Traditional healer-delivered point-of-care HIV testing versus referral to clinical facilities for adults of unknown serostatus in rural Uganda: a mixed-methods, cluster-randomised trial

Summary: Background: HIV counselling and testing are essential to control the HIV epidemic. However, HIV testing uptake is low in sub-Saharan Africa, where many people use informal health-care resources such as traditional healers. We hypothesised that uptake of HIV tests would increase if provided by traditional healers. We aimed to determine the effectiveness of traditional healers delivering HIV testing at point of care compared with referral to local clinics for HIV testing in rural southwestern Uganda. Methods: We did a mixed-methods study that included a cluster-randomised trial followed by individual qualitative interviews among a sample of participants in Mbarara, Uganda. Traditional healers aged 18 years or older who were located within 8 km of the Mbarara District HIV clinic, were identified in the 2018 population-level census of traditional healers in Mbarara District, and delivered care to at least seven clients per week were randomly assigned (1:1) as clusters to an intervention or a control group. Healers screened their clients for eligibility, and research assistants confirmed eligibility and enrolled clients who were aged 18 years or older, were receiving care from a participating healer, were sexually active (ever had intercourse), self-reported not having received an HIV test in the previous 12 months (and therefore considered to be of unknown serostatus), and had not previously been diagnosed with HIV infection. Intervention group healers provided counselling and offered point-of-care HIV tests to adult clients. Control group healers provided referral for HIV testing at nearby clinics. The primary outcome was the individual receipt of an HIV test within 90 days of study enrolment. Safety and adverse events were recorded and defined on the basis of prespecified criteria. This study is registered with ClinicalTrials.gov, NCT03718871. Findings: Between Aug 2, 2019, and Feb 7, 2020, 17 traditional healers were randomly assigned as clusters (nine to intervention and eight to control), with 500 clients of unknown HIV serostatus enrolled (250 per group). In the intervention group, 250 clients (100%) received an HIV test compared with 57 (23%) in the control group, a 77% (95% CI 73–82) increase in testing uptake, after adjusting for the effect of clustering (p