The Complexities of Family Caregiving at Work: A Mixed-Methods Study

Abstract

The current project examined the impact of caregiving and caregiving–work conflict on employees’ well-being. A sequential explanatory mixed-methods design (QUAN→qual) was utilized, and a total of 880 employees from a large health-care plan employer completed an online survey. Forty-five caregivers who completed the survey also participated in one of the five focus groups held 1 to 2 months later. Employed caregivers were significantly (p < .05) more likely to indicate poorer physical and mental health than noncaregivers; among caregivers (n = 370), caregiving–work conflict emerged as the most significant predictor of well-being and fully mediated the empirical relationship between burden and well-being. The focus group findings complemented the quantitative results; many of the challenges employed caregivers experience stem from their ability or inability to effectively balance their employment and caregiving roles. The results suggest the need to focus on caregiving–work conflict when constructing new or translating existing evidence-based caregiver interventions.

Various estimates suggest that 85% to 90% of all older adults receive needed help from family caregivers (Gitlin & Schulz, 2012). Caregiving refers to providing unpaid help to someone because her or his health needs are outside the norm of usual support exchange (Gaugler, Kane, & Kane, 2002; Schulz & Quittner, 1998). Caregiving is particularly challenging when family members must balance multiple life responsibilities such as employment (National Alliance for Caregiving and American Association of Retired Persons, 2015). In the United States, 60% of caregivers were employed in 2015 while providing help and 61% of caregivers indicated at least one job impact as a result of caregiving (e.g., disruptions to work schedule, losing one’s job entirely, having to use sick days to provide care). Relying on a corporate–community–university partnership and a mixed-methods design, the current study examined the psychosocial and health implications of employment on family caregivers, and in particular, the influence of caregiving–work conflict (i.e., the perceived conflict and time pressures of work due to care responsibilities) on caregiver well-being.

Background

Since the 1980s, a number of studies have found that the multiple, competing demands of employment and family caregiving often impede one or both of these roles (Barling, MacEwen, Kelloway, & Higginbottom, 1994; Scharlach, 1994; Scharlach, Sobel, & Roberts, 1991; Schulz & Martire, 2009; Stephens & Frank, 2009; Stone & Short, 1990; Thrasher, Zabel, Wynne, & Baltes, 2015; Wagner, 2006). Several recent studies have continued to shed light on the effects of employment on care-related stress and work (Carr et al., 2018; DePasquale, Polenick, Davis, Berkman, & Cabot, 2018; DePasquale, Polenick, Davis, Moen, et al., 2017; Fekete, Siegrist, Tough, & Brinkhof, 2018; Lerner et al., 2017; MacDonald, Fujishiro, Howard, Landsbergis, & Hein, 2017; Oldenkamp et al., 2018; Paulson, Bassett, Kitsmiller, Luther, & Conner, 2017). Family caregiving can also affect employers. For example, in 2010, employers lost $13 billion due to employees’ eldercare responsibilities (Albert, Schulz, & Colombi, 2010).

Because of methodological issues such as small sample sizes, inconsistent definitions of family caregiving, and conflation of employer and organization outcomes, a deeper understanding of the complex relationship between employment and caregiving outcomes remains limited (Calvano, 2013). The juggling of competing demands from family care, one’s employment, and other life contexts appears to determine whether employed caregivers manage distress effectively (Cole & Gary, 2012; Duxbury & Dole, 2015; Guberman & Maheu, 1999; Jolanki, 2015). Research has found that employed caregivers tend to indicate greater well-being when compared to nonemployed caregivers (Masuy, 2009; Scharlach, 1994; Stephens & Franks, 2009; Torres-Gil, 2009; van Campen, de Boer, & Iedema, 2013). At the same time, perceived organizational support for family caregiving may play a critical role in mediating caregiver strain (Zacher & Schulz, 2015). It is possible that the mixed effects of employment on caregiving outcomes evident in prior literature are due to treating employment as a dichotomous variable, which may oversimplify nuances such as hours worked, degree of caregiving–work conflict, and other factors. Fredriksen-Goldsen and Scharlach (2006) found that more demanding aspects of both caregiving and employment emerged as the strongest correlates of caregiving and work–family strain in a large sample of employed caregivers.

These and other findings (Giovannetti, Wolff, Frick, & Boult, 2009; Stephens, Franks, & Atienza, 1997; Swanberg, 2006) suggest the need for interventions that target not only domains of caregiving but also domains of work when attempting to alleviate the negative outcomes that may occur for employed family caregivers. More specifically, rarely does the clinical content of family caregiving interventions or the outcomes they consider focus on caregiving–work conflict; instead, evidence-based caregiving interventions across various conditions largely emphasize coping strategies, building of social support, or the provision of respite (Burgio, Gaugler, & Hilgeman, 2016). It remains unknown if caregiving–work conflict is as strong a predictor of employed caregivers’ well-being as more traditional variables such as care-related stress or burden; if this was the case, it would suggest the need to expand the scope of current interventions to address not just how caregivers’ effectively manage care responsibilities, but how they navigate employment concerns as well.

Conceptual Model

Conceptually, this study is framed within Stephens & Franks (2009) competing demands hypothesis, where the temporal, emotional, social, and financial demands of family care provision may interfere with work responsibilities and vice versa, thus resulting in greater strain and negative health for the caregiver. In the competing demands hypothesis, the time pressures of simultaneous work and family caregiving deplete an individual’s energy, thus placing her or him at risk of the negative psychological and physical health outcomes that are often associated with the provision of family care to a relative with a chronic illness (Trukeschitz, Schneider, Mühlmann, & Ponocny, 2013). Frameworks such as the Stress Process Model describe the overall mechanism of stress leading to negative outcomes on the part of caregivers as one of “proliferation,” whereby the emotional demands of care provision spread to other life domains (e.g., work), which thus leads to poorer mental and physical health outcomes (Pearlin, Mullan, Semple, & Skaff, 1990). Within the Stress Process Model, context of care and work indicators (e.g., sociodemographics of the caregiver and care recipient, duration of care, relationship of caregiver to care recipient, hours worked, job satisfaction) are noted as important factors that can potentially influence outcomes and thus should be incorporated into any analysis of the caregiving stress process. At the heart of the proliferation process are primary stressors or those linked to the demands of family care. These include primary objective stressors (e.g., care recipient condition) and primary subjective stressors (e.g., caregivers’ emotional reactions to or appraisals of objective stress). The exacerbation of primary stressors is hypothesized to spread to and influence other life domains, such as tension between work and family care responsibilities (i.e., caregiving–work conflict, or secondary stressors). This process is expected to negatively influence caregivers’ global well-being, including depression and perceptions of subjective health (Gaugler et al., 2008).

