According to South Africa's leading demographic model, the provision of highly active antiretroviral therapy (HAART) for people living with AIDS will have positive spin-offs in terms of limiting the spread of HIV/AIDS. The ASSA2000 Interventions Model (an offshoot of the ASSA2000 demographic model of South Africa) predicts that over a million new HIV infections could be averted in the first twelve years if a HAART treatment programme was added to a set of AIDS prevention interventions. In other words, not only is HAART assumed to prolong the lives of AIDS patients, but it is also assumed to save lives by preventing a significant number of new HIV infections. See Johnson and Dorrington (2002) for a discussion of the modelling assumptions and supporting evidence. This link between AIDS treatment and AIDS prevention occurs for two reasons: firstly, HAART lowers the viral loads of AIDS patients, thereby making them less infectious. Secondly, the model assumes that voluntary counselling and testing (VCT) is an integral part of a HAART programme, and that VCT is effective in encouraging people to practice safer sex. This modelled link between AIDS treatment and prevention is, however, contentious. According to a recent Joint Health and Treasury Task Team government report (on the economic of providing HAART in South Africa), there is 'no compelling evidence that antiretrovirals would reduce numbers of new infections' (2003: 18). This is a strange comment, given that the ASSA2000 Interventions model (which, as noted above, assumes the opposite) was used by the task team in their modelling work. Nevertheless, seeing as there is wide-spread concern about the nature of the link between HAART and fewer HIV infections, the matter requires serious consideration.