Predictors of participant retention in a community-based HIV prevention cohort: perspectives from the HPTN 071 (PopART) study

Abstract Introduction In 2020, there were 37.7 million people living with HIV (PLHIV) globally, of which 20,6 million (55%) were living in Eastern and Southern Africa. Longitudinal studies, inclusive of community randomized trial (CRTs), provide critical evidence to guide a broad range of health care interventions, including HIV prevention. In this study, we have used an individual-level cohort study design to evaluate the association between sex and other baseline characteristics and participant retention in HPTN 071 (PopART) in Zambia and South Africa. Methods HPTN 071 (PopART) was a community randomized trial (CRT) conducted from 2013-2018, in 21 communities. The PopART primary outcome was measured in a randomly selected population cohort (PC), followed up over 3 to 4 years at annual rounds. PC retention was defined as completion of an annual follow up survey. Baseline characteristics were described by study arm and Poisson regression analyses used to measure the association between baseline factors and retention. In addition, we present a description of researcher-documented reasons for study termination by study participants in the population cohort (Table 3). Results Overall, 48301 individuals aged 18-44 years were enrolled, of whom 15 225 (40%) were aged 18-24 years. Most participants were women (27 139, 71%). Of the 38 474 participants enrolled at during the first round of the trial (PC0), 73. completed at least one follow up visit. Retention was lower in men (RR: 0.895; 95% CI: 0.875, 0.914) and higher amongst older participants (RR: 1.220; 95% CI 1.188, 1.253) when comparing age>35 to 18-24 years. Retention was also higher amongst individuals with high socio-economic status (SES) when compared to medium SES (RR 1.161; 95% CI 1.135, 1.188) and low SES (RR 1.117; 95% CI 1.093, 1.142) . The most common reasons for study termination were study refusal (23%) and relocation outside the CRT catchment area (66%). Conclusion Despite challenges, satisfactory retention outcomes were achieved in PopART with limited variability across study arms. In keeping with other studies, age, sex and SES were associated with lower levels of retention, relocation outside of catchment area was the most common reason for non- retention in this CRT.