1146Assessing the uncertainty around age-mixing patterns in HIV transmission inferred from phylogenetic trees
Abstract Introduction Understand age-mixing patterns in HIV transmission is a key to design and implement HIV prevention interventions. As such, the use of phylogenetic tree seems promising since this approach is based on data from a transmission network. Methods Through a simulation study, we used transmission clusters computed from phylogenetic trees to investigate age-mixing patterns. From transmission clusters, we estimated a transmission network using pairings. We inferred measurements, which depict age-mixing patterns in transmission i.e. ‘proportions of men/women of a certain age-group paired with women/men of another age-group’, and mean, median, and standard deviation of average age difference between women/men with their respective pairs. We investigated the uncertainty around these measurements as a function of sampling coverage in different sequence missingness scenarios. Results According to the simulation set-up of age mixing in partnership, there are relationships, which are less likely and others highly likely. Explicitly, partnerships between younger men (<25 years) and older women (25 to 50 years) are less likely to happen in the sexual network, whereas partnership between younger women (<25 years) and older men (25 to 50 years) are more likely to happen as are age group relationship having intermediate magnitude in partnership. Conclusion Transmission clusters allow us to strive with more improved information regarding cross-generation transmission. It enables us to understand the importance of age mixing beyond describing the mean age difference, and that one needs to consider the variation of age differences between pairs of individuals as well as among the partners of a given individual.