LEVELS AND DETERMINANTS OF CHILD UNDERNUTRITION IN BANGLADESH

The study explores the predictors of child undernutrition according to weight-for-age index by utilizing the data of nationwide Bangladesh Demographic and Health Survey 2004. Multiple logistic regression models have been employed to identify the predictors of child undernutrition according to the groups of variables: basic, underlying, immediate and temporal. It is observed that the prevalence of undernutrition varied significantly by most of the variables under study. Near half of the children under-5 years of age were found malnourished with 13% severely underweight. Undernutrition was found higher among the children ages 12-23 months, illiterate and thin mothers, households with lowest wealth index and food deficient in whole year. Separate multiple logistic regression models identified several basic, underlying, immediate and temporal predictors including age of child, birth interval, region, mother's education, household's wealth & food insecurity status, maternal care, child feeding practices, receiving vitamin-A capsule, incidence of ARI and diarrhoea, mother's BMI and age at birth. The final model considering all the significant variables identified in the individual models indicates that children belonging to Chittagong division, ages 12-23 months, lower birth interval, poor child feeding practices, suffered from ARI and diarrhoea, households with poor wealth index, and mothers who were illiterate, thin and insufficient maternal care recipient were in great risk to be undernourished.