The effect of stress proliferation on depression in South African grandparent caregivers

Stress proliferation

Stress proliferation studies have been conducted primarily with samples from the United States, examining different types of primary and secondary stressors. For example, Pearlin et al. (1997) studied AIDS caregiving in a sample of unpaid, non-family caregivers in the United States, showing that the stressors of AIDS care proliferate beyond the immediate care relationship, including greater work stress and reduced social engagement. They studied urban, US AIDS caregivers who have higher than average education and adequate income, which is not typical for grandparent caregivers in South Africa. Unlike the US study, AIDS caregivers in South Africa are older, possess lower levels of education, and have lower income. LeBlanc et al. (2015) examined proliferation of stressors in minority, gay couples in the United States. Benson and Karlof (2009) found a relationship between anger, stress proliferation, and depressed mood in parents with children with autism in the United States. In China, grandparents who are caring for grandchildren who have migrated to work in the cities have additional responsibilities, such as to attend parent meetings at the school, take the children to the doctor, and the like (Baker and Silverstein, 2012).

Unlike the US studies, AIDS caregivers in South Africa are older, have less education, and have a lower income and therefore fewer resources to provide good care while protecting their own physical and emotional well-being. Elder South African grandparent caregivers experience a proliferation of stressors when caring for an adult child with AIDS, and they also experience stressors from caring for their child’s young children (the grandchildren), especially if the grandchild’s other family members are HIV positive, deceased, or not present in the child’s life. In this case, caring for an adult child with AIDS necessitates taking on another role as caretaker for the orphaned grandchildren, which has its own set of stressors that exist indirectly as the result of AIDS care, such as dealing with the grandchild’s behavioral problems due to grief and lack of adequate financial support to account for the additional members of the household (Oburu and Palmérus, 2005). The extent of these demands on the caregiver has the potential to intensify psychological distress in the form of anxiety (Casale et al., 2014; Sands and Goldberg-Glen, 2002), depression (Ku et al., 2013), and post-traumatic stress disorder (PTSD) (Absiubong et al., 2011; Nyirenda et al., 2013), among other stressors. The caregiver’s gender, health, race, family financial support, and financial resources have the potential to moderate or exacerbate the effects of this proliferation of stressors (Ku et al., 2013; Kuo et al., 2012).

Although there are studies of the primary and secondary stressors for orphans and their caregivers using South African samples, few use the theoretical framework provided by the concept of proliferation of stressors. Hlabyago and Ogunbanjo (2009) interviewed nine grandmothers in a clinical, qualitative study, who expressed pain and grief from caring for their adult children as well as for their orphaned grandchildren. These caregivers describe the emotional pain of the grief of losing an adult child from a long and difficult disease and the financial strain of caring, due to medical and funeral expenses and lost work time (primary stressors). This pain and grief are compounded by the expenses of providing food, clothing, and shelter for additional orphaned grandchildren (secondary stressors). This is especially true if the ailing adult child’s partner is also HIV positive, deceased, or not present in the child’s life.

Overall, previous studies have addressed the negative strains that develop from combined care-giving roles, which increase stress on the caregivers (Benson and Karlof, 2009; Pearlin et al., 1997; Penning, 1998). In addition, a proliferation of stressors (Hlabyago and Ogunbanjo, 2009; Pearlin et al., 1997) has the potential to negatively affect the mental health of caregivers by increasing depression.

Grandparents as caregivers

HIV/AIDS places considerable strain on the family support system (Dawson, 2013) by disrupting kin networks caused by large numbers of deaths due to AIDS (Shaibu, 2013). In Zimbabwe, it is estimated that kinship resources (adults available as caregivers) for children under the age of 10 declined by around 50 percent between 1990 and 2010 (Zagheni, 2011). However, research has found that the Black family norms have been flexible enough to adapt traditional family arrangements to enssure that family members receive care (Block, 2014), which includes greater reliance on the older generation.

