“Taking care of my own blood”: Older women's relationships to their households in rural South Africa

Abstract

Aim:

This paper examines financial, emotional, and physical responsibilities elderly women are being asked to take on due to the incapacity of their adult children to care for the next generation; such incapacity is likely to increase as the HIV/AIDS epidemic worsens.

Methods:

This paper combines quantitative and qualitative data. Census data from the Agincourt health and demographic surveillance system (AHDSS) describe the presence of the elderly (specifically women over the age of 60 and men over the age of 65) in households in the Agincourt study site. Semi-structured interviews with 30 female residents aged 60–75 complement the census data by exploring the roles that older women, in particular, are playing in their households.

Results:

An elderly man and/or woman lives in 27.6% of households; 86% of elders live with non-elders. Households with a woman over the age of 60 resident (as opposed to those without) are twice as likely to have a fostered child living in the household and three times as likely to have an orphaned child in the household. Elderly women face financial, physical, and emotional burdens related to the morbidity and mortality of their adult children, and to caring for grandchildren left behind due to adult children's mortality, migration, (re)marriage, and unemployment.

Conclusions:

Older women provide crucial financial, physical, and emotional support for ill adult children and fostered and orphaned grandchildren in their households. As more prime-aged adults suffer from HIV/AIDS-related morbidity and mortality, these obligations are likely to increase.

Introduction

Elders in developing countries face concerns about healthcare, social security, and old-age care. In addition, many elders (defined here as women over age 60 and men over age 65) in developing countries also face the unique consequences of high levels of prime-aged adult morbidity and mortality that affect both the support they receive from, and the contributions they make to, their families and households. In South Africa, adult children's migration, unemployment, and a high rate of non-marital births (often leading to children left with maternal grandparents when the mother marries another man) also contribute to the burdens faced by the elderly. Elders are being left without support and caregiving from their children, and they are taking on responsibilities for sick adult children, as well as fostered and orphaned young children.

In addition to age, gender plays an important role in shaping the burden of care in the African context. Women are more likely than men to take on caregiving activities relating to both the sick and those left behind. Elderly women contribute to their household through financial, emotional, and physical means. These contributions may be a continuation of duties that were always seen as an older woman's responsibility, or they may represent new commitments that she had shed when her own children were grown and now must take on again.

Aims

This paper explores the relationships that currently exist between elderly women and their households in the Agincourt subdistrict in South Africa's rural northeast. The purpose of this paper is to paint both a quantitative and a qualitative picture of older women's caregiving responsibilities in their households. Elderly women are playing important caregiving roles for the sick, as well as for fostered and orphaned children living in their households. Although this paper explores caregiving generally, increasing HIV/AIDS prevalence is likely to further intensify older women's contributions to their households.

Background

Elderly men and women in South Africa are giving care to fostered and orphaned grandchildren and ill adult children. These responsibilities are in part a continuation of the apartheid era “stretched” and “dispersed” families [1–3]. Migration, morbidity, mortality, and reorganization of household members between households, all contribute to household arrangements that leave the elderly “in charge”. The HIV/AIDS epidemic is likely to further exacerbate the quantity and quality of household contributions expected of the elderly [4]. Researchers are just beginning to explore the impact of HIV/AIDS on the caregiving roles the elderly take on, and the relationships they have with family members who are infected with and affected by HIV/AIDS [5–7].

As caregivers of the ill and left behind, the elderly experience emotional and financial stress [8,9], deteriorations in their health [10], increased loneliness and isolation due to changing family structure, and stigma from the community [11–13]. Many of these elders expected financial, emotional, or physical support from their adult children; as these children die from HIV/AIDS and other causes, the elderly lose their caregivers, while often simultaneously becoming caregivers for their orphaned grandchildren [14].

Prime-age adult morbidity and mortality have economic consequences by reducing income [8] and increasing expenditures for households in which the elderly live. When households lose an income-earner, the elderly are often left with the debts incurred from illnesses. The costs of morbidity are further compounded when a death occurs, resulting in substantial funeral-related expenditures [9].

