Advancing Sexual and Reproductive Health and Rights
International Perspectives on Sexual and Reproductive Health
Volume 35, Number 4, December 2009

South African Women Living in Areas with High Levels of Sexual Violence Are at Risk for HIV, Early Pregnancy

Young South African women residing in communities where sexual violence is common are more likely than other women to have been pregnant before age 18 and to be HIV-positive, according to a nationally representative study.1 In addition, young women in these communities are less likely than their peers to have used a condom at last sex. However, personal experience with sexual violence was not associated with these outcomes.

Forced sex and intimate partner violence have been linked to a variety of reproductive health problems in South Africa, where a third of women aged 20–24 are HIV-positive and a similar proportion of 15–19-year-old females have been pregnant. To assess whether sexual violence at the community level is associated with HIV status, condom use and adolescent pregnancy (i.e., before age 18), researchers analyzed data from 6,217 female respondents in a national 2003 survey of 15–24-year-olds from randomly selected households. In face-to-face interviews, respondents provided demographic information (including age, race, residence and level of education) and reported whether they had used a condom at last sex, whether they had been pregnant before age 18 and whether they had ever been coerced (threatened or physically forced) into having sex. Respondents' HIV status was determined by testing saliva samples. To create community-level variables for sexual violence, high school completion and sexual activity, the researchers aggregated respondents' data to determine the average level of these outcomes for each of the country's census enumeration areas. (About 4% of women were dropped from the analyses because the number of respondents who lived in their enumeration area was too low for the researchers to calculate community-level statistics.) Multivariate logistic regression analyses were used to assess the relationships among demographic variables, individual experience with sexual violence, community-level variables and the sexual health outcomes. The analysis of HIV status focused on 3,719 unmarried, sexually experienced women; the condom use analysis was further limited to 3,299 women who had had sex in the past 12 months, and the adolescent pregnancy analysis to 3,016 women aged 18 and older.

The mean age of young women in the full sample was 19, and eight in 10 were black. One-quarter (23%) of the women had a high school education and about two-thirds had ever had sex (65%); 11% of this last group had ever been coerced into having sex.

Among unmarried, sexually experienced respondents, 21% were HIV-positive; HIV status did not differ according to whether they had ever been forced to have sex. More than half of unmarried respondents who had had sex in the past year had used a condom at last sex (52%), but the proportion was lower among women who had ever had coerced sex (41%) than among women with no history of sexual violence (53%). Thirteen percent of unmarried, sexually active respondents 18 or older had been pregnant prior to age 18; again, the proportion differed between those who had ever been forced to have sex (16%) and those who had not (12%).

In multivariate analyses, the odds of condom use at last sex were reduced among women who lived in a community with high levels of sexual violence (odds ratio, 0.9 for each one- unit increase in z-scores). However, the association between condom use at last sex and having personally experienced sexual violence was only marginally significant (0.7) when community-level factors were taken into account. A personal history of sexual violence was not associated with being HIV-positive or with having had an adolescent pregnancy, although the odds of the two outcomes were significantly elevated among those living in a community where sexual violence was prevalent (1.2 for each one-unit increase in z-scores for both outcomes).

Several demographic variables were associated with the sexual health outcomes. The odds of being HIV-positive were higher among blacks (7.2) than among whites. High school graduates had twice the odds of those with less education of having used a condom at last sex (2.0), and they had reduced odds of being HIV-positive and having been pregnant during adolescence (odds ratios, 0.4 for both). Compared with their rural counterparts, urban women had higher odds of both having used a condom at last sex (1.9) and being HIV-positive (2.0).

The investigators acknowledge that the cross-sectional design of the survey prevented them from determining if coerced sex had direct negative effects on women's sexual health. Another limitation is that the community-level variables were based on data from the sample rather than from the entire community. Furthermore, communities with high levels of sexual violence may have had the attendant problems of increased HIV prevalence and reduced access to reproductive health services. Still, the researchers note, these findings "illustrate that social norms and community influences are important, especially in terms of reproductive health behaviors and outcomes." They encourage further qualitative and longitudinal research on this topic and increased programmatic efforts to "address sexual violence as part of effective prevention strategies."—S. Ramashwar


1. Speizer IS et al., Sexual violence and reproductive health outcomes among South African female youths: a contextual analysis, American Journal of Public Health, 2009, 99(Suppl. 2):S425–S431.