Relying on both the competing demands hypothesis as well as prior conceptual models of caregiving stress, we postulated that employed caregivers are more likely to indicate negative health and well-being even after adjusting for important sociodemographic and work characteristics than noncaregiving employees (Hypothesis 1). Among employed caregivers, perceptions of burden would be significantly associated with caregiving–work conflict and well-being even after controlling for key sociodemographic, work, and context of care characteristics (Hypothesis 2). Noting the mechanism of proliferation, the third and most consequential hypothesis (Hypothesis 3) was tested: Conflict and tensions at work (i.e., caregiving–work conflict) would fully mediate the empirical relationship between caregiver burden (i.e., feelings of being overwhelmed with the emotional, financial, social, and psychological aspects of care) and global measures of physical and mental health among employed caregivers.

Although the wealth of descriptive research on family caregiving and employment describes challenges related to balancing work and family care, insights as to why employees do or do not successfully balance family care responsibilities are not as evident. The current study further aimed to achieve a greater understanding of the challenges of family caregiving and employment, with a particular focus on how caregiving–work conflict influences the well-being of employed family caregivers. For this reason, two additional research questions guided subsequent mixed-methods analyses in concert with the hypotheses posed earlier:

1. What is the experience of family caregivers who are employed?

2. Based on these experiences, what approaches or potential interventions could better support employed family caregivers?

Methods

Procedure

A sequential explanatory mixed-methods design was used (QUAN→qual; Creswell & Plano-Clark, 2010) to examine the complexities of family caregiving and work in a large, local employer: Blue Cross and Blue Shield of Minnesota, or BCBS. Blue Cross and Blue Shield of Minnesota is the largest health-care plan in Minnesota and includes 2.6 million members throughout the United States. Although detailed data on work roles were not collected, employees at BCBS largely focus on oversight of health-care plans and are not direct care professionals. An online survey was collaboratively developed and reviewed by the first author (J. E. G), representatives of BCBS (R. S. and S. A. B), and the community-based project partner (Amherst Wilder Foundation; authors R. S., G. O., J. C., and M. K.) and then administered to all current employees of BCBS in Minnesota in March 2016 on its online survey platform. The survey was promoted by BCBS Human Resources through internal communications to employees. Employees were allowed and encouraged to complete the online survey during work hours. Those who completed the survey were entered into a $25 gift card drawing. All participants were assured that their responses would be anonymous and that BCBS would not receive any information that would allow them to identify individuals. Prior to data collection, approval from the University of Minnesota Institutional Review Board was obtained. At the conclusion of the online survey, those who indicated that they were caregivers were asked whether they were willing to participate in a follow-up focus group (the following dichotomous yes/no item was used to identify caregivers:

Are you currently providing unpaid care to a relative or friend 18 years or older to help them take care of themselves? This may include helping with personal needs or household chores. It might be managing a person’s finances, arranging for outside services, or visiting regularly to see how they are doing. This adult need not live with you. (National Alliance for Caregiving and American Association of Retired Persons, 2015)

The focus groups took place at an offsite location or at BCBS offices approximately 1 to 2 months following the completion of the online survey. The focus groups were 60 to 120 min each. Focus group sessions included 7 to 15 caregivers and were moderated by three of the investigators (JEG, RS, and GO). If not serving as the lead moderator, one of these investigators served as the assistant moderator to observe and take notes during the focus group session per recommended procedures (Morse & Field, 1995). All sessions were recorded and transcribed. Focus group participants were offered a $25 gift card for their time.

Sample

A total of 880 individuals participated in the online survey, 370 of which identified as caregivers. The survey was administered to all BCBS employees (N = 3,799). Of these, 1,188 opened the e-mail and clicked on the survey link. Most caregivers were either a spouse or adult child of their care recipient, although slightly over a quarter of respondents (26.3%) indicated “other” when indicating their relationship to the care recipient. Text responses indicated that these individuals were often caring for an adult child or grandchild with a health need. Ninety-three employees indicated that they were willing to participate in one of the five focus groups: 11 employees in a rural BCBS call center; 14 supervisors, managers, or leadership staff; and 68 nonmanagement employees. All employed caregivers in the rural BCBS call center were invited to participate in one focus group. All supervisors and leadership personnel were invited to participate in one of the two focus groups consisting of only management staff. Finally, to sample from the group of 68 nonmanagement employees, the sixth author reached out to those 68 individuals to arrange a focus group and the first 10 to 15 exempt (salaried) and 10 nonexempt (hourly) staff who responded and were available to meet for the focus group were selected to participate in two additional focus groups. Given the resources available, approximately 20 nonmanagement employees were invited to participate in two focus groups instead of including all 68 nonmanagement employees who expressed interest. Across all employee groups, a total of 45 caregiver respondents from the online survey participated in one of the five focus group sessions.

Measures/Data Collection

As BCBS desired a survey that would take, on average, less than 15 min for employees to complete, key variables representing background, work characteristics, and context of care were identified and included as covariates in subsequent quantitative analyses that also aligned with the Stress Process Model.

Background, work characteristics, and context of care.

Sociodemographic information, including age, gender, race/ethnicity, marital status, and education was collected from all survey respondents (see Table 1). Employment information was also collected from all respondents, including employee status (whether hourly or salaried), hours worked per week, length of employment with the organization, and job satisfaction. The 5-item Andrews and Withey Job Satisfaction Questionnaire (AW-JSQ) was administered to all survey respondents (Andrews & Withey, 1976). The AW-JSQ has shown good psychometric properties in prior research (Rentsch & Steel, 1992; van Saane, Sluiter, Verbeek, & Frings-Dresen, 2003). The AW-JSQ includes five items that measure global job satisfaction; item responses range from 1 = terrible to 7 = delighted and are averaged (example items include: “How do you feel about your job?;” “How do you feel about the workers you work with—your co-workers?;” and “How do you feel about the work you do on your job—the work itself?”). The AW-JSW demonstrated strong reliability in the BCBS survey (α = .88). Information specific to context of care, including duration of care, care recipient condition, relationship of caregiver to care recipient, and caregiver living arrangement are included in Table 2.

