AI tool can predict ADHD in children years before official diagnosis

New research shows that artificial intelligence can help identify children at risk of attention deficit hyperactivity disorder (ADHD) years before an official diagnosis is made. This was reported by Euronews. According to scientists, such an approach could allow for earlier intervention and support for children who need it.

The study was conducted by researchers from Duke Health, and the results were published in Nature Mental Health.

Artificial intelligence can help identify children at risk of ADHD years before a diagnosis is made, according to new data. ADHD is one of the most common mental disorders, affecting approximately 8% of children and teenagers, with symptoms including difficulties with concentration, restlessness, and impulsivity. However, many cases remain undiagnosed for years, which means missed opportunities for early support, even when warning signs are already present.

In the new study, scientists found that artificial intelligence systems can analyze routine electronic health records to calculate the probability of a child developing ADHD long before the standard time of diagnosis. The data shows that hidden patterns in daily medical information can help doctors identify children who would benefit from earlier assessment and follow-up.

“We have an extremely rich source of information in electronic health records,” says Elliott Hill, lead author of the study and a data analysis specialist at the Duke University School of Medicine.

“The idea was to see if hidden patterns in this data could help us predict which children would later be diagnosed with ADHD, long before it usually happens.”

Researchers analyzed the health records of over 140,000 children—both with and without ADHD—and trained an artificial intelligence system to recognize patterns from birth through early childhood.

The system learns to detect combinations of developmental, behavioral, and clinical events that often appear years before an ADHD diagnosis is made.

The results show high accuracy in risk assessment for children aged five and older, with results being consistent regardless of gender, race, ethnicity, or health insurance.

Experts emphasize that earlier detection can lead to earlier diagnosis and support, which is linked to better educational, social, and health outcomes for children with ADHD.

“Children with ADHD can face serious difficulties when their needs are not understood and the necessary support is not provided,” says Naomi Davis, associate professor in the Department of Psychiatry and Behavioral Sciences and co-author of the study.

“Connecting families with timely and evidence-based interventions is key so that children can achieve their goals and build a foundation for future success.”

The researchers emphasize that the tool is not designed to replace doctors or to make diagnoses independently.

“This is not artificial intelligence replacing the doctor,” says Matthew Engelhard of the Department of Biostatistics and Bioinformatics at Duke and senior author of the study.

“It is a tool that helps clinicians direct their time and resources more effectively, so that children who need help are not missed or left waiting for years for answers.”

The team adds that similar artificial intelligence approaches are also being explored to better understand the risks and causes of mental illness in adolescents.

According to the National Health Service (NHS), common symptoms of ADHD in children include being easily distracted, difficulties in listening, forgetting daily tasks, and high energy levels, such as constant movement or tapping of hands and feet.

It is also believed that the disorder often remains underestimated in girls compared to boys, as they more frequently exhibit symptoms related to inattention, which are harder to recognize.