Family medicine ·
Microsoft AI outdiagnoses generalist physicians in Science study; patients describe ambient AI scribe benefits; LLMs recommend some low-value care
Listen to this post
read by AI voices · all episodesTHE SHORT READ
Microsoft's MAI-DxO system correctly diagnosed 83.2% of complex New England Journal of Medicine cases versus 36.1% for generalist physicians using reference tools, according to a study published Oct. 22 in Science. Patients whose clinicians used ambient artificial intelligence (AI) scribes described more thorough visit notes in a qualitative study, and six large language models recommended low-value care in 5.4% of cases in an Annals of Internal Medicine study.
THE DETAILS
- AI Oct 22, 2026Microsoft's MAI-DxO system correctly diagnosed 83.2% of complex New England Journal of Medicine cases, compared with 36.1% for generalist physicians using search and reference tools and 22.3% without them, according to a study published Oct. 22 in Science. The physician comparison used a held-out set of 56 cases. On 325 emergency department admissions from the MIMIC-IV-ED dataset, it reached 80.3% accuracy versus 46.8% for physicians. The authors wrote that the text-only comparisons "do not establish superiority over specialists, multidisciplinary teams, or real-world clinical practice." All authors are Microsoft employees with equity, and Microsoft funded the research, Becker's Hospital Review reported.Sources: Microsoft AI, Becker's Hospital Review
- AI Oct 19, 2026Researchers at the University of Wisconsin at Madison interviewed 20 patients from ambulatory specialties, including primary care, whose clinicians were participating in a pragmatic randomized clinical trial, in a qualitative study of ambient artificial intelligence (AI) scribes published Oct. 19 in the Journal of General Internal Medicine. Patients described increased respect for their preferences, such as fewer distractions, improved thoroughness of visit notes and more time spent answering their questions. Comfort was tied to factors including exposure to similar technologies and positive pre-existing patient-clinician relationships. The authors said ambient AI scribes may improve the patient experience.
- AI Oct 19, 2026Six commercial large language models (LLMs) generated 4,050 responses to 225 scenarios built from 15 Choosing Wisely recommendations, in a study by a team from Clalit Health Services in Tel Aviv published Oct. 19 in Annals of Internal Medicine. Overuse was recommended in 42.8% of responses on thyroid ultrasound for an incidental subcentimeter nodule and 28.3% on routine electrocardiography in low-risk adults, according to AuntMinnie. False restraint occurred in 0.6% of cases and was highest for prostate-specific antigen screening, at 3.3%.Sources: Annals of Internal Medicine, AuntMinnie
The weekly innovation roundup for family medicine, published every Sunday for the week before; every item links to its sources. Latest family medicine news · All specialties