Most poor elderly South Africans receive a small means-tested non-contributory government pension, which was expanded to the black population in 1994 [15]. Women become eligible at age 60, men at age 65. Due to poverty and extremely high rates of unemployment, the elderly household member's pension may be the most stable and reliable income for many households [16,17]. This makes turning to the elderly for economic support a particularly viable strategy in South Africa. Pension income helps to diminish the impact of economic shocks, such as the cost of illness, funerals, and caring for fostered and orphaned children, and may contribute to families' reliance on the elderly [14,18].

Elderly men and women are likely to be impacted by the death of adult children. Caregiving responsibilities for the sick, fostered, and orphaned, however, usually leave mothers, wives, and sisters bearing the larger burden [7,19,20]. This research begins to explore the ways in which older women are contributing to their households – households that have suffered the loss of an adult member to HIV/AIDS, as well as those that have not. Their contributions are particularly vital in households with fostered and orphaned children, households with few working prime-aged adults, and households that have suffered economic shocks from the illness and death of members or outside kin. This paper focuses primarily on the physical contributions that older women make to their households.

Material and methods

Both the quantitative and the qualitative data come from the Agincourt health and demographic surveillance site. The Agincourt team conducts a yearly census to track births, deaths, and migrations. In addition, it collects verbal autopsy data, to identify a probable cause for each death occurring in the site. As of the annual census in 2003, the site was home to 70,272 people, from 11,665 households in 21 villages. All descriptive quantitative analyses use 2003 Agincourt census data. In the analysis below, age (60 years for women, 65 years for men) is used as a proxy for pension eligibility since the 2003 census does not have data on pension receipt.

The qualitative data come from in-depth interviews conducted June–August 2004, with 30 South African women aged 60–75 years selected from the 2003 Agincourt census. Of the 30 qualitative respondents, 29 were receiving pensions [18]. HIV/AIDS has been increasing in the site; the percentage of deaths attributable to HIV/AIDS increased from 1% in 1992 to 22% in 2003 [21]. Verbal autopsy data from 2001 to 2003 were used to define three strata of households by mortality experience: (1) an HIV/AIDS adult death had occurred, (2) a non-HIV/AIDS-related adult death had occurred, and (3) no adult death had occurred. Ten elderly women were randomly selected from each of the three strata. The intent of selecting deaths occurring between 2001 and 2003 was to facilitate recall regarding caregiving, but not to enter homes where a death had occurred in the previous year out of respect for the local mourning process.

Twenty-four of the originally sampled 30 women were interviewed. Reasons for non-response included: two women had had a recent death in their households; one refused; one had moved away; and two had died since the 2003 census. After selecting alternates, the team was able to interview 30 women (10 in each of the three strata) [18]. The interviews were conducted by three local women over the age of 40, trained in qualitative interviewing by the author. Each interviewer was responsible for conducting three interviews with each of her 10 designated respondents; in addition, each interviewer translated and transcribed her own interviews. While in the field, the author read each interview, reviewed queries with the interviewers, and wrote individual interview guides for later interviews to fill gaps, follow-up on interesting issues, and explore new questions.

This paper looks at how older women in all 30 households spoke about their relationships to other members of their families and households, and the roles that they play within their households. Respondents' narratives in the text below are followed by the respondent's pseudonym and strata into which her household fell. Although distinguished by strata in the text, there were fewer distinctions than expected. Regardless of strata, most of our respondents had current or past connections with a family member who was ill or had died of HIV/AIDS, and/or were taking care of fostered or orphaned children.

