Area under receiver operating characteristic curve at delivery exceeded 0.9 for 10 of 24 neonatal outcomes considered<\/p>\n<\/h3>\n
By Elana Gotkine HealthDay Reporter<\/b><\/p>\n
<\/b><\/p>\n
FRIDAY, Feb. 17, 2023 (HealthDay News) — A multitask deep learning model based on data from electronic health records (EHRs) can predict neonatal outcomes, according to a study published in the Feb. 15 issue of Science Translational Medicine<\/em>.<\/p>\n Davide De Francesco, Ph.D., from the Stanford University School of Medicine in California, and colleagues proposed a longitudinal risk assessment for adverse neonatal outcomes in newborns based on a deep learning model that uses EHRs to predict a range of outcomes starting shortly before conception and ending months after birth. A cohort of 22,104 mother-newborn dyads delivered between 2014 and 2018 was developed. A multi-input multitask deep learning model was trained using maternal and newborn EHRs to predict 24 neonatal outcomes. The model was validated using an additional cohort of 10,205 mother-newborn dyads delivered from 2019 to September 2020.<\/p>\n The researchers found that for 10 of the 24 neonatal outcomes, areas under the receiver operating characteristic curve at delivery exceeded 0.9; for seven additional outcomes, they were between 0.8 and 0.9. Multiple known associations between various maternal and neonatal features and specific neonatal outcomes were identified in a comprehensive association analysis.<\/p>\n “This is a new way of thinking about preterm birth, placing the focus on individual health factors of the newborns rather than looking only at how early they are born,” a coauthor said in a statement.<\/p>\n Several authors disclosed financial ties to the biopharmaceutical industry.<\/p>\n