.1	80.9		18.2	81.8
Proximity to care recipient			<.001			.185
In-home	48	52		47	53
Out of home	19	81		57	43
Three or more ADLs for care recipient			.0126			<.001
Yes	45.1	54.9		61.5	38.5
No	27.15	72.85		32	68
Care hours needed			.037			.002
0–4 hours	22.5	77.5		42.9	57.1
5–12 hours	47.4	52.6		44.3	55.7
13–23 hours	55.1	44.9		70.1	29.9
24 hours	40.6	59.4		75.6	24.4
Provide care to others			.381			.384
1 other person	42.1	57.9		48.2	51.8
2 other people	35.1	64.9		53.5	46.5
3–5 other people	31.7	68.3		61.9	38.1

Note: Findings are based on weighted samples per variable.

Is Experiencing Work Interference or a Change in Work Status Associated With Greater Emotional Stress?

Nearly 60% of nonworking caregivers and 40% of employed caregivers reported caregiving to be of the highest levels of stress (i.e., 4 or 5). Among the nonworking caregivers, quitting or retiring early because of caregiving was significantly associated with higher stress levels while controlling for relevant covariates (Table 3). Further, there is a monotonic increase as the number of caregiving hours needed went up, so did the reported stress. Specifically, individuals who provided care to someone needing more hours of care exhibited an increasing association with stress measured as 0.97 for care needed from 5 to 12 hours, 1.28 for 13–23 hours of care needed to a greater effect of 1.38 for 24 hours of care needed. Individuals with a high school education reported higher stress (1.51, p = .04) than those with less than a high school education. The findings also suggest that stress levels were lower (demonstrated by the negative coefficient of −1.03) with longer care duration among nonworking caregivers. Other variables included in the model were not significantly associated with nonworking caregivers’ stress levels.

Table 3.

Ordinal Logit Regression Analyses for Employed and Unemployed/Retired Caregivers and Emotional Stress

Nonworkinga				Employedb
Variable	Coefficient	SD	p > |t|		Coefficient	SD	p > |t|
Quit/retired				Work interference
No	Ref			No	Ref
Yes	1.07	0.31	.01	Yes	1.82	0.30	.00
Three or more ADLs				Three or more ADLs
No	Ref			No	Ref
Yes	0.43	0.43	.33	Yes	0.30	0.32	.36
Race				Race
Non-white	Ref			Non-white	Ref
White	0.46	0.36	.21	White	0.14	0.48	.77
Care hours				Care hours
0–4 hours	Ref			0–4 hours	Ref
5–12 hours	0.97	0.42	.02	5–12 hours	0.04	0.28	.88
13–23 hours	1.28	0.45	.01	13–23 hours	0.03	0.48	.94
24 hours	1.38	0.45	.003	24 hours	0.66	0.04	.07
Education				Education
Less than HS	Ref			Less than HS	Ref
HS or GED	1.51	0.72	.04	HS or GED	1.98	0.94	.04
Some college	1.17	0.80	.15	Some college	1.57	0.94	.10
BA/BS	1.35	0.84	.11	BA/BS	2.23	0.95	.02
Some graduate/graduate degree	1.53	0.81	.06	Some graduate/graduate degree	1.48	0.95	.12
Gender				Gender
Male	Ref			Male	Ref
Female	0.40	0.35	.24	Female	0.42	0.35	.24
Income				Income
<$10,000	Ref			<$10,000	Ref
$10,001–20,000	−0.44	0.76	.56	$10,001–20,000	0.25	0.99	.80
$20,001–30,000	−0.37	0.78	.64	$20,001–30,000	0.66	0.94	.48
$30,001–50,000	−0.61	0.84	.47	$30,001–50,000	0.57	0.86	.63
>50,000	−0.84	0.34	.69	>50,000	0.57	0.85	.50
Marital status				Marital status
Not married	Ref			Not married	Ref
Married	−0.48	0.34	.88	Married	−0.13	0.30	.65
Care duration				Care duration
<2 years	Ref			<2 years	Ref
2–5 years	−0.56	0.42	.19	2–5 years	0.77	0.43	.07
5–10 years	−0.29	0.31	.58	5–10 Years	1.23	0.40	.004
≥11 years	−1.03	0.53	.05	≥11 years	1.17	0.47	.02
Relationship to care recipient				Relationship to care recipient
Spouse	Ref			Spouse	Ref
Adult child/in-law child	−0.44	0.67	.50	Adult child/in-law child	0.45	0.52	.42
Other nonspouse relative or friend	−1.03	0.53	.27	Other nonspouse relative or friend	−0.10	0.70	.89
Care to others				Care to others
1 other person	Ref			1 other	Ref
2 other	−0.52	0.42	.22	2 other	0.08	0.34	.82
3 or more	−0.11	0.43	.80	3 or more	−0.20	0.44	.65
Proximity to CR				Proximity to CR
Out of home	Ref			Out of home	Ref
In-home	−0.70	0.40	.09	In-home	−0.20	0.44	.31
Caregiver’s age (years)				Caregiver’s age (years)
18–34	Ref			18–34	Ref
35–59	0.58	0.67	.50	35–59	0.16	0.63	.80
60–64	0.72	0.96	.45	60–64	0.97	0.76	.21