Results

Quantitative picture

In 2003, women over age 60 were just 4% of the total Agincourt population; however, almost a quarter of households (22.2%) included an older woman (see Table I). Men over the age of 65, who were only 1.6% of the Agincourt population, lived in 9.5% of households. Just over 5% of households had more than one pension-age-eligible member – the majority being husband–wife pairs (about 90% of multiple pension-eligible households), the remaining mainly included two or three female pensioners, most probably sisters. Data from the 2003 census update does not provide sufficient information to explain why the elderly live in certain households or how their presence relates to household-level morbidity and mortality. Future work will explore such relationships and the timing of other household members' movements.

Table I.

Age-eligible pensioners, Agincourt 2003.

	Individualsa (n=70,272)	Householdsb (n=11,665)
Women over 60	3.9% (2,752)	22.2% (2,671)
Men over 65	1.6% (1,121)	9.5% (1,112)
2+ pension-age-eligible adults	N/A	5.4% (630)
1 pension-age-eligible woman and 1 pension-age-eligible man	N/A	4.8% (565)

^aPercentage and number (in parentheses) of individuals in specified category. For example, 3.9% of the total population are women over the age of 60.

^bPercentage and number (in parentheses) of households with members in specified category. For example, 22.2% of all the households in the site are home to a woman over the age of 60.

Despite the small numbers of pension-age-eligible adults in the site, 30.3% of non-pension-age-eligible individuals in the site live in households with the elderly (results not shown in table). In the 2003 census, there were 25,727 children under the age of 15 (36.6% of the total population). Approximately a quarter of these children (26.4%) lived in a household with an elderly person; of these 85.2% lived with an elderly woman.

Table II presents mean household size for all households and those with and without a woman over the age of 60. The mean household size is 6.02. Households with no woman over age 60 have significantly fewer members (5.7 on average) compared with those with a woman over 60 (7.0 on average). Although small in number, the difference between the mean number of children in households with and without women over the age of 60 is significant at the 0.01 level.

Table II.

Mean household size (and size range), Agincourt 2003.

	Household size	Children in household	Number of households
Mean all households	6.02 (1–40)	2.21 (0–18)	11, 665
Mean with no woman over 60	5.73 (1–30)	2.17 (0–17)	8,994
Mean with woman over 60	7.00 (1–40)a	2.31 (0–18)b	2,671
Mean with 2 or more age-eligible pensioners	6.91 (1–40)a	2.25 (0–18)	630

^aDifference in mean size of households with and without a woman over 60 is statistically significant at the 0.001 level. The difference in mean household size with no elderly members and with two or more elderly members is also statistically significant at the 0.001 level.

^bDifference between mean number of children in household with and without a woman over 60 is significantly different at the 0.01 level.

As Table III shows, the percentage of households with one or more children differs little for households with a woman over age 60 and those without. Households with a woman over the age of 60, however, are much more likely than other households to host fostered children and maternal orphans. Agincourt defines a fostered child as a child whose mother is alive, but living elsewhere, and maternal orphans as children whose mother is dead. Prior to 2006, Agincourt does not capture paternal residence or status (alive/dead); without data on fathers' presence/absence, these data do not fully capture the number of orphaned and fostered children.

Table III.

Households with children, fostered children and orphans, Agincourt 2003.

	All households	Households with no woman 60+	Households with 60+ woman
Households with at least one child under 15	77.7%	77.8%	76.2%
Household with at least one fostered childa	15.4%	12.1%	26.5%c
Household with at least one maternal orphanb	5.5%	4.7%	15.9%c
Total n	11,665	8,994	2,671

^aThe mean number of fostered children in all households is 0.2; in a household with at least one foster child the mean number of fostered children is 1.4, in households with at least one foster child and a woman over 60, the mean is 1.5.

^bThe mean number of maternal orphans in all households is 0.1; in a household with at least one maternal orphan, the mean is 1.6, in households with at least one maternal orphan and a woman over 60, the mean is 1.7.

^cDifference between the percentage of fostered children/orphans in households with and without a woman aged 60+ is significant at the 0.001 level.

