The South African disability grant: Influence on HIV treatment outcomes and household well-being in KwaZulu-Natal

Abstract

This paper explores the implications of the disability grant for household members' well-being and adults' success on ART (antiretroviral therapy). It uses case studies based on data from an in-depth qualitative study of 10 households in KwaZulu-Natal. Receipt of the disability grant ensured that the basic needs of the HIV-infected adult could be met by other household members, especially when the grant was received when the person first met the qualifying criteria and in conjunction with ART. Where treatment was effective, HIV-infected adults were able to make substantial contributions to the well-being of other members in addition to the financial support provided by the grant itself. Thus, early access to financial support in conjunction with commencing ART may lead to improved health outcomes and reduce poverty and vulnerability associated with illness in poor households. This synergistic relationship between social welfare and treatment may in turn contribute to greater cost-efficiency.

Keywords:

• antiretroviral therapy
• disability grant
• HIV/AIDS
• KwaZulu-Natal
• social security systems

1. Introduction

Many South African households are faced with the dual challenges of HIV and poverty. The social security system in South Africa was designed to reduce deprivation and vulnerability in poor households by providing a range of means-tested cash grants (Lund, 2008). The child support grant and old age pension are the most common of these. The old age pension, in particular, is redistributed within households (Case & Deaton, 1998; Seekings, 2002; Duflo, 2003). However, despite their potential for redistribution and the benefits that grants have for poor households, including those affected by HIV, the design of the current system fails to account adequately for HIV.

The disability grant was originally designed to provide for those who were, as a result of severe disability, rendered unable to work and to supplement their household income (Seekings, 2002). Neither the Social Assistance Act of 1992 nor the 2001 amendment to it make any mention of HIV (Nattrass, 2005). Despite the high prevalence of HIV in South Africa, the ill-health and loss of work that results from this and the repercussions for households' socioeconomic status, the government currently does not provide a targeted grant for poor HIV-infected individuals or affected households. Those assessed as impaired or disabled because of chronic illness and therefore unable to work should, according to the guidelines, qualify for this cash transfer either on a ‘permanent’ (i.e. for five years) or ‘temporary’ (less than one year) basis before their eligibility is reviewed (Nattrass, 2005). The problem is that until 2008 when the policy was explicitly changed many provinces used a CD4 count of below 200 cells/mm³, regardless of impairment or disability, as the main criterion for eligibility (Simchowitz, 2004). Policymakers have therefore not adequately considered the differences between illness and disability and designed the grants system accordingly; this has led to criticisms (Simchowitz, 2004; Nattrass, 2005; Hardy & Richter, 2006).

The disability grant was worth R1200 in April 2012.4 Research in the Free State shows that, like the old age pension, which is the equivalent amount, it can contribute substantially to income and reduce the depth and severity of poverty in affected households (Booysen & Van der Berg, 2005). Despite not being designed to target those affected by HIV, the disability grant is likely to make a greater contribution to affected households, and particularly the sick person's socioeconomic status, than the smaller child support grant. While other cash transfers may play a role similar to that of the disability grant in affected households, because the disability grant is paid to an individual the focus in much of the literature is on whether receiving it may affect their adherence to antiretroviral therapy (ART). Specifically, the wish to meet the qualifying criteria for the disability grant may represent a disincentive to adhering to ART (Hardy & Richter, 2006; Leclerc-Madlala, 2006; Phaswana-Mafuya et al., 2009).

Instead of focusing on whether a negative relationship exists between the disability grant and ART, we explore the way in which access to both treatment and the disability grant can positively influence personal and therefore household well-being. Specifically, we investigate the impact of disability grant receipt on poor households affected by HIV in South Africa. We show that the timing of grant receipt plays a role in the contribution the grant makes to health outcomes for individuals on ART. The analysis compares case studies where the grant was received when the person qualified, a case where the receipt of the grant was delayed and another where, despite applying and qualifying for the grant, the applicant failed to receive it. The case studies also document the benefits that timely receipt of both the disability grant and treatment can have for the well-being of the rest of the household. The case study where receipt of the grant was delayed and the case where the grant was not awarded illustrate some of the barriers to more widespread uptake of the grant by eligible individuals.

