Azernews.Az

by Alimat Aliyeva

An artificial intelligence algorithm has shown the ability to predict the development of attention deficit hyperactivity disorder (ADHD) in children several years before a formal diagnosis. This conclusion was reached by researchers from the Duke University School of Medicine, AzerNEWS reports.

The approach is based on the analysis of electronic health records that track a child’s development from birth, including medical history, behavioral patterns, sleep data, and visits to healthcare specialists. Researchers found that long before an official diagnosis is made, subtle but consistent patterns linked to ADHD begin to appear in these records.

ADHD is typically characterized by symptoms such as inattention, hyperactivity, and impulsivity, which can significantly affect daily functioning. However, diagnosis often occurs only after these difficulties become more obvious in a school environment. The new AI-based method could allow clinicians to identify at-risk children much earlier.

To develop the model, scientists trained the algorithm on data from more than 720,000 patients and then tested it on a separate group of 140,000 children, whose medical histories were followed for up to nine years. By the age of five, the system was able to predict the likelihood of an ADHD diagnosis within the next four years with an accuracy of approximately 0.92 on a standardized evaluation scale.

The most significant predictive indicators included early developmental signs such as speech delays, learning difficulties, and early emotional or behavioral challenges.

Researchers emphasize that early identification of risk could make it possible to provide timely support, ranging from parental guidance to tailored educational interventions in schools. However, they also stress that the technology still requires further validation to ensure reliability and fairness across different populations.

An additional interesting aspect is that similar AI systems are now being explored for other neurodevelopmental conditions such as autism spectrum disorder and dyslexia, suggesting a future where early digital screening could become a routine part of pediatric healthcare.