Improving HIV pre-exposure prophylaxis persistence among adolescent girls and young women: Insights from a mixed-methods evaluation of community, hybrid, and facility service delivery models in Namibia

INTRODUCTION: Despite the potential for community-based approaches to increase access to pre-exposure prophylaxis (PrEP) for adolescent girls and young women (AGYW), there is limited evidence of whether and how they improve PrEP persistence. We compared PrEP persistence among AGYW receiving services through community and hybrid models in Namibia to facility-based services. We subsequently identify potential mechanisms to explain how and why community and hybrid models achieved (or not) improved persistence to inform further service delivery innovation. METHODS: Data were collected from PrEP service delivery to AGYW over two-years in Namibia's Khomas Region. We used Kaplan-Meier analysis to estimate survival curves for PrEP persistence beyond three-months after initiation and report the cumulative probability of persistence at one- and three-months. Persistence was defined as any PrEP use within three months after initiation followed by a PrEP refill or previously prescribed supply of at least 30 days at the three-month visit. Interviews were conducted with 28 AGYW and 19 providers and analyzed using a deductive-inductive thematic approach. RESULTS: From October 2017 through September 2019, 372 (18.7%) AGYW received services through a facility model, 302 (15.1%) through a community model, and 1,320 (66.2%) through a hybrid model. PrEP persistence at one- and three-months was 41.2% and 34.9% in the community model and 6.2% and 4.8% in the hybrid model compared to 36.8% and 26.7% in the facility model. Within the community and hybrid models, we identified three potential mechanisms related to PrEP persistence. Individualized service delivery offered convenience and simplicity which enabled AGYW to overcome barriers to obtaining refills but did not work as well for highly mobile AGYW. Consistent interactions and shared experiences fostered social connectedness with providers and with peers, building social networks and support systems for PrEP use. PrEP and HIV-related stigma, however, was widely experienced outside of these networks. Community-to-facility referral for PrEP refill triggered apprehension towards unfamiliar PrEP services and providers in AGYW, which discouraged persistence. CONCLUSION: Service delivery approaches that offer convenience and simplicity and foster social connectedness may reduce access barriers and increase social support enabling AGYW to self-manage their PrEP use and achieve improved PrEP persistence.