We first review the principal debates about the disability grant and HIV, and then provide some empirical evidence for the benefits of the disability grant from a small qualitative study of HIV-affected households in KwaZulu-Natal. We conclude with recommendations that may help to maximise the personal and household-level benefits and outcomes of access to a disability grant and ART.

2. ART and the disability grant in South Africa: Debates and developments

The literature dealing with both the disability grant and HIV/AIDS has, to date, largely focused on a perceived trade-off between access to ART and the disability grant (Hardy & Richter, 2006; Leclerc-Madlala, 2006; Phaswana-Mafuya et al., 2009; Venkataramani et al., 2010). Prior to 2008, the qualifying criteria for ART initiation and access to a disability grant were both based on a minimum CD4 count of 200 cells/mm³ or a medical assessment showing that the person was at stage 4 of the illness as defined by the World Health Organisation (Lawn et al., 2008). These identical eligibility criteria fuelled claims that the disability grant was a disincentive to continuing ART, i.e. that poor, largely unemployed people in receipt of a disability grant might not adhere to the therapy, or might even stop completely, to ensure that they would keep receiving the grant and not lose it (Nattrass, 2005). Despite this concern, however, no subsequent empirical research has provided evidence of a substantial trade-off. Recent research raises concerns instead about the impact of withdrawal of the disability grant from those whose health has improved with ART and points to the inappropriateness of the use of the CD4 count or the viral load as criteria for access to the grant (Hardy & Richter, 2006; Leclerc-Madlala, 2006; Phaswana-Mafuya et al., 2009).

Perhaps influenced by an outcry from researchers, civil society and the public (Heywood, 2008), the eligibility criteria for the disability grant were changed in 2008, disallowing its review or removal on the basis of the CD4 count (DSD, 2008). Instead, the major qualifying criteria for receipt are physical or mental impairment or disability that renders the person unfit for work or employment (SASSA, 2010). In addition, recent health policy changes mean that those with a CD4 count of less than 350 cells/mm³ qualify for ART (DoH, 2011). This change is recent and the bulk of adult patients currently on ART, including those enrolled in our study, qualified for treatment at 200 cells/mm³. These changes in eligibility criteria, along with little evidence for any disincentive, suggest that new research should abandon the focus on disincentives.

Despite changes to the review and renewal guidelines, concern does remain, however, about the consequences of grant loss for households. Some quantitative research in the Western Cape and Free State has indicated that the time limits on receipt of the grant have been poorly monitored and that grant receipt continued after the review period elapsed (Booysen & Van der Berg, 2005; Venkataramani et al., 2010). These findings suggested that fewer grants may have been terminated than anticipated on the basis of earlier literature (Hardy & Richter, 2006; Leclerc-Madlala, 2006; Phaswana-Mafuya et al., 2009). One third of those receiving the grant in the Western Cape sample were also working, demonstrating a possible failure to implement the means test which requires those applying to be unable to work (Venkataramani et al., 2010). In addition, the results show no change in self-reported health in those who lost the disability grant. The threat of grant loss and the disincentive effect associated with the loss may therefore be less significant than has been commonly assumed.

While some research points to the valuable role that the disability grant plays in the health, nutrition and income of households affected by HIV, these studies fail to consider when the grant is accessed and whether a potentially beneficial relationship exists between this grant and ART (Phaswana-Mafuya et al., 2009; Venkataramani et al., 2010). This paper therefore examines the relationship between receipt of the disability grant and successful ART and considers the benefits of timeous grant receipt.

3. Methods

3.1 Background

The households enrolled in this study were from rural and peri-urban communities in the Umkhanyakude district of northern KwaZulu-Natal and were part of the Africa Centre Demographic Surveillance Survey (DSS) population. The DSS population is very mobile and 33% were non-resident (Tanser et al., 2008). Fluid and multiple household memberships and a generational mix contribute to household complexity. Households are often poor and unemployment is high. Social grant receipt is fairly widespread and the DSS data from 2006 suggest that 90% of those eligible receive an old age pension. In addition, 27% of rural households and 43% of urban households with eligible children were getting at least one child support grant (Muhwava, 2008). For the purpose of this study, a ‘household’ was defined as people who consider themselves to be members of the same household; this includes resident members and, in some cases, non-resident members who live elsewhere but maintain social ties with the household.

