AI Model Uncovers Hidden Risk Factors for Stillbirth and Pregnancy Complications A groundbreaking artificial intelligence (AI) model has identified previously unknown risk factor combinations linked to stillbirth and severe pregnancy complications. Researchers from the Universities of Utah and Brown analyzed data from nearly 10,000 pregnancies, uncovering insights that could lead to more personalized pregnancy risk assessments and care. New Patterns in Pregnancy Risk The study, published in BMC Pregnancy and Childbirth , revealed that clinical guidelines currently treat certain pregnancies the same despite significant variations in risk, sometimes up to tenfold. The AI model detected unexpected patterns that even experienced clinicians had not recognized , offering a new way to assess pregnancy risks. Key risk factors identified by the AI include: [object Object], [object Object], [object Object] Senior author Nathan Blue , an assistant professor of obstetrics and gynecology at the University of Utah’s School of Medicine, highlighted the importance of these findings. “The AI model detected something that even highly skilled clinicians weren’t recognizing,” he said. Rethinking Risk in Low-Weight Fetuses The researchers were particularly interested in fetuses in the bottom 10% of the weight distribution, babies who are smaller than average but not critically underweight. Current clinical guidelines recommend intensive monitoring for these cases, often resulting in additional emotional and financial burdens for families. However, the study found that within this group, risk varied widely. Some of these pregnancies posed no greater danger than average, while others carried nearly ten times the By identifying specific factors that increase or decrease risk, the AI model could help doctors make more precise decisions about monitoring and interventions, reducing unnecessary procedures while ensuring high-risk pregnancies receive the care they need. The Power of Explainable AI in Medicine Unlike traditional AI models , which can function as “black boxes,” the research team used explainable AI, a model that not only predicts risk but also provides insights into how it reaches those conclusions. This transparency helps physicians understand and trust the model’s recommendations while also identifying potential biases. Blue emphasized the advantages of AI in clinical decision-making. “AI can estimate a risk specific to an individual’s situation, and it does so transparently and reproducibly, which even the best clinicians can’t always do,” he explained. The Future of AI in Pregnancy Care Before this AI model can be widely used, researchers must validate it in real-world clinical settings. However, the study offers a promising step toward more personalized pregnancy care. By combining large-scale data analysis with explainable AI, medical professionals could soon have a powerful tool to predict and prevent pregnancy complications more effectively than ever before. This research was supported by the One U Data Science Hub Seed Grant Program , the R Baby Foundation , and the National Institute of Child Health and Human Development .
