🤖 Patients ask AI: "What's wrong with me?"
🩺 Doctors ask AI: "What should I do about it?"
JAMA analyzed 271 million queries from 952,008 US clinicians on OpenEvidence (Jan 2024–June 2026): 51.9% were for clinical decision-making, 33.1% for treatment/management, only 13.9% for diagnosis.
Career stage drives use: <10 years out = 24.3% of the workforce but 61.8% of queries (ratio 2.54); 20+ years out = 53% of the workforce but only 16.5% of queries (0.31). With 90% monthly retention by 2026, this is habit, not experimentation.
📊 Cognitive fields lead — Oncology/Hematology 2.00, Internal Medicine 1.88, Family Medicine 1.80 — while procedural fields lag: Ophthalmology 0.23, Radiology 0.24, Orthopedics 0.24.
NPs/PAs averaged 44.5 queries per active user-month vs. 37.1 for physicians, and were >40% of users. Many queries also crossed specialties: 70% for surgeons, 73% for anesthesiologists, 35% for oncologists.
💡 AI is already embedded in treatment decisions, especially for early-career clinicians.
References:
https://jamanetwork.com/journals/jama/fullarticle/2855491