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
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.

Insights

Could a simple chatbot hold the key to solving medical mysteries that baffle top doctors?
What hidden dangers lurk when desperate parents upload their children's raw DNA into commercial AI?