Context: It is important to examine whether youth from disadvantaged households are less likely than others to use a condom at first sex, even after correcting for shared characteristics within communities.
Methods: Baseline survey data from the Transitions to Adulthood in the Context of AIDS in South Africa study in KwaZulu-Natal were used. Random effects logistic regression assessed the relationship between poverty and 14-22-year-olds' use of condoms at first sex, correcting for shared characteristics of adolescents within each community.
Results: Twenty-three percent of young people had used a condom at first sex. Poor and extremely poor females had about one-third the odds of nonpoor females of using a condom at first sex, even after adjusting for community clustering; among males; however, there was no association between poverty and condom use, after adjusting for background factors and community clustering.
Conclusions: The importance of community clustering of neighborhood-level characteristics differs by gender in South Africa. Poverty remains a central risk factor for HIV among young women, regardless of the surrounding context, but not among men.