Understanding the public health implications of self-reported condom use in HIV clinical trials: lessons learned in KwaZulu Natal, South Africa (2002-2016)

Aim To investigate the associations between self-reported condom use and HIV, sexually transmitted infections (STIs) and pregnancy rates among women who enrolled in HIV biomedical intervention trials in South Africa. Subjects and Methods Data from 9948 women were analysed. Cox regression models were used to assess the associations between condom use and primary outcomes. Results Overall HIV, STI and pregnancy incidence rates were 7, 15 and 9 per 100 person/year. Women reporting condom use (65%) at last sex were younger, had more schooling, had fewer children and were more likely to be single/non-cohabiting, and hormonal injectables were the most common family planning method among women who reported using a condom at last sex. They were at increased risk of HIV (adjusted hazard ratio (aHR) 1.23, 95% CI 1.07, 1.43), STI (aHR 1.20, 95% CI 1.05, 1.34) and pregnancy (aHR 1.23, 95% CI 1.10, 1.40). Conclusion In this population, self-reported condom users were identified as a proxy for a combination of high-risk sexual behaviours. They were associated with increased risks of HIV, STI and pregnancy incidence rates. These counterintuitive findings broadly indicate the women's correct perception about their risky sexual behaviours which may have led them to use condoms. Our findings can have significant implications in clinical research settings by assisting in recruitment and monitoring strategies.