Methodological challenges in the evaluation of community interventions

Abstract

The community is and will remain a common unit for intervention for several reasons: the impracticality of introducing the intervention at an individual level, avoiding contamination between intervention and control subjects or reducing costs. But the evaluation of community interventions poses methodological challenges concerning generalizability and validity. It is necessary to consider how the context may influence the measured outcomes: in the case of community interventions, how the characteristics of the communities themselves may affect the outcomes of the individuals living there. Although not necessary in the event that the communities are identical, this will be hard to achieve in practice. Adjustment for community characteristics will be as much a part of the analytical strategy as adjustment for individual characteristics and may provide important information regarding the possibility of replicating the intervention elsewhere. However, the importance of context must be considered at the time of the study design since adjustment for context can only be achieved if there are sufficient communities. The ability to prevent selection bias makes …