Assessing the differential impact of an hiv prevention intervention: who's putting the message into practice?

Recent data suggest that educational interventions aimed at reducing HIV risk behaviors have shown some success. Nonetheless, HIV risk behaviors are not always reduced by interventions and probably do not reduce risk behavior randomly. That is, the success of interventions may be related to participant characteristics. Identifying participant characteristics related to both intervention completion and reduction in risk behaviors may be useful for further developing explanatory models of health behavior and for targeting and customizing interventions. In this study differences between participants who completed an AIDS educational intervention (N = 741) and those who did not complete the intervention are first examined (N = 652) and then variables related to reducing drug and sexual risk behaviors among those who completed the intervention and follow-up interviews are examined. Results show that the majority of respondents report decreasing five out of six risk behaviors, with the smallest percentage (48.8%) decreasing rates of unprotected sex and the largest percentage (83.4%) decreasing frequency of drug injection. Different variables were found to be related to changes in the various risk behaviors. However, some relatively consistent results emerge. For all risk variables, the frequency of the specific behavior at baseline predicted the amount of change in that behavior, with those having higher levels of risk behaviors reducing their behavior the most. Positive HIV test results significantly decreased three of the four sexual risk behaviors examined, and living in a very rural area was found to be significantly related to three of the six risk behaviors. However, perceived chance of getting AIDS did not significantly reduce any of the risk behaviors. Gender and education level were also not related to changes in any of the risk behaviors. Implications include the importance of developing approaches to retain higher proportions of younger participants, males and homeless in interventions. It is particularly important to develop specific approaches to retain women in interventions. Because very rural participants were more likely to decrease crack use and alcohol or drug use with sex, rural interventions should target these behaviors at the outset of the intervention.