Duke study: AI model detects ADHD in kids as young as 5

Duke University researchers have developed an artificial intelligence model capable of detecting attention-deficit/hyperactivity disorder (ADHD) in children as young as five years old using electronic health record data. The findings were published in the journal *Nature Mental Health*. The study is part of ongoing efforts to use AI for early screening of neurodevelopmental disorders.

The model was trained on electronic health records from over 720,000 individuals and specifically on data from approximately 140,000 children. It identifies patterns in health data that correlate with a later diagnosis of ADHD. The researchers reported that the model accurately detected ADHD risk at age five, potentially enabling earlier intervention.

The study's authors suggest that early screening could lead to better outcomes by allowing treatment to begin sooner. However, they also note potential limitations, including biases inherent in electronic health records and the need for careful implementation. The research is seen as a step toward integrating AI into pediatric mental health screening.