Determinants of HIV infection among adolescent girls and young women aged 15–24 years in South Africa: a 2012 population-based national household survey

BACKGROUND: South Africa is making tremendous progress in the fight against HIV, however, adolescent girls and young women aged 15-24 years (AGYW) remain at higher risk of new HIV infections. This paper investigates socio-demographic and behavioural determinants of HIV infection among AGYW in South Africa. METHODS: A secondary data analysis was undertaken based on the 2012 population-based nationally representative multi-stage stratified cluster random household sample. Multivariate stepwise backward and forward regression modelling was used to determine factors independently associated with HIV prevalence. RESULTS: Out of 3092 interviewed and tested AGYW 11.4% were HIV positive. Overall HIV prevalence was significantly higher among young women (17.4%) compared to adolescent girls (5.6%). In the AGYW model increased risk of HIV infection was associated with being young women aged 20-24 years (OR = 2.30, p = 0.006), and condom use at last sex (OR = 1.91, p = 0.010), and decreased likelihood was associated with other race groups (OR = 0.06, p < 0.001), sexual partner within 5 years of age (OR = 0.53, p = 0.012), tertiary level education (OR = 0.11, p = 0.002), low risk alcohol use (OR = 0.19, p = 0.022) and having one sexual partner (OR = 0.43, p = 0.028). In the adolescent girls model decreased risk of HIV infection was associated with other race groups (OR = 0.01, p < 0.001), being married (OR = 0.07), p = 0.016], and living in less poor household (OR = 0.08, p = 0.002). In the young women's models increased risk of HIV infection was associated with condom use at last sex (OR = 2.09, p = 0.013), and decreased likelihood was associated with other race groups (OR = 0.17, p < 0.001), one sexual partner (OR = 0.6, p = 0.014), low risk alcohol use (OR = 0.17, p < 0.001), having a sexual partner within 5 years of age (OR = 0.29, p = 0.022), and having tertiary education (OR = 0.29, p = 0.022). CONCLUSION: These findings support the need to design combination prevention interventions which simultaneously ...