Multivariate Analysis of The Determinants of Undernutrition in Under-Five Children In Zimbabwe

Hugh Chikerema

Abstract

Background: Undernutrition is one of the leading causes of infant mortality in developing countries including Zimbabwe. The main objective for the study was to model the relationship between undernutrition and individual, household and community level determinants. Material and methods: Secondary data analysis of the 2019 Multiple Indicator Cluster Survey (MICS) was performed using the merged children and household datasets. A total of 6223 children’s data was used for the study. Univariate and bivariate statistics were used to describe the data and trends of undernutrition among children 6-59 months in Zimbabwe. Three indices of undernutrition used as the dependent variables were height for age, weight for height and weight for age z-scores. Multivariate analysis techniques employed to analyze the data. Principal component analysis was used for variable reduction where components with eigen values more than one were used. Multivariate regression analysis was carried out to determine the determinants of undernutrition and model building using ten principal components. Multivariate Analysis of Variance (MANOVA) was used to check for the goodness of fit of the model. Results: Of the children measured and weighed, 23.5% were stunted, 9.7% were underweight and 2.9% were wasted. The overall KMO value for sampling adequacy check was 0.72. The three undernutrition categories had different predictor variables. Area of residence, sanitation and wealth status and diarrhoea were common significant factors for all the three undernutrition categories. Conclusion: The determinants of undernutrition are different for each undernutrition category. Six factors were significant for each category at individual, household, and community levels.