Does serosorting reduce HIV incidence?

The evidence is very mixed on whether serosorting works as an HIV risk-management strategy. Linking serosorting with HIV incidence poses several methodological problems.

In ecological studies, one of the biggest problems lies in ruling out other variables. For instance, in the San Francisco figures for HIV incidence cited below, declines in incidence were initially ascribed to serosorting. Latterly, however, the tendency has been to look much more at whether treatment is reducing the overall infectivity of HIV-positive men. Without a direct link between specific gay men and their risk of acquiring HIV, it is very difficult to establish exactly what contribution serosorting makes to incidence.

Another factor, explored later in ‘Serosorting or Seroguessing?’,is that most surveys among gay men ask respondents about their perception of their partners’ HIV status, and studies have established that this is often highly inaccurate. Few studies (one is cited below) have only looked at the behaviour of gay men in situations where they have specifically asked and received an answer about their partner’s status.

Finally, it is methodologically and ethically very difficult to perform a study of HIV-positive gay men’s risk of transmitting HIV. Although one carefully constructed series of studies has managed to establish a link between HIV-positive gay men’s viral load and their risk of transmission,¹ the same has not been done for serosorting behaviour.