Abstract
The field of HIV risk and prevention research has labored under the assumption that the randomized trial is the single best way to test the efficacy of behavioral interventions. Medical model conventions for trial design and analysis have been adopted whole cloth. However, the usual approaches used to analyze randomized trial data focus on identification of overall treatment effects in aggregate samples. This universal, mechanistic approach is contrary to the daily realities encountered by prevention practitioners and investigators, that different people change in different ways, that group characteristics influence participant outcomes, and that what people get out of an intervention depends upon individual, historical and cultural differences. This paper presents the case for re-examining data from randomized prevention studies, using methods to discern diversity in the ways that people respond to programs and to identify contextual factors that may influence outcomes. Specific methods include the examination of individual outcome trajectories, the identification of dynamic interactions among determinants of risk behavior, the analysis of group-level influences on individual outcomes, and the incorporation of historical data on availability of prevention services in analysis of treatment trials. A meta-analytic framework is proposed to synthesize findings generated by these and other exploratory approaches. HIV prevention science may be best served by methods optimized to discover, highlight and weigh the implications of diversity, rather than relying on approaches that emphasize central tendencies and simple, linear effects.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call