Notes: HS = high school.

Statistically significant findings are in bold text.

^aFindings are based on a weighted sample of 358 employed caregivers.

^bFindings are based on a sample of 321 unemployed/retired caregivers.

Among employed caregivers, work interference was significantly associated with higher levels of emotional stress (Table 3). Providing care to an individual needing round-the-clock care (24 hours/day) was also associated with higher emotional stress. Those with a high school education and a bachelor’s degree reported more stress than those without a high school diploma. Finally, providing care for a longer period of time was also associated with higher stress.

Discussion

In this national sample of working-age caregivers, our findings demonstrate that some caregivers reported ending paid employment before traditional retirement age, whereas others report that caregiving has interfered with their work. Indeed, nearly 40% of nonworking caregivers below retirement age reported quitting or retiring early because of caregiving demands. And among employed caregivers, more than half reported that caregiving responsibilities had interfered with their job. As a result, these findings have important implications for workplace productivity, caregivers’ financial well-being, and overall society benefit (e.g., social security savings) (Gonzales, Lee, & Brown, 2015). The potential exists that informal caregivers might have less resource reserves, which could adversely impact long-term quality of life. Future research should model the implications of work impact on the long-term financial stability of caregivers and related quality of life.

Our findings also highlight the need to continue to monitor the experience of providing informal care among women in the workforce. Previous research shows that caregiving can be physically and emotionally burdensome to women working full-time (Juratovac & Zauszniewski, 2014) and that lacking work–family support has been found to be more strongly associated with exhaustion in women than in men (Canivet et al., 2010). In the current study, a higher proportion of women reported having had to quit or retire early due to caregiving demands. As a result, this could lead to more financial instability among women. Indeed, previous research on female caregivers notes that providing care earlier in life increases poverty risks later in life (Lee, Tang, Kim, & Albert, 2014; Wakabayashi & Donato, 2006). Similarly, our findings support prior work demonstrating that objective burden-related variables, such as duration of care, ADLs, and hours of care, are directly associated with a caregiver’s ability to work (Van Houtven et al., 2013).

Interestingly, caregivers with higher education were more likely to report that caregiving interfered with their work compared with caregivers with lower education levels. Perhaps caregivers with higher education also hold positions that entail greater job responsibilities or longer working hours; hence, they may perceive that caregiving is interfering with their ability to fully carry out their job. The identification of subgroups (e.g., women, those in management positions) that are at particular risk for poor work outcomes is integral to the development of novel programs to assist informal caregivers in maintaining desired levels of employment while balancing caregiving demands.

