Parents Use AI to Help Diagnose 1-in-100,000 Okur-Chung Syndrome in 7-Year-Old Daughter
Updated
Updated · The Boston Globe · Sep 12
Parents Use AI to Help Diagnose 1-in-100,000 Okur-Chung Syndrome in 7-Year-Old Daughter
1 articles · Updated · The Boston Globe · Sep 12
Summary
Low-antibody test results in January 2025 pushed Hilary and Matt Eaton to feed their daughter Olivia’s symptoms and lab data into ChatGPT after years of inconclusive doctor visits.
ChatGPT first suggested common variable immunodeficiency, giving the family a lead to discuss with doctors; later, after the Eatons spent about $4,500 on whole-genome sequencing, it helped them narrow genetic clues behind Olivia’s broader symptoms.
Additional testing ordered by physicians pointed to Okur-Chung neurodevelopmental syndrome — a disorder estimated to affect 1 in 100,000 people — and doctors believe the mutation may also impair Olivia’s antibody production.
Olivia began monthly antibody infusions last fall and later started anti-seizure medication after an EEG; her parents say she has since missed just three days of school and is finally sleeping through the night.
The case highlights both AI’s growing role in medicine — 25% of Americans report using chatbots for symptom checks — and its risks, as the Eatons also encountered false alarms from erroneous genome data and relied on doctors to confirm findings.