Antenatal HIV prevalence in the district in 2008 was 40%, the fourth highest rate in KwaZulu-Natal (DoH, 2009). ART has been available since 2004, through government clinics, supported by an HIV Treatment and Care Programme run by the Africa Centre, and an estimated 7500 people in the area were receiving treatment by the end of 2008 (Hontelez et al., 2011). The community is fairly well-serviced by clinics staffed weekly by Africa Centre doctors. Cause-specific mortality data collected using verbal autopsies5 between 2000 and 2009 showed that 50% of adult deaths in the DSS were HIV-related (Herbst et al., 2011). Access to ART has resulted in a substantial reduction in mortality – approximately a 22% reduction for women and 29% for men between 2002 and 2006 (Herbst et al., 2009). But despite increased access to treatment, improved health and reduced mortality, late presentation for HIV testing and delayed initiation of ART because of the CD4 count guidelines means that many households nevertheless have some experience of illness or death.

3.2 Data collection and analysis

The study was conducted in 10 households that included either an adult HIV death or a member living with symptomatic or treated HIV. All the households were poor, with limited access to formal employment, and relied heavily on social grant income. A local home-based care project and the verbal autopsy staff at the Africa Centre were identified as being able to share information about this type of household. These sources were useful because they could put us in contact with households where people were aware of the HIV status of their members and were comfortable with talking about the sensitive issues relating to HIV. The study enrolled five households identified by the local Catholic Church home-based care programme, one household identified by verbal autopsy staff, and one that was a chance contact. The study also included three households that had been part of an earlier Household Dynamics study investigating the socioeconomic impact of HIV/AIDS on households and their members in the same district prior to the advent of public ART (Montgomery et al., 2006; Hosegood et al., 2007).

For logistical reasons all households selected had to have an adult respondent who was present during the day to be interviewed. This meant that available respondents were likely to be too old or sick to be working, unemployed, studying, on holiday or doing shift work. Our respondents were therefore mostly women and mostly 55 years and above, potentially introducing gender and age bias. On the other hand, the 10 households reflected a variety of social and economic situations and were broadly representative of other households in Umkhanyakude. All respondents provided informed consent and the Humanities and Social Sciences Research Ethics committee at the University of KwaZulu-Natal and the Research Ethics Committee at the London School of Hygiene and Tropical Medicine granted the study ethical approval.

We collected in-depth ethnographic, retrospective and contemporary information about the experience of illness and death. The data were collected using multiple semi-structured interviews and non-participant observation in the household over a six-month period beginning in January 2008. Each interview dealt with a particular topic but prospective data about the current household situation were also collected at each visit. Household genograms and event maps were used to collect information about household composition, relationships and important episodes and events relating to illness and death (Adato et al., 2007).

Data from genograms, household event maps, transcripts of translated interviews and observation field notes were used to create household case studies covering the five years preceding fieldwork and the six-month period of research. All names used in the analysis and this paper were changed to pseudonyms. The case study approach enabled a detailed description of household experience in context and the analysis of socioeconomic changes at household and individual level (Mikkelson, 1995; Russell, 2005). The case studies were compared and interview transcripts and field notes were also analysed cross-sectionally using framework analysis (Mason, 2002). Although limited in their generalisability to different contexts, these in-depth qualitative case studies revealed the complexity of the respondents' experience and they offer important insights into social policy (Ritchie & Spencer, 1994). We coded the research questions according to themes and added to these as further themes were identified in multiple reading and coding iterations.

4. Findings

The focus of this paper is on adult illness, ART and the disability grant. We chose four case studies to summarise for this paper, as examples of four typical scenarios. Two had members receiving both ART and a disability grant during the study period. The Dlamini household had a member benefiting concurrently from the disability grant and ART. The Nkosi household had a member who had started treatment before the study began but received her disability grant later – this provided an example of the effect of a new grant on the well-being of a person on ART and their household. The other two had members receiving ART but no disability grant. The Dube household had a member on ART who had received a temporary disability grant but lost it before the study began – this was an example of the effect of grant loss and also of the importance of timeous receipt of both the grant and ART. The Bhengu household had a member who had made an incomplete application for a grant – this was an example what can happen to a household and its sick members when disability grant applications fail.

