Steep HIV incidence declines among female sex workers in Cotonou, Benin: reconstructing finely stratified HIV incidence estimates from HIV prevalence 1993-2022

Background: HIV prevalence among female sex workers in Benin has declined steeply over the last thirty years, driven by successful HIV prevention and treatment programming among sex workers and their clients. To maintain progress in an era of constrained funding, HIV prevention programming should be targeted towards those most at risk of HIV seroconversion. We present a model-based approach that produces finely stratified HIV incidence estimates from widely available HIV prevalence data from serial cross-sectional surveys among female sex workers. Methods: We analysed participant-level data from twelve cross-sectional surveys among FSW in Cotonou, Benin from 1993-2022. We created a compartmental model representing women transitioning into and out of sex work and acquiring HIV. The model was stratified by single year of age, duration at-risk, calendar year, and time since HIV seroconversion and calibrated to HIV prevalence by single-year age and duration at-risk, and age and duration distribution data. Results: HIV prevalence among sex workers aged 15-49 in Cotonou declined from 44.9% (95%CI 41.2-48.9) in 1995 to 8.6% (95%CI 6.4-11.8%) in 2022 (Figure 1). HIV incidence declined by 86% (95%CI 82-90%) between 1994 and 2013 from 19.7/100py (95%CI 9.5-26.9) to 2.5/100py (95%CI 1.2-3.5), and remained stable thereafter through 2022 (Figure 2A). Throughout the study period, new infections were concentrated among women who recently started selling sex: in 2022, 72% of infections occurred in the first year of sex work (95%CI 64-81%; Figure 2B). Incidence patterns by age varied less than by duration and was similar at around 1.5/100py for all sex workers under age 30, rising to 4/100py among those over age 40. Modelled incidence exceeded empirical cohort estimates, highlighting sensitivity to assumptions about HIV prevalence at sex work initiation, episodic sex work, duration misclassification, and population size. Conclusion: We estimated large declines in HIV incidence among sex workers in Benin from 1994-2013, but slower progress from 2013-2022. New sex workers across all age groups, who may be weakly linked to programmes, should be the highest priority for HIV prevention interventions and community outreach efforts. This modelling approach to empirically monitor incidence among sex workers should be applied across high burden epidemic settings with serial survey data to guide prevention efforts, including the allocation of limited provision of long-acting pre-exposure prophylaxis.