Risk factors and co-occurring patterns of low birth weight in Bangladesh: Insights from logistic regression and association rule mining

Low birth weight (LBW) remains a major public health concern in South Asia, including Bangladesh, contributing significantly to neonatal morbidity and mortality. This study aimed to identify individual risk factors for LBW using logistic regression (LR) and to explore co-occurring patterns among these risk factors through association rule mining (ARM). Analyzing the Bangladesh Demographic and Health Survey (BDHS), 2022 data with 1,435 participants, LR identified division, twin status, wealth index, place of delivery, duration of breastfeeding, and birth order as significant individual risk factors for LBW. The ARM revealed that infants in the Dhaka division with multiple births exhibited a higher risk of LBW, and this risk further increased when delivery occurred at a private facility. In Sylhet, LBW is more likely among 2nd born children from low-wealth households who are not currently breastfeeding. In Chittagong, infants from single births who are not currently breastfeeding, delivered at home, and from low-wealth households are also at higher risk. Across all divisions, low-wealth households and lack of breastfeeding appeared as co-occurring patterns, indicating the combined influence of socioeconomic disadvantage and postnatal vulnerability among LBW infants. Combining LR and ARM provides a comprehensive understanding of individual and interacting LBW risk factors, supporting targeted interventions to lower LBW prevalence and neonatal mortality in Bangladesh, thereby contributing to SDG 3.