Advancing Sexual and Reproductive Health and Rights
International Perspectives on Sexual and Reproductive Health
Volume 39, Number 3, September 2013

Home-Based Services May Increase HIV Testing In Rural South Africa

Rural South Africans who lived in areas where counselors offered home-based HIV testing as part of an intervention were more likely to take an HIV test than were those who could obtain services only at clinics (prevalence ratio, 1.5), according to a cluster-randomized, controlled trial.1 Among individuals who underwent testing, those who had a home test were twice as likely as those tested at a facility to receive couples-based counseling and get tested together (prevalence ratio, 2.2). In addition, at the end of the study, residents of the intervention area were less likely than those of the control region to report having had multiple partners or a casual partner in the last three months (0.5 and 0.6, respectively).

Although the prevalence of HIV is very high in South Africa (17% in 2009), only a small proportion of residents have been tested for the virus, particularly in rural areas, where poverty and HIV stigma are common. To assess whether access to home-based HIV counseling and testing increases the prevalence of testing, researchers conducted an intervention in the uMzimkhulu subdistrict of KwaZulu-Natal province.

In 2008, researchers undertook a baseline survey of all of the subdistrict’s households and used the resulting information to demarcate 16 clusters, half of which were randomly assigned to receive the intervention and half to a control group. More than 85% of households in each group completed the survey. Residents were eligible for the survey (and subsequent intervention) if they were at least 14 years old; minors were required to obtain parental permission. The 4,710 respondents had an average age of 41; about two-thirds were women (many men in the area migrate for work), one-third had ever had an HIV test and roughly half had no more than a primary education. Infrastructure was somewhat better in the control areas than in the intervention areas (e.g., higher proportions of residents had piped water and electricity in their home).

In the intervention clusters, researchers trained women who were experienced in community outreach to become HIV counselors. From September 2009 to November 2010, these counselors visited every household in the intervention area, offered information about HIV and rapid HIV tests, and provided pre- and posttest counseling. Residents who tested positive received follow-up counseling and a referral to a nearby health facility. In control clusters, no counselors visited the households, and HIV services were available only at area health facilities and mobile units; however, as part of an unrelated national HIV awareness campaign, the South African government began promoting testing in this area (but not the intervention area) midway through the trial period.

To assess changes in the prevalence of HIV testing and to measure the prevalence of risky sexual behavior after the intervention, researchers conducted a follow-up survey from February to May 2011; again, all residents of the two areas were invited to participate. The 4,154 respondents were split equally between the control and intervention clusters. The proportion of residents who had ever been tested rose from 32% at baseline to 69% in the intervention clusters, and from 31% to 47% in the control clusters. Among those who underwent an HIV test between the baseline and follow-up surveys, 58% of intervention area residents and 1% of control area residents reported having been tested at their home; 85% of those who obtained home-based services rated the quality of counseling as good or very good.

In multivariate analyses that adjusted for clustering, the likelihood of having been tested for HIV during the trial period was higher among respondents of the intervention areas than among those in the control areas (prevalence ratios, 1.5); the findings were similar when the analytic sample was limited to only men or only women. Among those who underwent testing, residents of the intervention areas were more likely than those who lived in the control areas to have been tested and counseled with their partner (2.2), and less likely to have experienced intimate partner violence following disclosure of their HIV status (0.6). Moreover, respondents of the intervention areas were less likely than their counterparts elsewhere to report having had multiple or casual sexual partnerships in the last three months (0.5 and 0.6, respectively). No differences between groups were found for measures of HIV knowledge, condom use and experiences with HIV stigma.

Limitations of the study include the use of self-reported data from cross-sectional surveys, which could be subject to recall and social desirability bias, and its focus on a single small, rural subdistrict. However, the investigators point out that this type of study has never before been conducted in a rural location where HIV is prevalent, and that the findings may be relevant for the many other rural Sub-Saharan African communities where HIV stigma is high and male migration for work is common. Furthermore, by encouraging couples testing, the intervention could help reduce the likelihood of transmission between serodiscordant partners. Considering that the vast majority of intervention area respondents gave high ratings to the services they received at home, the researchers recommend that the South African government train community health workers in administering home-based tests and counseling to “achieve high uptake of HIV testing.” Moreover, the lower likelihood of certain risk behaviors in the present study’s intervention group suggests that wide-scale implementation of the program might reap “broader benefits beyond the actual testing.”—S. Ramashwar


1. Doherty T et al., Effect of home based HIV counseling and testing intervention in rural South Africa: cluster randomized trial, BMJ, 2013, doi: 10.1136/bmj.f3481, accessed Aug. 19, 2013.