When the 'treatment is (or as) prevention' (or 'test and treat' and various other names) hypothesis was first mooted, some wondered how it would work. The plan is to test everyone in a population for HIV regularly and treat everyone found to be infected immediately, rather than waiting for them to reach a particular clinical stage. Adherents of the strategy have vaguely suggested testing 80% or so of a population but this has not been achieved in any high prevalenc country. But if such levels of testing are achieved, how often would testing need to be carried out, and what would be the feasibility of testing such large numbers of people that often?
These issues are still fuzzy. But recent research suggests that many people are unaware of their HIV positive status, even where high rates of testing have been achieved. According to Salim Abdool Karim, this illustrates "the need for frequent repeat testing and comprehensive prevention efforts".
Whatever 'frequent' testing means in high resource, low HIV prevalence countries, it seems an unlikely option in high prevalence countries, which are all poor. Health services are generally not able to cope with relative simple health conditions and in many countries are simply too expensive to afford or too distant to reach. Karim's study shows that rates of new transmissions, which are unbelieveably high in parts of South Africa, are also high among those who tested negative only a short time before.
Getting everyone to test once for HIV, even where 'everyone' means 80%, is hard enough, but getting them to test every year would be a whole lot harder. And every year is not enough in the South African study area in question. So test and treat still raises more questions than answers; are there any high prevalence countries that can meet the challenge of testing so many people so frequently?