Inside the black box: Process evaluation of a peer group intervention for HIV prevention in Malawi

BACKGROUND: Few studies have looked inside the black box to examine the processes through which peer group interventions change HIV risk behaviors. The purpose of this process evaluation was to identify changes in group member engagement, peer leader (PL) facilitation skills, and other aspects of group process across 6 sessions of an HIV-prevention intervention for adults and health workers in Malawi. METHODS: Trained observers used our collaboratively developed observation guide to rate standardized descriptors of engagement and facilitation behaviors, critical elements of the peer group approach. To capture change within each 2 hour session and across sessions, we observed a convenience sample of 235 sessions (49 groups, mean 10.39 members per group, observed 2 or more sessions each), in 30-minute blocks within each session and across 6 sessions. RESULTS: Engagement increased significantly (p<.05) from the initial 30-minute block to subsequent blocks within each session. Engagement (p<.05) and facilitation skills (p<.05) increased significantly from session 1 to session 6. Groups were increasingly more like a peer group than classroom from session 1 to 6. Group engagement and PL facilitation were moderately correlated (.409). Regression analysis examined and compared adults' and health workers' group process: Engagement and PL facilitation together explained 65.3% of the variance in “more like a peer group than classroom” for adults but only 16.1% for health workers. CONCLUSIONS: Process evaluation builds the evidence base for peer group interventions by documenting the effectiveness of PLs and the group process that occurs during the intervention and by validating intervention fidelity.