HIV and AIDS prevention programmes often have little impact because such programmes have traditionally had a biomedical focus and an emphasis on individual behaviour. Yet social and community level factors influence the rate and method of HIV transmission and can affect the success or failure of prevention programmes.

In this paper we discuss the importance of these factors with reference to a South African prevention programme. We argue that HIV prevention efforts are affected by the participation of a wide range of groups and processes at local, national, and international level. The scope of research, interventions, and policies must be widened to take this into account.

Among marginalised groups in poor countries, providing information about health risks changes the behaviour of, at most, one in four people—generally those who are more affluent and better educated.¹ This is because health related behaviours (such as condom use) are determined not only by individual choice but also by the extent to which social conditions enable such behaviours. Health promoters need to develop policies and interventions that promote “health enabling” social conditions (conditions that enable and support health enhancing behaviour).^2,3

Concepts such as community participation and stakeholder partnerships are increasingly included as articles of faith in HIV prevention policies and interventions worldwide. Community led peer education and participation by local stakeholders are widely used strategies.⁴ Our understanding of the processes through which participation might bring health benefits, however, is still limited.⁵

We explored community participation through our involvement in the evaluation and implementation of the Carletonville project. This project is a community led HIV intervention programme in a South African gold mining community of about 200000 people (including 70000 migrant miners). The project has three arms: the control of sexually transmitted infections; community led peer education and condom distribution; and local multistakeholder management.⁶ The project was initiated by a grassroots township group and some academics; lengthy negotiation brought together local stakeholders—including the township group, representatives of mine management, the trade unions, provincial and national government, and several researchers. This group set up a non-governmental organisation to run the project, which began in 1997 and employs three full time workers.

We investigated the project's peer education component from mid-1997 to mid-2000.⁷ We did this through a longitudinal qualitative study of sexuality and community led HIV prevention among miners, commercial sex workers, and young people; we also examined the effectiveness of the project's local stakeholder management process.^8,9

Baseline research showed that despite high levels of knowledge about the dangers of HIV, people still engaged in high risk sex. One key social obstacle to safe sexual behaviour was how men and women defined their sexuality in the context of poverty, unemployment, and violence (social construction of sexuality). Miners were competing for a decreasing number of jobs. They coped with the stresses and dangers of underground work by assuming a macho masculine identity. Such an identity militated against using condoms as it was associated with the notions of (a) an uncontrollable sexual desire driving men to multiple sexual partners; (b) a lack of caution in high risk situations; (c) a biological need for the pleasures of “flesh to flesh” sex; and (d) the desire to father many children.¹⁰

Impoverished women flocked to the mines to sell sex from chaotic and violent shack settlements. Poverty made it difficult for the women to refuse sex if a client refused to use a condom—a situation aggravated by the women's lack of confidence in a male dominated community that accorded them little respect. Competition for clients undermined the development of a unified response to men who refused to use condoms. Women had often experienced physical and emotional abuse, and they lacked the economic or psychological resources to take control. A resulting sense of fatalism often reduced their motivation to protect their sexual health.¹¹

Young women who carried condoms were considered to have “loose morals.” Young men who used condoms were often subject to ridicule from peers. A degree of physical or emotional pressure on women was sometimes considered acceptable in sexual encounters, particularly in early sexual encounters, where foreplay was often minimal, and many women described their first encounter as painful and unpleasant.¹²

These findings highlight how simply giving people information about HIV and AIDS and telling them to use condoms is unlikely to have optimal impact unless efforts are made at the same time to promote local conditions that support condom use. The Carletonville project sought to pool the efforts of local stakeholders through two forms of community participation—community led peer education and condom distribution, and multistakeholder project management—to try to create health enabling conditions in the community.

Traditionally health education information is provided by outside experts to a passive target group to promote behaviour change in individuals. In community led peer education, however, local people are trained to disseminate information through debate and discussion. Assuming that sexuality is socially constructed, rather than being the product of individual decisions,¹³ peer education enables people to make collective decisions to change their behaviour.

