Institutional Delivery Coverage and Maternal Mortality: A Global Country-Year Fixed-Effects Analysis

Background: Maternal mortality remains a major marker of health-system performance and inequity, and global progress has slowed despite substantial declines since 2000. Institutional delivery is a core policy target for reducing preventable maternal deaths, but service coverage alone might not translate into survival benefit when facility capability, referral readiness, and quality of care are weak. We aimed to estimate the within-country association between institutional delivery coverage and maternal mortality over time, and to translate this association into policy-relevant absolute mortality differences.<br><br>Methods: We did a longitudinal ecological panel analysis of country-year observations from 195 countries between Jan 1, 2000, and Dec 31, 2022. We assembled a harmonised panel of annual country-level indicators for maternal mortality ratio (MMR; deaths per 100 000 live births), institutional delivery coverage, anaemia prevalence among women, and adolescent birth rate. The primary outcome was log-transformed MMR, with winsorisation of MMR at the 1st and 99th percentiles in the primary analysis. The primary exposure was institutional delivery coverage lagged by 1 year and scaled per 10 percentage-point increase. We estimated two-way fixed-effects models with country and year fixed effects, adjusting for lagged anaemia prevalence and lagged adolescent birth rate, and used two-way clustered standard errors by country and year. Prespecified secondary analyses included spline modelling for non-linearity, effect modification by anaemia, a falsification model using a placebo lead of institutional delivery coverage, region-by-year fixed-effects sensitivity analysis, and a reduced-sample robustness model additionally adjusting for skilled birth attendance and antenatal care with at least four visits. We translated the primary coefficient into country-specific expected absolute MMR changes associated with a 10 percentage-point increase in institutional delivery coverage.<br><br>Findings: The analytic panel included 4485 country-year observations with non-missing MMR from 195 countries. In the prespecified primary fixed-effects model, estimated on 1231 complete country-year observations, a 10 percentage-point increase in institutional delivery coverage lagged by 1 year was associated with lower subsequent maternal mortality (β −0·072, SE 0·016; p&lt;0·001). In the sensitivity model with region-by-year fixed effects (n=1224), the association remained negative (β −0·049, SE 0·020; p&lt;0·05). In the reduced-sample robustness model additionally adjusting for skilled birth attendance and antenatal care coverage (n=462), the institutional delivery coefficient attenuated towards the null (β 0·004, SE 0·045; p&gt;0·90), whereas skilled birth attendance was negatively associated with maternal mortality (β −0·117, SE 0·042; p&lt;0·05). In the falsification analysis using a placebo lead of institutional delivery coverage (n=1267), the lead coefficient was also negative (β −0·079, SE 0·017; p&lt;0·001). In the effect-modification model, the interaction between institutional delivery coverage and centred anaemia was negative but imprecisely estimated (β −0·0165, SE 0·0090; p=0·082). Policy translation of the primary model suggested the largest expected absolute MMR reductions in countries with the highest baseline burden, including Central African Republic (predicted reduction 114·3 [95% CI 64·5 to 162·5] deaths per 100 000 live births), Nigeria (72·5 [40·9 to 103·1]), and Chad (66·3 [37·4 to 94·4]) for a 10 percentage- point increase in institutional delivery coverage.<br><br>Interpretation: Higher institutional delivery coverage was associated with lower maternal mortality within countries over time, with the largest implied absolute gains in high-burden settings. Yet attenuation in more fully adjusted models and a non-null placebo lead indicate that this association is unlikely to represent the isolated causal effect of facility birth alone. Institutional delivery instead appears to function as a marker of broader maternal health- system strength, including emergency obstetric capability, referral readiness, postpartum surveillance, and quality of care. Progress towards SDG 3.1 will depend not only on expanding facility-based childbirth, but also on converting facility birth into effective, equitable, quality-assured survival care.