Article Text


Social and behavioural aspects of prevention poster session 4: Health Services Venue Attendees
P2-S4.08 Patterns and correlates of condom use at last sex episode in a population presenting for voluntary HIV counselling and testing in Soweto, Johannesburg
  1. T Nshimiyimana1,
  2. E Thandekile1,
  3. I Katz2,
  4. N Martinson1,
  5. G Gray1,
  6. G de Bruyn1
  1. 1Perinatal HIV Research Unit, Johannesburg, South Africa
  2. 2Harvard Medical School, USA


Background Promotion of condom use remains an important public health intervention to prevent HIV transmission and acquisition. Routine data from voluntary counselling and testing (VCT) can inform evaluation of prevention efforts that emphasise condom use. This study uses a Johannesburg township VCT data to characterise the correlates of condom use at the sexual episode that preceded VCT visit.

Methods This retrospective analysis uses a service level data on 2829 individuals seeking VCT at a Soweto-based testing center between 2008 and 2010. Multivariate models were used to compare the correlates of condom use with clients' HIV-1 serostatus.

Results Women were 1800 (63.6%) with a mean age of 32 while men were 1029 (36.4%) with a mean age of 35. Out of 1,862 clients reporting sexual intercourse within the month that preceded VCT visit, 27.8% were HIV-1 infected. Among 1,916 HIV-1 seronegative clients, 51.8% reported not using condom at last sexual episode compared to 44% among 911 clients in the HIV-1 seropositive subgroup. HIV-1 seropositive clients reporting not having a sex partner but who described their most recent sexual episode as casual were less likely to have used a condom during that sex act compared to those who reported having their last sex with their regular partners (Adjusted Odd Ratios [AOR], 0.4; 95% CI, 0.2 to 0.6). This difference was not observed in the HIV-1 seronegative subset. In the HIV-1 seropositive subgroup, those reporting previous exposure to VCT were more likely to have used a condom during the sex episode that preceded their VCT visit compared to those reporting it was their first encounter with VCT (AOR, 2.6; 95% CI 1.6 to 4.3). However, the effect of previous exposure to VCT on condom use was not significant in the HIV-1 seronegative subset. For both the HIV-1 seronegative and seropositive subsets, condom use likelihood was higher among men compared to women (AOR, 1.6; 95% CI 1.2 to 2.0 and AOR 1.7; 95% CI 1.1 to 2.7), respectively.

Conclusion The findings underline a need for distinguishing HIV risk reduction counselling targeted at the HIV-1 infected from those aimed at the HIV-1 uninfected, especially in Soweto context.

Statistics from

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.