Duke University researchers trained an algorithm on the electronic medical records of 45,000 children to detect signs of autism in infants.
The algorithm predicts which babies would become autistic, differentiating them from those who developed attention-deficit hyperactivity disorder (ADHD) or other neurodevelopmental conditions.
Duke's Matthew Engelhard said the researchers focused on the model's performance in groups of children that traditional screening methods often miss, including girls, children of color, and children with combined autism/ADHD diagnoses.
Duke's Geraldine Dawson said a computer collecting information as a child receives healthcare could alert a pediatrician that an autism diagnosis is more likely "based on the child's pattern of utilization."
From USA Today
View Full Article
Abstracts Copyright © 2023 SmithBucklin, Washington, D.C., USA
No entries found