Table 1.

Background and Work Characteristics (N = 880).

Descriptive variable	Total sample	Caregivers	Noncaregivers
Background characteristics
Female*	82.8%	87.9%	79.0%
Age**	M = 45.59 SD = 10.99	M = 47.77 SD = 10.53	M = 43.99 SD = 11.06
Race/ethnicity (White)	89.2%	87.2%	90.7%
Marital status (married or living with partner)	70.0%	67.9%	71.5%
Education received (bachelor’s degree or higher)	53.5%	50.5%	55.8%
Caregiver	42.3%
Work characteristics
Employee status
Hourly	39.6%	38.8%	40.2%
Salaried	60.4%	61.2%	59.8%
Average hours worked per week	M = 43.14 SD = 7.80	M = 43.21 SD = 8.66	M = 43.10 SD = 7.12
Length of employment at BCBS**
Less than 1 year	17.1%	16.8%	17.3%
1–3 years	19.4%	15.9%	21.9%
4–5 years	9.9%	9.9%	9.8%
6–10 years	15.6%	14.8%	16.3%
11–15 years	16.5%	14.8%	17.7%
16–25 years	12.6%	17.6%	9.0%
25+ years	9.0%	10.2%	8.1%
Job satisfaction questionnaire (range: 1 = terrible to 7 = delighted)	M = 5.42 SD = 0.92	M = 5.38 SD = 0.98	M = 5.46 SD = 0.87

Note. BCBS = Blue Cross and Blue Shield of Minnesota; M = Mean; SD = standard deviation.

^*p < .01.

^**p < .001.

Table 2.

Context of Care and Caregiver Outcomes (N = 370).

Descriptive variable
Context of care
Duration of care (5 years or more)	30.1%
Primary caregiver	58.1%
Care recipient needs care because of a:
Long-term physical condition	61.9%
Short-term physical condition	19.5%
Memory problem	29.5%
Emotional/mental health problem	30.0%
Behavioral issue/developmental/intellectual disorder	15.7%
Relation to care recipient (daughter or son)	48.2%
Sum of care recipient needs	M = 1.66 SD = 0.91
Where care recipient lives (at home alone)	31.2%
Caregiver well-being
Global physical healtha,* (range: 4–20)	M = 15.67 SD = 2.34
Global mental healtha,* (range: 4–20)	M = 14.10 SD = 3.17
Caregiver well-being
Caregiver burden
How burdened the caregiver feels (1 = not at all; 5 = extremely)	M = 2.38 SD = 1.02
Caregiving-work conflict	M = 2.68 SD = 0.91

BCBS = Blue Cross and Blue Shield of Minnesota; M = Mean; SD = standard deviation.

^aA score of 16 or above on the physical health subscale represents self-reported physical health that is comparable or better to the general U.S. population. A score of 15 or above on the Mental Health score suggests similar comparability.

^*p < .001.

Caregiver burden.

Caregivers were asked one item that assessed their overall feeling of burden due to care responsibilities. The item response ranges from 1 = not at all to 5 = extremely. This item was extracted from the Zarit Burden Interview (Zarit, Todd, & Zarit, 1986) and was selected due to its high factor loading demonstrated in prior analyses (Ballesteros et al., 2012).

Physical and mental health.

The 10-item global health measure developed as part of the Patient-Reported Outcome Measurement Information System (PROMIS) project was utilized. The 10 global health items reflect core PROMIS domains of health (physical function, pain, fatigue, emotional distress, and social health domain). PROMIS scoring guidelines were utilized (Hays, Bjorner, Revicki, Spritzer, & Cella, 2009) to create two validated subdomains of well-being: physical and mental health. Both the physical and mental health domains were converted to T scores to determine how well-being of caregiving respondents compared to population-normed well-being scores (see Cella, Gershon, Bass, & Rothrock, 2013; Hays et al., 2009).

Caregiving–work conflict.

A 5-item measure was administered to family caregivers to determine the extent of caregiving impact on their work experience (Pearlin et al., 1990). Example items include “You have less energy for work;” “You have missed too many days;” and “You have been dissatisfied with the quality of your work.” Item responses range from 1 = strongly disagree to 5 = strongly agree and are averaged. The caregiving–work conflict measure showed appropriate internal reliability in the current study (α = .80).

Post-survey focus group.

Focus group facilitators utilized a series of questions to guide conversations, as shown in the Online Supplemental Material. The semi-structured focus group questions aimed to understand how family caregiving was perceived by both employees and supervisors at BCBS, the experiences of being a caregiver while employed at BCBS, and recommendations about how caregivers could be better supported by BCBS.

Analysis

Descriptive data (frequencies and means) and bivariate analyses (T tests and chi-square analyses) were initially examined to explore differences between caregiving and noncaregiving employees. A series of multiple regressions were also conducted to examine Hypothesis 1 (the effects of caregiving status on global physical and mental health, controlling for sociodemographic characteristics of all survey respondents); Hypothesis 2 (the effects of burden on caregiving–work conflict and well-being among caregiving respondents, adjusting for sociodemographic, context of care, and work indicators); and Hypothesis 3 (whether caregiving–work conflict mediated the relationship between caregiving burden and well-being).

The classic Baron and Kenny approach was used to test the mediational effect postulated in Hypothesis 3 (Baron & Kenny, 1986). First, the separate effects of burden and caregiving–work conflict on physical and mental health were examined in multiple regression models (controlling for background, context of care, and work indicators). Then, the effects of burden on caregiving–work conflict were considered per the test of Hypothesis 2 discussed earlier. A final regression model (also part of the Hypothesis 2 test) then served to ascertain whether caregiving–work conflict mediated the empirical relationship of burden on physical and mental health among employed caregivers.

Analyses were performed using SPSS version 21 (IBM Corporation, 2012). Normality, linearity, and homoscedasticity of residual assumptions were examined and met per standard screening criteria (Tabachnick & Fidell, 2007). Similarly, assumptions of collinearity were met according to tolerance and variance inflation factor collinearity statistics.

