Reprod Health. 2026 May 29. doi: 10.1186/s12978-026-02367-0. Online ahead of print.
ABSTRACT
BACKGROUND: Short birth interval (SBI) is a major public health concern, associated with adverse outcomes such as preterm birth, low birth weight, and maternal depletion. Identifying its influential predictors is essential for improving family planning and maternal health interventions. The study aimed to predict SBI and identify its influential predictors using stacked machine learning and SHAP explainability.
METHODS: This study used data, obtained from the Bangladesh Demographic and Health Survey (BDHS), 2022. The dataset comprised 11,872 respondents, of whom 2,137 experienced SBI (18.0%). Class imbalance problem was addressed by applying class weighting during model training. Boruta and least absolute shrinkage and selection operator-based methods were applied to identify the most important predictors of SBI. Subsequently, multiple machine learning models (including logistic regression, random forest, extreme gradient boosting, categorical boosting, artificial neural network, and stacking ensemble) were used to predict SBI and evaluated their model performances using accuracy, precision, sensitivity, F1-score, and the area under the curve (AUC). Finally, SHAP explainability were employed to identify the most influential predictors of SBI.
RESULTS: The stacking ensemble model achieved the highest predictive performance compared to the individual models, with an accuracy of 65.8%, precision of 42.6%, sensitivity of 72.3%, F1-score of 53.6%, and modest AUC of approximately 0.667. The SHAP analysis showed that no educated respondents, higher parity, and do not intend to use contraceptive method were the most influential predictors of SBI.
CONCLUSION: Interventions could therefore focus on improving female education, expanding access to contraceptives, and promoting awareness of optimal birth spacing. Policymakers may incorporate SHAP explainability to support data-driven reproductive health strategies to reduce SBI in Bangladesh.
PMID:42216193 | DOI:10.1186/s12978-026-02367-0
