AI Predicts ADHD Risk Years Before Diagnosis by Mining Electronic Health Records

AI Turn "Hidden Patterns" Into ADHD Insights - Neuroscience News

Summary: A new study demonstrates that artificial intelligence can accurately estimate a child's risk of developing ADHD years before a clinical diagnosis occurs. By mining "hidden patterns" in routine Electronic Health Records (EHR) from birth through early childhood, the AI identifies combinations of developmental and behavioral markers that human clinicians might overlook during brief visits. This tool is designed to act as a "clinical safety net," ensuring at-risk children receive early evaluations and support during critical developmental windows. Attention-deficit/hyperactivity disorder (ADHD) affects millions of children, yet many go years without a diagnosis, missing the chance for early support that can change long-term outcomes even when early signs are present. In a new study, Duke Health researchers found that artificial intelligence tools can analyze routine electronic health records to accurately estimate a child's risk of developing ADHD years before a typical diagnosis. By reviewing patterns in everyday medical data, the approach could help flag children who may benefit from earlier evaluation and follow-up. The research, published in Nature Mental Health on April 27, highlights how powerful insights can come from information already collected during regular health care visits to help support early decision making by primary care providers. "We have this incredibly rich source of information sitting in electronic health records," said Elliot Hill, lead author of the study and data scientist in the Department of Biostatistics & Bioinformatics at Duke University School of Medicine. "The idea was to see whether patterns hidden in that data could help us predict which children might later be diagnosed with ADHD, well before that diagnosis usually happens." To arrive at the findings, researchers analyzed electronic health records from more than 140,000 children, with and without ADHD. They trained a specialized AI model to look at medical history from birth through early childhood. The model learned to recognize combinations of developmental, behavioral, and clinical events that often appeared years before an ADHD diagnosis was made. The model was highly accurate at estimating future ADHD risk in children age 5 and older, with consistent performance across patient characteristics like sex, race, ethnicity, and insurance status. Importantly, the tool does not make a diagnosis. It identifies children who may benefit from closer attention by their pediatric primary care provider or an earlier referral for ADHD assessment by a specialist. "This is not an AI doctor," said Matthew Engelhard, M.D., Ph.D., in Duke's Department of Biostatistics & Bioinformatics, and senior author of the study. "It's a tool to help clinicians focus their time and resources, so kids who need help don't fall through the cracks or wait years for answers." The researchers note that earlier identification for screening could lead to earlier diagnosis and therefore earlier support, which is linked to better academic, social, and health outcomes for children with ADHD. They also emphasize the need for further studies before such tools are used in clinical settings. "Children with ADHD can really struggle when their needs aren't understood and adequate supports are not in place," said study author, Naomi Davis, Ph.D., associate professor in the Department of Psychiatry and Behavioral Sciences. "Connecting families with timely, evidence-based interventions is essential for helping them achieve their goals and laying a foundation for future success." Hill and Engelhard have also researched the use of AI models in predicting potential risks and causes for mental illness in adolescents. In addition to Hill Engelhard, and Davis, the authors for this study include De Rong Loh, Benjamin A. Goldstein, and Geraldine Dawson. Funding: The study was supported by grants from the National Institute of Mental Health (K01-MH127309, UL1 TR002553) and National Center for Advancing Translational Sciences. Author: Stephanie Lopez Source: Duke University Contact: Stephanie Lopez - Duke University Image: The image is credited to Neuroscience News Original Research: Closed access. "Fetal and postnatal metal metabolism-related changes in brain function are associated with childhood behavioral deficits" by Elliot D. Hill, De Rong Loh, Naomi O. Davis, Benjamin A. Goldstein, Geraldine Dawson & Matthew Engelhard. Nature Mental Health DOI:10.1038/s44220-026-00628-2 Abstract Attention deficit hyperactivity disorder (ADHD) is a common neurodevelopmental condition that can negatively impact long-term outcomes for individuals. Early diagnosis is critical, yet demographic and clinical disparities can delay detection. Using electronic health records (EHRs) from a cohort of over 720,000 patients, we pretrained an EHR foundation model. We then fine tuned it to predict the likelihood of ADHD diagnosis and timing from birth until age 9 years in a pediatric cohort of over 140,000 patients. By age 5 years, the model achieved a time-dependent area under the receiver operating characteristic curve of 0.92 at a 4-year time horizon. Overall, the model maintained its performance across patients with differing demographics, including sex, race, ethnicity and insurance status. Our feature importance analysis found that ADHD was strongly associated with developmental, behavioral and psychiatric conditions. Our results suggest that EHR-based predictive models could help providers reliably identify children with ADHD in a timely manner.

