Implementation fidelity of a peer-led self-help group intervention for female sex workers in Zimbabwe: A qualitative process evaluation

Abstract Background Various factors exacerbate female sex workers’ (FSWs) HIV vulnerability. FSWs are also underserved by health services. Self-help groups (SHGs) can build FSW agency and support health-services linkage. We explored implementation fidelity of the Adapted Microplanning to Eliminate Transmission of HIV in Sex Transactions (AMETHIST) trial SHG component to better understand trial outcomes. Methods December 2020 to August 2021, we conducted 64 semi-structured interviews (SSIs) and 24 natural group discussions (NGDs) across three trial sites in two rounds of research activities. Participants (FSWs, microplanners (peer empowerment workers) and staff) were purposively sampled, with new individuals per round. Each round included 30 SSIs (n = 5 FSWs; n = 5 microplanners per site). We also conducted SSIs (n = 4) with staff, resulting in N = 64 SSIs altogether. We also conducted 24 NGDs over two rounds (n = 12 per round). Trained female social scientists led data collection and analysis, guided by a fidelity framework assessing peer-led governance, structured meetings, internal savings/lending ( mukando ), safe spaces and health linkages. Results Fidelity was dynamic and negotiated rather than rigid blueprint adherence. Sixty-five (31.3%) of the planned 208 SHGs were formed across 11 intervention sites. Thirty (46.2%) remained active after 28 months of establishment. High-fidelity components; ISAL/ mukando and safe spaces directly addressed FSWs' economic and psychosocial needs. Safe spaces enabled discussions of HIV status, antiretroviral therapy/pre-exposure prophylaxis adherence. In contrast, components requiring ongoing external facilitation (structured curriculum meetings and transition to peer-led governance) had low fidelity. A fifth of microplanners (20/104, 19.2%) were not fully trained on establishing and supporting SHGs. Some SHGs struggled to relinquish microplanner leadership after the required six months. COVID-19 undermined fidelity, with members prioritising survival needs over SHG participation. Conclusions The AMETHIST trial SHG model was partially implemented, particularly regarding coverage, structured meetings and transition to independent peer governance. This limits interpreting the trial as a definitive test of the full model. Future delivery should ensure full facilitator training, adequate supervision and communication resources, readiness-based transition to member leadership, and contingency plans preserving core functions during disruptions, while allowing flexible meeting formats and locally-selected activities. Trial registration The trial was registered in the Pan African Clinical Trials Registry on 2 July 2020 (PACTR202007818077777), after randomisation but before data collection.