Horizons Report | December 2005 | A Targeted Intervention Falls Short

Population movement has long been identified as a factor in the spread of HIV throughout southern Africa. Today, migrant workers remain a key population to target for HIV prevention efforts not only because of their role in its transmission, but also because of their own vulnerability to the disease as a result of high-risk sexual behavior.¹ In South Africa, mine workers comprise a large number of the migrant workers who fall into this category.

Carletonville, a community in South Africa built on the gold mining economy, depends heavily on migrant work. In 2002, Carletonville and the neighboring township of Khutsong had more than 70,000 men working in 12 gold mining shafts, many from other rural areas of South Africa and neighboring countries. At the time, most mine workers lived in single-sex hostels near the mine compounds and visited their homes only a few times a year. “Hotspots,” establishments where beer and sex are sold, surrounded the compounds. The existence of such hotspots, coupled with long absences from home, has been linked to risky sexual behavior among mine workers.²

To prevent the spread of HIV within this migrant community, the Horizons Program, in collaboration with the Center for Scientific and Industrial Research, the South African Institute for Medical Research, and the London School of Economics and Political Science, conducted an intervention study in the Carletonville area. A key objective of the study was to assess the impact of an HIV and STI prevention program targeted to mine workers and sex workers on the larger community.

In 1998, the study partners launched the Mothusimpilo (“working together for health”) Intervention Project (MIP) to reduce community prevalence of HIV and other STIs and to sustain those reductions through enhanced prevention programs and STI treatment services. The interventions focused on two populations—female sex workers and male mine workers. It was assumed that focusing on the sexual contacts between sex workers and mine workers, who also have sex with others in the community, would have an impact on the larger community. MIP involved three components: HIV/STI peer education, condom promotion and distribution, and improved management of STIs.

For the peer education program, sex workers were recruited and trained in community work, hygiene, HIV/AIDS, and signs and symptoms of STIs. They provided information to other sex workers, and to a limited extent, the community at large. A similar program was designed for mine workers, who were perceived to be the primary clients of sex workers. Peer educators were also trained in the promotion and distribution of condoms for HIV and STI prevention.

STI management in public and private health facilities was strengthened by training local providers in the syndromic approach and reinforced through implementation of a periodic presumptive treatment (PPT) program for sex workers. Those who enrolled in the program received one gram of azithromycin and two grams of metronidazole once per month to treat most curable STIs present in the community. In addition, participants in the program were taught about preventive strategies, including condom use. The administration of PPT to sex workers was facilitated by the use of two mobile clinics in and around the hotspots.

To study the effects of the intervention, researchers conducted two surveys in 1998 and 2001—randomly sampling mine workers, sex workers, and men and women in the larger community—and a qualitative study investigating sexual networks among sex workers in 2001. Blood and urine samples were collected from the survey respondents to assess the prevalence of syphilis, chlamydia, gonorrhea, and HIV infection. Participants found to be positive for STIs were referred to the health system for treatment, whereas HIV testing was unlinked and anonymous.

Knowledge of HIV transmission on the whole was high among all groups at baseline and increased further by 2001. But overall, the findings on behavior change were uneven among the study populations. For example, researchers found that from 1998 to 2001 the percentage of mine workers who reported one or more casual partners in the 12 months prior to each survey decreased (53 to 43 percent). However, among men in the community who reported having casual partners, there was an increase (37 to 45 percent), and among women in the community, there was a significant increase (24 to 41 percent).

Reported consistent condom use with casual partners also showed mixed results. Regular use of condoms increased from 1998 to 2001 among mine workers (19 to 24 percent) and men in the community (28 to 37 percent), and increased slightly among women in the community (22 to 25 percent). But, the proportion of sex workers who reported consistent condom use with casual partners remained unchanged (59 percent).

The results were less encouraging for condom use with regular partners. Very few mine workers reported always using a condom with their regular partner and the percentage of sex workers using condoms consistently with regular partners decreased from 26 to 12 percent. Consistent condom use with regular partners among men and women in the community remained low.

