Effective HIV prevention and testing planning requires an understanding of where and among whom new infections are occurring. Sub-Saharan Africa remains the region most affected by the epidemic, but as a result, it also has accumulated a large quantity of data that allow better prediction of HIV incidence patterns. UNAIDS, through their “know your epidemic, know your response” initiative, has been assisting countries in characterizing their epidemic through the use of mathematical modeling tools. Among those is the “Modes of Transmission” (MoT) model, which predicts the distribution of new infections in the next year by groups, according to their main mode of HIV exposure. We developed a new version of this model, the Incidence Patterns Model (IPM), that maximizes the use of the Demographic and Health surveys (carried out routinely in most sub-Saharan African countries) to predict short term incidence patterns by group, according to key HIV risk factors relevant programmatically, and by geographical location. The model is embedded in a Bayesian framework, allowing rigorous uncertainty estimation, and was validated and trained with data from four cohort studies in the region. Used in conjunction with contextual epidemiological studies, the IPM represents a valuable tool to support programmatic planning and data collection towards the implementation of effective interventions to curb incidence and identify new infections at the local level.
See other Director's Choice papers.