This implementation study is testing whether community members can use a 3-step model (prepare, roll-out and sustain) to implement an efficacious peer group HIV prevention intervention with fidelity and effectiveness in a single district in Southern Malawi. A collaborative partnership including university researchers and district government, health system and traditional leaders will support community volunteers in implementation. Both the implementation process and the effectiveness of the intervention when implemented by communities will be evaluated.

In response to the National Institutes of Health's call for empirically-supported models to implement research-tested health behavior change interventions, the investigators are testing a 3-step (prepare, roll-out, and sustain) Community Implementation Model (Model) in low-resource African communities. The Model is adapted from a theory-based model that was effective in implementing a hospital-based intervention in South Africa. Its participatory approach, clear steps, observable benchmarks and multimedia Toolkit will facilitate community ownership and build capacity to use the Model with fidelity. The investigators test whether community members can use the Model to effectively implement a peer group HIV prevention intervention called Mzake ndi Mzake (Friend to Friend) with fidelity and effectiveness. Developed and tested among several populations in Malawi (R01-NR08058), This intervention was efficacious in rural, central Malawi. Intervention communities had significantly better outcomes, including higher condom use and HIV testing for both adults and adolescents, than control communities. The investigators evaluate the Model's implementation success in Phalombe District, in southern Malawi, where the HIV prevalence of 14.5% is twice that of other regions. Our implementation partnership builds on the strengths and contributions of four sectors. Traditional community leaders mobilize their communities' assets, including volunteers to implement the peer group intervention. The District health system has formal authority in health organizations, contributes health knowledge, and provides local District health data. The District political administration supports this effort by negotiating donated space for an HIV/AIDS Resource Centre and by pledging funds to sustain and expand implementation of the intervention if communities demonstrate that they can use the Model. Researchers from both US and Malawi universities contribute the Model and the Mzake ndi Mzake peer group intervention, provide supportive technical assistance for program implementation and conduct evaluations. Specific aims and their evaluation are: Aim 1) Prepare and support 3 communities in using the Model to implement Mzake, evaluated with community self-rated benchmarks. Aim 2) Identify implementation patterns across sites and over time, evaluated using mixed-methods to integrate benchmarks with observations, focus groups, and interview. Aim 3) Assess communities' effectiveness in using the Model to implement Mzake, evaluated with a stepped wedge design (N = 864). Analysis uses multi-level hierarchical models to detect improvements over time in HIV-related behavioral outcomes and a symptoms interview with a nurse for sexually transmitted infections (STIs) in participants who have received the intervention compared to the delayed-intervention control group. (Our original plan to use rapid STI tests appropriate for community use had to be modified because of new data indicating unacceptable reliability and withdrawal of one test from the market). Aim 4) Evaluate whether the Model is feasible, acceptable, effective and sustainable when used by communities to implement Mzake, examined by integrating data from Aims 1-3. If successful, this study will advance implementation science by providing a replicable evidence-based model for implementation of HIV prevention interventions and other health interventions by low-resource communities.

Inclusion Criteria:

• reside in specified community,
• parental permission if under age 18

Exclusion Criteria:

• not in designated community

De-identified data will be made available to other researchers for secondary analyses after the primary outcome publications have been accepted for publication, approximately 3 years after the grant ends.