Chapter 4 reported the results of a demographic modelling exercise which assumed that an aids treatment intervention not only extends the lives of those on haart, but also results in fewer new hiv infections. This is because (as modelled by the assa2000 Interventions Model) hiv-positive people are less infectious when they are on haart, and because more of them will participate in a vct programme when aids treatment interventions are available. The assa2000 Interventions Model assumes that people who have experienced vct subsequently modify their sexual behaviour, although this improvement is assumed to wear off over time for those who test hiv-negative.

The idea that a haart programme has a strong preventive element is often greeted with scepticism and surprise. The government's Joint Health and Treasury Task Team reflected the conventional wisdom on the topic by stating that there is ‘no compelling evidence that antiretrovirals would reduce numbers of new infections’ (2003: 18). Any argument to the effect that haart is likely to prevent new hiv infections thus needs to be developed in more detail.

As noted above, the assa2000 Interventions Model relied on several behavioural studies indicating that there is a positive relationship between behaviour change and vct. But is this a reliable basis for modelling the net behaviour change resulting from haart? vct can facilitate shifts towards safer sexual practices, but might not the very availability of haart itself result in behaviour change in the opposite direction, i.e. by encouraging people to relax their guard and practise less safe sex?