Community mobilisation is a term akin to community development, but it is used here to distinguish it from community development methods used in other health areas because it has a specific history in the HIV field.

Community mobilisation has been studied in several well–controlled trials in the United States. The most famous example is the Kelly study of opinion leaders; a randomised, controlled trial conducted amongst gay men in paired towns in four US states. The study tested the argument of social diffusion theorists, that if 15–20% of a population adopts an innovation, then it will be conveyed through natural social networking to cause community wide change. Many historians of safer sex have argued that this is how safer sex came to be adopted by gay men during the mid–1980s.

The study utilised `opinion leaders' who were identified as the people most likely to influence their peers. Opinion leaders were identified by asking bartenders to nominate people they considered to be most liked and trusted by patrons. The opinion leaders selected were then invited to attend group sessions at which they were trained in ways of delivering HIV risk reduction messages to their peers. These methods included endorsing the benefits and timeliness of risk behaviour change, recommending strategies for implementing change and correcting risk misperceptions amongst their friends. The men were then invited to initiate twenty conversations amongst their friends over the following weeks.

The study was carried out in four states, with two cities in each state compared, one serving as a control and one as an intervention site. Three months and again six months later, surveys of all men entering bars in the intervention cities and control cities were carried out. The response rate was 85%, and unprotected anal intercourse in the intervention cities declined from 33% at baseline to under 25% at nine months, a reduction of 28%. No change was found in the control group.

The reported incidence of unprotected insertive anal sex fell from 27% to 18% over the same time period in the intervention cities, but the incidence of unprotected receptive anal sex fell more slowly. There was no change in the control cities.

When researchers analysed the results to determine the factors accounting for the changes in behaviour they had observed, they found that men who spent the most time in bars had the highest levels of risk behaviour at baseline. But they also showed the greatest reduction in such behaviour following intervention. Since most peer conversations probably took place in bars, the researchers hypothesised that there was a dose response effect. This was corroborated by men's own reports: those who reported being engaged in the greatest number of risk reduction conversations also reported the greatest reduction in risk behaviour.

Another US project, similar to Gay Men Fighting AIDS and a number of other UK projects in its design, was the empowerment project. This was tested by US National Institutes of Mental Health researchers in two West Coast cities, Santa Barbara, California, and Eugene, Oregon, in 1992.

Eugene received the intervention first whilst Santa Barbara served as control, and longitudinal cohorts of gay men aged 18 to 29 were recruited and assessed before and after by mail–back survey.

The project tested the hypothesis that involvement with the process of peer education would have an effect on risk behaviour. By allowing the community to take responsibility for behaviour change efforts, it was hoped to foster a deeper and longer lasting effect than one achieved by professionals.

A core group of 15 self–selected volunteers was recruited and given a brief to design the project, with advice from local academics, public health professionals and AIDS service organisations. The project developed several arms: peer outreach conducted by volunteers in bars, community events and events organised by the group itself. This outreach work also served to recruit other peer educators in a rolling programme of recruitment.

The group also organised a storefront drop–in centre and small group sessions focussing on safer sex information and skill development. The latter were attended by 170 young men in the town (estimated to represent at least 15% of the local gay population).

One year after the intervention a 26% reduction was reported in the number of individuals who reported any instance of unprotected anal intercourse. This included a 45% reduction in respondents reporting any unprotected anal intercourse with non–primary partners and a 24% reduction in those reporting an unprotected anal intercourse with primary partners. In contrast there was no change in these variables in the control city (Kegeles).

A study amongst injecting drugs users provides evidence that such methods need not be confined to gay men. The US National Institutes of Drug Abuse has studied the use of peer educators amongst injecting drug users, and concluded that a project employing volunteer drug users as peer educators and community development facilitators not only reduced the sharing of injecting equipment, but led to fall in HIV incidence in Chicago (Wiebel). However, this is the only controlled study amongst IDUs.

A problem with all of these examples of community mobilisation is the difficulty we have in judging the relationship contexts in which unprotected sex was taking place in these studies. Whilst the empowerment study did ask questions about unprotected sex with regular partners, the Kelly studies did not, and critics of the community mobilisation approach have argued that such interventions are only likely to have an effect on those having unprotected sex with casual partners. Those in regular partnerships may see themselves as immune from such messages and may need to be targeted specifically. However, these criticisms ignore the fact that all studies of community mobilisation have been studies of a methodology, not a message. In order to prove that community mobilisation is an inappropriate way to target men in relationships, it is necessary to design a study which looks at the diffusion of messages about unprotected sex in relationships. This has not yet happened.

Clearly these studies do not answer all the questions which might arise about the distribution of resources in a community mobilisation type project, but they do suggest that the level of contacts achieved need not be analogous to the total population of a district. However, the Kelly studies suggest that dosage and exposure were important elements in the success of the project.

Both studies support the need for knowledge of local sexual and social networks, and sensitivity to the customs and idiom of the local community. The selection of local informants is a crucial first stage in a community mobilisation project: if this goes wrong the project could veer off course.

There is also a danger that culturally specific problems could arise in replicating community mobilisation–type efforts in other countries. Evidence of this was supplied at a recent meeting of the British Psychological Society by Graham Hart, who reported on a study being conducted by the Medical Research Council's Medical Sociology Unit in Edinburgh and Glasgow, intended to replicate some elements of the Kelly and Kegeles studies. Hart reported that local informants were none too keen on the idea of encouraging popular local figures to discuss intimate details of people's sex lives, because their popularity tended to be proportional to their store of local gossip! Many people would be unwilling to discuss intimate sexual matters with such popular but indiscreet figures!

Nevertheless, such evidence does suggest that the general concept of community mobilisation has a valuable contribution to make to HIV prevention activities. However, design and evaluation of future projects should proceed with an awareness of the limitations of current evidence, and should seek to add to the body of research knowledge about community mobilisation strategies. For example, is it adequate to reach 15–20% of the population to diffuse a message, and how would you measure this? What is the best balance of activities within community mobilisation?

References

Kegeles S, et a. The Mpowerment project: a community level HIV prevention intervention for young agy and bisexual men, Am J Public Health, in press, 1997.

Kelly JA, et al. HIV risk behaviour reduction following intervention with key opinion leaders of the population: an experimental analysis, Am J Public Health 81: 168–171, 1991.

Kelly JA, et al. Community AIDS/HIV risk reduction: the effects of endorsements by popular people in three cities, Am J Public Health, 82: 1483–1489, 1992.