A hierarchical competing risks survival model in the study of child mortality in Bangladesh

Under-five child mortality remains a significant public health issue in Bangladesh and other developing countries. Identifying key risk factors within a hierarchical data structure is crucial for improving health system performance. This study employed the Fine-Gray Frailty (FGF) model to Bangladesh Demographic and Health Survey (BDHS) data to assess child mortality from disease-related causes, treating non-disease-related deaths as competing risks in hierarchical framework. The model was also applied by considering non-disease deaths as the main event in a subsequent analysis. Findings reveal that maternal age, child sex, birth order, and regional variation remain significant determinants of mortality after adjusting for the hierarchical data structure. Children of mothers older than 30 years face a significantly higher risk of non-disease deaths compared with those aged 20-30 years. Male children experience higher mortality than females for both disease and non-disease causes. Higher birth order is associated with a lower risk of non-disease mortality. Regionally, Khulna shows a significantly increasing hazard of non-disease deaths than Barisal. This study underscores the value of hierarchical competing risk models for guiding targeted public health interventions.