Peer education seeks to empower lay people through placing health related knowledge in their hands. This increases the likelihood that people will feel they have some control over their health.¹⁴ Peer education succeeds to the extent that it promotes two forms of social capital (“bonding” and “bridging”).¹⁵ Bonding social capital refers to trusting and cooperative relationships in homogeneous peer groups, which we believe are necessary for collective debate and empowerment—for example, in groups of sex workers. Bridging social capital refers to collaboration among diverse groups of stakeholders who might not otherwise have had contact and whose collaboration increases the likelihood of the programme being successful. An example of this is an alliance between the women who sell sex to miners and the representatives of the powerful mining houses that employ sex workers' clients. These two groups both want mine workers to use condoms, albeit for different reasons.

Our research has highlighted the complexities of implementing community participation in real life settings. In collaborative project management, the stakeholders often varied in the quality and quantity of their commitment, as well as in their understanding of the causes and management of HIV. Thus, for example, the mine managers' primary commitment was to productivity, whereas trade unions were committed to individual human rights.⁶ Project scientists were motivated by the need to use rigorous survey procedures; such procedures were not always liked by township people, who argued that with high unemployment it was inappropriate for scientists to control the selection of research staff. Such projects become, like ours did, the focus of personal and political conflicts and controversies, unrelated to HIV issues, absorbing time and energy and sometimes undermining collaborative efforts. The stakeholder committee was often dogged by lack of trust between different parties.

The peer education programmes threw up a range of challenges. Although the project made great strides in building committed teams of peer educators in marginalised communities of sex workers, women led programmes were resented by men in the shack settlements. These men were unfamiliar with the concept of female leadership and sought to discredit the programme. Rather than seeking to understand the pressures that clients placed on their colleagues to have unprotected sex, some sex worker peer educators advocated ostracising or using violence against those caught not using condoms; these methods are contrary to the spirit of peer education, which emphasises mutual tolerance and support.¹⁶

Some school based peer education programmes, rather than creating awareness of the way in which traditional gender roles put young people at risk of HIV infection, reproduced those same gender roles at peer group meetings. The meetings were often dominated by young men, with women reluctant to assert themselves. Authoritarian and didactic school settings were not the ideal place for young people to debate their sexuality critically and openly. Some teachers insisted that the peer educators should promote abstinence, a strategy that few young people were likely to take seriously.¹⁷

A major constraint on the success of the project's peer education initiative was the over-representation of biomedical experts on the stakeholder committee. The primary commitment of these experts was to implement the part of the project aimed at controlling sexually transmitted infections by biomedical means; they made few attempts to address the community and social conditions that lead to disease transmission and hinder prevention efforts. Although, in the original proposal, peer education was to have had the same importance as infection control, the key stakeholders—particularly the biomedical researchers and mining house representatives—dismissed peer education as “vague social science.”⁸

Such attitudes, combined with a lack of commitment by the miners' trade union to the project management, resulted in peer education activities being conducted in a patchy way. Thus, for example, although there was much success in building peer education teams among sex workers in informal shack settlements (bonding social capital), implementing peer education in the formal setting of the mines was met with obstacles.⁹ In our three year research period, most Carletonville miners were not exposed to peer education. Yet it was male miners who held economic and psychological power in encounters with female sex workers. This is an example of one of the ways in which the project failed to build the types of bridging social capital that are essential for peer education to succeed.

Local HIV prevention programmes are unlikely to flourish in a regional or national vacuum,^18,19 particularly in communities with highly mobile groups such as migrant miners. South Africa has had a lack of government leadership, with the president's well publicised doubts about the links between HIV and AIDS causing disunity in local HIV/AIDS circles.

The country's national AIDS plan, formulated in 1994 by representatives of government, industry, and civil society has not been fully implemented. This has left open ground for international development agencies and externally funded non-governmental organisations, and the country is now besieged with overseas experts, often with access to sizeable sums of foreign funding. HIV prevention strategies are informed by the assumptions of Western science and policy, with insufficient assessment of whether these are appropriate for local conditions. Proposals for projects funded by overseas bodies may be written by external consultants and presented to local groups for implementation. Local people may therefore have little sense of “ownership” of the proposals and may lack the conceptual understanding, technical skills, or training staff to implement them properly.

HIV prevention cannot be regarded simply as a biomedical or behavioural issue,²⁰ located “out there in the townships” or “out there in Africa.” Frameworks are needed that treat HIV transmission and prevention as social issues. Despite the proliferation of community based approaches to HIV prevention, few studies have examined the role of community and social factors in a programme's success. Our work in Carletonville points to the urgent need for such research.

We thank Brian Williams (principal investigator, Carletonville Project) and Catherine MacPhail for their contribution to the research.