Despite the administration of PPT for curable STIs among sex workers during the intervention period and training of providers for improved STI services, STI prevalence among sex workers, mine workers, and men in the community remained constant or increased slightly. Significant increases were noted among women in the community. Although very high at baseline, all groups, except for men in the community, experienced an increase in HIV prevalence by 2001. For mine workers and women in the community, the increase was statistically significant (see Table 1).

Shortcomings of the Intervention

The intervention had a limited impact for a variety of reasons. Baseline findings indicated that the prevalence of HIV infection was already high in all study groups. This suggests that the HIV epidemic was already generalized and therefore targeting only mine workers and sex workers was not enough to reduce the levels of STIs and HIV in the larger community.

A group of miners gather in front of their two story hostel.

One of the male hostels in the mining town of Carletonville.

Photo credit: Johannes van Dam

The sexual network analysis, conducted in 2001, revealed that sexual networks in Carletonville linking mine workers, sex workers, and men and women in the community were more complex than originally thought. Mine workers had casual sexual relationships not only with sex workers but also with local women living outside of the targeted hotspots. The women living in these other areas did not self-identify as sex workers and were overlooked in the intervention design, despite the fact that many engaged in transactional sex.

Eight months after the start of PPT, more than 900 sex workers had enrolled to receive treatment and monitoring. While rates of STIs decreased significantly among these women, no significant reduction was noted among the sex workers who participated in the cross-sectional surveys. Several reasons likely explain the lack of effect of PPT at the population level: insufficient program coverage of sex workers and other women having transactional sex, lack of STI treatment for the regular partners of sex workers, inconsistent condom use, and a high risk of re-infection.

The peer education program targeted to mine workers also experienced difficulties. Only a handful of mine workers were trained by MIP as peer educators and these few received little ongoing support and follow-up training. They also did not receive paid time off from the mining companies to educate their peers, which resulted in the educational sessions being taught inconsistently. In addition, trade unions did not actively support the MIP program beyond identifying the initial participants to be trained as peer educators.

On the other hand, the peer education program for sex workers was very strong and well-supervised. Since the program began, more than 200 peer educators have been trained, reaching approximately 2,000 sex workers. Despite this achievement, sexual behavior among sex workers changed little. High levels of unemployment; limited job opportunities for women in the community; and difficulty in negotiating condom use due to fears of abandonment, loss of economic support, and physical violence, contributed to the lack of change in sexual behavior among sex workers.

What We Learned

Although the intervention had mixed results, this project nonetheless provides valuable lessons for the implementation of other large-scale HIV/STI interventions, particularly in settings with high HIV and STI prevalence rates.

First, it is important to analyze the sexual networks and the pattern of HIV and STIs that exist in the community prior to the introduction of interventions, in order to understand exactly who to target. The existence of several different sub-populations that play key roles in HIV/STI transmission in the community requires that the intervention address each group’s particular needs and constraints.

“Overall, appropriate diagnostic and formative work, and a clear understanding of sexual networks, are extremely important for the tailoring and thus success of any intervention,” said Dr. Johannes van Dam of Horizons/Population Council, one of the study’s principal investigators.

Second, to have an appreciable impact on STI prevalence in the community, PPT interventions must have sufficient coverage. PPT can be an effective intervention to reduce STI rates among sex workers and other high-risk individuals, but there may not be a noticeable effect within the general population if enough individuals and their partners do not participate in the program.

Lastly, changing sexual behavior is difficult without an enabling environment. In this intervention, peer education was implemented for sex workers, but it failed to achieve universal condom use among them because little was done to help support women’s negotiation of condom use with their male partners. The absence of a sustained peer education program for men in the community and in the mines meant that they were receiving limited information and motivation for behavior change. Therefore, peer education programs for all populations involved in sexual networks need to be comprehensive and coordinated to help create a supportive environment for adopting preventive behaviors.

References ¹International Organization for Migration [IOM] and UNAIDS. 2003. “Mobile populations and HIV/AIDS in the Southern African region: Desk review and bibliography on HIV/AIDS and mobile populations,” IOM-UNAIDS Reports on Mobile Populations and HIV/AIDS Series 1.

²Campbell, C. and Y. Mzaduime. 2002. “How can HIV be prevented in South Africa? A social perspective,” British Medical Journal 324(26): 229–232.