AI News
01 May 2026
Read 8 min
Predicting childhood ADHD from medical records speeds care
Predicting childhood ADHD from medical records helps clinicians flag at-risk children for earlier care
New Duke Health research shows that predicting childhood ADHD from medical records is possible years before a formal diagnosis. An AI model scanned routine electronic health records for patterns and accurately estimated risk in children age 5 and older, helping pediatric teams flag kids who may need earlier evaluation and support.
Attention-deficit/hyperactivity disorder affects many children, yet signs often go unnoticed for years. This new study shows that everyday health data already stored in the clinic can help. By reading long-term patterns in checkups and notes, AI can point primary care teams toward kids who may benefit from early screening and follow-up.
Prompt timely conversations between parents and pediatricians
Support earlier referrals for behavioral or developmental assessments
Help clinics focus limited time and resources on children at higher risk
Reduce the chance that symptoms are missed or dismissed over years
Developmental milestone notes and growth charts
Behavior or attention concerns recorded during well visits
Referrals to therapy or behavioral health
Sleep, nutrition, and activity reports from caregivers
School or caregiver observations documented in the chart
Family medical history and prior conditions
Medication history, allergies, and follow-up patterns
By predicting childhood ADHD from medical records, the model gives primary care teams a head start. It highlights who might benefit from screening now, not years later.
Share clear examples of attention or behavior concerns with your pediatrician
Bring school notes or caregiver observations to visits
Keep a simple log of sleep, routines, and challenging moments
Ask if your clinic uses or plans to use tools that scan records for risk
Review longitudinal notes for repeated attention or activity concerns
Use standardized screeners when concerns appear
Consider early referrals to specialists or school-based services
Discuss privacy, consent, and family preferences when using AI risk tools
Validate performance in diverse clinics and regions
Monitor fairness across demographic groups over time
Offer clear explanations and opt-in consent for families
Integrate tools into workflows without adding burden
Track outcomes to ensure earlier flags lead to earlier, effective support
The study, published April 27, 2026, in Nature Mental Health, was led by Elliot Hill with senior author Matthew Engelhard and colleagues Naomi Davis, De Rong Loh, Benjamin A. Goldstein, and Geraldine Dawson. It was supported by the National Institute of Mental Health and the National Center for Advancing Translational Sciences.
In short, predicting childhood ADHD from medical records can help kids get evaluated sooner, reduce waiting years for answers, and connect families to proven supports when they matter most.
(p(Sou Source: https://www.news-medical.net/news/20260427/AI-tools-can-predict-ADHD-risk-years-before-a-formal-diagnosis.aspx)
How the study’s AI model reads routine care
Data and training
Researchers at Duke Health trained a model on electronic health records from more than 140,000 children, both with and without ADHD. The model reviewed information from birth through early childhood. It learned patterns of developmental, behavioral, and clinical events that often appear before a formal diagnosis.Performance and fairness
The model accurately estimated future ADHD risk for children age 5 and older. Importantly, it worked consistently across sex, race, ethnicity, and insurance status. It does not diagnose. Instead, it helps clinicians decide who might need closer monitoring or a referral to a specialist sooner.Why early flagging changes outcomes
Early support can improve school progress, peer relationships, and overall health for children with ADHD. Long waits for answers can add stress for families and widen learning gaps. A risk tool that scans records during routine visits can:Predicting childhood ADHD from medical records in everyday care
This approach uses information that clinics already collect, which can make it easier to adopt at scale. While the study did not list specific inputs, such tools can draw on routine record elements, such as:What parents and clinicians can do today
For parents
For clinicians
Guardrails and next steps
The research team stresses that this is not an “AI doctor.” It is a decision support tool meant to augment clinical judgment. Before broad use, health systems should:For more news: Click Here
FAQ
Q: What did the Duke Health study find?
A: Duke Health researchers found that artificial intelligence tools can analyze routine electronic health records to estimate a child’s risk of developing ADHD years before a typical diagnosis. The study used data from more than 140,000 children and showed the approach could help pediatric teams flag kids for earlier evaluation and follow-up.
Q: How did the AI model identify children at risk?
A: The researchers trained a specialized model on longitudinal electronic health records from birth through early childhood so it could recognize combinations of developmental, behavioral, and clinical events that often appear before a formal diagnosis. This approach demonstrates predicting childhood ADHD from medical records by reading long-term patterns in routine care data.
Q: Does the AI tool provide a formal diagnosis of ADHD?
A: No, 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. It is intended as decision support to help clinicians focus time and resources, not as an “AI doctor.”
Q: At what age did the model perform well and was it equitable?
A: The model was highly accurate at estimating future ADHD risk in children age 5 and older and showed consistent performance across sex, race, ethnicity, and insurance status. The researchers nonetheless emphasize the need for further studies and validation before broad clinical use.
Q: How could predicting childhood ADHD from medical records change care for families?
A: Earlier identification for screening could lead to earlier diagnosis and therefore earlier support, which the study notes is linked to better academic, social, and health outcomes for children with ADHD. The tool can prompt timely conversations, support earlier referrals, and help clinics focus limited time and resources so children don’t wait years for answers.
Q: What can parents do now if they have concerns about attention or behavior?
A: Parents can share clear examples of attention or behavior concerns with their pediatrician, bring school notes or caregiver observations to visits, and keep a simple log of sleep, routines, and challenging moments. They can also ask whether their clinic uses or plans to use tools that scan records for risk to support earlier screening.
Q: What should clinicians consider when using AI-based risk tools?
A: Clinicians should review longitudinal notes for repeated attention or activity concerns, use standardized screeners when concerns appear, and consider early referrals to specialists or school-based services as appropriate. They should also discuss privacy, consent, and family preferences when implementing tools that scan records.
Q: What guardrails and next steps do researchers recommend before wide adoption?
A: The team recommends validating performance in diverse clinics and regions, monitoring fairness across demographic groups over time, and offering clear explanations and opt-in consent for families. They also advise integrating tools into workflows without adding burden and tracking outcomes to ensure earlier flags lead to earlier, effective support.
Contents