Data from the focus groups were coded inductively through an iterative process by two of the investigators (DLP and HD) using NVivo version 11 (QSR International Ltd, 2012). Each of the investigators independently read and coded one of the transcripts and then compared codes to develop an initial codebook. As the remaining transcripts were analyzed, the investigators met periodically to discuss any new and emergent codes until a final codebook was formed and used to code all transcripts. When discrepancies in coding occurred, the investigators (DLP and HD) discussed the rationale for adding, deleting, moving, or modifying codes until they mutually agreed on all codes. Transcripts were then recoded using the agreed upon final codebook. The data were coded by applying initial coding, a process whereby the data are compared to form unique codes, and subcoding (Saldaña, 2013). Following principles outlined by Saldaña (2013), coded data were then formed into categories and finally grouped into themes. To enhance validity of the findings, another investigator (J. E. G.) served as a peer reviewer to provide “an external check of the research process, much in the same spirit as interrater reliability in quantitative research” (Creswell, 2013, p. 251).

Integration of the qualitative and quantitative study strands occurred during the interpretation of results. The themes identified from the focus group sessions were compared to the descriptive analyses and multiple regression findings to demonstrate how family caregiving, caregiving–work conflict, and caregiver outcomes intersect or diverge (see Discussion section; Creswell & Plano-Clark, 2010).

Results

Sample Characteristics

Background and work characteristics of the sample are included in Table 1. Over 40% of respondents were or had provided care to someone in the past year (n = 372; 42.3%). Most respondents were women (82.8%), with a significantly greater proportion of female caregivers than noncaregivers (87.9% vs. 79.0%, respectively; p < .01). Caregivers were also significantly older than noncaregivers (M = 47.77 years vs. 43.99 years, respectively; p < .001). Caregivers tended to be employed longer at BCBS than noncaregivers (p < .01).

Context of care characteristics are presented in Table 2. Over half of caregivers considered themselves the person most responsible for providing help to the care recipient (i.e., the “primary” caregiver; 58.1%). The principal reason assistance was provided to care recipients was due to a long-term physical condition (61.9%), and on average care recipients had close to two care needs (M = 1.66; SD = 0.91). Nearly half of caregivers were adult children of the care recipient and had been providing care from 1 to 4 years. Approximately two thirds of care recipients lived at home in the community, either alone (31.2%) or with the caregiver and other relatives (35.7%).

Quantitative Analyses: Hypothesis Tests

Hypothesis 1: Health and well-being of caregivers and noncaregivers.

As shown in Table 2, caregivers had slightly but significantly lower physical health scores than noncaregivers (M = 15.32 vs. 15.92, respectively; p < .001). Even more pronounced, caregivers had significantly lower mental health than noncaregivers (M = 13.43 vs. 14.60, respectively; p < .001). Caregivers scored below population-normed scores of physical and mental health (below 16 on the physical health subscale and below 15 on the mental health score; Cella et al., 2013; Hays et al., 2009).

Multiple regressions were conducted to examine the impact of caregiving on global physical and mental health, adjusting for sociodemographic, background, and work characteristics of survey respondents (n = 880; see Table 3). Caregivers reported significantly lower physical and mental health when compared to noncaregivers (β = −.10, p = .002; β = −.17, p < .001, respectively). The adjusted R² of .14 in the physical health model indicates that caregiving status accounted for a small amount of variance in this outcome, although the adjusted R² of .20 for mental health suggested that this model predicted more moderate overall variance. Additional employment-related variables in these models were significantly associated with physical and mental health and proved more potent as predictors of well-being when compared to caregiving status. Respondents who indicated greater job satisfaction reported more positive physical and mental health (β = .23, p < .001; β = .36, p < .001) while those who were hourly employees indicated poorer physical health and mental health than salaried employees (β = −.23, p < .001; β = −.16, p < .001), respectively.

Table 3.

Multiple Regression Analysis-Global Physical and Mental Health (N = 880).

	Physical healtha				Mental healthb
Variable	B	SE (B)	β	T	B	SE (B)	β	T
Respondent is caregiver	−.49**	.16	−.10	−3.11	−1.09***	.21	−.17	−5.26
Average hours worked per week	.00	.01	−.01	−.41	.00	.01	−.01	−.23
Age	.00	.01	−.02	−.46	.02*	.01	.08	2.11
Gender	−.16	.20	−.03	−.76	−.22	.27	−.03	−.80
Caucasian	−.02	.25	.00	−.09	−.33	.33	−.03	−1.00
Married/living with partner	.14	.17	.03	.81	.51*	.22	.08	2.31
Bachelor’s degree or higher	.44*	.17	.09	2.54	.23	.23	.04	1.03
Hourly employee	−1.01***	.18	−.22	−5.58	−.99***	.24	−.16	−4.16
Job satisfaction	.61***	.08	.24	7.25	1.26***	.11	.37	11.47
Length of employment at BCBS	.12*	.05	.10	2.48	.04	.06	.02	.62

Note. BCBS = Blue Cross and Blue Shield of Minnesota; B = unstandardized regression coefficient; SE = standardized error; β = standardized regression coefficient; Sig. = significance.

^aAdjusted R² =.14.

^bAdjusted R² = .20.

^*p < .05.

^**p < .01.

^***p < .001.

Hypothesis 2: Predictors of caregiving–work conflict, physical health, and mental health among caregivers.

Table 4 shows the results from multiple regression models that identified predictors of caregiving–work conflict, physical health, and mental health among BCBS employees who were caregivers. These models featured additional background variables related to context of care, care recipient needs, and caregiving burden and accounted for moderate amounts of variance (caregiving–work conflict adj. R² = .35; physical health adj. R² = .26; mental health adj. R² = .31). The most potent predictor of caregiving–work conflict was caregiver burden; BCBS caregiving employees who indicated greater care-related burden also reported more tension between work and care responsibilities (β = .45, p < .001). In both the physical and mental health models, three variables emerged as significant predictors. Caregivers who were hourly employees were more likely to indicate less positive physical and mental health (β = − .21, p < .001; β = −.13, p < .05, respectively). Those who indicated higher job satisfaction also expressed more positive physical and mental health (β = .11, p < .05; β = .20, p < .001). Caregivers who indicated greater caregiving–work conflict reported less positive physical and mental health (β = −.32, p < .001; β = −.33, p < .001); caregiving–work conflict was the most potent predictor of caregiver well-being in these models.

Table 4.

Multiple Regression Analysis-Predictors of Caregiving–Work Conflict, Physical Health, and Mental Health Among Caregivers (N = 370).

