Coital dilution, the reduction in the coital frequency per partner when an additional ongoing partner is added, may reduce the transmission potential of partnership concurrency for HIV and other sexually transmitted infections. Empirical estimates of dilution, especially dilution of sexual acts unprotected by condoms, are needed to inform prevention research.
Sexually active adults in Accra, Ghana were recruited in a multistage household probability sample. Degree (number of ongoing partners), total acts, and unprotected acts were measured retrospectively for each month in the past year through an event history calendar. Random-effects negative binomial models estimated the association between degree and coital frequency.
Compared with person-months with a single partner (monogamy), 2.06 times as many total acts and 1.94 times as many unprotected acts occurred in months with 2 partners. In months with 3 partners, 2.90 times as many total acts and 2.39 times as many unprotected acts occurred compared with monogamous months. Total acts but not unprotected acts also declined with partnership duration.
No dilution was observed for total acts with up to 3 concurrent partners, but a small amount of dilution was observed for unprotected acts for months with multiple concurrencies. This suggests moderate selective condom use in months with multiple concurrencies. The implications of the observed dilution for future HIV transmission must be investigated with mathematical models.
*Department of Epidemiology, University of Washington, Seattle, WA;
†Regional Institute for Population Studies, University of Ghana, Legon, Ghana;
‡Department of Virology, Noguchi Memorial Institute for Medical Research, University of Ghana, Legon, Ghana;
§Department of Sociology, Pennsylvania State University, University Park, PA; and
‖Department of Geography, University of California, Santa Barbara, CA.
Correspondence to: Samuel M. Jenness, MPH, Department of Epidemiology, University of Washington, 1959 Pacific Street NE, Box 357236, Seattle, WA 98195 (e-mail: firstname.lastname@example.org).
Supported in part by the NICHD (R00 HD057533) and the UW Center for AIDS Research SPRC (P30 AI027757). Computing support was provided by an NICHD research infrastructure grant (5R24HD042828) to the UW Center for Studies in Demography and Ecology.
The authors have no funding or conflicts of interest to disclose.
Received August 13, 2014
Accepted December 01, 2014