South Africans rely on the support of grandmothers to care for grandchildren, such as in the case of other parental illnesses, work, or educational needs (Dolbin-Macnab and Yancura, 2017; Madhavan, 2004; Mtshali, 2015; Schatz and Ogunmefun, 2007). However, grandparents also benefit from the support of older grandchildren who may live in grandparents’ households to help grandparents with household chores and other tasks (Madhavan, 2004).

It was found that AIDS orphans are more likely to live with grandparents if their mother passes away from AIDS or if they are double orphans, which means that both parents are deceased (Beegle et al., 2010). In the case of HIV/AIDS care, unlike other types of custodial grandparenting, adult children are not available to provide economic support because of poor health or being deceased (Chazan, 2008), which intensifies the indirect financial stressors associated with care. Studies have not examined the effect of caring for orphans after having cared for those orphans’ parents, which increases the proliferation of stressors.

Emotional stress and depression

Studies of adults caring for children with AIDS in Southern Africa have found high levels of depression and emotional distress (Casale et al., 2014; Kuo et al., 2012; Mhaka-Mutepfa et al., 2014; Nyirenda et al., 2013; Tamasane and Head, 2010). For example, Hamad et al. (2008) found a high percentage of South African caregivers reporting clinical levels of depression (64.5% women and 50.4% men). Also, Kuo et al. (2012) found that, among grandparents caring for grandchildren who are AIDS orphans, almost one-third are depressed (Muliira and Muliira, 2011). Research has found that caregivers for individuals with HIV/AIDS have a greater risk of PTSD (19%), which is even higher if they are caring for orphans whose parents are deceased (28%) (Blustein et al., 2004; Kuo et al., 2013).

Research is mixed as to the impact of providing care for adult children or grandchildren on grandparents’ mental health. According to Kuo et al. (2012), having young children in the household or caring for adult children who are ill increases depression in grandparent caregivers. On the other hand, researchers using the Cape Area Panel Study (Ardington et al., 2010) found no relationship between depression and caring for a grandchild or having an adult child who had died. In these studies, researchers examine either the independent effects of caring for an adult child with AIDS or the independent effects of raising a grandchild on depression rather than estimating the combined effects of engaging in both types of care, sequentially or concurrently (Ardington et al., 2010; Casale et al., 2014; Kuo et al., 2012; Mhaka-Mutepfa et al., 2014; Nyirenda et al., 2013; Tamasane and Head, 2010). We suggest that the strongest relationship of caring to depression occurs when the grandparent is caring for both child and grandchild/children, and therefore it is necessary to estimate not only the independent effects of each type of care, but also the combined effects of caring for both.

In addition, there is a secondary stigma experienced by caregivers. Due to fear of the disease, neighbors family and friends may avoid contact with the caregiver as well as the person with AIDS. This creates isolation and loneliness, which increases emotional distress in adult caregivers (Ogunmefun et al., 2011) by reducing the amount of emotional and instrumental support available from family and the community. Emotional support is an important contributor to coping in elderly caregivers, which is diminished because of this stigma (Gerard et al., 2006). In one study, 60 percent of men had not told others that their status was positive, 40 percent of men reported discrimination, and 20 percent lost a place to stay because of others learning about their HIV/AIDS status (Simbayi et al., 2007).

Research has shown that some of the most important contributing factors that increase or reduce distress as a result of caregiving fall in the following four categories: economic resources (DeSilva et al., 2013; Kuo et al., 2012; Tamasane and Head, 2010), education (Nyirenda et al., 2013; Tomita and Burns, 2013), health and physical disability (Ice et al., 2008; Tomita and Burns, 2013), and family structure (Shaibu, 2013; Tomita and Burns, 2013).

A prolonged illness for indigent caregivers and the responsibility for raising orphans can exacerbate poverty. A Ugandan study found that grandparents caring for orphans are more likely to live in larger households with less income (UNICEF South Africa, 2014). Consequently, researchers have found that households with orphans have high rates of food insecurity, with one household in five not having had enough food during the past month (DeSilva et al., 2013; Kuo et al., 2012). Often the elder caregiver goes hungry so that the grandchildren can be fed. South Africa provides a means-tested pension for the elderly, which is often too small to provide adequate food and nutrition for grandmothers and their households (Chazan, 2008; Kaseke, 2010). The added economic strain has the potential to increase stress and lead to greater depression (Dolbin-Macnab and Yancura, 2017).

