The effect of six-month PrEP dispensing supported with interim HIV self-testing on PrEP continuation at 12 months in Kenya: a randomized implementation trial


BACKGROUND: In Kenya, HIV pre-exposure prophylaxis (PrEP) is primarily delivered at HIV clinics where client barriers to continuation include privacy concerns, long wait times, and setting-associated stigma. Six-month PrEP dispensing supported with interim HIV self-testing (HIVST) would reduce the number of clinic visits and potentially address some of these continuation barriers.
METHODS: We conducted a non-inferiority trial to test this model of PrEP delivery in Thika, Kenya. Eligible participants were PrEP clients '¥18 years who returned for their first one-month follow-up visit. We randomized participants 2:1 to: 1) six-month PrEP dispensing with semiannual clinic visits and interim HIVST at three months, or 2) standard-of-care (SOC) PrEP delivery with quarterly clinic visits and clinic-based HIV testing. Pre-specified outcomes at 12 months included HIV testing (any in past six months and '¥2 times since enrollment) and PrEP refilling (at 12 months and both six and 12 months). We used binomial regression models, adjusted for sex and serodifferent partnership status, to estimate risk differences (RDs) and interpreted one-sided 95% confidence interval (CI) lower bounds (LB) '¥-10% as non-inferior.
RESULTS: From May 2018 to February 2020, we enrolled and followed 495 participants. At 12 months, 73.3% (241/329) in the intervention and 72.3% (120/166) in the SOC arm returned to clinic. In the intervention arm, 69.9% (230/329) tested for HIV in the past six months and 72.3% (238/329) tested '¥2 times since enrollment, compared to 69.9% (116/166, RD -0.3%, 95% CI LB -7.4%) and 71.7% (119/166, RD 0.2%, 95% CI LB -6.8%) in the SOC arm, respectively. Additionally, 59.6% (196/329) in the intervention arm refilled PrEP at 12 months compared to 62.7% (104/166) in the SOC arm (RD -3.3%, 95% CI LB -10.8%, thus failing to demonstrate non-inferiority). However, 56.5% (186/329) in the intervention and 56.6% (94/166) in the SOC arm (RD -0.2%, 95% CI LB -7.9%) refilled PrEP at both at six and 12 months.
CONCLUSIONS: Six-month PrEP dispensing with interim HIVST resulted in high PrEP continuation at one year. HIV testing and PrEP refilling were generally comparable to SOC PrEP dispensing. This novel model has the potential to optimize PrEP delivery in Kenya and similar settings.