South African Intervention Improves Likelihood Of Consistent Condom Use

First published online:

An intervention aimed at reducing the risk of HIV and other STIs increased the likelihood of condom use among South African men, according to a neighborhood-randomized trial conducted between 2007 and 2010.1 Compared with peers who received a general health intervention unrelated to sexual risks, men who received the HIV risk-reduction intervention had 30% higher odds of having consistently used condoms during vaginal intercourse and 40% higher odds of having used them at last vaginal intercourse. These differences were apparent for both sex with steady partners and sex with casual partners.

For each of 22 matched pairs of neighborhoods in Eastern Cape Province, South Africa, investigators assigned one neighborhood to an HIV risk-reduction intervention and the other neighborhood to a control intervention targeting diet, physical activity and alcohol intake to improve health. The former aimed to strengthen behavioral beliefs that support condom use, increase skill and self-efficacy to use condoms, and improve knowledge of how to reduce HIV risk. Both interventions were delivered in three weekly sessions to groups of 9–15 men by trained facilitators and used interactive exercises, games, brainstorming, role playing, take-home assignments, group discussions and locally filmed videos.

Participants were recruited from a variety of community venues and were eligible for the study if they were aged 18–45 and had had vaginal intercourse in the past three months. They answered questions about their social and demographic characteristics and sexual behaviors during audio computer-assisted self-interviews at baseline and at six and 12 months. The primary outcome of interest was consistent condom use during vaginal sex (i.e., every time) in the past three months, although other outcomes were also assessed. The investigators compared outcomes between groups over time using descriptive statistics and logistic, linear and multinomial generalized estimating equation models.

The 1,181 men included in analyses were 27 years old, on average, at baseline. Only 6% were married, but 80% had had a steady partner and 51% a casual partner in the past three months. Forty-four percent had completed high school. Roughly seven in 10 were unemployed and six in 10 were alcohol dependent.

At baseline, participants in the HIV intervention group had consistently used condoms in the past three months with 34% of their steady partners and 55% of their casual partners; at the two follow-up surveys, consistent condom use had risen to 39–40% with steady partners but had not increased with casual partners (51–54%). In the control group, levels of consistent condom use showed little variation across the three time points, with either steady partners (32–34%) or casual partners (50–52%).

In multivariate analyses that combined data for the two follow-up surveys and adjusted for consistent condom use at baseline, men in the HIV risk-reduction intervention group had higher odds of reporting consistent condom use in the past three months compared with their peers in the general health intervention group (odds ratio, 1.3). They also were more likely to report having used a condom at last vaginal intercourse (1.4) and having talked with their steady partner about such use (1.5), and their frequency of condom use—measured on a five-point scale—was higher (1.4). In addition, the proportion of vaginal intercourse acts in which they had used a condom increased to a greater extent (mean difference, 0.1). Tests of potential interactions indicated that these associations were similar for behavior with steady partners compared with casual partners and for the 6-month follow-up compared with the 12-month follow-up. The control and risk- reduction groups were statistically indistinguishable with respect to the odds of reporting unprotected vaginal intercourse, heterosexual anal intercourse and vaginal intercourse with multiple partners.

Study limitations included reliance on self-reported behavior and a possible lack of generalizability to all South African men, according to the investigators. Nonetheless, they contend, the study is noteworthy for being the first large trial to find significant improvements in sexual risk behaviors associated with an HIV intervention among South African men, and the findings demonstrate that targeting men can be an effective strategy in this setting. "Additional research might strengthen the impact of the intervention on multiple partnerships and address the generalizability of the present findings to biological outcomes," the investigators conclude.—S. London


1. Jemmott JB et al., Cluster-randomized controlled trial of an HIV/sexually transmitted infection risk-reduction intervention for South African men, American Journal of Public Health, 2014, 104(3):467–473.