This research adds to the literature by exploring whether experiencing work interference or a change in work status as a caregiver is associated with emotional stress. A high proportion of nonworking (59.7%) and employed (40%) caregivers perceived caregiving for an older adult to be at the highest levels (i.e., 4 and 5). For nonworking caregivers, change in work status due to caregiving demands was associated with higher emotional stress. Other variables, including duration of care and caregiver education level, were also associated with emotional stress. For example, among nonworking caregivers, providing care for more years was associated with lower stress. This finding may indicate either adaptation and/or acceptance of the informal caregiver role. It is plausible that relatively new caregivers experience greater stress as they struggle to cope with their new responsibilities, whereas over time, caregivers become more familiar with their caregiving responsibilities and develop effective coping strategies. Caregiving research in the cancer context has also shown a decrease in distress among caregivers over time (i.e., time since patient diagnosis) (Longacre, Ridge, Burtness, Galloway, & Fang, 2012). Alternatively, caregivers who have been providing care for a longer period of time may have greater acceptance of their new role. Gordon and Rouse (2013) suggest that greater acceptance or prioritization of a particular role can be stress reducing. This concept is worth exploring in future studies to examine whether attaching greater importance to the caregiving role leads to lower stress levels.

Among employed caregivers, work interference was associated with higher emotional stress. Similar to nonworking caregivers, caregiver education level and duration of care were associated with emotional stress. Employed caregivers with a high school or college education experienced higher stress compared with those without a high school degree. Higher education was not significantly associated with elevated stress. As indicated by Hobfoll (2002), greater resource reserve (education, income) might be protective against perceiving caregiving as stressful. Caregivers with lower education levels may be more likely to have jobs with less flexibility in terms of work hours or work location (e.g., having the ability to work from home). Findings indicate that a proportion of the U.S. workforce experiences irregular or uncertain work schedules and that this experience is particularly evident for those in retail sales work and those with lower education (Golden, 2015). In comparison, individuals with more education may be in positions earning not only more money but may have more control over their schedule. However, it is important to note the small percentage of caregivers with less than a high school degree, and this finding should be explored more fully with a larger, diversely educated population.

Also, in contrast with nonworking caregivers, longer duration of care provided was associated with higher stress among employed caregivers, perhaps reflecting the cumulative burden of caregiving over time with work demands. Further, as indicated in the results, there was a monotonically increasing effect of the hours of care needed and increased stress. In addition to the cumulative burden of caregiving with respect to time, it is also worthy to explore if more job-related opportunities may be missed given time constraints. As such, they face higher opportunity costs in terms of lost potential income via job advance or promotion. This again relates to having a lower resource reserve (Hobfoll, 2002). This finding of association between time spent caring and stress is corroborated by Gordon and Rouse (2013) in which time given to care led to some form of work costs and family–work conflict. Whereas Gordon and Rouse (2013) focused on hours per week, time commitment may also be measured in terms of number of years providing care. For nonworking individuals, stress was reduced by the number of years providing care, but stress increased with greater number of hours of care needed per day. Conversely, individuals who were still participating in the workforce had higher stress levels with more years providing caregiving services. Therefore, it is important to be specific regarding how time commitment is measured because the time reference may be different—hours per day or years of care, it appears that there is a family–work stress that is exhibited by both nonworking and working caregivers. However, despite the obvious strain on employed caregivers, various aspects of employment (e.g., purpose-finding, social interactions, financial stability) that might serve as protective factors for an individual’s well-being should be considered in future research.

Finally, whereas the literature details other demographic factors that could explain variation in stress levels among either nonworking or working caregivers including race and gender (Dunkin & Anderson-Hanley, 1998; Etters et al., 2008; National Alliance for Caregiving, 2009; Yee & Schulz, 2000), we did not find a statistically significant effect of race or gender on stress in the multiple regression model. Similarly, our regression findings did not support income as a mitigating factor for stress (higher incomes would ensure future financial viability for caregivers) as suggested it might in the literature (Carr, 2002).

Going forward, the findings from the present analysis provide an important perspective for understanding caregiving and work and highlight key subgroups of caregivers that may benefit from additional support via organizational and social (work-based) policies. In the United States, the Family Medical Leave Act (FMLA) remains the most viable federal option for a caregiver who wishes to temporarily leave work to provide informal care. However, this policy lacks universality for all workers and leaves states to modify or supplement the policy to address shortcomings (e.g., unpaid leave) (Yang & Gimm, 2013). To date, no national federal policy exists with respect to preserving work for all working caregivers who temporarily provide care. Given that emotional strain among caregivers is known to be associated with poorer outcomes among caregivers and care recipients (Damjanovic et al., 2007; Mittelman, Haley, Clay, & Roth, 2006; Redwine et al., 2004), efforts to preserve work choice among caregivers may have positive implications for both caregivers and care recipients. Subsequent research should also explore whether such approaches contribute to greater benefit for organizations in terms of increasing productivity, and for society more generally, in terms of enhancing emotional well-being and reducing economic hardship among the growing number of informal caregivers.