Of all Agincourt households, 15.4% are home to at least one fostered child. Over a quarter of the households with an older woman, however, have one or more fostered children, compared with 12.1% of households without an elderly woman. This pattern is more dramatic for maternal orphans; nearly three times as many households with an older woman (15.9%) as those without (5.5%) have at least one maternal orphan.

Although one-dimensional, i.e. without controls for wealth, household mortality profiles, or other potentially important correlates, these simple cross-tabulations begin to show a surprising presence of older women in Agincourt households. This is particularly notable in households with potentially vulnerable members, such as fostered and orphaned children. Future work will explore these statistics by household mortality experience.

Qualitative picture

The quantitative data create a picture of the presence of vulnerable children living in households with older women. The census data cannot, however, describe the role that older women are playing in their grandchildren's lives. It is possible that older women are living in multi-generational households in order to receive support, for instance assistance with household tasks and financial maintenance. As the qualitative data show, for some women, this is the case. For most of our respondents, however, surrogate-mothering of grandchildren and caring for ill adult children extends into their pension years. It is not possible to say from these data whether this is a new phenomenon, but it is accurate to describe this as an overwhelming reality for our respondents.

Children supporting parents

There is perhaps a certain level of expectation among women over the age of 60 that the children they raised will support them in their old age. For respondents whose children were gainfully employed, this was often the case. Grown children living with our respondents provided some of them with financial assistance. Other respondents received remittances from children who migrated outside the area. Support came not just in the form of cash but also through in-kind assistance with food and other subsistence needs, like home improvements, and in the form of physical support by taking over chores. Household chores like cooking, cleaning, and collecting firewood and water may become difficult for older women, particularly the latter two which include walking long distances and carrying large loads. Living with younger household members who conducted these chores made a significant difference in the lives of some of our respondents.

Emily, a divorced 65-year-old, lives with four unmarried daughters and three grandchildren. One of these grandchildren is an 18-year-old fostered grandson. A married daughter who lives elsewhere sent her son to live with and assist Emily: “I asked his parents if [my grandson] could stay with me so that I can send him to fetch water, [which is difficult for me to do] because I am old. My other grandchildren are still young [so they cannot help yet]” (Emily, No Death Household). Similarly, Maria's son who lives in Johannesburg sent one of his sons to help: “They said that he must stay with me and help me with some difficult jobs, for example going to fetch water, because we walk for kilometers to fetch water. He also protects me and the house” (Maria, Other Death Household). Despite talking about financial, in-kind, and physical assistance they received from kin, many more respondents spoke about the financial and physical responsibilities they continue to shoulder despite their ageing and sometimes ailing health.

Caring for the sick

The theme of caregiving is something about which we asked each respondent explicitly. We asked if the respondent had been or was currently the primary caregiver for anyone who was sick in the household. Although we did not designate a time period, most of the caregiving, particularly for grown children, had occurred in the last five years. As Table IV shows, over two-thirds of our respondents mentioned caring for an ill adult child (presently or in the past), one-third mentioned having taken care of an ill husband, and two-thirds mentioned helping to care for other kin like grandchildren, daughters-inlaw, siblings, and parents. More women living in households where there had been an HIV/AIDS death than women from the other two strata reported having cared for an ill adult child. More of the women in the other two groups reported taking care of someone other than an adult child. It is likely that the women in HIV/AIDS households also took care of husbands and other kin, but we concentrated on learning about the adult child's illness resulting in less time to ask about other cases.

Table IV.

Gogo^* project sample, caregiving, Agincourt 2003.

	Took/taking care of ill adult child	Took/taking care of ill husband	Took/taking care of other ill kin
HIV/AIDS Households (n=10)	9	2	5
Other Death Households (n=10)	6	5	7
No Death Households (n=10)	7	3	8
Total (n=30)	22	10	20

^*Gogo: ‘grandmother’ in the vernacular.

