Exclusive: OpenEvidence aims to be 'Disney+' with new JAMA deal

Screenshot 2025-06-05 at 5.39.50 AM

OpenEvidence this morning announced it has a new multi-year deal with JAMA to make full-text content from the American Medical Association's 13 journals available in its artificial intelligence-powered search engine for medical evidence. For example, a doctor searching for the efficacy of xanomeline-tropsime chloride (a schizophrenia medication) will see detailed results, including charts, pulled from a recent JAMA publication. The company earlier this year announced a similar deal with the New England Journal of Medicine. The company's CEO and founder Daniel Nadler told me the company's approach here is like Disney+. Doctors who want to search for information published by top journals and specialty societies will need to use OpenEvidence, just like people who want to stream Moana will need to go to Disney.

A streaming service isn't the comparison that immediately comes to mind for OpenEvidence. Founded in 2022, the service uses a conversational prompt interface and generates responses much like OpenAI's ChatGPT, and OpenEvidence is competing with products like UpToDate that doctors use to search for evidence when making clinical decisions. From the earliest days, Nadler said the big differentiator for OpenEvidence, compared to generative AI chatbots, was that it cites its sources. He said that company has taken care to design the system so that it doesn't invent answers the way other generative AI products do — in many cases, the service will respond to queries with verbatim text from studies. He said OpenEvidence is fundamentally a search product designed for the very narrow use case of doctors looking for medical evidence. He said the company's "most valuable secret sauce" is a machine learning search algorithm, and he compared the company's technology to "a really phenomenal colored number sticky system" that a student might use to label a textbook for an open book exam. The product is free for doctors and supported by advertising. The search is HIPAA compliant.

So far, OpenEvidence has produced impressive numbers. Nadler said that about 350,000 doctors have registered and on average the company fields "7.2 million clinical consultations a month." OpenEvidence recently announced a $75 million fundraise led by Sequoia at a $1 billion valuation. It was recently reported that the company is in talks Google Ventures and Kleiner Perkins to raise over $100 million more — Nadler didn't confirm the raise but confirmed that those firms are investors in OpenEvidence.

Are you a doctor using OpenEvidence or other AI search tools in your practice? We're hoping to speak to as many as possible. Reach out if so.

policy

Takeaways from Medicare' big DC tech party

A who's who of health IT leaders descended on the Hubert Humphrey building in DC on Tuesday for an invite-only event about how the federal health department's Medicare agency and its health IT regulator might better facilitate the flow of data in the health care system. In recent weeks, Medicare officials have been beating the tech modernization drum. They hope to stimulate the development of more advanced tech products, including AI tools to help people manage their health.

My invite must've gotten lost in the mail, but a few takeaways from what I heard:

CMS spelled out five priority tech areas it hopes to work on, including a few I hinted at in my story last week — a simpler digital identity authentication system for Medicare, and a national provider directory. The agency said will expand its Blue Button API that allows developers to help patients access their data, and it will plow forward with Data at the Point of Care, an initiative to make Medicare claims data easily accessible to providers so they can better understand a patient's medical history. Finally, and most provocatively, the agency said it will participate in "trusted data exchange."
That last one is kind of a doozy! As Lisa Bari, an executive at Innovaccer and a former CMS official, pointed out on LinkedIn, it was a Rorschach Test— people saw what they wanted to see. If you are in favor of TEFCA (the government supported data sharing framework that's been inching forward), you saw this as an endorsement of TEFCA. If you are against TEFCA and prefer to see the government support open standards that allow all manner of data sharing frameworks to flourish, you saw the failure to spell out T-E-F-C-A as a very telling omission. As I reported in STAT Health Tech earlier in the week, a health department budget document for 2026 suggested that ASTP/ONC, the HHS IT regulator, will continue to support TEFCA. Is there a disconnect between ASTP and CMS here?
It's notable that CMS announced its plans as health care organizations are mulling responses to an e xpansive request for information meant to inform its thinking. I'm not saying that the RFI won't be carefully read by people at CMS, and that it won't guide policy. But it's been clear from the very beginning that the super detailed RFI came from leaders who very much have plans in mind.

I'm following the Medicare RFI and its tech modernization agenda closely. I'd love to hear your thoughts about it. Please reach out.

artificial intelligence

CHAI descends on Stanford, teasing NIH news

Brittany Trang writes with updates related to the Coalition for Health AI:

Earlier this week, the organization announced that Pacific AI, an AI governance and compliance organization, will go through CHAI's new assurance service provider certification process and join CHAI's list of approved vendors. What's assurance service provider certification? Check out Brittany's recent AI Prognosis to learn about CHAI's pivot from "assurance labs" to this broader certification for services across the AI lifecycle. Interestingly, Pacific AI will help CHAI write the standards that will be used to CHAI-certify AI governance providers, including itself — which seems like a conflict of interest. When asked about this, a CHAI spokesperson said that vendors, health systems, attorneys and others are helping them write these standards and that it's "consistent with all of our other working groups and thought processes around gaining consensus."
Also, CHAI's big summit is taking place today and tomorrow at Stanford, and National Institutes of Health principal deputy director Matthew Memoli is slated to "share AI-related news" at the conference. Stay tuned for an update.

FDA AI updates, an Eli Lilly app, and more

On Monday, Brittany Trang broke the news that FDA was set to release it generative AI tool, Elsa, to workers. Brittany has fresh intel on the tool, how the rollout is going, and what industry thinks of it in her AI Prognosis newsletter.
Welldoc announced that it's the developer behind a new app meant to support people on Eli Lilly's GLP-1 medications. It includes features like medication reminders, medication information, and educational content. Welldoc is best known for its apps that support people with diabetes.
Yahoo Finance reported that CVS Health plans to spend $20 billion (!) on a broad effort to improve the consumer experience of health care. Executive Keith Shah spelled out some of the ambitions on LinkedIn.
H1, a platform for information about providers, acquired Veda Data Solutions, which works on ensuring accuracy of provider data. H1 earlier this year acquired Ribbon Health, another company working in the provider data space.
TeleTracking, a platform that helps hospitals manage operations, has a new deal with Palantir Technologies to develop analytics for applications like optimizing staffing.
Rune Labs, developer of an app for tracking Parkinson's disease symptoms for the Apple Watch, is working with the Parkinson's Foundation on research to correlate genetic data of people with the disease to digital data. A few years ago, I wrote about Rune and others using the Apple Watch to help people manage their Parkinson's.
Two former Ginkgo Bioworks employees, including Renee Wegrzyn, the former head of federal health moonshot agency ARPA-H — are starting on a new AI venture called Transfyr. The company is building AI tech to help facilitate the "transfer of discoveries from lab bench to real-world application" and is hiring both AI/machine learning folks as well as biomedical scientists and product developers.

More around STAT

Check out more exclusive coverage with a STAT+ subscription

Read premium in-depth biotech, pharma, policy, and life science coverage and analysis with all of our STAT+ articles.

What we're reading

A cautious culture cost Novo Nordisk its lead in the obesity drug race, former employees say, STAT
Internal document shows how UnitedHealth executives prepared to tamp down investor unrest, STAT
Opinion: A Culture War is Brewing Over Moral Concern for AI, UNDARK