Artificial Intelligence in Maternal Health Forecasting in Low- and Middle-Income Countries: A Systematic Narrative Review of Enablers, Barriers, and Policy Recommendations

Background Maternal mortality remains a major concern, with over 90% of the 295,000 maternal deaths in 2017 occurring in low- and middle-income countries (LMICs). Health systems in these contexts often lack predictive capacity, continuity of care, and timely response to complications. Artificial intelligence (AI) has emerged as a tool to strengthen maternal health forecasting and decision-making where access to quality care is limited. Objectives To review current AI applications for maternal health forecasting in LMICs, evaluate effectiveness in predicting and preventing complications, and propose an ethical, context-sensitive framework for integration into health systems. Methods A systematic search of PubMed, WHO Global Health Library, and IEEE Xplore (2018–2025) was conducted using terms related to “artificial intelligence,” “machine learning,” and “maternal health.” Grey literature from NGOs, donor agencies, and ministries of health was included. Eligible studies applied AI to maternal health in LMICs and reported implementation or pilot outcomes. Data extraction followed PRISMA 2020. Results Of 436 studies, 42 met inclusion criteria. Applications included predictive modeling for preeclampsia and preterm birth, cesarean risk stratification, ultrasound interpretation, and decision support. AI-based prediction showed AUROC 0.75–0.93. Interventions in Kenya, India, Ethiopia, and Guatemala demonstrated improved risk stratification and patient engagement. Key enablers were data quality, connectivity, and workforce training; barriers included ethical, equity, and infrastructure challenges. Conclusions AI-enabled maternal health forecasting shows potential to reduce maternal and perinatal risk in LMICs. Scalable impact requires integration into health systems, ethical governance, and outcome-based evaluations to ensure equity and sustainability. Registration: PROSPERO ID: 1127055