Equity-Adjusted Safe-Delivery Thresholds: A Global Causal Analysis of Institutional Birth Coverage and Maternal Mortality in 182 Countries, 2000–2022

Abstracts Background Maternal mortality remains far above the Sustainable Development Goal (SDG) target in many settings, and global strategies lack an empirically derived coverage threshold for “safe” delivery. Methods We assembled a longitudinal panel of 2 184 country–year observations from 182 countries (2000–22), linking UN Maternal Mortality Estimation Inter-agency Group estimates with WHO and World Bank indicators. The primary exposure was institutional-birth coverage (% of livebirths in health facilities); the outcome was maternal mortality ratio (MMR; deaths per 100 000 livebirths). Using two-way fixed-effects models with restricted cubic splines, adjusted for anaemia prevalence and adolescent birth rate, we estimated the within-country association between institutional births and log(MMR). We then applied Hansen-type panel threshold regression to identify coverage levels at which the slope of this association changed, and repeated analyses by WHO region. Findings Median institutional-birth coverage was 72·4% (IQR 54·2–90·7) and median MMR 173 deaths per 100 000 livebirths (81–328). Globally, each 10–percentage-point increase in institutional births was associated with a 7·8% (95% CI 6·1–9·5) reduction in log(MMR). Threshold analysis identified a single global inflection at 70·2% coverage (95% CI ∼68–72). Below this threshold, each 10–point increase in institutional births was associated with a 12·6% (10·2–14·8) reduction in log(MMR), compared with 4·1% (2·5–5·7) above the threshold—an almost three-fold difference in marginal effect. Region-specific thresholds ranged from 65% in the African Region to over 90% in the Western Pacific. Interpretation Seventy per cent institutional-birth coverage represents a “safe-delivery” threshold: below this level, expanding facility births yields large reductions in maternal mortality; above it, further gains require investments in quality of care, emergency obstetric capacity, and equity. Embedding this empirically derived threshold into SDG 3·1 and universal health coverage monitoring could sharpen accountability, guide resource allocation, and accelerate progress towards ending preventable maternal deaths. Funding None.