Research on the effect of education on depression in caregivers is mixed. Studies have found that the more education a person has, the less they are depressed (Nyirenda et al., 2013; Tomita and Burns, 2013), while others have found no effect of education on depression among caregivers (Beegle et al., 2010; Musil et al., 2013).

Older adults have an increased risk of poor health and disability (Ice et al., 2008), which affect their ability to provide care and affect their mental well-being. A national study of the health of older South Africans found that lower quality of life, functional disability, and chronic conditions are positively related to depression (Peltzer and Phaswana-Mafuya, 2013).

Research in Kenya, Europe, and Taiwan has found that grandparents who are caring for grandchildren are healthier than other grandparents (Di Gessa et al., 2016; Hughes et al., 2007; Ice et al., 2008; Kagotho and Ssewamala, 2012; Ku et al., 2013). It is suggested that the satisfaction of caring for grandchildren may have a positive effect on well-being (Hughes et al., 2007; Ice et al., 2008; Ku et al., 2013). It is also possible that grandparents who are healthier are more able to take on the care of children and grandchildren. However, those elder caregiving grandparents who have cared for their ailing adult child and the adult child’s children have experienced considerable economic, psychological, and physical stressors, which could affect their mental health in a negative direction.

Caregivers can derive support from marital partners and other living children when available, which reduces the risk of depression (Tomita and Burns, 2013). Available family members can help with chores and financial support (Casale et al., 2014; Casale and Wild, 2015; Hayslip et al., 2015) as well as provide emotional support. Also, resident grandchildren can be a source of support with household chores for their grandparents (Madhavan, 2004; Rutakumwa et al., 2015). Therefore, a person who is married and/or has a partner in the household and/or has living, healthy adult children in the household has the potential for receiving support.

Women are the primary caretakers in families in South Africa as well as globally (Mtshali, 2015). In a study of older caregivers in Durban, Port Elizabeth, and Cape Town, it was found that women are more likely to be caregivers and more likely to be depressed than men (Hamad et al., 2008). However, there is evidence that South African men are involved in caring to a greater degree than has been acknowledged (Montgomery et al., 2006). Race and ethnicity are also strong predictors of providing care, with Black women being more likely to provide care for family members with AIDS (Chen et al., 2015).

Research aims

The goal of this research was to examine the combined caregiving responsibilities (adult child and grandchildren) on depression in the grandparent who is providing care, and the variables that exacerbate or alleviate those stressors. Using the concept of proliferation of stressors, we estimate the effect on depression when a grandparent caregiver is simultaneously or sequentially caring for an adult child with AIDS or other long-term serious disease and his or her orphaned grandchildren. In addition, we will examine the role that contextual variables play in mediating the impact of proliferated stress on depression. These contextual variables include financial resources, health and disability, family structure, and demographic variables of gender and race/ethnicity.

Therefore, our analyses are informed by the following two research questions:

Sample

We analyzed data from the Cape Area Panel Study, Wave 4 interview of older adults (Lam et al., 2008). This is a four-wave longitudinal study conducted in the Cape area of South Africa during 2002–2006. We used Wave 4, which contains a cross-sectional sample of older adults, aged 50 and over, who lived in the originally sampled households but were not surveyed in the earlier data collections. The longitudinal data were limited to youths aged 14–22 and their households. Although there are household data that could be used to measure change over time, the key variables for older adults, such as depression, are only available in Wave 4.

The Wave 4 older adult study has a sample size of 3557 (from roughly 5250 in Wave 1 with 1609 male respondents and 1948 female respondents), in which only one older adult was interviewed in each household. The data included information from the four largest ethnic groups in South Africa. These groups include White (many are descendants of British and African settlers), Black (the two largest tribal groups are Xhosa and Zulu), Colored (mixed-race South Africans), and Asian Indians. We restrict our analyses to the Black and Colored older adults from the sample. We did not include Whites and Indians or the racial category ‘other’ in the analyses because there were no multiple caregivers (caring for adult child and grandchildren) in these racial groups.

