Physician in the Loop
Subscribe

Utah lets AI write first acne prescriptions; CommonSpirit credits AI with $200 million; most pathology AI lacks public data; AI intake beats eye residents

A prescription pad with a pen standing over it, no hand holding it, its amber nib trailing a glowing amber line, beside a tube of skin cream.
THE SHORT READ

Utah approved a pilot Oct. 5 that lets Nolla Health's software issue first-time prescriptions for acne, with physicians reviewing each prescription in advance only for the first 100 patients, MobiHealthNews reported.

The state's earlier sandbox pilots let AI renew existing prescriptions; Nolla's is the first to let AI start a treatment, according to Becker's. STAT reported that Utah also plans to require third-party audits of the claims companies make in its AI sandbox.

Elsewhere, CommonSpirit Health's chief financial officer told investors the system has captured $200 million in financial value from AI. A Helsinki review found no public performance data for 61% of authorized AI products in pathology and hematology morphology, and a randomized trial in China found a language model agent took more complete eye-clinic histories than ophthalmology residents.

THE DETAILS
  • Regulation Oct 5, 2026
    Utah lets Nolla Health's AI issue first-time acne prescriptions from a physician-approved list of topical drugs
    The pilot, approved Oct. 5 under Utah's AI regulatory sandbox, covers adults 18 and older with mild to moderate acne who complete an intake questionnaire and a face scan; the app costs $4.99 a month, with medicines extra, and oral drugs and isotretinoin are excluded, MobiHealthNews reported. Two physicians review prescriptions for at least the first 100 patients before they are sent, physicians then review the next 500 or more weekly after issuance for at least eight weeks, and later a monthly review covers at least 10% of patients, according to the report. "The AI can't improvise," said Dr. Zaid Fadul, the company's chief medical adviser. Clinicians agreed with the AI's recommendations in more than 96% of real-world cases, according to the company's figures reported by Becker's; no peer-reviewed results have been published. The state's earlier pilots, Doctronic and Legion Health, renew existing prescriptions, and STAT reported that Utah plans third-party audits of sandbox companies' claims.
    Sources: MobiHealthNews, Becker's Hospital Review, STAT
  • Money Oct 2, 2026
    CommonSpirit says AI has produced $200 million in financial value while AI spending stays under 0.75% of its IT budget
    Michael Browning, CommonSpirit Health's chief financial officer, said on the system's fiscal 2026 investor call Oct. 2 that it has captured $200 million in financial value from AI, tracked from pipeline review through post-implementation audits, Becker's reported. Direct AI spending is under 0.75% of the IT budget, and enterprise token costs run about $25,000 a month, according to the report. A 36-member governance committee has reviewed more than 122 enterprise AI uses and rejected 20 over bias, safety risk or vendor overlap, and the clinical ethics team holds a veto over any deployment, Browning said. He said health care remains "roughly 80% people, 15% process and only 5% technology" and that AI does not change the ratio. The system did not break down the $200 million by tool in the report.
    Source: Becker's Hospital Review
  • Evidence Oct 5, 2026
    No public performance data found for 47 of 77 authorized pathology and hematology AI products, Helsinki review finds
    Researchers at Helsinki University Hospital and the University of Helsinki reviewed 77 AI diagnostic products from 29 manufacturers that hold a CE mark or FDA authorization, 68 in digital pathology and nine in hematology morphology, and found publicly available performance data for 30, or 39%, according to the paper in npj Digital Medicine published Oct. 5. Evidence was available for all eight products authorized in both the European Union and the United States, for 34% of pathology products and for 78% of hematology morphology products, the authors reported. Published evaluations used 15 different performance metrics, which limits comparison between products. The authors noted that performance evaluation results do not have to be published and described "a gap between evidence availability and regulatory authorization."
    Source: npj Digital Medicine
  • Evidence Oct 5, 2026
    Language model agent takes fuller eye-clinic histories than ophthalmology residents in 172-patient randomized trial
    At Zhongshan Ophthalmic Center in Guangzhou, China, 172 patients were randomized to pre-consultation history-taking by a large language model agent or by ophthalmology residents, and the agent scored 17.7 points higher on a history-quality scale after adjustment, according to the trial published Oct. 5 in npj Digital Medicine. Patients rated the agent higher for empathy, and its interviews ran a median 11.2 minutes against 3.1 for residents, the authors reported. Diagnoses made from the history alone were more accurate after the agent's interviews (F1 score 0.85 vs 0.68), while a physical examination improved the residents' accuracy more. The trial ran at one tertiary hospital in 2025, and the authors said results may not carry over to other specialties or settings.
    Source: npj Digital Medicine