Caregiving-work conflicta
Variable	B	SE (B)	β	T
Average hours worked per week	−.01	.01	−.07	−1.48
Age	.00	.01	.02	.35
Gender	−.05	.13	−.02	−.35
Caucasian	−.11	.13	−.04	−.81
Married/living with partner	.03	.09	.02	.50
Bachelor’s degree or higher	.05	.09	.03	.50
Hourly employee	.07	.10	.04	.69
Care recipient care need
Long-term physical condition	.29	.29	.16	1.02
Short-term physical condition	.39	.30	.17	1.32
Memory problem	.11	.29	.06	.39
Emotional/mental health problems	.19	.30	.10	.65
Behavioral issue/developmental-intellectual disorder	.34	.34	.14	1.01
Sum of care recipient needs	−.17	.27	−.17	−.63
Caregiver lives with care recipient	−.10	.09	−.05	−1.07
Spousal caregiverd	.18	.18	.05	1.02
Adult child caregiverd	.01	.09	.01	.12
Primary caregiver	.08	.10	.04	.79
Duration of care	.01	.04	.02	.35
Burden	.41***	.05	.45	8.79
Caregiving-work conflict	-	-	-	-
Job satisfaction	−.25***	.04	−.27	−5.73
Length of employment at BCBS	.02	.02	.03	.64

	Mental healthb				Physical healthc
Variable	B	SE (B)	β	T	B	SE (B)	β	T
Average hours worked per week	−.03	.02	−.08	−1.53	−.01	.01	−.04	−.83
Age	.02	.02	.07	1.17	.00	.01	−.02	−.33
Gender	−.76	.49	−.07	−1.57	−.26	.35	−.04	−.72
Caucasian	−.45	.49	−.05	−.90	.17	.36	.02	.46
Married/living with partner	.59	.34	.08	1.76	.20	.25	.04	.82
Bachelor’s degree or higher	.03	.35	.01	.10	.47	.25	.10	1.85
Hourly employee	−.85*	.38	−.13	−2.25	−1.00***	.28	−.21	−3.58
Care recipient care need
Long-term physical condition	1.28	1.10	.18	1.17	.89	.80	.18	1.11
Short-term physical condition	1.67	1.13	.20	1.48	1.35	.83	.23	1.63
Memory problem	1.58	1.11	.22	1.42	.67	.81	.13	.83
Emotional/mental health problems	.93	1.17	.13	.83	.53	.82	.11	.65
Behavioral issue/developmental-intellectual disorder	−1.17	1.28	.13	.91	.98	.93	.15	1.05
Sum of care recipient needs	−1.64*	1.02	−.45	−1.61	−.77	.74	−.30	−.104
Caregiver lives with care recipient	.11	.35	.02	.30	.10	.26	.02	.40
Spousal caregiverc	−.99	.66	−.08	−1.49	.15	.49	.02	.31
Adult child caregiverc	.40	.34	.06	1.19	.40	.25	.09	1.63
Primary caregiver	−.10	.36	−.02	−.27	−.23	.26	−.05	−.85
Duration of care	.13	.15	.04	.86	−.09	.11	−.04	−.83
Burden	−.29	.20	−.09	−1.46	−.22	.14	−.09	−1.52
Caregiving-work conflict	−1.23***	.22	−.33	−5.71	.82***	.16	−.32	−5.22
Job satisfaction	.68**	.17	.20	3.87	.27*	.13	.11	2.12
Length of employment at BCBS	.02	.09	.01	.20	.12	.07	.10	1.83

Note: B = unstandardized regression coefficient; BCBS = Blue Cross and Blue Shield of Minnesota; SE = standardized error; β = standardized regression coefficient; Sig. = significance.

^aAdjusted R² = .35.

^bAdjusted R² = .31.

^cAdjusted R² = .26.

^dReference category: other relationship to care recipient.

^*p;<.05.

^**p;<.01.

^***p;<.001.

Hypothesis 3: The mediational effects of caregiving–work conflict.

The results also support the hypothesized mediational effect, whereby caregiving–work conflict mediates the empirical relationship between caregiving burden and well-being among employed caregivers. Using the Baron and Kenny approach (Baron & Kenny, 1986) described in detail earlier, two subsequent regression models were tested that were identical to the models in Table 3 but removed caregiving–work conflict (not presented here due to space considerations; please see Online Supplemental Material). In both regression models, caregiving burden emerged as a highly significant predictor of physical (β = −.24, p < .001) and mental health (β = −.24, p < .001). As caregiving burden was nonsignificant in Table 4 when caregiving–work conflict was included as a predictor of physical and mental health, the quantitative results indicate that caregiving–work conflict fully mediates the empirical association between caregiving burden and physical and mental health in employed caregivers (burden→caregiving–work conflict→physical and mental health).

Qualitative Analysis

Four overarching themes emerged from the five focus groups held with BCBS employee caregivers. These themes included balancing work and caregiving, challenges and effects faced by working caregivers, caregiver support, and areas of opportunity to enhance workplace support (Table 5).

Table 5.

Qualitative Themes With Their Definitions and Codes.

