INTRODUCTIONThe number of HIV trials in Africa is increasing, and they target population groups with high HIV incidence, such as sex workers. Little information, however, is available about the adherence to long-term therapy among such marginalized groups with few economic resources and poor social and family support. A project called "Yerelon" ("know herself" in the Dioula language) began in 1998 in Bobo-Dioulasso to improve the health of women involved in commercial sex through STI/HIV prevention and care adapted to them. This study was conducted before introducing long-term treatment to the population, to assess the effect of communication with those around them on the capacity of these vulnerable women to adhere to drug prescriptions.METHODSThe study was based on interviews conducted during the pilot phase of a 3-month trial of vitamins with potential participants. It concerned two groups of women: one group was infected with HIV (N = 22), the other was not (N = 20); all women in both groups were infected by HSV-2, however. For 5 weeks, the two psychologists of the study team in charge of adherence assessment carried out weekly in-depth interviews with the participants. The qualitative data analysis was organised around several themes. The data were related to aspects of communication with family and friends, serologic results, and adherence.RESULTSAccording to our definition of communication about treatment, 20 participants communicated with their family and friends; adherence was good for all but three of them. Women who reported that they were married or living with someone (7/42) nearly all spoke about the study treatment (06/07) with him. Of 16 participants living in a family, 10 communicated with them about the treatment. On the other hand, as seems logical, single women who lived alone spoke less often about the treatment with family and friends (04/19). Talking about the treatment did not appear to involve the family or friends in the treatment; no one reminded any participant, whether she lived alone or in a family, to take her medicine. Nor did this discussion seem "helpful" to any of the women. Twenty-two participants hid the study treatment from family and friends; adherence was good for all but two of these. Social management of the treatment was related to HIV serologic status and relationships with family and friends. Concern about gossip about HIV status made it difficult to integrate the treatment into conversation. Those who did not agree to communicate with their family about the treatment did not even take the drug in the sight of the others. Sometimes, refusal to communicate was aimed at avoiding disapproval when the family did not have a favorable perception of prolonged treatment. Hiding the treatment was not an obstacle to good adherence. Adherence was related to perception of the treatment more than to communication about it.CONCLUSIONAdherence was similar in cases with and without communication. It appeared that these marginalized women, without social networks, were able to adhere correctly to a long-term treatment. To minimize the risks of non-adherence, the support system planned must take into account the factors influencing perceptions of the drug. Specific psychological support centered on the relation with the drug appears necessary during treatment initiation and follow up.