Cost-effectiveness of HIV/AIDS interventions in South Africa.

Abstract

Despite some evidence that the effectiveness of HIV/AIDS interventions depends on the context of their implementation, there is a paucity of evidence on the cost-effectiveness (CE) of these interventions in South Africa. The objective of this study is therefore to compare the CE of major HIV/AIDS interventions in epidemiological and socio-economic contexts in South Africa using a methodology which takes into account the effect of the interaction between the context and HIV/AIDS interventions on the costs and effectiveness of such interventions. In epidemiological contexts, the CE of HIV/AIDS interventions is compared across a low HIV prevalence context (LPC) and a high HIV prevalence context (HPC) while in socio-economic contexts the comparison is done across a rural context and an urban context. The comparison of the CE of HIV/AIDS interventions requires the follow-up of patients in HIV/AIDS progression states (non-infected, infected, AIDS, death) over time. However because this follow-up is costly, the thesis models hypothetical populations of HIV/AIDS patients in each context, using two types of models, namely, Markov models and population projection models. These models simulate and project patients in the above-mentioned HIV/AIDS states over time and the cost and effectiveness data, systematically collected from South African literature, are applied to simulated and projected patients. The study finds that in epidemiological contexts, modelled HIV/AIDS interventions are generally more cost-effective in a LPC than they are in a HPC. In socio-economic contexts, the pattern of the CE of modelled HIV/AIDS interventions across a rural and an urban context is not specific and depends on the type of intervention. Prevention of mother-to-child transmission (PMTCT) is more cost-effective in the rural context than it is in the urban context while highly active antiretroviral treatment for adults and children is more costeffective in the urban context than it is in the rural context. The study also finds that the extent of CE varies across HIV/AIDS interventions in any context. Therefore policy makers should allocate resources in accordance with these CE variations.