Document Type


Publication Date


Department 1



The notion that concurrent sexual partnerships are especially common in sub-Saharan Africa and explain the region’s high HIV prevalence is accepted by many as conventional wisdom. In this paper, we evaluate the quantitative and qualitative evidence offered by the principal proponents of the concurrency hypothesis and analyze the mathematical model they use to establish the plausibility of the hypothesis.

We find that research seeking to establish a statistical correlation between concurrency and HIV prevalence either finds no correlation or has important limitations. Furthermore, in order to simulate rapid spread of HIV, mathematical models require unrealistic assumptions about frequency of sexual contact, gender symmetry, levels of concurrency, and per-act transmission rates. Moreover, quantitative evidence cited by proponents of the concurrency hypothesis is unconvincing since they exclude Demographic and Health Surveys and other data showing that concurrency in Africa is low, make broad statements about non-African concurrency based on very few surveys, report data incorrectly, report data from studies that have no information about concurrency as though they supported the hypothesis, report incomparable data and cite unpublished or unavailable studies. Qualitative evidence offered by proponents of the hypothesis is irrelevant since, among other reasons, there is no comparison of Africa with other regions.

Promoters of the concurrency hypothesis have failed to establish that concurrency is unusually prevalent in Africa or that the kinds of concurrent partnerships found in Africa produce more rapid spread of HIV than other forms of sexual behaviour. Policy makers should turn attention to drivers of African HIV epidemics that are policy sensitive and for which there is substantial epidemiological evidence.


This work is licensed under a Creative Commons Attribution 3.0 License.



Required Publisher's Statement

Original version is available from the publisher at: