Background Concurrent partnerships (CPs) are hypothesised to be a risk factor for transmitting HIV. Some have noted that their impact on the epidemic depends upon how common they are in populations, the average number of CPs an individual has, and the amount of time they overlap. However, estimates of prevalence of CPs in Southern Africa vary widely, and the length of overlap in these relationships is inadequately described.

Methods We conducted a sexual behaviour survey (n = 878) in Cape Town, South Africa, using Audio Computer-Assisted Self-Interviewing to collect sexual relationship histories on partners in the previous year. Using the beginning and end dates for the partnerships, we calculated the point prevalence, 1-year cumulative prevalence and degree distribution of CPs, as well as the duration of overlap for relationships begun in the previous year. Linear and binomial regression models were used to quantify race and gender differences in the duration of overlap and relative risk of having CPs in the past year.

Results The overall point prevalence of CPs 6 months before the survey was 8.4%: 13.4% for black men, 1.9% for coloured men, 7.8% black women, and 5.6% for coloured women. The 1-year cumulative prevalence for all sexually active participants was 25.4%. The median duration of overlap in CPs was 7.5 weeks. Women had less risk of CPs in the previous year than men (RR 0.43; 95% CI: 0.32–0.57) and black participants were more at risk than coloured participants (RR 1.86; 95% CI: 1.17–2.97). Overall, of those who had 2 or more relationships in the past year, 85.9% had concurrent as opposed to serially monogamous partnerships.

Conclusions Our results indicate that in this population the prevalence of CPs is relatively high and is characterised by overlaps of long duration, implying there may be opportunities for HIV to be transmitted to CPs.

Disclosure of interest statement We have no commercial contributions to disclose.