AI tools can predict ADHD risk years before a formal diagnosis

Duke University Medical CenterApr 27 2026 Attention-deficit/hyperactivity disorder (ADHD) affects millions of children, yet many go years without a diagnosis, missing the chance for early support that can change long-term outcomes even when early signs are present. In a new study, Duke Health researchers found that artificial intelligence tools can analyze routine electronic health records to accurately estimate a child's risk of developing ADHD years before a typical diagnosis. By reviewing patterns in everyday medical data, the approach could help flag children who may benefit from earlier evaluation and follow-up. The research, published in Nature Mental Health on April 27, highlights how powerful insights can come from information already collected during regular health care visits to help support early decision making by primary care providers. We have this incredibly rich source of information sitting in electronic health records. The idea was to see whether patterns hidden in that data could help us predict which children might later be diagnosed with ADHD, well before that diagnosis usually happens." Elliot Hill, lead author of the study and data scientist, Department of Biostatistics & Bioinformatics, Duke University School of Medicine To arrive at the findings, researchers analyzed electronic health records from more than 140,000 children, with and without ADHD. They trained a specialized AI model to look at medical history from birth through early childhood. The model learned to recognize combinations of developmental, behavioral, and clinical events that often appeared years before an ADHD diagnosis was made. The model was highly accurate at estimating future ADHD risk in children age 5 and older, with consistent performance across patient characteristics like sex, race, ethnicity, and insurance status. Importantly, the tool does not make a diagnosis. It identifies children who may benefit from closer attention by their pediatric primary care provider or an earlier referral for ADHD assessment by a specialist. "This is not an AI doctor," said Matthew Engelhard, M.D., Ph.D., in Duke's Department of Biostatistics & Bioinformatics, and senior author of the study. "It's a tool to help clinicians focus their time and resources, so kids who need help don't fall through the cracks or wait years for answers." The researchers note that earlier identification for screening could lead to earlier diagnosis and therefore earlier support, which is linked to better academic, social, and health outcomes for children with ADHD. They also emphasize the need for further studies before such tools are used in clinical settings. "Children with ADHD can really struggle when their needs aren't understood and adequate supports are not in place," said study author, Naomi Davis, Ph.D., associate professor in the Department of Psychiatry and Behavioral Sciences. "Connecting families with timely, evidence-based interventions is essential for helping them achieve their goals and laying a foundation for future success." Hill and Engelhard have also researched the use of AI models in predicting potential risks and causes for mental illness in adolescents. In addition to Hill Engelhard, and Davis, the authors for this study 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 National Center for Advancing Translational Sciences. Duke University Medical Center Journal reference: Hill, E. D., et al. (2026). Early attention deficit hyperactivity disorder prediction from longitudinal electronic health records. Nature Mental Health. DOI: 10.1038/s44220-026-00628-2. https://www.nature.com/articles/s44220-026-00628-2

AI Tool May Spot ADHD Years Before Children Are Diagnosed | Newswise

Newswise -- DURHAM, N.C. -- Attention-deficit/hyperactivity disorder (ADHD) affects millions of children, yet many go years without a diagnosis, missing the chance for early support that can change long-term outcomes even when early signs are present. In a new study, Duke Health researchers found that artificial intelligence tools can analyze routine electronic health records to accurately estimate a child's risk of developing ADHD years before a typical diagnosis. By reviewing patterns in everyday medical data, the approach could help flag children who may benefit from earlier evaluation and follow-up. The research, published in Nature Mental Health on April 27, highlights how powerful insights can come from information already collected during regular health care visits to help support early decision making by primary care providers. "We have this incredibly rich source of information sitting in electronic health records," said Elliot Hill, lead author of the study and data scientist in the Department of Biostatistics & Bioinformatics at Duke University School of Medicine. "The idea was to see whether patterns hidden in that data could help us predict which children might later be diagnosed with ADHD, well before that diagnosis usually happens." To arrive at the findings, researchers analyzed electronic health records from more than 140,000 children, with and without ADHD. They trained a specialized AI model to look at medical history from birth through early childhood. The model learned to recognize combinations of developmental, behavioral, and clinical events that often appeared years before an ADHD diagnosis was made. The model was highly accurate at estimating future ADHD risk in children age 5 and older, with consistent performance across patient characteristics like sex, race, ethnicity, and insurance status. Importantly, the tool does not make a diagnosis. It identifies children who may benefit from closer attention by their pediatric primary care provider or an earlier referral for ADHD assessment by a specialist. "This is not an AI doctor," said Matthew Engelhard, M.D., Ph.D., in Duke's Department of Biostatistics & Bioinformatics, and senior author of the study. "It's a tool to help clinicians focus their time and resources, so kids who need help don't fall through the cracks or wait years for answers. The researchers note that earlier identification for screening could lead to earlier diagnosis and therefore earlier support, which is linked to better academic, social, and health outcomes for children with ADHD. They also emphasize the need for further studies before such tools are used in clinical settings. "Children with ADHD can really struggle when their needs aren't understood and adequate supports are not in place," said study author, Naomi Davis, Ph.D., associate professor in the Department of Psychiatry and Behavioral Sciences. "Connecting families with timely, evidence-based interventions is essential for helping them achieve their goals and laying a foundation for future success." Hill and Engelhard have also researched the use of AI models in predicting potential risks and causes for mental illness in adolescents. In addition to Hill Engelhard, and Davis, the authors for this study 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 National Center for Advancing Translational Sciences.