Article Text
PDF
Abstract
Background The influence of turnover of female sex workers (FSW), resulting from different durations spent in sex work or migration patterns remains partly understood but is believed to be an important determinant of HIV/STI spread among FSW and overall populations and intervention impact.
Methods A deterministic transmission dynamics model, parameterised using serial cross-sectional data, was used to simulate the course of HIV and gonorrhoea (Gc) in FSW/clients and the general population of Cotonou. Different turnover patterns were produced by varying the time spent in sex work (SW duration=lifelong to 1 year) or, equivalently, time spent in the location by migrants (sojourn duration=lifelong, 5, 1 year). Each FSW leaving the local population was replaced by a new FSW with the same risk behaviour and with FSW HIV prevalence, for the SW duration scenarios, set to that of the low risk population and, for the migration scenarios, to 0%, 5% or 50%.
Results Shortening SW duration (or sojourn duration with initial 0% FSW HIV prevalence) from lifelong to 5- and 1-year delayed and reduced FSW HIV peak prevalence from ∼75% in 1999 to 56% in 2000 and ∼25% in 2020, respectively. The 1-year scenario increased long-term HIV prevalence of the overall female population by twofold, as the rapid turnover reduced the decline in Gc prevalence caused by AIDS differential mortality. With 1-year sojourn duration, local FSW HIV prevalence was mostly determined by the HIV prevalence of new FSW and FSW HIV incidence always exceeded prevalence (Abstract P1-S2.04 figure 1). Sudden increases in turnover of new HIV negative FSW (or sojourn duration=1 year & 0% HIV), in a maturing epidemic (1993), could reduce FSW HIV prevalence by >50% within 2 years but increase HIV incidence, in absence of intervention, compared to no turnover (ie, lifelong duration). The different SW and migration duration patterns did not significantly reduce the general population impact of a FSW targeted condom intervention per se, after controlling for epidemic stages, assuming similar exposure to the intervention by local and new FSW.
Cotonou.
Conclusions The impact of FSW turnover on HIV is complex but relatively modest if occurring at a slower rate than 0.20 per person-year (ie 1/5 years). Monitoring change in FSW turnover is important to interpret HIV prevalence trends over time, especially following an HIV intervention. Targeted FSW intervention remains effective in presence of rapid turn-over if new FSW can be reached rapidly.