Weighted Random Effects Multinomial Model with Application to Anemia and Malnutrition Comorbidity among Under five Children in Nigeria

Abstract

This study develops multinomial models with weighted random effects to analyze the spatial pattern and risk factors associated with anemia, malnutrition, and their co-occurrence among children under the age of five in Nigeria. A Bayesian hierarchical multinomial model with weighted random effects and adjusted Intrinsic Conditional Auto regressive (ICAR) prior for the random effects, was used to account for the comorbid patterns of anemia and malnutrition among young children in Nigeria. The study utilized data from the 2018 Demographic and Health Survey. The structured random effects were weighted to reflect state-level variation in precipitation, a climatic factor considered to influence child health outcomes. The results of fixed effects indicated that area of residence, maternal education level, and household wealth status were significant predictors of anemia and malnutrition co-occurrence. The generated map identified the north eastern region of the country with low average precipitation as a high-risk region for anemia and malnutrition co-morbidity. These findings emphasize the need for targeted interventions to mitigate precipitation-related health risks and public health campaigns focusing on maternal education on child nutrition, hygiene, and disease prevention.