Tag: Clinical Decision Support
Reasoning models catch more disease and cry wolf more often
A Hong Kong-led benchmark of 12 medical specialties finds that step-by-step reasoning models beat standard language models on diagnosis while…
Chatbots order extra tests mainly where doctors already disagree
A Clalit and Ben-Gurion study finds six leading AI models diverge on unnecessary testing precisely where clinicians themselves disagree.
OpenEvidence moves inside the Epic chart at NYU Langone
NYU Langone Health will embed OpenEvidence's patient-aware clinical AI inside Epic, reaching clinicians across 12.4 million outpatient visits a year.
Nine AI models missed doses a doctor would not
Doximity tested nine leading AI models on 8,500 dosing questions and found serious errors in roughly one answer in fifty.
Breast ultrasound AI could spare six in ten benign biopsies
A retrospective study finds an FDA-cleared AI tool put 60.6% of benign BI-RADS 4 breast lesions in lower-risk tiers, though…
Vera Health raises $28.5M to put evidence grades on clinical answers
Vera Health closed a $28.5M Series A on October 27, with Threshold Ventures in the lead and Optum Ventures, Mantis…
Hopkins hands OpenEvidence its curriculum, not just its clinicians
Johns Hopkins Medicine and OpenEvidence will redesign medical education and build evidence-grounded clinical tools together, the latest in a string…
A single mistyped probability nearly broke a transparent cancer model
Leipzig's Bayesian network matched tumor board decisions 95.1% of the time, and an audit of its own tables explains why.