Although the households in these studies were in receipt of other grants, particularly child support grants, these grants were smaller and targeted at other members of the household. In addition, all these other grants had been accessed for some time before the study and their use was well-established prior to both the illness of the adult and the receipt of the disability grant.

4.1 Dlamini household

The Dlamini household was a very large four-generation household. Gugu, the eldest living daughter, aged 33, was HIV-positive but had been on treatment since late 2006 and was well. Although her mother was the household head, the responsibilities for financial decision-making and ensuring the household met its basic needs fell to Gugu. The household survived on the income from a number of social grants. Seven of the younger children in the household qualified for child support grants and this income was spent on their basic needs and their mothers'. The elderly household head, Gugu's mother, received an old-age pension, but Gugu explained that the bulk of this was spent on her mother's health care expenses for diabetes and high blood pressure. Despite a fairly large total monthly household income, the size of the household meant that it was quickly spent on the basic needs, particularly food for the children and the health care of the household head.

Gugu received a disability grant for herself one month after starting ART. She attributed the improvement in her health as much to the grant as to the treatment:

When I started taking treatment, I went to register for my disability grant. Everything went quickly and I got it in November. I have seen a big difference. I can see that my body is much better.

The improvements that Gugu observed in her own health and her psychological and financial welfare were not only crucial for herself and her child but had implications for the rest of the household. As the decision-maker, she controlled the purchase of food, ensured that the household members' basic needs were met and cared for her mother. The disability grant income meant that she could invest some capital in the informal family business, which had stopped operating while she was severely ill. Gugu therefore made substantial contributions to the household's income and well-being as a result of receiving the disability grant.

Limited data are available from this study about the possible disincentive that receiving a disability grant presents to adherence to ART. The respondents seldom spontaneously initiated reports of their concerns about the loss of a disability grant. It was only when asked that they mentioned such concerns. Gugu was one respondent who articulated her fear of losing the grant:

They told us that we could lose the [disability] grant. I don't know how I feel about that. That is why I want to register [my son for a care dependency grant], because when [my disability grant] gets cut off I will need help. … I went to look for a job before but I had a problem because [he] was not staying with me and he didn't take his treatment well. I am afraid to leave him again [to look for work] because maybe he won't take it again and will get sick.

When asked to explain her concerns in more detail, Gugu said she was particularly concerned about losing the grant because she was caring for her sick child and would have to leave home to seek employment if she lost this grant.

4.2 Nkosi household

Thobela and Bheki were a cohabiting couple living with their two children. Both were HIV-positive and Thobela, aged 36, was extremely ill, unable to walk and required ongoing physical care. She had been on ART for about four months at the start of fieldwork. Although Thobela and her family believed that her health had improved slightly because of the treatment, they continued to struggle to afford adequate food and transport to the clinic. They were also concerned about how cut off they were becoming from their neighbours and community. These social and economic problems experienced by the household appeared to contribute to Thobela's failure to make a better recovery.

The household was poor before Thobela started treatment: Bheki had lost his job and was neither employed nor looking for work because he needed to care for Thobela and Thobela had stopped informal trading because of her illness. This meant that they were surviving financially on a child support grant, worth R210 in 2008,6 for their youngest child and a very small piecemeal income that Bheki earned doing herding or weeding when Thobela could be left alone. Thobela and Bheki both spoke about being poor and their feeling of desperation about their financial situation. The household had no alternative source of income or support and were isolated from both family and the community. Their financial situation made it difficult to buy food and other necessities and get Thobela to the clinic:

It's not enough [money] … to go to the clinic I need money, it's all credit [she had to borrow money to pay for transport to the clinic]. If I got a [disability] grant it would be better.

