Predicting preemie health

Machine learning approach helps identify which infants are vulnerable to complications

Featured Media for Predicting preemie health

Scientists have developed a machine-learning algorithm that analyzes electronic health records of mothers and their babies to predict which premature newborns are likely to develop health problems in their first two months of life.

Knowing which infants are vulnerable could enable targeted preventive measures.

“This is a new way of thinking about preterm birth, focusing on newborns’ individual health factors rather than only how early they are born,” said Nima Aghaeepour, PhD, associate professor of anesthesiology, perioperative and pain medicine and of pediatrics.

He was the senior author of a study describing the method, published in February 2023 in Science Translational Medicine. The study used data from the electronic health records covering 32,354 births between 2017 and 2020.

Adobe Stock image by xreflex

Read more here.

Author headshot

Erin Digitale

Erin Digitale is the pediatrics senior science writer in the Office of Communications. Email her at digitale@stanford.edu.

Email the author