Could the key to earlier autism diagnoses be hidden in the digital trail of pediatric health checkups?

A new study from the Duke Center for Autism and Brain Development explores how electronic health records (EHRs) might help clinicians identify signs of autism spectrum disorder (ASD) earlier — long before a formal diagnosis is made.

The Research at a Glance:

• Researchers analyzed over 14,000 EHRs from a diverse population of children.

• Subtle patterns emerged in early health visits — such as motor delays, feeding issues, or certain parental concerns — that could correlate with later autism diagnoses.

• These early flags typically appeared years before the average diagnosis age of 4.

• By using machine learning to identify these trends, the team hopes to accelerate early detection, especially for children who may otherwise fall through diagnostic gaps.

Why It Matters:

Early diagnosis of autism opens the door to critical early intervention, which can improve communication, behavior, and long-term developmental outcomes. This research could help transform pediatric health systems into more proactive autism detection networks — especially in underserved or rural areas with limited specialist access.

The future may see AI-driven tools embedded in health systems, quietly scanning EHRs for patterns that prompt earlier autism screenings and support referrals.