Theme	Definition	Codes	Exemplary quotes
Balancing work and caregiving	Perception that employed caregivers constantly juggle various tasks and responsibilities that come from formal work and providing care. The strategies people use to be able to perform both their caregiver and employee duties.	Challenges to balancing work and caregiving: pressures at work, not wanting to let others down, feeling that coworkers or managers won’t understand Strategies for balancing work and caregiving: being flexible, prioritizing, setting boundaries, working outside of the office (e.g., taking work home or to doctor’s appointments), and using vacation/PTO for caregiving Work a reprieve	Challenges: “I would not expect my manager to understand what I need to do in a day. She doesn’t have the perspective. I haven’t shared it and so how would she know the strains and stresses that are on me?” Strategies: “I’d get up in the morning while she was still sleeping and work a few hours, take care of her throughout the day and then work after she went to bed. Maybe some days I didn’t necessarily give the whole 8 hours in, but by the end of the week I had 40 hours.”
Challenges and effects	Difficulties and impacts on the caregiver that stem from their role as a caregiver, as well as from balancing work and caregiving roles.	Challenges: caregiving from a distance, family dynamics, navigating the medical system, caregiving as thankless work, being “tied to the job” for benefits, being overwhelmed or feeling that responsibilities are too much Impacts of caregiving: financial consequences, personal and professional sacrifices, sense of obligation, negative effects on caregiver’s health and well-being: anxiety/worry, exhaustion, guilt, stress	Challenges: “Sometimes I just feel overwhelmed because I feel like she calls me for everything. I help her. She’s in a care center right now, and we’re trying to get her into an apartment or something, so we’re going through public housing, we’re talking with an attorney, trying to get disability for her. And I mean I’m glad I’m there, but some days, it’s hard. Impacts: “He didn’t have anybody to take care of him so I let him move back in with me. It’s sort of put a strain on my relationship with my own children because my time, my energy, and everything. ‘Oh, I can’t do that with you because I got to take [name] here and do this and do that.’”
Support	Assistance and support caregivers receive from family and friends as well as within the workplace that contributes to or detracts from their ability to balance work and caregiving.	External support Positive: assistance with caregiving (external to BCBS): spouse, other family, church, friends and community; paid caregivers (e.g., home health aide) Negative: limited or nonexistent family support Workplace: Organization support Positive: benefits, flexibility Negative: lack of support Workplace: Leadership support Positive: attributes of good leaders, benefits of supportive leaders, flexibility, strategies to support caregivers Negative: perceived discrimination, generation gap Workplace: Team support Positive Negative	External support—positive:” If it weren’t for my siblings or the other caregivers I don’t know what I would’ve done,”External support—negative: “My sister lives in [another state] and she’s not here to see what’s going on with my parents and she has very different opinions on how they should be cared for and I hesitate to say it, but I want to … I feel like saying, ‘Get your butt up here and see for yourself.’” Workplace support—positive: “If you’ve got people around you whether it’s your own team, colleagues, friends, other people in the department or division it’s amazing how decent people will step up and fill that gap … it’s really that support system around you with colleagues that to me has made all the difference.” Workplace support—negative: “I had a supervisor tell me once that I didn’t have 40 hours of PTO, so I couldn’t use my PTO because I only had 8 hours left. I’m like, ‘I can use it. I’ve got it.’ She goes, ‘What if you have to be out for a long time?’ I said that’s what FMLA is for. She did not like that at all.”
Areas of opportunity	Things that could be done within the workplace to enhance the support of caregivers.	Promoting a supportive workplace culture Improved policies and procedures: FMLA, constricting or outdated organizational policies Organizational resources: caregiver support groups, information and guidance, trainings or seminars More flexibility More consistency across departments or job classes	Organizational resources—information and guidance: “It would be helpful to have a case manager or someone to help us with the legalities of things or even like some long-term care benefits that we could have … I’m going to call my parents tonight and talk long-term care insurance.” More flexibility: “I think we all feel that we can do an exceptional job if we are allowed to pick and choose when we get that done or if we are allowed to organize our schedule around shuffling when I need to be at this event or this appointment and work on that presentation … The ability to be able to login and be with your mom—everybody wins. There’s no loss there and I’m sure your work quality is as good if not better because you’re there with her and you’re not worrying. You’re not distracted.”

Note. BCBS = Blue Cross and Blue Shield of Minnesota; FMLA = Family Medical Leave Act; PTO = paid time off.

Balancing work and caregiving.

Focus group participants held two distinct roles: employee and caregiver, and they discussed how they must constantly balance the two. This balance included maintaining boundaries between work and caregiving, keeping up with work responsibilities, and occasionally requesting that colleagues take on additional work to help them through a deadline or a difficult situation. One participant described the delicate balancing act, stating,

I think sometimes you are involved in the situation where it’s been a challenge to figure out which ball we are going to drop. Sometimes a lot of things coincide. You sit there and you figure out what suffers. I think we’ve all been in the position where we consciously have made a decision … in order to take care of what we thought was more important. The ball is made out of glass as opposed to rubber.

Caregivers described various strategies they used to balance work and caregiving responsibilities. These included being flexible, prioritizing, and setting boundaries when necessary. Many caregivers reported taking work with them as they provided care (such as to the hospital while sitting with a sick parent), working during off hours (such as late at night or on the weekends), or using paid time off (PTO) for caregiving responsibilities. One participant said,

I took conference calls when we were [working on a large] bid in a parking lot and walking around the streets of [city] because my mom had a meeting at the TCU (transitional care unit) and I needed to be there but I needed to be on this conference call … I’ve done conference calls in her apartment bedroom. I have done things with my laptop on the counter with headphones on while I’m in the clinic waiting.

Challenges and effects faced by working caregivers.

Participants also discussed challenges and effects that result from balancing work and caregiving. Challenges associated with caregiving included dealing with difficult family dynamics, providing caregiving support from a distance, or not feeling appreciated for caregiving efforts. Participants also noted financial hardships and professional sacrifices they had made as a result of their caregiving responsibilities, including shying away from promotional opportunities or other career advancements. Other participants described giving up needed vacations because of their caregiving responsibilities, having little time for hobbies or relaxation, and not having adequate time to spend with other family members or friends. One participant described wanting to spend time with her husband saying,

I have a great husband and we’ve been married for [many] years. I think we’re there for the long term. But sometimes I think he would really like it if I would sit on the couch at night. I don’t have time …. But even just to be a good wife to sit next to him, and that happens maybe every two months there will be a movie I like to sit down for.

Many caregivers described experiencing stress, exhaustion, guilt, and other negative impacts on their health and well-being as a result of managing their employment and caregiving responsibilities. Some described neglecting their own well-being, such as one participant who said, “it’s kind of yourself that you leave out … you don’t take care of yourself.”

Caregiver support—positive and negative.

Participants described support as being important both within and outside the workplace. When support was present, caregivers noted that it enabled them to better manage the complex balancing act of work and caregiving. When support was lacking, this detracted from the person’s ability to manage both roles. Whereas some employees noted receiving considerable workplace support and flexibility related to their caregiving responsibilities, others had less positive experiences, with some participants fearing negative consequences at work. In general, managers and salaried employees tended to speak more favorably of organizational support than hourly employees, particularly in terms of workplace flexibility. Whereas one manager said, “Flexibility has been critical in my life. If I didn’t have that I wouldn’t be able to do what I do,” an hourly employee commented,

I know other people who are not in [a supportive] situation, where they are afraid to call in or they’re afraid of what repercussions they are going to have from their coworkers or from their boss because they need to use that time [for caregiving].

Relationships and interactions with managers were described by many caregivers as being vital to their ability to effectively manage their dual roles. One participant remarked,

That’s the ultimate linchpin in my view. If your leadership is flexible and supports that kind of creativity [coming up with flexible ways to complete work tasks to support caregiving], life is good. And if your leadership is ignorant and does not support that, life is unbearable. It’s actually unbearable.

