Leveraging Foundation Models in Maternal and Child Health: A Systematic Review

Maternal and child health (MCH) represents a critical domain requiring accurate, timely, and data-driven decision-making to optimize outcomes from pregnancy through early childhood. Foundation models (FMs) are large pre-trained artificial intelligence models that offer potential for clinical support in diagnostics, medical adherence, and reducing disparities. We conducted a systematic review to identify recent studies leveraging FMs in MCH published between 2020 and 2025. Of 785 studies, 63 met the inclusion criteria. FMs demonstrated strong potential to generalize across clinical tasks by integrating multimodal data, including text, electronic health records, imaging, and temporal data to support disease diagnosis, streamline clinical documentation, and generate high-quality medical responses throughout maternal, neonatal, and pediatric care. Moving forward, rigorous validation and close collaboration with clinicians will be essential for the safe, equitable, and effective deployment of FMs in MCH care.