AI Applications in Maternal Care Expand Globally
New applications of artificial intelligence in maternal care are emerging, with India's AI Mission integrating AI for nutrition tracking and Kenya deploying AI-supported ultrasound. In the U.S., Human Longevity launched a personal AI health app, while a recent podcast discussed the critical need for safety and governance protocols as these technologies are adopted.
- The AI-supported ultrasound project in Kenya is an eight-month research collaboration between Google and Jacaranda Health, a local non-profit, aimed at addressing a significant care gap, as only 16% of pregnant women in the country currently have access to this service. The AI-enhanced devices are designed to be portable and simplify image acquisition and interpretation, enabling nurses and technicians to perform examinations without extensive sonography training. - AI models are being specifically developed to expand the capabilities of midwives and other frontline health workers in low-resource settings. One such tool, funded by the Gates Foundation for use in Malawi and Uganda, allows a provider to calculate gestational age through a simple "blind-sweep" scan, requiring no advanced training in image interpretation to get a rapid, accurate estimate. - In India, the AI-powered "Kilkari" program, the world's largest mobile health information service, delivers weekly audio messages to over 60 million women. AI models are used to predict when a user might stop listening to the calls, allowing for targeted, personalized outreach that has retained approximately 30% of those at high risk of disengaging. - Another application in Andhra Pradesh, India, the Janani Mitra app, uses AI to screen for anaemia by analyzing photos of a pregnant woman's eyes, tongue, and nails, sending an alert to health workers if early signs are detected. - Predictive modeling in maternal care has demonstrated high accuracy in identifying risks; one systematic review of 31 articles found that an AI model predicted neonatal mortality with 99.7% accuracy, while another achieved 95.7% accuracy for predicting premature birth. These models primarily analyze data from electronic medical records, medical images, and genetic markers to forecast complications. - For midwives, the integration of AI tools could significantly reduce time spent on administrative tasks. It's estimated that implementing electronic records and other digital technologies could increase the proportion of time midwives spend on direct clinical care from 30% to as much as 70%. - Professional organizations are responding to the growth of AI in maternal care with calls for careful governance. The Australian College of Midwives, for instance, has advocated for legislation to ensure AI serves as a support tool and does not replace the clinical decision-making of human healthcare professionals. [14,
