Chapter 3 argued that an mtctp programme would more than pay for itself because fewer children would become hiv-positive – thus resulting in a lower overall cost to the health sector. Can a similar case be made for the provision of chronic highly active antiretroviral therapy (haart) for those already living with aids? Could a haart intervention ‘pay for itself’ by averting aids-related morbidity and by encouraging people to be tested and counselled – thereby contributing to behaviour change and lower rates of hiv transmission?

This is a much tougher question to answer because the direct costs are so high, and because the benefits are in large part dependent on the validity of the assumptions used to model the impact of the intervention. The analysis presented in this chapter integrates a detailed costing exercise with the output of the assa2000 Interventions Model. It draws on work presented in more detail in Geffen et al. (2003). This study is one of several publicly available costing studies of antiretroviral provision in South Africa (see review by Boulle et al. 2003 and jhttt 2003). Although some of the input data and modelling assumptions occur in all studies, they are sufficiently different to prevent direct comparisons. The study presented here is the most detailed and produces the highest cost estimates. It can thus be regarded as the ‘worst-case scenario’ from a cost perspective.

The assa2000 Interventions Model, which is an adaptation of the assa2000 model, was designed to estimate the likely impact of various hiv/aids health policy interventions.