Thobela applied for a disability grant soon after starting ART but did not receive it immediately. Despite visiting the Department of Social Development and the South African Social Security Agency to follow up on the progress of her application, she failed to find out why she was not receiving it, although some officials suggested her application might have been lost. Then, after four months of investigation, with no explanation, Thobela's disability grant was suddenly paid out. The reaction of the household was one of great joy – the grant of R1140 almost quadrupled their monthly income.

Thobela's receipt of the disability grant subsequently had clear benefits for both the household's well-being and her own health. Bheki confirmed that he was under less pressure to provide resources for his family and felt generally better, more rested and healthier. The first disability grant payment enable the family to purchase new school uniforms for the children, thereby ensuring they could remain in school. Thobela felt a dramatic improvement in her health:

It's not like it was before. We are not suffering now. You can see that now I can walk. A person can die, if they are not eating and taking pills [ART]. I will be fine now.

Thobela attributed the improvement in her health to being able to afford different food. Whereas before the household survived largely on staples such as maize meal, after the grant they were able to expand their diet to include some tinned fish for protein and more vegetables.

Thobela's return to health meant she was able to resume her domestic chores. Towards the end of the study period, she was walking, albeit slowly and with some pain. She was able to work in her garden and was growing spinach for her family. The homestead became cleaner and tidier as Thobela was able to do more and the other members of the household had time to devote to activities other than her care and support.

4.3 Dube household

The Dube household was headed by a 90-year-old woman who lived with her 34-year-old daughter-in-law, Thembilihle, her daughter Lungile and three grandchildren. Thembilihle was HIV-positive, had been on treatment since 2004 and was well in 2008. Like Gugu, she found that her health improved on the combined receipt of ART and the disability grant:

When I went on treatment I became better, I gained weight again and the marks I had on my body are not there anymore.

Like all the sick household members in our sample, she was severely ill and unable to work when she started ART and the disability grant supported her and her household financially while she recovered.

Thembilihle's disability grant was discontinued. She had recovered on treatment and at a review with her doctor in 2007 she was considered well enough to work and her doctor therefore refused to fill in another application for the grant. She had since requested another review but was informed that her health had improved, her CD4 count had normalised and that she was able to work and therefore would not be considered. Since losing the grant she had obtained a formal job working full-time at a grocery shop in town. She continued taking her medication and had negotiated working hours with her employer enabling her to attend the clinic or hospital without jeopardising her employment.

Thembilihle, like Gugu, was the primary decision-maker in her household. Her income from her job went into the communal pot, along with her mother-in-law's old age pension and extra income from selling beer, to support the rest of the household, none of whom contributed financially.

Thembilihle planted and tended the household garden and fruit trees that provided the household with seasonal produce. She was also a member of a stokvel7 along with other women in the community. This stokvel was not merely a saving scheme: the money pooled was lent to others and the interest generated added to the group's income.

4.4 Bhengu household

Nomsa, aged 56, was the head of a multi-generational household of 11. Her adult daughter, Zinhle, was HIV-positive and extremely ill. Prior to her illness Zinhle had been working in Richards Bay, which enabled her to remit money home to build a structure on her mother's property. When she became very ill she stopped working and returned to live there. No other household member was employed and they relied on the income from three child support grants (which in 2008 had a combined value of R630) collected for Nomsa's youngest child and two grandchildren. Nomsa described the household's financial situation since the loss of Zinhle's support:

I try, with this R600, to buy groceries, pay the school fees and also the school transport for R100. I am left with R500, I went and bought 25 kg [of maize meal], it doesn't last the month, I also buy a bag of beans and I pay for [funeral] insurance … As I have explained, we are hungry. Last month I had to get R200 credit because I didn't have any maize meal left.

Zinhle was unable to walk and needed company when leaving the household. This increased the household's transport costs, which they struggled to afford, and made access to treatment and health care difficult. Nomsa expressed her despair at her household's situation and concluded that to pay for the required trips to the clinic for Zinhle she would need to borrow the money. Transport problems also affected Zinhle's ability to access a disability grant. Although she had applied at the clinic, her approval required assessment by the medical examiner at the Department of Social Development offices, as Nomsa explained:

I haven't been able to take her [to the medical examiner] because I don't have any money left. I have to hire a car [to get to town].

