HIV self-testing (HIVST) enables novel strategies, but needs new approaches to maximise linkage to care. We investigated: 1) whether financial incentives for community-selected volunteers (CVs) who distributed HIVST-kits improved timely linkage, and 2) if community-based HIVST increased facility-based ART initiation.
Trained CVs distributed HIVST kits door-to-door for 4-6 weeks in 38 rural wards in Zimbabwe. Using a cluster-randomised design with 1:1 randomisation of wards, CVs were allocated to receive stipend-only (one-off payment of US$50) or stipend-plus: US$50 + US$0.20 incentive per client linked to confirmatory HIV testing, non-communicable disease screen, family planning or male circumcision. Client self-report of HIVST uptake and linkage was assessed 6 weeks later by population survey. The primary outcome was linkage to any post-HIVST service, analysed with random-effects logistic regression, adjusted for imbalance between arms. We used a difference-in-differences quasi-experiment to investigate trends in ART initiation at public facilities in both HIVST and non-HIVST communities, from 6 months before to 3 months after HIVST distribution. Generalised estimation equations (GEE) were used to analyse the relationship between campaign period and trial arm on ART initiation. .
A total of 39,205 HIVST kits were distributed by 445 CVs in the stipend-plus arm (mean/CV 88; 95%CI 85-92) and 41,173 by 447 CVs in the stipend-only arm (mean/CV 93; 95%CI 89-96). Overall 7,146/ 8,566 (83.4%) household members responded at 6 weeks; 50.3% had self-tested, 46.5% in males and 46.2% in young people <25 years old. Self-test HIV prevalence was 8.0%; 36.3% of self-testers were first-time testers. Incentives had no effect on the primary outcome, but confirmatory testing by newly diagnosed/untreated HIVST+ clients was significantly higher in the stipend-plus arm, 25/33 (75.8%) versus 20/40 (50.0%), adjusted risk ratio 1.59, 95%CI 1.05-2.39 (Table). GEE modelling of 12,808 ART initiations from 168 clinics (1192 clinic-months) showed a 27% increase in ART initiation in HIVST versus non-HIVST communities (95%CI 14-43%), with no difference by incentive arm (Table).
Community-based HIVST campaigns achieved high uptake, including among youth, men and first-time testers, and increased demand for ART. A small linkage incentive to distributors may have increased timely linkage to care in HIV-positive participants not already on ART. Funding: UNITAID-PSI STAR, PACTR20160700170178