AI Identified a 25‑Year Missed Sleep‑Apnea Case
An Economic Times write‑up described how an AI conversation spotted undiagnosed sleep apnea after 25 years of missed diagnoses, illustrating AI’s diagnostic reach—but MedCity cautions outcomes depend on follow‑up and human coordination to act on those predictions. AI can find the signal; systems must close the loop. (economictimes.indiatimes.com) (medcitynews.com)
The Reddit thread identified Anthropic’s Claude as the chatbot used in the exchange, a detail reported by The Economic Times and Moneycontrol on March 28–29, 2026. (economictimes.indiatimes.com) (moneycontrol.com) News accounts describe the patient as a 62‑year‑old in India with kidney failure on regular dialysis, diabetes, hypertension and a stroke history, and specifically note headaches that occurred only when lying down. (ndtv.com) (economictimes.indiatimes.com) Several outlets report the chatbot’s follow‑up questions prompted the family to acknowledge decades of loud snoring and daytime sleepiness, which led them to arrange a formal polysomnography (sleep study). (newsbytesapp.com) (thecsrjournal.in) Published coverage states the sleep study confirmed obstructive sleep apnea and clinicians prescribed CPAP therapy, and the Reddit poster said the patient’s positional headaches resolved after treatment. (thecsrjournal.in) (dev.to) Reporting across India and international news aggregators on March 28–29, 2026 prompted clinicians and commentators to underline that consumer chatbots can synthesize multi‑specialty histories but cannot substitute for confirmatory testing or clinician assessment. (economictimes.indiatimes.com) (msn.com) MedCity’s March 2026 commentary echoed those cautions, arguing that realizing AI’s access benefits requires operational changes — routing model‑generated signals into validated clinical workflows, human review, and referral pathways so that “systems close the loop” on predictions. (medcitynews.com)
