Through a glass darkly?': An enquiry into HIV prevalence on Stellenbosch wine farms

Master Thesis


Permanent link to this Item
Journal Title
Link to Journal
Journal ISSN
Volume Title

University of Cape Town

Despite the complex and often highly specific nature of the social aspects of the HIV/AIDS pandemic, many projects working in the field do not base their strategies on local evidence, given the paucity of suitable local-level data as well as the presences of organisational constraints. A project offering HIV testing to farm-based communities in Stellenbosch is a case in point. While no prevalence data exists for this sub-population, the assumption was that there may be high levels of infection, following the organisation's experience of AIDS-related illnesses on these farms and the social conditions on wine farms which were thought to produce vulnerability for infection. Some in the organisation also thought that farm-based communities battled to access healthcare. During the first year of providing voluntary counselling and testing (VCT) on wine farms, however, the Stellenbosch Hospice's Farms Project consistently found lower than expected levels of HIV infection. This gave rise to the question being addressed in this thesis - which is what can be 'known' about HIV prevalence in a sub-population for whom there is no evidence-based prevalence data. In practical terms, if modestly-funded local-level organisations1 were able to undertake accessible forms of research, what would they be able to surmise about HIV prevalence among proposed beneficiaries? Taking an unusual approach to research on prevalence, this study employs a minimally positivist approach to investigate what can be 'known' about HIV prevalence on wine farms in the Stellenbosch area. It does so by 1 This term is used to include various forms of organisations - be they nongovernmental, non-profit or service organisations - which are small, relatively survivalist organisations. It may be a church-based organisation, a large communitybased healthcare organisation or a service organisation like a hospice. I do so to differentiate it from the larger, professionalised non-governmental organisations (NGO) which frequently have research capacity. My notional organisation is also not a community-based organisation (CBO), however, which are largely membership-based and whose access to their locations is usually more organic and embedded, while NPOs are invariably staffed by people who do not necessarily live in the locations in which they are intervening. vi triangulating data from the four sources that such an organisation might use, had they the capacity. These sources are published statistics and published articles, the opinions of local 'experts', and their own organisational data - in this case the first year of Farms Project's results. Significantly this does not include the more conventional surveys and statistical modelling, which is beyond this kind of organisation's capacity. After reviewing publicly available prevalence data and showing that there are none for this sub-sector, this study probes the HIV 'risk' and related prevalence data associated with issues of poverty, gender relations, 'race' and alcohol consumption on Stellenbosch wine farms. In addition it presents prevalence data from a sample of farms as well as reviews HIV 'risk' and prevalence in rural areas nationally. In doing so, it critiques the causal links often made between the kinds of social conditions found on farms and HIV infection. On the basis of the data considered and the methods used, the study finds that levels of HIV infection on farms could be expected to be lower than the average prevalence in the Stellenbosch health sub-district. It cautions, however, that this finding is not conclusive, not least as it was unable to consider some significant social conditions - like the movement of people, and effects of socially conscious farmers and the services they provide. In addition it is not generalisable to other South African farms, given the particularity of wine farms and of the Western Cape. The study concludes by noting the limited value of prevalence data to project design, given the range of factors that can affect it at any time, and that it necessarily masks variation within an area or sub-population. While prevalence is useful as a starting point in project design, it is important to disaggregate where infection lies through an analysis of key social conditions. The study concludes by highlighting the importance of this finer analysis for project design in order to avoid strategies founded on poor assumptions, while recognising the difficulty of this for modest organisations.