Geographical accessibility to hospital care amid armed conflict in Myanmar: a population-weighted analysis of compounding health inequality, 2021–2024

INTRODUCTION: Armed conflict disrupts health systems, but whether it compounds pre-existing geographical barriers to hospital care has not been quantified. We aimed to estimate population-weighted travel time to public hospital care across Myanmar's 330 townships; test whether conflict following the February 2021 coup concentrated in populations with the worst baseline access, and whether attacks on health facilities followed the same pattern; and measure the cumulative person-year burden of simultaneous poor access and intense conflict. METHODS: We conducted a longitudinal ecological analysis of 330 townships over 16 quarters from 2021 to 2024. Travel time from each populated square kilometre to the nearest of 68 general hospitals, representing the surgical-care tier, was estimated using the pre-coup facility registry, OpenStreetMap road data and WorldPop 2020 population estimates. Townships were classified by population-weighted median travel time. Quarterly conflict intensity was calculated from 35 647 Armed Conflict Location and Event Data Project battle and explosion or remote-violence events; 655 Safeguarding Health in Conflict Coalition-reported facility attacks were analysed separately. The concentration index assessed whether conflict events, fatalities and facility attacks were disproportionately concentrated among populations with worse baseline access. Double-burden exposure (poor or very poor access combined with high or extreme conflict) was accumulated in person-years. Because access was held at pre-coup baseline, estimates should be read as optimistic bounds. RESULTS: Before the coup, 5.8 million people (12%) lived over 2 hours from a general hospital. Double-burden exposure grew from 0.4 million (Q1 2021) to 3.1 million (Q1 2024). Conflict was significantly concentrated in worse-access populations: concentration index 0.126 (95% CI 0.053 to 0.199) for events, 0.169 (95% CI 0.078 to 0.263) for fatalities. Facility attacks showed no significant concentration (0.049, 95% CI -0.041 to 0.145), as remote townships lack targetable facilities. An estimated 8.4 million person-years of double-burden exposure accumulated over 2021-2024, rising roughly eightfold quarterly, from about 100 000 (Q1 2021) to 780 000 (Q1 2024). CONCLUSION: Conflict in Myanmar has compounded pre-existing geographical inequality in hospital access, with populations furthest from surgical care bearing a disproportionate burden. The concentration curve framework used is transferable to other conflict-affected settings.