Older women described their caregiving as including a diverse array of activities and financial responsibilities. These included feeding, bathing, fetching and preparing treatments, washing soiled clothing and blankets, and helping the ill person to the pit latrine (none of our respondents had flush toilets or piped water). Thandizile cared for her ill son, assisting his wife: “I used to wake up in the morning and boil water for him to bathe, cook soft porridge for him” (Thandizile, HIV/AIDS Household). Sinah took care of her daughter, who she admitted probably had AIDS: “[My daughter], when she came here, she was seriously ill. I had to wash her, I had to carry her to the toilet [pit latrine behind the house]; she couldn't walk a long distance…. When she finished, I had to carry her back to the house. Then, I had to go back, take a spade and throw the faeces in the toilet” (Sinah, HIV/AIDS Household).

Older women's caregiving responsibilities also included traveling with the sick person to the traditional healer, clinic, private doctor, or hospital to receive care and treatment. Sometimes they helped the patient walk to these places, and sometimes they paid for the transport – for an older woman the former has physical costs, whereas the latter has financial costs. Constance alludes to both issues: “It was difficult in case of money because of moving [my daughter] up and down to hospitals, hiring cars to send her to hospitals. I even borrowed from my neighbors to send her to hospitals…. My heart is painful, I suffered a lot about my daughter moving up and down on clinics, hospitals, and doctors. But [still], my daughter passed away after suffering for a long time” (Constance, No Death Household). Despite the fact that Constance's daughter's death did not occur in her household, she clearly played an important role in caring for her while she was sick.

A large number of the respondents' adult children were not living in the household when they became ill, but were brought “home” for care. This seemed to be particularly true for children suffering from HIV/AIDS. Some of our respondents spoke of caregiving as a mother's responsibility, others denigrated daughters-in-law who did not properly care for their ill sons, forcing the respondent to take over these responsibilities. Auphrey took care of an HIV-positive daughter and also is taking care of her son who lives near her. She said: “Any child who becomes sick in our culture, while their parents are still alive, the mother must take care of her child” (Auphrey, HIV/AIDS Household). Grace had a different opinion. Her son, who recently returned from Johannesburg without his wife, was thin, coughing, and tested HIV-positive. She complained that her daughter-in-law is relying on her to care for him: “A city woman runs away when her husband is sick. She only comes back when her husband is okay” (Grace, No Death Household).

Caring for grandchildren

On average, the women with whom we spoke lived with four to five grandchildren. The maximum number of children under age 15 in a household was 13, the most common was just 2. The configurations of households varied greatly. Most households had three generations, but very few were simply a grandmother living with an intact family unit (i.e. parents and their children). Some respondents lived with grandchildren whose parents were not resident in the house due to mortality, migration, or non-marital births (when daughters marry men other than previous children's fathers, they often leave these children behind in their own mother's care), as well as with other of their grown children.

Regardless of kin configurations, many of our respondents took on physical and financial caregiving responsibilities for their grandchildren [18]. These grandmothers purchased and washed their grandchildren's clothes, bought food and cooked for them, paid their school fees and took them to school, fetched water and firewood for them, and bathed young ones. Essentially, these older women acted as surrogate parents to their children's children. Our respondents did not always clearly differentiate fostered or orphaned children in their care from those who had parents living in the household. Even when children's parents were absent and our respondents were the primary caregivers, they sometimes claimed there were no orphans or fostered children in a household. These contradictions further emphasize how fully these respondents viewed their parenting roles.

Anna lives in a household with her grown son and daughter, one school-aged daughter, and three grandchildren. She is the foster parent to two of these grandchildren; Anna takes full responsibility for these children. She explains: “[My daughter] gave birth to the children while she was staying here at home. She was not married. The children grew up in my house. I take them as my children” (Anna, Other Death Household). Now Anna's daughter lives with a different man and is unemployed. Although Anna is proud of her role in her grand-children's life, she sometimes finds caregiving challenging:

It's difficult, but I am bound to look after them because there is no one else to look after them. Since they were born, [my daughter] hasn't looked after them. She used to move up and down with the boys. I was the one who took care of them. It's difficult to look after kids. Children need good care…. These are my grandchildren. If they are hungry and ask for food, I buy it for them. If I don't give to them [what they need], my heart is painful…. (Anna, Other Death Household)

Many of our respondents were in similar situations, raising fostered grandchildren belonging to their migrant, (re)married, or unemployed children.

