Duke Health researchers have developed an artificial intelligence tool capable of identifying the risk of Attention-Deficit/Hyperactivity Disorder (ADHD) in children years before a standard diagnosis is typically made. Published in Nature Mental Health, the study demonstrates how analyzing routine electronic health records can reveal hidden patterns in developmental, behavioral, and clinical data that signal future risk. This approach aims to enable earlier evaluation and intervention, addressing the critical issue where millions of children remain undiagnosed despite exhibiting early signs, often missing out on support that could significantly improve long-term outcomes. The research team analyzed electronic health records from over 140,000 children, both those with and without ADHD. They trained a specialized AI model to review medical history from birth through early childhood. The system learned to recognize specific combinations of events that frequently precede a formal diagnosis. The model demonstrated high accuracy in estimating future ADHD risk for children aged five and older. Notably, the tool maintained consistent performance across diverse patient characteristics, including sex, race, ethnicity, and insurance status. Dr. Elliot Hill, the lead author and a data scientist at the Duke University School of Medicine, emphasized the potential of the vast amount of data already present in health records. The goal was to determine if patterns hidden within this information could predict ADHD well in advance. However, the researchers clarified that the tool is not a diagnostic device. Instead, it serves as an alert system to help primary care providers identify children who may benefit from closer monitoring or an earlier referral to a specialist. Dr. Matthew Engelhard, the study's senior author, stressed that the AI does not act as a doctor. Its primary function is to assist clinicians in focusing their time and resources, ensuring that children who need help do not fall through the cracks or endure years of uncertainty. Early identification facilitates earlier diagnosis, which is linked to better academic, social, and health outcomes. Dr. Naomi Davis, a study author and associate professor in psychiatry, highlighted that connecting families with timely, evidence-based interventions is essential for helping children achieve their goals and laying a foundation for future success. The team, which also includes Dr. De Rong Loh, Benjamin A. Goldstein, and Geraldine Dawson, noted that while the findings are promising, further studies are necessary before such tools can be widely adopted in clinical settings. This project is part of a broader effort by Hill and Engelhard to explore the use of AI models for predicting risks and causes of mental illness in adolescents. By leveraging existing data, the researchers hope to transform routine health visits into opportunities for early detection and proactive care for children at risk of ADHD.