Longitudinal patterns and predictors of opioid and stimulant use initiation and cessation among female sex workers living with HIV in South Africa

Abstract

Female sex workers (FSW) are disproportionately impacted by HIV and substance use. Substance use has been linked to poor HIV treatment outcomes, necessitating exploration of substance use patterns — including polysubstance use — and predictors among FSW living with HIV. Data were obtained for 777 FSW living with HIV who were not virally suppressed and previously randomized to Siyaphambili, a trial of HIV treatment support strategies implemented through TB HIV Care in eThekwini, South Africa. FSW were asked about recent marijuana, opioid, stimulant, and hazardous alcohol use at enrollment and semi-annually for 18 months from June 2018-January 2022. We estimated incidence of substance use initiation/cessation post-enrollment and used Kaplan-Meier plots and lasagna plots to visualize trends. Cox proportional hazards models assessed baseline predictors of substance use initiation/cessation. Overall, 454 FSW (58.4 %) reported any opioid and/or stimulant use. Prior visit hazardous alcohol use (aHR: 0.20, 95 % CI 0.09–0.41) and prior-visit stimulant use (aHR: 2.80, 95 % CI 1.23–6.37) were negatively and positively associated with opioid initiation, respectively. Prior visit marijuana use (aHR: 1.75, 95 % CI 1.11–2.75) and opioid use (aHR: 5.31, 95 % CI 3.32–8.51) were positively associated with stimulant initiation. We found a high prevalence of substance use among FSW living with HIV that was dynamic over time, including a shared relationship between opioid and stimulant use that suggests intertwined substance use. Further investigation into the impact of polysubstance use patterns on success of HIV support strategies is needed to inform HIV treatment and harm reduction programming.

HIGHLIGHTS • Opioid and stimulant use are prevalent among female sex workers in South Africa. • Opioid use at prior study visit predicts stimulant initiation and vice versa. • Psychosocial factors (e.g. stigma) predict substance use changes over time.