A modeling framework to inform PrEP initiation and retention scale-up in the context of Getting to Zero Initiatives.

OBJECTIVE(S): "Getting to Zero" (GTZ) initiatives aim to eliminate new HIV infections over a projected time frame. Increased PrEP uptake among populations with the highest HIV incidence, such as young black men who have sex with men (YBMSM), is necessary to accomplish this aim. Agent-based network models (ABNMs) can help guide policymakers on strategies to increase PrEP uptake.

DESIGN: Effective PrEP implementation requires a model that incorporates the dynamics of interventions and dynamic feedbacks across multiple levels including virus, host, behavior, networks and population. ABNMs are a powerful tool to incorporate these processes.

METHODS: An ABNM, designed for and parameterized using data for YBMSM in Illinois, was used to compare the impact of PrEP initiation and retention interventions on HIV incidence after 10 years, consistent with GTZ timelines. Initiation interventions selected individuals in serodiscordant partnerships, or in critical sexual network positions, and compared to a controlled setting where PrEP initiators were randomly selected. Retention interventions increased the mean duration of PrEP use. A combination intervention modeled concurrent increases in PrEP initiation and retention.

RESULTS: Selecting HIV negative individuals for PrEP initiation in serodiscordant partnerships resulted in the largest HIV incidence declines, relative to other interventions. For a given PrEP uptake level, distributing effort between increasing PrEP initiation and retention in combination was approximately as effective as increasing only one exclusively.

CONCLUSIONS: Simulation results indicate that expanded PrEP interventions alone may not accomplish GTZ goals within a decade, and integrated scale-up of PrEP, ART and other interventions might be necessary.