Peer support may be important for increasing PrEP use. However, little is known about the influence of social networks (connections between and among individuals) on PrEP use in sub-Saharan Africa. We aimed to assess whether social network contacts predicted PrEP uptake in rural Kenya and Uganda after accounting for known predictors of PrEP initiation.
The SEARCH study (NCT01864603) offered TDF/FTC PrEP in 16 communities during population-level HIV testing starting in 2016-2017. Universal PrEP access with rapid or same-day start was offered to HIV-uninfected adults ?15 years, with enhanced PrEP counseling for those at elevated HIV risk (based on serodifferent partnership, risk score, or self-identified risk). During population-level testing, persons were asked to name social contacts in 5 domains: health, money, emotional support, food, and free time. Named contacts were matched to community residents to build community-specific, sociocentric networks of 56,770 persons and 124,054 connections. Using targeted maximum likelihood estimation, we evaluated social network predictors of PrEP uptake within 1 year of population-level testing among persons at elevated HIV risk who had ?1 matched first-degree contact, accounting for clustering by community and adjusting for sociodemographic factors (sex, age, serodifferent partner, polygamous marriage, mobility, occupation).
Among 13,159 persons at elevated HIV risk, 8,898 (68%) had ?1 matched network contact. Of the 8,898, 49% were women, 34% ages 15-24 years, 11% had a serodifferent partner, 14% had ?1 contact who started PrEP, and 18% had ?1 contact living with HIV (LHIV). Overall, 2,570/8,898 (29%) started PrEP. Persons with ?1 contact who started PrEP were 57% more likely to start PrEP (adjusted risk ratio [aRR] 1.57, 95% CI 1.44-1.70, p<0.001) than those with contacts who did not start PrEP (Figure). Results were similar when stratified by sex and for same-sex and opposite-sex social contacts. Having an opposite-sex contact LHIV was associated with PrEP uptake in unadjusted analyses (RR 1.30, 95%CI 1.05-1.61, p=0.009), but not after adjustment for serodifferent partners and other factors (aRR 0.93, 95%CI 0.77-1.12, p=0.39).
Persons with a social contact who initiated PrEP were more likely to themselves start PrEP within 1 year of PrEP offer during population-level HIV testing. Interventions to facilitate peer support and strengthen social connections to other PrEP users should be considered to foster PrEP uptake.