Research

Labor-AI integrates real-world clinical data from over 130,000 deliveries to develop, validate, and deploy AI-driven decision support tools in obstetrics. Our research spans three core domains:

• Labor & Delivery Prediction Models: Development and external validation of models to predict unplanned cesarean delivery (uCD), labor progression, and risk of fetal/maternal complications.
• Perinatal Epidemiology: Large-scale cohort analyses linking maternal and neonatal outcomes, including studies on labor timing, weight estimation errors, and risk factors in special populations.
• Clinical ML Methodology: Research on explainable AI (XAI), reinforcement learning for labor decisions, and transportability of machine learning models across hospitals.

Our work includes ORACLE-AI, an explainable model for predicting uCD at the time of admission; MRI-based pelvimetry for delivery planning; unsupervised clustering of partograms; and AI-enhanced triage tools.

All models are developed with an emphasis on clinical relevance, transparency, and potential integration into hospital systems.