AI Powered Monitoring and Risk Prediction for Maternal Health to Ensure Fetal Well-Being

Health complications due to maternal and fetal mortality are a worldwide concern. Traditional medicine collects no data (or very little) in real-time, making it impossible to intervene in a timely manner. Here, we introduce a Stacked Ensemble combining TabNet, XGBoost, and Bayesian Optimization for the prediction of maternal health outcome risk. Additionally, TabNet retrieves complex data, XGBoost increases accuracy and Bayesian Optimization optimizes hyperparameters. In contrast to hospital-based models, our pipeline is designed for continuous risk assessment, which permits the detection of pregnancy complications such as gestational hypertension at an early stage. This AI-based framework holds the potential to mend maternal care gaps to enable more proactive, accessible monitoring that can save lives and improve maternal and fetal health, particularly among underserved populations.