Despite their difficulties getting access, health care was an unspoken priority for the household and the journey to the clinic was shorter and thus cost less than getting to town. Zinhle therefore failed to complete her application and receive a grant.

It is possible that had Zinhle been able to access a disability grant the household's experience of illness would have been different. Zinhle should also have qualified for food parcels from the clinic and these, had they managed to get them, might have made a small difference to the household. Access to a disability grant would have more than doubled the household's small income from the child support grant, thus dramatically improving not only the household income but also Zinhle's personal situation. Unfortunately, although Zinhle started treatment, her condition did not improve, she was eventually hospitalised and died without receiving a disability grant. A combination of health and financial problems that a disability grant might have gone a long way to ameliorating trapped Zinhle in illness and her household in poverty.

5. Discussion

Our case studies appear to indicate an important relationship between ART and the receipt of a disability grant. It seems that, in similar situations, access to targeted financial support, in this case the disability grant, can help to ensure successful health outcomes for those on ART, and also contribute to household's well-being more generally. To recover their health while receiving treatment, household members living with HIV depend on the ability of the household to provide sufficient care, nutrition and other necessities (Zachariah et al., 2006). Like many poor households in South Africa with limited access to other sources of income or sustainable livelihood activities, those in our study depended on the disability grant, and other grants, to meet basic needs, particularly for food and health care.

Quantitative research in the Eastern Cape (Phaswana-Mafuya et al., 2009) produced similar findings, particularly about the food needs of the sick. Households with members receiving the disability grant reported significantly fewer shortages of food in the preceding 12 months than HIV-affected households without the grant. Additional household income in the form of a disability grant has major implications for both the sick person and their household. Evidence from Malawi suggests that inadequate nutrition is linked to greater mortality within the first three months of ART (Zachariah et al., 2006). Botswanan and South African research shows that where household income is lost as a result of illness, replacement income, such as social grants, is key for meeting basic needs, food and economic security, thereby supporting individuals' well-being (Booysen, 2004; Rajaraman et al., 2006). Other research agrees that household poverty and hunger may hinder long-term adherence to treatment, suggesting that the ability to provide for the basic needs of sick adults may support adherence to ART (Hardon et al., 2007; Coetzee et al., 2011).

Besides helping to meet basic needs and thus improving nutrition and health, the different outcomes in our case studies suggest that the relative timing of receipt of the disability grant and beginning ART is important. In contrast to those on ART who received the grant in good time, those who received it late recovered health more slowly. Our case studies therefore suggest that a synergistic relationship may exist between timeous receipt of this grant and improved health outcomes on ART. Several studies have shown that the first six months to a year of treatment is crucial and determines whether the treatment will be successful; those who survive this period have a reduced risk of both morbidity and mortality (Bussmann et al., 2008; Lawn et al., 2008). Our case studies show that timely receipt of the disability grant enabled household members to respond to the nutritional and health care needs of the sick member during this decisive period.

In the case study households where sick members received a disability grant and ART in good time, the respondents spoke about the substantial contribution to household livelihood that the sick member was able to make on returning to health: bringing money into the household through formal or informal employment, investing time and labour in diversifying the household's activities, and taking on additional responsibilities in the home so as to free other members for additional domestic and income generating activities. The timeous receipt of the grant thus had benefits not just for the sick member but also for the household as a whole.

The HIV/AIDS and STI National Strategic Plan 2007–2011 proposes the creation of a new grant targeted at those with chronic illness (DoH, 2007). This has yet to be implemented, despite hearings for an amendment to the Social Assistance Bill in 2010, and is not addressed in the 2012–2016 plan (PMG, 2010; DoH, 2011). Although the plan does not supply much detail about the proposed grant, civil society organisations have actively supported the proposal and suggest further that a chronic disease grant should be designed and targeted towards all those who are chronically ill, not just persons living with HIV (Silber, 2009; Black Sash, 2010). We support further exploration and consideration of such a grant. By improving access to basic needs, it would promote adequate access to health care and better health outcomes for individuals and households, and possibly longer-term adherence to ART (SANAC, 2008). The need to remove remaining barriers to accessing the disability grant and the confusion about qualifying criteria, as well as the need for timely access in combination with ART, support the case for a better targeted and managed cash transfer for severely ill individuals. A chronic illness grant that was separate from the disability grant would be easier to administer and would obviate some of the problems with review and adherence raised in the literature and assumed by general opinion (PMG, 2010), and it would fill the current gap in social welfare provision for those with HIV and other chronic illness such as tuberculosis.

