summary: A new study has demonstrated that artificial intelligence can accurately estimate a child’s risk of developing ADHD years before a clinical diagnosis is made. By mining “hidden patterns” in routine electronic health records (EHRs) from birth to infancy, AI identifies combinations of developmental and behavioral markers that human clinicians may miss during brief visits.
The tool is designed to act as a “clinical safety net”, ensuring at-risk children receive early assessment and support during critical developmental periods.
important facts
- Dataset: The researchers analyzed the following medical history: over 140,000 childrento create a large-scale comparative baseline of people with and without ADHD.
- Early detection: AI models analyze data from birth to accurately estimate future risk at each age. 5well before the average age of diagnosis.
- Fair performance: A key finding is that model accuracy is consistent across different demographics, including: Gender, race, ethnicity, insurance statussuggesting that it may help reduce existing disparities in ADHD care.
- Support, not diagnosis: This tool is clearly not an “AI doctor”. The purpose is to flag children who should be prioritized for testing by a primary care provider or specialist.
- Better results: Early detection enables evidence-based interventions before children fall behind, which is directly linked to improved academic, social, and long-term health outcomes.
sauce: duke university
Attention-deficit/hyperactivity disorder (ADHD) affects millions of children, but it goes undiagnosed for years, missing out on early support that could change long-term outcomes, even if early signs are present.
In a new study, Duke Health researchers found that an artificial intelligence tool can accurately estimate a child’s risk of developing ADHD years before a typical diagnosis is made by analyzing routine electronic health records. By reviewing patterns in routine medical data, this approach could help flag children who may benefit from early assessment and follow-up.
This study natural mental health The April 27 paper highlights how powerful insights can be derived from information already collected during routine medical visits to support early decision-making by primary care providers.
“We have this incredibly rich source of information stored in electronic medical records,” said Elliott Hill, lead author of the study and a data scientist in the Department of Biostatistics and Bioinformatics at Duke University School of Medicine.
“The aim was to see if patterns hidden in that data could help predict which children were likely to later be diagnosed with ADHD, usually well before the diagnosis was made.”
To arrive at this finding, researchers analyzed electronic health records of more than 140,000 children with and without ADHD. They trained a specialized AI model to examine medical history from birth to infancy. The model has learned to recognize a combination of developmental, behavioral, and clinical events that often appear years before an ADHD diagnosis.
The model was highly accurate in estimating future ADHD risk in children aged 5 years and older, with consistent performance across patient characteristics such as gender, race, ethnicity, and insurance status.
Importantly, this tool is not diagnostic. This identifies children who may benefit from close attention by a pediatric primary care provider or early referral for professional ADHD evaluation.
“This is not an AI doctor,” said Matthew Engelhard, MD, PhD, of Duke University’s Department of Biostatistics and Bioinformatics and senior author of the study. “This is a tool that allows clinicians to focus their time and resources so that children in need don’t have to fail or wait years for answers.”
Researchers note that early identification through screening may lead to earlier diagnosis and, in turn, earlier support, which may lead to better academic, social, and health outcomes for children with ADHD. They also stress that further research is needed before such tools can be used in clinical practice.
“Children with ADHD can really struggle if their needs aren’t understood and the right supports aren’t in place,” said study author Dr. Naomi Davis, associate professor in the Department of Psychiatry and Behavioral Sciences. “Connecting families with timely, evidence-based interventions is essential to helping families achieve their goals and laying the foundation for future success.”
Hill and Engelhardt also studied 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.
Funding: This research was supported by grants from the National Institute of Mental Health (K01-MH127309, UL1 TR002553) and the National Center for the Advancement of Translational Science.
Answers to key questions:
answer: What the AI sees is timing and combination Certain developmental delays, sleep disturbances, and frequent hospital visits for behavioral concerns may not seem important in themselves, but together they form a “risk sign” for ADHD.
answer: no. Lead author Dr. Matthew Engelhard emphasizes that this is a resource management tool. This helps pediatricians know which children need further testing so they don’t “fail” while waiting years for standard evaluations.
answer: Primary care providers often have very limited time with patients. AI scans thousands of pages of a child’s medical history in seconds and highlights relevant clinical trends that may have occurred years ago or by a different doctor.
Editorial note:
- This article was edited by the editors of Neuroscience News.
- Journal articles were reviewed in full text.
- Additional context added by staff.
About this AI and ADHD research news
author: Stephanie Lopez
sauce: duke university
contact: Stephanie Lopez – Duke University
image: Image credited to Neuroscience News
Original research: Closed access.
“Changes in brain function related to fetal and postnatal metal metabolism are associated with behavioral disorders in childhood” by Elliott D. Hill, de Ron Loh, Naomi O. Davis, Benjamin A. Goldstein, Geraldine Dawson, and Matthew Engelhard. natural mental health
DOI:10.1038/s44220-026-00628-2
abstract
Attention-deficit/hyperactivity disorder (ADHD) is a common neurodevelopmental disorder that can negatively impact an individual’s long-term outcomes. Early diagnosis is critical, but demographic and clinical disparities can delay detection.
Electronic health records (EHRs) from a cohort of over 720,000 patients were used to pre-train an EHR foundation model. We then fine-tuned it to predict the likelihood and timing of an ADHD diagnosis from birth to age 9 in a pediatric cohort of more than 140,000 patients.
By age 5, the model achieved a time-dependent area under the receiver operating characteristic curve of 0.92 over a 4-year time horizon. Overall, the model maintained performance across patients with different demographics, including gender, race, ethnicity, and insurance status.
Our feature importance analysis revealed that ADHD is strongly associated with developmental, behavioral, and psychiatric conditions. Our results suggest that EHR-based predictive models may help healthcare providers reliably identify children with ADHD in a timely manner.