Measures

Descriptive analysis

Table 1 shows the results of the mean comparisons between the two samples of those older adults who reported being caregivers of one or more orphaned grandchildren in their household, compared to those who were not caregivers of orphaned grandchildren. The results show that caregivers for orphans are more likely to be female, Black rather than Colored, and older. Caregivers are 80 percent Black and 74 percent female with a mean age of 65 compared to non-caregivers whose mean age is 60. Although the percent of Coloreds reporting caring for orphans in the household is much smaller than that of Blacks, this suggests that further research on the experiences of Colored caregivers is important. The same holds true for men. Although 26 percent of those caring for orphans in the household are men, the experiences of these men are important to understand the nature of care being received.

OPEN IN VIEWER

Table 1.

Comparison of descriptive statistics for the characteristics of older adults who have orphaned grandchildren in the household with those who do not have orphans in the household.

Variable	Orphans in HH				Orphans not in HH
	Mean	Standard deviation	Min.	Max.	Mean	Standard deviation	Min.	Max.
Demographics
Age	65.16***	9.53	47	97	60.22	8.42	47	97
Female	0.74**	0.438	0	1	0.57	0.495	0	1
Black	0.80***	0.404	0	1	0.36	0.479	0	1
Economic security
Food insecurity	0.351***	0.480	0	1	0.146	0.353	0	1
Gives money outside the HH	0.447*	0.49983	0	1	0.342	0.475	0	1
Has a flushing toilet	0.48***	0.502	0	1	0.74	0.438	0	1
Number of rooms in the home	4.52	1.683	1	14	4.61	1.739	1	14
Income	4001.84**	3101.61	1	43,087	5659.63	4885.25	1	62,000
Government pension	0.52***	0.502	0	1	0.29	0.456	0	1
Education	6.53	3.175	0	14	7.13	3.182	0	14
Physical and emotional well-being
Health	2.29*	0.713	1	3	2.15	0.653	1	3
IADL	1.29***	0.557	1	6	1.14	0.450	1	6
Depression	10.74***	4.818	6	30	8.43	4.646	6	30
Social support
Married	0.51	0.503	0	1	0.57	0.495	0	1
Number of living children	5.57***	2.584	1	15	4.13	2.13	1	15
Received income from outside HH	0.83	0.378	0	1	0.84	0.364	0	1
Stressors
Orphans of adult children in the household	1.69	1.068	1	6	–
Cared for child dying of long-term illness	.574***	0.497	0	1	0.045	0.208	0	1
N	94				2051

HH: household; IADL: instrumental activities of daily living.

*

p < 0.05; **p < 0.001; ***p < 0.0001.

These older adult caregivers for orphans have fewer financial resources. They have lower income, with caregivers for orphans having an average of 4001 rand per year in total household income compared to 5659 rand per year for non-caregivers. They are more likely to qualify for the means-tested government pension (52% vs 25%). Also, caregivers for orphans are more likely to report food insecurity (35% compared to 14% of non-caregivers) and to live in homes without flushing toilets (76% compared to 50%). There is no significant statistical difference between the means in the number of rooms in the homes of both groups.

Caregivers for orphans have poorer physical and emotional well-being. Respondents with orphans in the home are more depressed (8.43 vs 10.74) and report greater disability (1.14 vs 1.29), and declines in health (2.15 vs 2.29).

In terms of potential and actual family support, there is no statistically significant difference in being married or cohabiting (0.57 and 0.51) but caregivers have more children (5.57 and 4.13). Caregivers for orphans are less likely to receive money from others outside the households (0.07% compared to 0.12% of caregivers not caring for orphans; Table 1). In terms of caregiving, 57 percent of those with orphans also cared for a dying adult child, while only 4.5 percent of those without orphans did so. These descriptive findings support the need for understanding the totality of the caring experience on depression, rather than examining each type of care separately.