The establishment of a new grant would make it possible to develop eligibility criteria that would enable the applicant to receive it at diagnosis or the start of treatment, thereby acting as an incentive for early VCT (voluntary counselling and testing) and timely treatment. The amount currently provided by the disability grant could also be reduced to ensure its longer-term fiscal viability. Booth & Silber (2008) argue that these changes could remove any current threat of loss currently associated with the disability grant and the possible temptation to stop treatment in favour of the grant; themes that dominate the literature and rhetoric around this grant. A chronic illness grant could also reduce stigma and public misunderstanding about the way in which people with chronic illness qualify for the current disability grant and highlight the differences between chronic long-term illness and disability (Silber, 2009). In addition, it would be possible to make ART adherence a condition of grant receipt, therefore maximising health outcomes. This condition would also support contact with testing, treatment and health care services.

A limitation of this study is that the selection of households was limited to those who were available, had some experience of illness and consented to be interviewed. The convenience sample that resulted is of course not representative of all households. Despite the potential bias towards disclosed households8 and those willing to participate, the sample reflects a range of household types and contexts. While the method of selecting the sample potentially undermines the external validity of the study, it is likely that the findings would be similar in similar contexts.

Another limitation of the research is the small sample size. Case studies of a few households have limited generalisability, but can provide valuable in-depth insight into the experiences of households and are recognised as an important method of analysing data in qualitative research (Mikkelson, 1995). Case studies allow for the collection of retrospective and prospective data over a long time and we would argue that this detailed contextual and historical household information helps us to understand how processes work and therefore develop explanations that help us to suggest possible causality (Mason, 2002).

6. Conclusions

The findings of our four case studies show that the disability grant helps households to provide for their sick members who are on ART, and can enable those members to attend clinics. It can successfully cushion households against the loss of the sick person's employment and help them to respond to the sick person's needs until that person returns to health and can contribute once more. Those with successful ART outcomes are able to make substantial contributions to the well-being of the household. The most important finding is that combined access to the disability grant and ART, especially where the grant is received in good time, has particular benefits for the health of those receiving treatment. The importance of prompt receipt of the disability grant by those on ART suggests that measures to ensure its prompt delivery could improve the cost-effectiveness and long-term efficiency of South Africa's ART programme.

These findings demonstrate the role that social welfare payments can play in supporting the provision of other services, in this case the provision of health care by the Department of Health. We recommend that government further consider and explore the feasibility of a chronic illness grant, thereby providing a cash transfer tailored to the needs of this client group instead of conflating their needs with those of the disabled. Although this proposal needs further investigation, it has the potential to better target poor individuals and households affected by HIV, to act as an instrument to reduce the problems of late diagnosis and initiation of treatment and to allow for more efficient and effective administration of access to the grant.

Acknowledgements

This research was conducted while L Knight was a PhD student at the London School of Hygiene and Tropical Medicine. The research was funded by a UK Economic and Social Research Council (ESRC) postgraduate studentship linked to an ESRC/Department of International Development funded research project (grant number RES-167-25-0076). The authors are grateful to the Africa Centre for Health and Population studies for logistical and administrative support. Comments on a draft of the paper were provided by the facilitators of a SANPAD-funded Writing for Scientific Publication workshop. An early version of this paper was presented at the 4th SA AIDS Conference in 2009.

Notes

⁴1 USD = 8.17 ZAR (on average) in 2012.

⁵Verbal autopsies analyse symptoms to determine cause of death and are particularly useful where cause-of-death data are poor (Herbst et al., 2011).

⁶1 USD = 8.25 ZAR (on average) in 2008.

⁷A stokvel is an informal communal savings or rotating credit union.

⁸These are households where the members have disclosed their status to one another and in some cases to people outside the household.