Methods for community-engaged data collection and analysis in implementation research.

Background: Community engagement is widely recognized as critical to successful and equitable implementation of evidence-based practices, programs, and policies. However, there are no clear guidelines for community involvement in data collection and analysis in implementation research.

Methods: We describe three specific methods for engaging community members in data collection and analysis: concept mapping, rapid ethnographic assessment, and Photovoice. Common elements are identified from a case study of each method: 1) selection and adaptation of evidence-based strategies for improving adolescent HPV vaccine initiation rates in disadvantaged communities, 2) strategies for implementing medication for opioid use disorders among low-income Medicaid enrollees during natural disasters, and 3) interventions to improve the physical health of adults with severe mental illness living in supportive housing.

Results: In all three cases, community members assisted in participant recruitment, provided data, and validated preliminary findings created by researchers. In the Photovoice case study, community members participated in both data collection and analysis, while in the concept mapping, community members also participated in the initial phase of organizing and prioritizing evidence-based strategies during the data analysis.

Conclusions: Community involvement in implementation research data collection and analysis contributes to greater engagement and empowerment of community members and validation of study findings. Use of methods that exhibit both scientific rigor and community relevance of implementation research also contributes to greater community investment in successful implementation outcomes. Nevertheless, the case studies point to the importance and efficiency of the division of labor embedded in community-engaged implementation research. Building capacity for community members to assume greater roles in obtaining and organizing data for preliminary analysis prior to interpretation is recommended.