Breaking down relationship barriers to increase PrEP uptake and adherence among adolescent girls and young women in Kenya: safety and preliminary effectiveness results from a pilot cluster‐randomized trial

Abstract Introduction Oral pre‐exposure prophylaxis (PrEP) has the potential to reduce HIV acquisition among adolescent girls and young women (AGYW) in sub‐Saharan Africa, a priority population for epidemic control. However, intimate partner violence (IPV) and low relationship power can create significant challenges to PrEP use. The Tu'Washindi intervention aimed to increase PrEP use by addressing relationship‐ and violence‐related barriers among AGYW enrolled in the DREAMS Initiative in Siaya County, Kenya. Methods Our multi‐level, community‐based intervention was piloted in a cluster‐randomized controlled trial conducted at six DREAMS sites from April to December 2019 (NCT03938818). Three intervention components were delivered over 6 months: an eight‐session empowerment‐based support club, community sensitization targeted towards male partners and a couples’ PrEP education event. Participants were ages 17–24, HIV negative and either eligible for, or already taking, PrEP. Over 6 months of follow‐up, we assessed IPV (months 3 and 6) and PrEP uptake and continuation (month 6) through interviewer‐administered questionnaires; PrEP adherence was assessed with Wisepill electronic monitoring devices. These outcomes were compared using adjusted Poisson and negative binomial regression models. Results We enrolled 103 AGYW with median age of 22 years (IQR 20–23); one‐third were currently taking PrEP and 45% reported IPV in the past 3 months. Retention was 97% at month 6. Compared to the control arm, intervention arm participants were more likely to initiate PrEP, if not already using it at enrolment (52% vs. 24%, aRR 2.28, 95% CI 1.19–4.38, p = 0.01), and those taking PrEP had more days with device openings (25% of days vs. 13%, aRR 1.94, 95% CI 1.16–3.25, p = 0.01). Twenty percent of participants reported IPV during follow‐up. There were trends towards fewer IPV events (aIRR 0.66, 95% CI 0.27–1.62, p = 0.37) and fewer events resulting in injury (aIRR 0.21, 95% CI 0.04–1.02, p = 0.05) in the intervention versus control arm. Conclusions Tu'Washindi shows promise in promoting PrEP uptake and adherence among AGYW without concomitant increases in IPV; however, adherence was still suboptimal. Further research is needed to determine whether these gains translate to increases in the proportion of AGYW with protective levels of PrEP adherence and to evaluate the potential for the intervention to reduce IPV risk.