Stunting, wasting and underweight are the three widely recognised indicators of child undernutrition. This study aimed to simultaneously model all indicators while accounting for their association using a joint modelling technique to identify their risk factors.

This was a cross-sectional study design.

The anthropometric data of children were elicited from the Bangladesh Multiple Indicator Cluster Survey (MICS) 2019.

Stunting, wasting and underweight were the main outcome measures of child undernutrition. Initially, a generalised linear mixed model (GLMM) was developed for each indicator separately to identify the underlying risk factors by considering children within the cluster (district level) as hierarchically nested. Finally, a joint model was developed by combining the separate GLMMs with the condition of correlated cluster-specific (district-specific) random effects.

The developed joint model provided precise effects of the risk factors and quantified the association among stunting, wasting and underweight. The joint correlations of underweight with stunting and underweight with wasting were significantly positive based on the predictors in the joint model. The age of children, household wealth status, and maternal education were risk factors for both stunting and underweight. The mother’s age at first birth and the episode of fever were significantly correlated with both underweight and wasting, while the episode of fever was significantly associated with underweight. The region, residence and sex of children were significantly associated with all undernutrition indicators.

This study demonstrates the application of a joint model to simultaneously identify the risk factors associated with indicators of child undernutrition. The study findings reveal a substantial positive association between stunting and underweight, as well as between underweight and wasting, with shared risk factors contributing to the disparity in the prevalence of all forms of child undernutrition in Bangladesh.