Geostatistical Analysis of Under-Five Children Mortality and Associated Factors Across Sub-Saharan African Countries

The important point in analyzing the spatial data is accounting for the spatial autocorrelation through the weight matrix. In the Bayesian context, the intrinsic conditional autocorrelation prior distribution has been used to model spatial autocorrelation. We used the most recent Demographic and Health Surveys (DHS) data from 33 sub-Saharan African (sSA) countries. The Bayesian generalized geo-additive mixed effects model was fitted to incorporate the linear, nonlinear, and spatial effects of childhood mortality. The spatial components were smoothed by the two-dimensional spline, and the continuous variables were modeled by the penalized splines. The model inference was made based on integrated nested Laplace approximation (INLA) and the parameter sensitivity for priors was validated. A total of 352,322 under-five children were included in the study. The overall prevalence of under-five mortality in sSA was 6.04% with countries Chad (9.79%), Nigeria (9.68%), and Sierra Leone (8.99%) recording the highest prevalence. The likelihood of being dead was prominently high among children decreasing in urban residences (AOR = 0.94; 95% CI = 0.90–0.99). Similarly, significant association with the risk of under-five mortality (U5M) includes child-level covariates (sex, birth order, place of delivery, and dietary diversity) and household-level covariates (mother education, wealth index, autonomy of mothers, source of water, sanitation facilities, and cooking fuels). We found that incorporating the spatial components in the analysis improves the performance of the model. The deviance information criteria revealed that the geo-additive generalized mixed model is an optimal choice for achieving a good model fit. Sex, weight, and age of a child, educational attainment of mothers, source of water for the household, type of toilet facility, and fuel use have contributed significantly to child’s under-five mortality across sSA. Hence, the government of each country should work on the significant contributors to the under-five mortality among under-five children.