Gender asymmetry in concurrent partnerships and HIV prevalence

Ka Yin Leung*, Kimberly A. Powers, Mirjam Kretzschmar

*Corresponding author for this work

Research output: Contribution to journalArticleAcademicpeer-review


The structure of the sexual network of a population plays an essential role in the transmission of HIV. Concurrent partnerships, i.e. partnerships that overlap in time, are important in determining this network structure. Men and women may differ in their concurrent behavior, e.g. in the case of polygyny where women are monogamous while men may have concurrent partnerships. Polygyny has been shown empirically to be negatively associated with HIV prevalence, but the epidemiological impacts of other forms of gender-asymmetric concurrency have not been formally explored. Here we investigate how gender asymmetry in concurrency, including polygyny, can affect the disease dynamics. We use a model for a dynamic network where individuals may have concurrent partners. The maximum possible number of simultaneous partnerships can differ for men and women, e.g. in the case of polygyny. We control for mean partnership duration, mean lifetime number of partners, mean degree, and sexually active lifespan. We assess the effects of gender asymmetry in concurrency on two epidemic phase quantities (R 0 and the contribution of the acute HIV stage to R 0) and on the endemic HIV prevalence. We find that gender asymmetry in concurrent partnerships is associated with lower levels of all three epidemiological quantities, especially in the polygynous case. This effect on disease transmission can be attributed to changes in network structure, where increasing asymmetry leads to decreasing network connectivity.

Original languageEnglish
Pages (from-to)53-60
Number of pages8
Early online date20 Jan 2017
Publication statusPublished - Jun 2017


  • Concurrency
  • Gender asymmetry
  • HIV prevalence
  • Mathematical model
  • Polygyny


Dive into the research topics of 'Gender asymmetry in concurrent partnerships and HIV prevalence'. Together they form a unique fingerprint.

Cite this