AI tools can predict ADHD risk years before formal diagnosis

Attention-deficit/hyperactivity disorder (ADHD) affects millions of children, yet many go years without diagnosis, missing the opportunity for early help, which can alter long-term outcomes even when early symptoms are present.

In a new study, Duke Health researchers found that artificial intelligence tools can analyze routine electronic health records to accurately predict a child’s risk of developing ADHD, years before a typical diagnosis. By reviewing patterns in everyday medical data, the approach can help identify children who may benefit from earlier assessment and follow-up.

Research, published in nature mental health On April 27, it highlighted how powerful insights can come from information already collected during routine health care visits to help primary care providers make quicker decisions.

We have this incredibly rich source of information in the electronic health record. The idea was to see if hidden patterns in that data could help us predict which children might later be diagnosed with ADHD, before that diagnosis typically occurs.

Elliot Hill, lead author of the study and data scientist, Department of Biostatistics and Bioinformatics, Duke University School of Medicine

To reach the findings, researchers analyzed the electronic health records of more than 140,000 children with and without ADHD. They trained a special AI model to look at medical history from birth to childhood. The model learned to recognize combinations of developmental, behavioral and clinical phenomena that often appear years before an ADHD diagnosis is made.

The model was highly accurate in predicting future ADHD risk in children ages 5 and older, with performance consistent with patient characteristics such as gender, race, ethnicity, and insurance status.

Importantly, this tool does not diagnose. It identifies children who may benefit from close attention by their pediatric primary care provider or earlier referral for ADHD evaluation by a specialist.

“This is not an AI doctor,” said Matthew Engelhardt, MD, PhD, in Duke’s department of biostatistics and bioinformatics and senior author of the study. “It’s a tool to help physicians focus their time and resources, so that kids who need help don’t have to suffer or wait years for answers.”

The researchers note that early identification for screening can lead to earlier diagnosis and therefore earlier support, which is associated with better academic, social, and health outcomes for children with ADHD. They also emphasize the need for further studies before using such devices in clinical settings.

“Children with ADHD can really struggle when their needs are not understood and do not get adequate support,” said study author Naomi Davis, PhD, associate professor in the Department of Psychiatry and Behavioral Sciences. “Connecting families to timely, evidence-based interventions is essential to helping them achieve their goals and laying the foundation for future success.”

Hill and Engelhardt have also researched the use of AI models in predicting the potential risks and causes of mental illness in adolescents.

In addition to Hill Engelhard and Davis, the study’s authors include De Rong Loh, Benjamin A. Goldstein and Geraldine Dawson.

The study was supported by grants from the National Institute of Mental Health (K01-MH127309, UL1 TR002553) and the National Center for Advancing Translational Sciences.