Daily Briefing for October 14, 2026

Show Notes

OpenEvidence joins Johns Hopkins on medical training; EHR vendors' AI disclosures fall short; Agilent buys Proscia; Suki raises $90 million

The voices on this episode are generated, and so is the script, from the day's post at Physician in the Loop. Every fact in the post links to its source.

Read the post and its sources

Get the free email briefing

Johns Hopkins Medicine and OpenEvidence said Tuesday they will redesign medical education and build clinical AI tools together, a day after Northwestern Medicine opened OpenEvidence inside its Epic record to nurses and pharmacists.

All 35 electronic health record developers reviewed in a study published Wednesday in npj Digital Medicine were certified as meeting federal AI transparency requirements, although their public risk disclosures were inconsistent and often incomplete, the authors said. Agilent Technologies agreed to buy the digital pathology software company Proscia for up to $270 million, and the ambient documentation company Suki raised $90 million.

Separately, authors in NEJM AI said Medicare's repeal of the breakthrough-device route to new technology add-on payments will likely disadvantage independent AI developers, and researchers reported retrieving 1,000 stored conversations from a patient chatbot without logging in.

In this episode

• Johns Hopkins and OpenEvidence plan to redesign medical training; Northwestern opens the tool in Epic to nurses and pharmacists
• All 35 EHR developers won certification despite inconsistent, often incomplete AI risk disclosures, npj Digital Medicine study finds
• Agilent agrees to buy Proscia, whose digital pathology software serves more than 4,000 labs, hospitals and clinics, for up to $270 million
• Suki raises $90 million and says encounters on its AI documentation tools grew more than 190% this year
• Repeal of Medicare's breakthrough-device pathway to add-on payments will likely disadvantage independent AI developers, authors say
• About 75% of Intermountain Health radiologists use Microsoft AI that drafts report impressions, system says
• Researchers retrieve 1,000 stored conversations from a public patient chatbot without a login, NEJM AI paper says
• Character.AI seeks Chapter 11 sale; Pennsylvania sued the chatbot company in May over bots claiming to be psychiatrists
• Function lab results will flow into the Google Health app, where an AI coach can use them for wellness guidance

What is Physician in the Loop?

How technological innovation is changing medicine for patients and the people who practice on the front lines.

Voice 1: This is Physician in the Loop, the daily briefing on AI in medicine. It's Wednesday, October 14th, 2026.

Voice 1: Johns Hopkins Medicine and OpenEvidence said Tuesday they will redesign medical education around pathophysiology, evidence-based medicine and clinical reasoning, and build AI that pulls together published evidence, a patient's history and Johns Hopkins practices to support care decisions.

Voice 2: Is that an actual product, or still a plan on paper?

Voice 1: Still early. Both sides say they haven't defined the specific tools yet, and they plan to evaluate effects on clinical decisions, quality and patient outcomes. OpenEvidence Chief Executive Daniel Nadler called it at minimum a research, development and product collaboration, and didn't disclose terms.

Voice 2: And a day earlier, Northwestern opened the door a lot wider.

Voice 1: Right. Northwestern Medicine said OpenEvidence now launches from its Epic record with hospital sign-in, and nurses, pharmacists and other clinical staff have access for the first time. Its own release described no evaluation plan.

Voice 2: How big is OpenEvidence's user base at this point?

Voice 1: More than 1.1 million verified U.S. clinician users as of September, according to Fierce Healthcare.

Voice 2: So one tool, inside one record system, now reaching a lot of people who aren't physicians at all.

Voice 1: The EHR industry got its own report card this week, and it wasn't flattering.

Voice 2: All 35 electronic health record developers reviewed in a new study were certified as meeting federal AI transparency requirements, even though their public risk disclosures were inconsistent and often incomplete.

Voice 1: Inconsistent how, exactly?

Voice 2: Researchers at Georgetown University School of Medicine and MedStar Health's National Center for Human Factors in Healthcare reviewed required disclosures covering risk analysis, risk mitigation, governance and eight other attributes. Reporting varied widely in structure and detail across the 35 developers, the authors said.

Voice 1: And every one of them still got certified.

Voice 2: Every one. The authors, led by corresponding author Raj Ratwani of MedStar Health, called for standardized reporting and more rigorous certification review.

Voice 1: So the checklist passed even where the disclosures themselves didn't actually tell you much.

Voice 2: Different kind of announcement out of pathology: Agilent is buying Proscia.

Voice 1: For how much?

Voice 2: Up to $270 million: $240 million in cash at closing, plus up to $30 million tied to performance milestones. Proscia's Concentriq platform links slide scanners, lab systems and image-analysis and AI tools, and it serves more than 4,000 labs, hospitals and clinics.

Voice 1: Is the AI part actually cleared for diagnosis?

Voice 2: Partly. Proscia's Concentriq AP-Dx product is FDA cleared as an aid to pathologists reading scanned surgical slides for primary diagnosis, but the AI applications elsewhere in its portfolio are offered for research use only, the company says.

Voice 1: Agilent's chief executive, Padraig McDonnell, said digital pathology is quickly becoming the foundation for pathology data generation. The deal is expected to close early in Agilent's 2027 fiscal year.

Voice 2: Suki just raised $90 million, and its chief executive wants to be clear documentation was never the point.

Voice 1: Who put the money in, and what's the growth story?

Voice 2: New investors Healthier Capital, Crosslink Capital and Hercules Capital, alongside existing backers including Venrock and Flare Capital Partners. Suki said revenue doubled in 2025, and patient encounters on its AI documentation tools grew more than 190% year over year in 2026. It didn't disclose a valuation.

Voice 1: What do they want to build with it?

Voice 2: “Clinical documentation was never the destination. It was the entry point,” Chief Executive Punit Soni said. The company said it will build AI agents for health systems and health technology companies, bringing total funding to $250 million.

Voice 1: A documentation company calling documentation a stepping stone is a line worth remembering.

Voice 2: Before we go, something to watch.

Voice 1: Utah's prescribing sandbox just expanded again. Nolla Health's software can now issue first-time topical acne prescriptions, with two physicians reviewing each before it's sent, moving to sampling of at least 10% of enrollment monthly once it's established.

Voice 2: What about August AI, and who's actually checking any of this?

Voice 1: August AI's software can recommend chronic-disease renewals for $4, with a physician reviewing each one. The state named six outside evaluators, including the Coalition for Health AI and Stanford, to inspect decision logs and outcomes. Doctronic's refill agreement ends this month, and Utah's Office of Artificial Intelligence Policy is discussing a 12-month extension.

Voice 2: And next week the AAFP FMX AI hackathon runs in Nashville, October 20th.

Voice 1: No agenda details yet, just the date on the calendar.

Voice 2: A sandbox adding two more AI prescribers while still deciding whether to keep the first one running is a strange kind of expansion.

Voice 1: Sources and corrections are at physicianintheloop.org, where you can sign up for the free email briefing. This briefing was read by AI voices.