Regression analysis

To more fully account for the effects of multiple caregiving, we estimate a set of regression equations, predicting the level of depression on older caregiving adults while accounting for the effects of contextual variables. In Model I, we estimate the separate caregiving variables and the combined effects of being an orphan caregiver and a caregiver for an ailing adult child who passed away (stress proliferation). In Model II, we add the financial variables; in Model III, we add the health and disability variables; and in Model IV we add the family structure variables. We entered each variable in a stepwise fashion, so we could capture the effects of each set of variables in each model on depression in older adults (see Table 2).

OPEN IN VIEWER

Table 2.

Linear regression estimates of the effect on depression of elders with controls.

Variables	I – Caregiving w/controls	II – Economic	III – Health	IV – Family structure
	B (SE)	B (SE)	B (SE)	B (SE)
Orphan in household	−0.949 (0.683)	−1.117 (0.666)	−0.685 (0.597)	−1.209 (0.667)
Cared for adult deceased child w/long illness	−0.202 (0.479)	−0.272 (0.475)	−0.343 (0.469)	−0.307 (0.471)
Cared for child* raising orphaned grandchild	2.952 (1.022)**	2.776 (1.010)*	2.706 (0.469)**	2.768 (0.997)**
Gives money	0.467 (0.227)*	0.643 (0.212)**	0.724 (0.226)**	0.725 (0.226)**
Female	0.939 (0.216)***	0.743 (0.241)**	0.643 (0.212)**	0.613 (0.230)**
Black	3.190 (0.235)**	2.286 (0.263)***	2.184 (0.260)***	2.209 (0.261)***
Pension		0.703 (0.233)**	0.396 (0.236)	0.402 (0.238)
Log income		−0.472 (0.234)***	−0.451 (0.121)***	−0.435 (0.122)***
Hunger		1.760 (0.122)***	1.628 (0.312)***	1.570 (0.312)***
Education		−0.007 (0.034)	0.009 (0.034)	0.003 (0.034)
IADL			1.377 (0.239)***	1.39 (0.239)***
Health			0.613 (0.160)***	0.615 (0.160)***
Marital status				0.109 (0.238)
Children				−0.024 (0.049)
Received $				−1.106 (0.355)**
R 2	0.129	0.160	0.183	0.187
R2 change	0.007	0.031	0.023	0.004
F change	3.447***	16.941***	26.019***	3.289*
N	1840	1836	1834	1831

IADL: instrumental activities of daily living; SE: standard error.

*

p < 0.05; **p < 0.001; ***p < 0.0001.

In Model I, we estimate separately the two sources of stress (having orphans in the household and caring for an adult child), with the combined effect, which is the only variable that is statistically significant. In the final model (IV), when estimating the combined effect of caring for an adult child and their children together the level of depression is increased by 2.8 points when the provided multiple caregiving for a dying adult child and that adult child’s orphans is compared with any other combination of these variables.

Of the controls for gender and race, Model IV also shows that being Black increases the level of depression by 2.2 points over a Colored older adult, while being a female older adult increases the level of depression by 0.613 points over men (Table 2).

While Model II shows that the pension variable is statistically significant and positively predicts depression, still notice that, after controlling for the health-related variables on Model III, the government pension variable lost its significance, suggesting that health variables account for the impact of much of the financial effects. Model IV shows that being food insecure increases depression by 1.57 level points. Also notice that as we added more variables to each of the models, the strength of the coefficient for food insecurity slightly decreased. In addition, for each unit increase in the level of log income, the level of depression in the individual decreases by 0.402 level points, suggesting that the lower the income of the older adult, the higher his or her level of depression.

As shown in Model IV, those who say their health was worse in the year of the survey than in the past year experience a 0.615-point increase in depression. Also, those who experience more decrements in instrumental activities of daily living (IADL) are also more depressed by 1.39 points.

