Spatio-temporal mapping of stunting and wasting in Nigerian children: A bivariate mixture modeling

Abstract

Studies have shown that stunting and wasting indicators are strongly correlated among children, with the potential of concurrently affecting their physical and cognitive development. However, the identification of subpopulations of children with varying risks of stunting and wasting could be valuable for targeted intervention. This work proposed a bivariate spatio-temporal mixture model within a Bayesian framework to describe the spatial behavior of subpopulations of the children within the wider population of children under five years of age in Nigeria. The model assumes that each sub-population follows a Gaussian distribution, and therefore, the overall population is modeled by combining Gaussian sub-spatial models probabilistically. Inferences were based on the Markov chain Monte Carlo algorithm, that draw samples from the joint posterior distribution. The model was applied to data from four waves of the Nigerian Demographic and Health Survey. We identified a significant negative correlation between stunting and wasting among subpopulations with a negative spatial correlation between the spatial patterns of both illnesses. The findings demonstrate varying risk factors between the subpopulations with an evidence of spatio-temporal disparity in the likelihood of stunting and wasting. The findings underscore the need for a comprehensive national intervention program with attention given to high-burden states in a manner that involves communities and subpopulations. The maps could serve as a valuable tool for intervention planning.