Prevalence of Childhood Stunting and Its Associated Factors in Bangladesh: Quantile Regression Modeling Approach.

Background: Stunting is a key public health concern among under-five children in Bangladesh. It affects approximately one-quarter of children under 5 years of age, which is higher than global averages and insufficient to meet the goal of the World Health Organization (WHO) for reducing stunting to 15%. So, further research is needed to better understand the current dynamics of stunting and to develop pragmatic policies based on the most recent nationwide data which will provide valuable insights into the factors influencing stunting in Bangladesh. This study investigates the prevalence and risk factors of stunting among children under five using the most recent national data. Methods: Data were analyzed from the Bangladesh Demographic and Health Survey (BDHS) 2022. The data collection for the 2022 BDHS took place from June 27, 2022, to December 12, 2022. This survey used stratified sampling and selection is made in two stages. In the first stage, 675 enumeration areas (EAs) were picked from each stratum and a fixed number of 45 households per EA were selected in the second stage. The final data contain 8784 children under 5 years. After cleaning the flagged cases and missing values of related variable, the dataset comprises information on weighted sample of 3664 children. In bivariate analysis, a nonparametric Kruskal-Wallis test was employed for violating the parametric assumptions. Subsequently, robust quantile regression models were utilized to evaluate the impact of various independent variables on height-for-age Z-score (HAZ) across different quantiles. Results: Results reveal that the prevalence of stunting was 23%, with 17.8% and 5.2% classified as moderately and severely stunted, respectively. The quantile regression analysis shows significant heterogeneity in the determinants of child growth across the HAZ distribution. Predictors with stronger effects among moderately and severely stunted children (lowest HAZ quantile) include older child age, high birth order, twin births, low maternal BMI, low maternal education, lack of media exposure, low paternal education, and poor household wealth. Older child age was consistently associated with lower HAZ across all quantiles. The positive effects of maternal education, maternal overweight status, media exposure, and household wealth index are strongest at the lower HAZ quantiles. Children with higher birth orders have lower HAZ through all quantiles compared to first-born children. Twin births are associated with lower HAZ in the lower and middle quantiles. Several predictors, particularly maternal education, maternal nutritional status, and household wealth, exhibit higher prognostic value for severe and moderate stunting than for children in higher HAZ quantiles. Conclusions: Stunting remains a significant public health concern among children in Bangladesh. The effects of child, maternal, household, and health-related factors varied across the HAZ distribution, with several predictors showing stronger associations among children with lower HAZs. Quantile regression provided important insights beyond conventional analyses by identifying factors with stronger effects among children in the lower tail of the HAZ distribution.