In South Africa, the proportion of women aged 15-24 years who are infected with HIV is about three times that of men this age (16% vs. 5%), according to a nationally representative survey.1 Among sexually experienced youth, men's odds of HIV infection were elevated if they had had a genital ulcer in the past year, and women's odds were elevated if they had had an unusual vaginal discharge during that period, if their most recent partner was older than they were or if they did not use condoms consistently. The odds for each sex increased with the number of partners. In contrast, young people who had taken part in a national HIV prevention program were less likely to be infected.

Between March and August 2003, researchers conducted a nationally representative household survey among 15-24-year-olds in South Africa. Young people were asked about demographics, sexual behavior, symptoms of sexually transmitted infections and exposure to programs of a national HIV prevention campaign (loveLife). Samples of saliva were collected and assayed for HIV antibodies. The researchers used chi-square analysis to compare HIV status and characteristics between men and women, and they used multivariate logistic regression analysis restricted to sexually experienced youth to identify risk and protective factors for HIV infection. The multivariate analyses included both variables that were found significant in bivariate analyses and factors theorized as important. Analytic methods accounted for sampling strata, primary sampling units, and population weights.

The total sample consisted of 11,904 young people; 77% of eligible youth participated. As in the South African population, 82% of young people were of black African race. About half resided in rural areas; one-fourth lived in households lacking electricity. Roughly 38% of 20-24-year-olds had completed high school, and 74% of 15-19-year-olds attended school. Small proportions—1% of men and 3% of women—were married.

More than one-third of young people had attended a loveLife program, and about two-thirds said they had changed their behavior because of HIV/AIDS, with no differences between the sexes. A larger proportion of women than of men had previously been tested for HIV (25% vs. 15%).

The proportion of young women infected with HIV was more than three times that of young men (16% vs. 5%). For females, HIV prevalence rose rapidly from 4% at ages 15 and 16 to a high of 31% at age 21; for men, however, prevalence was relatively stable at 2-3% between ages 15 and 19, and then increased to 11-12% at ages 23-24.

Two-thirds of youth reported ever having had vaginal or anal sex. In this sexually experienced subset, the average number of lifetime sexual partners was greater among men than women (five vs. two), as was the average number of partners in the past year (two vs. one). A larger proportion of women than of men had had an unusual genital discharge in the past year (19% vs. 9%), whereas similar proportions had had genital ulcers or sores (7% and 6%).

Roughly 83-89% of sexually experienced youth had had sex in the past year. In this subset, the majority had not consistently (always) used condoms with their most recent partner; inconsistent use was more common among women than among men (71% vs. 61%). The most recent partner had been the same age or younger for 88% of men, but had been older for 90% of women. For the majority of each sex, the last partner had been their main partner; however, for a larger proportion of men than women, this partner had been a casual one (18% vs. 2%).

A sex-stratified multivariate analysis examined possible risk factors for HIV infection. In this analysis, young men and women who were black had significantly higher odds of HIV infection than did their counterparts of other races (adjusted odds ratios, 2.6 and 8.3, respectively). The odds of HIV infection were higher among men and women from urban settings than among those from rural settings (2.0 for each), among those aged 20-24 than among those aged 15-19 (2.6 and 4.3), and among those who had not completed high school relative to those who had (1.9 and 2.3). The odds increased with each additional sexual partner for both men and women (1.03 and 1.09, respectively).

Women had a significantly elevated risk of being infected with HIV if they had been sexually active for longer than a year (adjusted odds ratio, 2.4), if they had not always used a condom with their most recent partner (1.5) and if they had had an unusual vaginal discharge in the past year (1.8). Among 15-19-year-old women, the odds were higher for those whose recent partner was five or more years older than for those whose partner was the same age or younger (3.2); similarly, among 20-24-year-old women, odds were higher among those whose most recent partner was 1-4 years older than among those whose partner was the same age or younger (2.3). In contrast, women who were married were less likely to be infected than their unmarried counterparts (0.6). Men who had experienced genital ulcers in the past year had higher odds of HIV infection than did men who had not experienced ulcers during this period (1.9), while men who were circumcised had lower odds than did their uncircumcised counterparts (0.6).

A second sex-stratified multivariate analysis examined possible HIV protective behaviors among sexually experienced youth, taking into account other factors potentially influencing the likelihood of HIV infection. In this analysis, young men and women who had participated in a loveLife prevention program were significantly less likely than nonparticipants to be infected with HIV (adjusted odds ratios, 0.6 for each). However, men who had ever been tested for HIV were more likely to be infected than untested men (1.9). Having known someone who had died of AIDS and having participated in youth programs was not associated with the risk of infection.

According to the researchers, although levels of behaviors that protect against HIV infection may be rising among South African youth, they are still not high enough to substantially curtail the epidemic. Given the likely interaction among the risk factors studied, they argue, prevention efforts must be comprehensive. "Programs for youth must continue to promote partner reduction, consistent condom use and prompt treatment for sexually transmitted infections while also addressing contextual factors... that make it difficult for youth, particularly young women, to implement behavior change...," they conclude. —S. London


1. Pettifor AE et al., Young people's sexual health in South Africa: HIV prevalence and sexual behaviors from a nationally representative household survey, AIDS, 2005, 19(14):1525-1534.