ABSTRACT
Background: Differentiated service delivery (DSD) models provide flexibility for patients accessing antiretroviral treatment (ART) in sub-Saharan Africa, and decongest health facilities. With the global COVID-19 pandemic, DSD models, which promote social distancing and facilitate easier access to ART in the community are critical. We investigated the clinical effectiveness of community-based multimonth ART provision in high HIV prevalence settings in Zimbabwe and Lesotho. Methods: Individual-level patient data were pooled from two clusterrandomized noninferiority trials that compared 3 and 6-monthly communitybased ART provision vs. standard of care facility-based ART dispensing for stable HIV patients, which were conducted between 2017-2019. Both trials had three-arms: ART collected 3-monthly at facilities (3MF, control);ART provided 3-monthly in community ART refill groups (CARGs) (3MC);and ART provided 6-monthly in either CARGs or at community-distribution points (6MC). Stable adults with viral suppression receiving ART ≥6 months were recruited. The primary outcome was retention in ART care 12 months after enrolment, and secondary outcomes were viral suppression and number of unscheduled clinic visits between months 0-12. Individual-level regression analyses were conducted by intention-to-treat specifying for clustering and adjusted for country. Results: 60 randomized clusters were included with 3817, 2893 and 3426 participants enrolled in arms 3MF, 3MC and 6MC, respectively. After 12 months, retention in 3MF, 3MC and 6MC was 95.0%, 95.7% and 95.1%, respectively;adjusted risk differences 0.3 (95% CI: -0.8 to 1.4);-0.2 (95% CI: -1.4 to 1.0) and -0.5 (95% CI: -1.7 to 0.6) for 3MC vs. 3MF, 6MC vs 3MF, and 6MC vs 3MC, respectively (Figure 1). All comparisons achieved the prespecified noninferiority margin of -3.25%. Viral suppression after 12 months was high in all arms;97.8%, 98.6% and 97.9% in 3MF, 3MC and 6MC, respectively, adjusted risk ratios=1.0 (95% CI: 0.99-1.01);1.0 (0.97-1.01) and 1.0 (0.97-1.01) for 3MC vs. 3MF, 6MC vs 3MF, and 6MC vs 3MC, respectively. No differences in the number of unscheduled clinic visits between arms were apparent. Conclusion: Community-based ART provision at both 3 and 6-monthly intervals were noninferior vs. 3-monthly facility-based dispensing in high HIV-prevalence settings for stable HIV patients. These DSD models are now more critical than ever, and can be scaled-up to promote social distancing and clinic decongestion to mitigate the impact of COVID-19.