Artificial Intelligence (AI) Implementation in Maternal and Child Health: A Scoping Review
TL;DR
Current evidence suggests AI holds promise as an assistive tool in maternal and child healthcare, but given the limited number and heterogeneity of included studies, this evidence should be interpreted as preliminary.
Abstract
Abstract Artificial intelligence (AI) has increasingly been applied in maternal and child health. However, current evidence remains largely focused on model development, while reports on real-world clinical implementation are limited. This scoping review aimed to map the implementation of AI in maternal and child health settings and summarize reported outcomes. This scoping review followed the PRISMA extension for Scoping Reviews (PRISMA-ScR) guideline. Searches were conducted in PubMed and Scopus using Boolean operators combining terms related to artificial intelligence, maternal and child health, and implementation. Studies published in English between 2021 and 2026 were included if they reported clinical or community-based implementation of AI involving real patients or healthcare providers within an ongoing care pathway. Studies focused solely on AI model development or technical validation, reviews, conference abstracts, and editorials were excluded. Of 191 records identified, 156 were screened after duplicate removal, 64 full-text reports were assessed, and seven studies met all inclusion criteria. Identified AI applications were grouped into three themes: maternal support and community-based interventions, neonatal and pediatric monitoring, and screening and diagnostic support. Included studies reported promising outcomes, including improved monitoring accuracy, maternal engagement, and image quality standardization in low-resource settings. However, the seven included studies were highly heterogeneous, and most remained limited to feasibility studies or early-stage implementations. Current evidence suggests AI holds promise as an assistive tool in maternal and child healthcare. However, given the limited number and heterogeneity of included studies, this evidence should be interpreted as preliminary. Organizational, regulatory, financial, and workforce-related barriers, along with ethical considerations such as data privacy and algorithmic bias, remain to be addressed. Further large-scale, long-term implementation studies are needed to evaluate the integration and sustainability of AI in routine maternal and child healthcare practice.