Other respondents, like Mumsy, take care of orphans. Mumsy cares for orphans who live in her household as well as orphans who live elsewhere. One of her daughters died in 1999 leaving three children. A son died more recently leaving four orphans for whom she helps care; in addition, she assists with three of her sister-in-law's orphaned grandchildren:

[My daughter] was staying [near Johannesburg] selling fruits, and vegetables; that's where she got these children. We don't know who the father of these children is. The children were staying [near Johannesburg] with their mother. Now their mother is dead. No one was going to look after them; that is the reason they came and stayed with me…. I am [also] taking care of my grandchildren; the four who are not staying here…. Now my son and his wife both died…. If they have problems, they come to me [and] I help them. I am also staying with my sister-inlaw. She has three grandchildren. Their father died, no one is taking care of them. I use my pension money to help them. (Mumsy, Other Death Household)

Mumsy does not live in an “HIV/AIDS household”, but it is not unlikely that her children's deaths were due to AIDS. By taking care of these orphans, she is affected by the HIV/AIDS epidemic.

With so many people relying on her, Mumsy feels stretched and worries about making ends meet:

When my grandchildren come to me and ask for money, if I don't have any, I worry. Like now, one of my daughter's children is sick. No one is helping me to take care of her. One [grandchild] wasn't going to school because she didn't have shoes. [My grandchildren whose] father died don't get any money from the company where their father was working. If I think about all of this, I worry a lot. (Mumsy, Other Death Household)

Mumsy takes care of more grandchildren than most of our respondents, but the feelings of being “bound” to care for the children and stretched thin financially resonates with most of our respondents' narratives.

The one way in which our respondents felt that they would be “rewarded” for their caregiving was through their grandchildren returning the assistance when the older woman needed it in the future and paying for their grandmother's burial. Sister did not see any disadvantage to taking care of orphaned or fostered children because: “The advantage is that they will grow up, work, and support me when I will be too old, and also take care of me” (Sister, HIV/AIDS Household). Thembi, who takes care of two grandchildren, said: “To take care of my grandchild is good because I know when I die my grandchild is going to bury me” (Thembi, No Death Household). Regardless of whether these rewards ever come to fruition, in the present our respondents saw taking care of their grandchildren as their responsibility. Several respondents, like Dorah who cares for two grandchildren, put this responsibility in the following terms: “I don't have any problem because I'm supporting my own blood” (Dorah, No Death Household).

Discussion

Despite being less than 5% of the Agincourt population, the quantitative data show that older women are dispersed over nearly a quarter of the households in the site. And they are playing important roles in these households. The qualitative interviews point to the fact that older women are not just being supported by household members in the households in which they live, although sometimes working adult children send financial assistance or grandchildren as “helpers” for their elderly mothers. Instead, these elderly women, who as a collective would usually be considered a “vulnerable population”, are themselves taking care of other vulnerable household members – ill adult children, and fostered and orphaned grandchildren. This paper begins to outline the roles that they play and their importance in sustaining the multi-generational households in which they live.

Although older women were taking care of kin in nearly all the households in our sample, it was clear that the responsibilities associated with caring for those sick with HIV/AIDS and the children left behind are great. These responsibilities spread beyond households in which an HIV/AIDS death occurred in the recent past, as designated by the strata in the qualitative study. Many of our respondents would hesitate to call their responsibilities “burdens” because they are simply “taking care of their own blood”; however, this paper suggests that older women might need further physical, emotional, and financial support as the HIV/AIDS epidemic escalates and they continue to be “bound” to take on increasing responsibilities.