Sandrah Rahantanirina

Background: Stockouts of essential medicines remain a persistent challenge across the Global South, undermining healthcare access. Unmanned aerial vehicles or drones have emerged as a promising solution to improve delivery. Methods: We implemented a two-arm cluster randomised trial in eastern Madagascar. Hard-to-reach public-sector facilities were randomised to drone delivery or standard land-based delivery of medical commodities. Preintervention and postintervention data included audits, provider and client interviews and population-based surveys of reproductive-age women. Our primary outcomes were (1) current modern contraceptive use and (2) the probability of stockouts. Secondary outcomes included aligned and preferred contraceptive use, vaccine and malaria outcomes and perceived quality. For each outcome, we compared intervention and control clusters using analysis of covariance (ANCOVA) regressions. Facilities randomised to drone delivery received continuous service from September 2023 to October 2024; outcomes were measured at baseline (June–August 2023) and endline (October 2024–March 2025). Results: Between June and August 2023, 109 facilities were randomised to drone intervention (n=55) or control (n=54). Within the corresponding communities, a representative sample of 1199 reproductive-age women (n=605 treatment, 594 control) completed a baseline survey. In primary ANCOVA analyses, drone delivery had no measurable impact on current modern contraceptive use, stockouts of most commodities, malaria rapid diagnostic test stockouts or most client-reported quality measures. Supporting difference-in-difference analyses suggested possible reductions in malaria rapid diagnostic test stockouts and modest improvements in aligned and preferred contraceptive use, but these findings were not consistently replicated in the primary specification. Conclusion: In this cluster randomised trial, drone delivery did not measurably improve most stockout, service readiness, utilisation or client-reported quality outcomes. These findings suggest that drone delivery alone may be insufficient to improve commodity availability where upstream supply-chain constraints remain binding. Prior estimates of drone impact across a broad range of supplies may have been inflated by methodological limitations. Rigorous studies in additional settings are needed before further integration of drones into national health systems in the Global South. Trial registration number: AEARCTR-0011710.