Organizational policies and leadership behavior were often described as having a trickle-down effect. When management staff set a positive tone toward family caregivers, other team members may be more likely to demonstrate support and encouragement toward fellow employees in turn. As one participant said when describing her coworkers,

It moves mountains when you can have that support to be, “I’ve got to go.” There’s no questions asked. There’s no doubt at work. There’s no, “I did this for you.” There isn’t any negativity once you return. That in itself is a huge elephant off my shoulders.

In addition to support received from within BCBS, many participants described assistance with caregiving they received from family members, members of the community, and paid caregivers (such as home health aides) as contributing to their ability to continue caregiving responsibilities while employed. A manager commented:

For me it’s my sisters and brothers. I have one sister out of town that will come in once a month and stay a week. She has her home health aides and my mother has tons of grandkids. My kids, my sister’s kids will come over and do stuff.

Although many focus group participants described receiving crucial support in their caregiving roles from family and friends, others discussed ways in which family members in particular could make caregiving more difficult. A number of caregivers described family conflict or difficult family dynamics as detracting from their ability to effectively manage work and caregiving. For example, several participants talked about feeling “dumped on” or abandoned by siblings. One exempt employee said,

The [other siblings] that don’t [help] we’re not even going to put them in on the email chain anymore. We’ll just let them go because they let us down or they’ve not followed through or they’ve done things for my mom, but they haven’t done them happily or willingly. They do it, but they’re not happy about it and it’s just like, “Then just forget it. You’re not part of this anymore.”

Areas of opportunity to enhance workplace support

Participants discussed several changes to organizational policies and resources that could enhance workplace support for family caregivers. Many expressed confusion about the Family Medical Leave Act (FMLA), suggesting that more formal guidance and support in navigating FMLA would be helpful. Similarly, some participants indicated that BCBS should provide more centralized guidance to employees in navigating other complex policies and systems, such as long-term care insurance. A few participants recommended that BCBS offer seminars and other educational opportunities on topics related to caregiving and self-care, and several others recommended that BCBS could help to organize caregiver support groups. Several employees described support received through BCBS’s Employee Assistant Program as being extremely helpful. Others said their manager had been instrumental in giving them information about caregiver-related resources and supports.

A consistent theme throughout the focus groups was a desire for more workplace flexibility, particularly related to work hours. Some discussed the organization’s rules related to maintaining full-time positions, with a number of caregivers noting a desire to temporarily work reduced hours (such as working a 0.8 appointment without losing the right to return to full-time when caregiving responsibilities relaxed). Similarly, many caregivers discussed a need for greater consistency in the application of company policies and procedures. Several caregivers discussed a perception that some departments or employee groups received more flexibility than others, and others described policies they perceived to be outdated or overly punitive, such as those related to attendance.

Nonmanagerial employees in particular expressed a desire for greater flexibility, such as being able to work 8 hr on Saturdays, or longer days in general, in order to take time off at other times during the week. Others described wanting more flexibility related to PTO, including being able to use “volunteer PTO hours” (hours provided by BCBS to support employee volunteerism) for caregiving or the ability to donate PTO to other employees who may be caring for a family member.

Finally, a number of participants discussed a desire for the organization to more actively promote a workplace culture that is supportive of family caregivers. Suggestions included providing training to managers so they better understood the needs and experiences of family caregivers, encouraging managers to more clearly and consistently communicate with employees about company policies and resources available for caregivers, and promoting opportunities for team members to support and encourage each other (such as through donated PTO time). Some expressed concerns that younger or newer leaders coming into the organization were less likely to understand or support employees who are caregivers. Others talked about wanting to feel more appreciated by BCBS management in general.

Discussion

The objective of this study was to generate mixed-methods data to empirically describe the effects of caregiving on well-being for those who are employed as well as processes (e.g., caregiving–work conflict) that may help to explain well-being among employed caregivers. Qualitative data were utilized to expand upon these findings and to examine how employees experience family caregiving, and how their employer does or does not facilitate their caregiving role. The quantitative and qualitative data expand upon the extensive research on family caregiving and employment (Schulz & Martire, 2009; Stephens & Franks, 2009) and provide new insights into the key role of caregiving–work conflict in driving caregiver well-being. The results emphasize the need to focus on the work–family caregiving nexus when crafting new or delivering existing evidence-based interventions.

As summarized earlier, many prior studies of caregiving and employment exist. However, one of the key findings to emerge from our quantitative analyses was the robust influence of caregiving–work conflict on caregiver mental and physical health. In this regard, the results align with the competing demands hypothesis, whereby the tension of navigating work and care responsibilities has potentially negative ramifications for employed family caregivers (Stephens & Franks, 2009). The majority of descriptive and intervention studies on family caregiving often select caregiver burden, stress, or depression as outcomes of interest (Burgio & Gaugler, 2016; Gitlin & Hodgson, 2015), with less attention placed on the “secondary” stressors that reflect the spread of stress beyond the immediate caregiving situation and into other life domains, such as caregiving–work conflict. The integrated data here emphasize that this is an important conceptual and clinical omission. Single studies of secondary stressors in caregiving, such as gender role conflict (Bai, Liu, Baladon, & Rubio-Valera, 2017), financial and family conflict (Butterworth, Pymont, Rodgers, Windsor, & Anstey, 2010), the reconciliation of work and family care responsibilities (Wang, Shyu, Chen, & Yang, 2011), and the implications of work–family conflict and schedule control among men and women employed in long-term care settings (DePasquale, Davis, et al., 2016; DePasquale, Zarit, et al., 2018) have emerged. However, the present study demonstrates that, when considering overall caregiver health and well-being, caregiving–work conflict appears to supersede burden as an important potential clinical target for services and supports.

A secondary yet intriguing finding that emerged was the challenges of hourly employed caregivers. Prior research suggests that full-time employed caregivers appear less likely to indicate distress and negative outcomes when compared to part-time employees (Hansen & Slagsvold, 2015; Pavalko & Woodbury, 2000; Stephens & Franks, 2009; Stone & Short, 1990). However, to our knowledge, no studies have examined the variations between salaried and hourly employment arrangements. For example, the lack of flexibility that is often apparent in hourly employment may further complicate work and family caregiving arrangements. This association emerged in the qualitative focus group data as well; although multiple participants emphasized positive aspects of the organizational culture at BCBS, such sentiments appeared more prominent among salaried workers. For hourly employees, the strain of balancing caregiving and work responsibilities and the lack of flexibility to do so may have resulted in greater difficulties.

