Designing evaluations for community based participatory research interventions using existing publicly available datasets

Local ‘hot spots' of poor health outcomes present a challenge for CBPR and developing evidence-based interventions for specific communities because information allowing population-based characterization of health disparities is often derived from larger national or regional data, and evidence about intervention effectiveness is often developed outside the local setting. As part of a community-university partnership using CBPR to design physical activity and nutrition interventions to decrease obesity in a designated neighborhood in San Francisco, CA, we sought to determine the utility of publicly available datasets to evaluate interventions targeting a specific neighborhood. A review of datasets was conducted using internet search engines, academic literature, and consultation with university and community partners, the Clinical and Translational Science Institute Community Engagement and Health Policy Program which works with local community based organizations, and the San Francisco Department of Public Health. Datasets that included the target neighborhood in its catchment area and measured proximal or distal variables related to obesity outcomes were included. Of 33 databases identified to date, only 12 (36%) had enough granularity to examine neighborhood, zip code, or school- specific levels of the data. Next steps include working with community partners to identify, evaluate and select databases and variables to design an evaluation. As databases become more widely available, participatory methods to identify, evaluate and quantify data sets' usefulness to a particular, geographically focused at-risk population, such as those presented here, will be in increasingly greater demand.