HIV spread and partnership reduction for different patterns of sexual behaviour - a study with the microsimulation model STDSIM

We studied how sexual behaviour affects population HIV spread simulating stylized risk profiles: (1) prostitution, no short relationships (resembling settings in South-East Asia); (2) prostitution, concurrent short relationships (resembling South-America and urban sub-Saharan Africa); (3) no prostitution, concurrent short relationships (resembling rural sub-Saharan Africa); (4) prostitution, serial short relationships (a generic low-risk setting). We explored the impact on HIV prevalence of prevention programs accomplishing postponement of sexual debut, reduction in partner change rate and in prostitution. We described the representation of sexual behaviour in the microsimulation model STDSIM, comparing it to non-individual-based models. The profiles generate markedly different time courses of HIV spread. Concentration of risk causes a rapid initial spread (Profiles 1 and 2), whereas the final prevalence depends more on the overall extent of risk behaviour in the general population (highest for Profiles 2 and 3). Effects of partnership reduction are strongly context dependent. Small decreases in numbers of partners reduce HIV spread considerably if they reflect decreases in the contacts of highest risk in that setting. In settings with risk behaviour dispersed over a large part of the population (Profiles 2 and 3), indirect effects can cause the impact on HIV to be disproportionately large compared to the magnitude of behaviour change.