Trends patterns and predictors of child mortality in India

Abstract Background Child survival is a fundamental indicator of population health and socioeconomic development. Although child mortality has declined substantially in India over the past two decades, considerable regional and socioeconomic disparities persist. This study examines trends, patterns, and determinants of neonatal, infant, and under-five mortality using mixed-effect binary logistic regression and multivariate decomposition analysis to identify the factors underlying observed reductions in child mortality. Methods This study utilized data from three rounds of the National Family Health Survey (NFHS): NFHS-3 (2005–06), NFHS-4 (2015–16), and NFHS-5 (2019–21). Trends and spatial patterns in child mortality indicators were illustrated using graphs and thematic maps. Mixed-effect binary logistic regression models were employed to identify determinants of neonatal, infant, and under-five mortality while accounting for clustering effects. The intra-class correlation coefficient (ICC) and likelihood ratio (LR) test were used to assess clustering. Model performance was compared using deviance, Akaike Information Criterion (AIC), and Bayesian Information Criterion (BIC). Additionally, logit-based multivariate decomposition analysis was performed to quantify the contributions of compositional (endowment) and coefficient effects to the decline in child mortality over the 16-year period. Findings Child mortality declined substantially between NFHS-3 and NFHS-5. Neonatal mortality decreased from 34.73 to 24.85, infant mortality from 51.08 to 35.11, and under-five mortality from 66.94 to 43.05 deaths per 1,000 live births. Mixed-effect binary logistic regression showed that higher birth order (≥ 4), maternal overweight, inadequate antenatal care (ANC) utilization, and low birth weight (< 2.5 kg) were significantly associated with increased odds of neonatal, infant, and under-five mortality in NFHS-5. Multivariate decomposition analysis indicated that compositional (endowment) changes explained 23.08% of the decline in neonatal mortality, 27.49% of the decline in infant mortality, and 24.64% of the decline in under-five mortality. Improvements in maternal age, maternal education, birth order, and full ANC utilization were the major contributors within this component. Coefficient effects accounted for the remaining proportion of the observed decline, accounting for 76.91%, 72.50%, and 75.35% of the declines in neonatal, infant, and under-five mortality, respectively. Within this component, maternal age, household religion, mass media exposure, birth interval, and full ANC utilization emerged as important contributors to the coefficient component of the observed mortality decline. Conclusion The substantial decline in child mortality in India over the past 16 years reflects changes in both population characteristics and the estimated associations between observed characteristics and child mortality over time. Strengthening maternal and child health programs targeting rural populations and multiparous women, promoting institutional deliveries, expanding access to quality antenatal care, and improving maternal nutrition are essential for sustaining reductions in neonatal, infant, and under-five mortality and achieving the Sustainable Development Goal (SDG) targets.