As shown in Model IV, only receiving money from someone outside the household is significantly related to the level of depression. Receiving money is related to a one-unit reduction in the level of depression. The R² increases in the stepwise analysis also tell an interesting story. There was a 0.031 increase in R² with the introduction of economic variables in Model II and a 0.023 increase in R² with the introduction of health variables in Model III. As shown in Model IV, there was an increase of only 0.004 with the introduction of family structure variables, suggesting that these factors may not be as important for increasing or reducing depression in older adults as much as the economic and health factors are. The overall R² in Model IV is 0.187, explaining less than 20 percent of the variance with all the models included in the analyses.

Limitations of the research

We found that using marital status and number of children as a proxy for social support does not provide for direct measurement of the aspects of support that provide a mediator for depression. Although the dataset we are using has multiple measures of instrumental support, there is less information about emotional support, which has been demonstrated to be important for coping and resisting depression.

Also, our dataset lacked other important context variables. For example, we do not know the status of the grandchild’s other parent. The other parent could be an important source of financial and emotional support, even if he or she is not directly caring for the grandchild. Also, as mentioned previously, neither do we know the exact cause of death for the adult child or the sequencing and timing of those two proliferators of stressors (adult child care and orphan care). We know that the grandparent cared for an orphan of a child who died, but all the other details are lacking in the dataset. Still, this research provides some evidence of the existence of a form of proliferation of stressors and how this can have an important impact on the mental health of older adults who find themselves in situations of caring for extended family.

Implications for social work practice

There has been a tendency to adopt Anglo-American social work practices in developing countries, which has proven to be inadequate because these practices do not account for the unique economic, social, cultural, and political conditions in countries such as South Africa (Gray and Mazibuko, 2002; Tsu and Yan, 2010). Social work practice should have the goal of designing holistic interventions that recognize the traditional practices, available resources, and values of a community as well as the structural conditions that impinge on individual lives (Taylor, 1999). Ideally, theory and practice should be integrated with local wisdom with the goal of increasing empowerment and autonomy of the vulnerable members of the population (Taylor, 1999). For example, a support program for elder caregivers that uses existing relationships of family, friends, and neighbors could help reduce the stress of elder caregivers in situations where they are overburdened with caring for seriously ill children and their minor grandchildren. An emphasis on organizing community networks not only would help by providing support when there are few economic resources, but also would help ensure that the support would include accommodating indigenous or traditional knowledge and cultural practices.

In order to design such a program, we suggest that, after doing an assessment of a community’s needs and cultural context, social workers call on community leaders to encourage the establishment of local support groups that could advance networking and reduce stress for elder caregivers (Gigabel, 2015). The support relationships with other elder caregivers as well as with other social and community workers would allow elder caregivers to share strategies for locating additional resources, providing respite care by sharing childcare, and resisting the stigma associated with HIV/AIDS. This support could be helpful in providing strategies for overcoming economic stress, reducing isolation, and improving self-esteem, which protect a caregiver from depression.

The results of this study provide a framework that could be applied in a variety of global settings where elders experience economic, social, and emotional stress as a result of caring for grandchildren. For example, grandparents are often responsible for grandchildren when parents are not available because of drug and alcohol addiction (Burton, 1992), incarceration (Hanlon et al., 2007), migration that results in leaving children behind in villages in order to seek employment elsewhere (Baker and Silverstein, 2012), or parental casualties in war-torn countries (Levy et al., 2017). These grandparents experience a proliferation of stressors, such as worrying about the well-being of drug-addicted children, which compounds the stress felt by grandparent caregivers (Burton, 1992).

In conclusion, future research should focus on collecting more qualitative and quantitative data on the grandparent caregiver to investigate how the caregiver role leads to depression when they experience a proliferation of stressors not just in the case of AIDS, but of other types of care as well. Social workers who design and implement programs need to be aware of the ways that stressors proliferate as a result of secondary obligations indirectly related to AIDS care that occurs in specific social and cultural contexts. These stressors influence mental health, which can have also a damaging effect on the quality of life of these grandparents. We suggest that social workers engage in monitoring and organizing elderly populations in order to build local support networks to provide early interventions to avoid any of the negative consequences for these vulnerable older caregivers.