HIV Sexual Transmission Is Predominantly Driven by Single Individuals Rather than Discordant Couples: A Model-Based Approach

David Champredon, School of Computational Science and Engineering, McMaster University, Hamilton, CanadaSteve Bellan, Center for Computational Biology and Bioinformatics, University of Texas at Austin, Austin, Texas, United States of AmericaJonathan Dushoff, Department of Biology, McMaster University, Hamilton, CanadaUnderstanding the relative contribution to HIV transmission from different social groups is important for public-health policy. Information about the importance of stable serodiscordant couples (when one partner is infected but not the other) relative to contacts outside of stable partnerships in spreading disease can aid in designing and targeting interventions. However, the overall importance of within-couple transmission, and the determinants and correlates of this importance, are not well understood. Here, we explore how mechanistic factors – like partnership dynamics and rates of extra-couple transmission – affect various routes of transmission, using a compartmental model with parameters based on estimates from Sub-Saharan Africa. Under our assumptions, when sampling model parameters within a realistic range, we find that infection of uncoupled individuals is usually the predominant route (median 0.62, 2.5%–97.5% quantiles: 0.26–0.88), while transmission within discordant couples is usually important, but rarely represents the majority of transmissions (median 0.33, 2.5%–97.5% quantiles: 0.10–0.67). We find a strong correlation between long-term HIV prevalence and the contact rate of uncoupled individuals, implying that this rate may be a key driver of HIV prevalence. For a given level of prevalence, we find a negative correlation between the proportion of discordant couples and the within-couple transmission rate, indicating that low discordance in a population may reflect a relatively high rate of within-couple transmission. Transmission within or outside couples and among uncoupled individuals are all likely to be important in sustaining heterosexual HIV transmission in Sub-Saharan Africa. Hence, intervention policies should be broadly targeted when practical.DC and JD were supported by grants from the James S. McDonnell Foundation, the Canadian Institutes of Health Research, and the Canadian Natural Sciences and Engineering Research Council. SEB was supported by a National Institute of General Medical Sciences MIDAS grant U01GM087719 to Lauren A. Meyers and Alison P. Galvani. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.Center for Computational Biology and BioinformaticsEmail: champrd@math.mcmaster.c