Sexual behaviour in a rural high HIV prevalence South African community: time trends in the antiretroviral treatment era

OBJECTIVES Data from generalized epidemic settings have consistently found that patients on antiretroviral therapy (ART) reduce sexual risk behaviours, but how sexual behaviour changes in the general population in response to ART availability, including amongst HIV-uninfected and undiagnosed adults, has not been characterized in these settings. DESIGN General population open cohort. METHODS We report trends in sexual behaviour indicators for men aged 17-54 years and women aged 17-49 years in rural KwaZulu-Natal province, based on annual sexual behaviour surveys during ART scale-up from 2005 to 2011. Estimates are adjusted for survey nonparticipation and nonresponse to individual survey items using inverse probability weighting and multiple imputation. Trends are presented by HIV status, knowledge of status, age and marital status. RESULTS Reports of condom use at last sex with a regular partner increased by 2.6% points per year [95% confidence interval (CI) 1.5%, 3.7%] for men and 4.1% per year (3.0%, 5.3%) for women. Condom use at last sex with a casual partner was high and did not change significantly over the period for both sexes. There were statistically significant declines in the percentage reporting multiple partnerships in the last year and the point prevalence of concurrency. Trends within subgroups were generally consistent with overall estimates. CONCLUSION We find no evidence of increased sexual risk-taking following ART availability and protective changes in some behaviours, suggesting that general trends in sexual behaviour are not counter-acting preventive effects of HIV treatment. Continued monitoring of population-level sexual behaviour indicators will be essential to interpret the success of combination-prevention programmes.