In countries with similar experiences of a growing older adult population and dwindling informal care structure (e.g., Germany, Japan), social policies have been enacted to address concerns of caring for an aging population (Campbell, Ikegami, & Gibson, 2010; Frericks, Jensen, & Pfau-Effinger, 2014). Examples include varied financing structures for family leave, such as reduced pay while on leave and for a designated timeframe upon returning to work (Frericks et al., 2014), movements to “formalize” long-term care through direct payment and social security benefits for caregivers as well as legal “employment” rights for caregivers, and national long-term care insurance to cover services (Japan) (Campbell et al., 2010; Frericks et al., 2014; Yang & Gimm, 2013). Though only one study, research from Japan concluded that their long-term care system benefits caregivers in terms of reduced burden (Arai & Zarit, 2011). In light of the potential benefits, further study of these various approaches upon the emotional response of caregivers would be highly informative. In time, we could also learn through varied approaches whether caregivers experience better immune function and, subsequently, better overall health (Sherwood et al., 2008), which could be a financial benefit to society over time.

Limitations

We acknowledge that there are several limitations with the current study. First, this sample of caregivers may not be representative of informal caregivers in the United States given that these caregivers actively sought assistance through the AAA through the NFCSP. Caregivers with high demands and lower resources have greatest access via this program, and, thus, these caregivers are not representative of all caregivers. However, it is possible that this sample of caregivers might have better outcomes because of the support provided to them. In addition, this sample of caregivers was predominantly women and of non-Hispanic white background; thus, it is unknown whether the findings will pertain to male caregivers or caregivers of ethnic/racial minority background. Second, the cross-sectional nature of the data does not permit any causal inferences to be made. Future studies involving longitudinal data collection will enable us to examine whether work interference or changes in work status precede or contribute to caregiver stress. Third, because this is a secondary analysis utilizing an existing data source, we were constrained by the existing measures collected. The use of a single-item indicator of emotional stress of caregiving is not ideal, given concerns about the validity of a single-item measure under various conditions. However, single-item measures of stress have demonstrated measurement reliability and validity comparable with longer stress questionnaires (Littman, White, Satia, Bowen, & Kristal, 2006). We also did not have instrumental activities of daily living within the analyses. Lastly, we also did not know the work history of those categorized as nonworking, including those who reported not having to quit or retire early because of caregiving. It is possible that these individuals were never employed, and thus employment history must be captured in future caregiving work. Future research should include more comprehensive measures of psychosocial well-being and social functioning. In sum, the findings from the present study suggest that additional research on caregiving-related work interference, productivity, and change in work status is warranted in order to improve caregiver emotional well-being and work outcomes.

Conclusions

This analysis is timely given that the landscape of informal caregiving will continue to change over the next several decades. Namely, the growth in the older adult population will lead to an increased need for caregivers (Redfoot et al., 2013). Caregivers will also likely perform increasingly demanding caregiving tasks in light of continuing health care changes and industry preferences for greater outpatient care. All the while, many caregivers will feel pressure or desire to remain employed for financial reasons. Hence, these findings will contribute to discussion about work-based policy approaches, organizational and/or social, that can support employed caregivers and lead to improved mental health outcomes.

Specifically, the present findings increase our understanding of caregiving and work by noting that a high proportion of informal caregivers experienced either work interference or a change in work status due to their caregiving responsibilities. Further, experiencing work interference or a change in work status was each associated with heightened emotional stress. Future research should explore whether strategies, policies, or programs that help to sustain desired levels of employment will yield improved caregiver outcomes and other work-related benefits.

Funding

This work was funded through the National Institute of Health grants R41CA144100 and P30CA006927.

Conflict of Interest

The authors declare no conflicts of interest with respect to the authorship and/or publication of this article.