Although many prior studies of caregiving and employment have compared indicators of well-being between caregivers and noncaregivers, the current study builds upon this work to utilize measures that were population normed (Cella et al., 2013; Hays et al., 2009). The results implied that caregivers at BCBS experienced poorer mental and physical health than noncaregivers at BCBS as well as adults in the United States. These results were expounded upon by focus group participants, who described the difficult balancing act of negotiating employment and caregiving responsibilities and the emotional and physical consequences they often experienced as a result. Taken together, these findings confirm the well-documented health implications of family caregiving for those who work (Fredriksen-Goldsen & Scharlach, 2006; Giovannetti et al., 2009; Stephens et al., 1997; Swanberg, 2006) and that exposure to the demands of family care can result in negative mental and physical well-being. At the same time, this study adds to our understanding of the importance of perceived workplace support for caregivers’ overall well-being. Our findings also suggest that perceived workplace support may contribute to employee morale and loyalty to the organization.

Similar to well-being, BCBS caregivers experienced acute, negative influences of caregiving which were made apparent in the qualitative data. Many focus group participants described experiencing limited formal or informal support within the workplace, as well as challenges they experienced in managing both caregiving responsibilities and employment. Such findings imply the need for greater organizational support for family caregivers, including work flexibility (Berecki-Gisolf, Lucke, Hockey, & Dobson, 2008; Doty & Crown, 1998; Henz, 2006; Mutschler, 1994; Pavalko & Artis, 1997). Interestingly, although caregivers experienced challenges when balancing the time demands of work and caregiving, many BCBS caregivers described being reluctant to reduce their work hours or modify their job responsibilities. The focus group data emphasized that BCBS caregivers were dedicated to carrying out their jobs as expected regardless of family care responsibilities. At the same time, several focus group participants described wanting to temporarily reduce their work hours because of caregiving but were unable to do so due to workplace policies. Others noted feeling considerable pressure to keep up with job responsibilities, often to avoid being perceived as less productive than other employees. Although many employees suggested that their own internal expectations pushed them to excel at work, improved workplace support and flexibility could perhaps contribute to better physical and mental well-being on the part of employee caregivers.

There are several important limitations to this study. Although the sequential explanatory mixed-methods design provided greater insight as to the reasons why caregivers experienced greater distress and negative well-being and also yielded a number of recommendations, the study was cross-sectional, and thus causal relationships were not possible to discern. The study took place at a single, large employer with its own history and workplace culture; whether these findings are generalizable across employment settings is unknown. Because of the need to limit respondent burden and improve overall response rates, it is possible that several key covariates were not included (e.g., number of living children or where children were currently living), and the precision of certain items was lacking (e.g., care recipient health needs). Due to the region and the type of organization where this study took place, the sample was largely Caucasian and well educated, further limiting its generalizability.

The results emphasize that even for an employer that offers a robust set of caregiver-friendly programs and policies, there are opportunities to enhance workplace support. Moreover, the ability of employed caregivers to balance work and caregiving may be affected, perhaps dramatically, by the support they receive from the workplace. Further, the ability of these employees to access and use caregiver support programs effectively is largely mediated by management. Simply offering policies such as flexible time arrangements or location of work (Beauchamp, Irvine, Seeley, & Johnson, 2005; Utz, Lund, Caserta, & Wright, 2012) is insufficient. As heard in the focus groups, perception of supports varied based on specific work assignments, as well as whether an employee was salaried or hourly. In general, salaried employees described more workplace supports, including greater flexibility. Focus group participants emphasized the need for greater consistency in the application of flexibility across work units and job classes, as well as a desire for better communication within the organization about policies and services that could be used to address caregiving needs. Key recommendations that emerged from both the quantitative and qualitative data emphasize: (a) the need to link employed caregivers to some type of navigator within the organizational structure who can help caregivers better access employer programs (e.g., an occupational health nurse that can provide employed caregivers with a “Caregiving Worksite Toolkit” that provides information on attendance and leave policies, educational resources, and appropriate community resources; see Pitsenberger, 2006); (b) moving beyond services and building community among existing family caregivers through peer-led support groups, “lunch and learn” activities, or similar initiatives (in several of the focus groups, participants often indicated surprise at the number of other BCBS employees in similar situations as themselves); (c) more effective management training; and (d) taking steps to ensure consistent application of caregiver support policies across work units and employee job classes, as well as enhanced workplace flexibility, particularly for hourly employees.

The findings also emphasize the importance of the caregiving–work conflict stressor and should inform the content of new family caregiving interventions as well as the delivery of existing evidence-based interventions to family caregivers. As noted earlier, the large majority of family caregiving interventions’ clinical content is focused on the stress of providing care to a relative and how to mitigate such stress (e.g., via improved coping strategies, respite, behavioral management approaches, building social support). However, given the importance of caregiving–work conflict, the need to address how to navigate family caregiving demands with employment-based ones may offer a critical and novel clinical focus. In addition, the findings here indicate that employment settings may be fertile environments for delivering caregiving interventions. Establishing partnerships with employers to deliver and evaluate evidence-based programs to support family caregivers would advance research and align with recent recommendations from the National Academies of Sciences, Engineering, and Medicine (2016).

Author Biographies

Joseph E. Gaugler, PhD, is long-term care professor in Nursing at the University of Minnesota.

Deborah L. Pestka, PharmD, is a PhD candidate in Social and administrative pharmacy in the College of Pharmacy at the Universityof Minnesota.

Heather Davila, MPA, is a PhD candidate in Evaluation Studies and a research fellow in the Division ofHealth Policy & Management at the University of Minnesota.

Rebecca Sales, MPH, is a research associate at the Amherst H. Wilder Foundation.

Greg Owen, PhD, is a consulting scientists at the Amherst H. Wilder Foundation.

Sarah A. Baumgartner is a human resources partner at Blue Cross Blue Shield of Minnesota.

Rocky Shook, PhD, is a manager, Market Research and Insights at Blue Cross Blue Shield of Minnesota.

Jane Cunningham is a program developer, Community Services for Aging and Early Childhood Services at the Amherst H. Wilder Foundation.

Maureen Kenney, MPA, is director of Aging Services at Amherst H. Wilder Foundation.