You’re reading the web edition of STAT’s Health Tech newsletter, our guide to how technology is transforming the life sciences. Sign up to get it delivered in your inbox every Tuesday and Thursday.
Brittany Trang writes: I’m in Las Vegas at the Healthcare Information and Management Systems Society conference. When I haven’t been talking to people at HIMSS about AI, I’ve been talking about interoperability and data exchange — or, how patient information gets shared across the health care system. Here are big themes that have emerged:
What’s happening to TEFCA?
CommonWell executive director Paul Wilder calls the government-supported health data exchange framework, TEFCA, the “adult in the room” because even though it’s voluntary to participate in it, the framework has authority from its Congressional mandate and the participation of federal health agencies. But with a new administration shaking things up, it’s unclear whether that adult is going to leave the room. Will the government pull TEFCA’s funding? Is that funding currently frozen? Will the government not prioritize supporting the exchange? No one knows, but everyone is anxious to find out.
The future of Carequality
Carequality is one of the leading alternatives to TEFCA. Remember the Epic Systems-Particle Health dispute last spring over alleged bad behavior on the Carequality data-sharing network that ended up in an ongoing antitrust lawsuit? I haven’t heard anyone say the name “Particle” directly at the conference, but multiple people wondered if the community had lost some trust in the ability of networks like Carequality to ensure that there weren’t any bad actors trying to unlawfully access patients’ data. That would be a setback for making data exchange easy, because these exchanges are based on trusting that other people are going to use your patients’ data lawfully.
The looming CMS-0057 interoperability rule
Last year, the Centers for Medicare and Medicaid Services issued a rule requiring insurers that administer government-funded health insurance programs to use newer health data exchange technology. This upgrade will speed up processes like prior authorization and will overall smooth out the transfer of patient data. Speakers at the conference kept reminding everyone that no matter how much organizations want to drag their feet, the health care system will have to implement these tools — called FHIR APIs — by 2027, and that there’s no point in dilly-dallying because it will make everyone’s lives easier by decreasing the number of faxes the health care system sends.
The Neighborhood Watch
In his closing keynote at the interoperability pre-conference forum — which had the feel of a preacher riling up congregants at the pulpit — former U.S. chief technology officer and Arcadia chief strategy officer Aneesh Chopra proposed a “Neighborhood Watch” idea where the community proactively calls out vendors who aren’t doing a good job of facilitating data exchange. There are supposedly over 300 electronic health records that are certified by the Office of the National Coordinator as compliant with certain data requirements, but that doesn’t mean that they are actually usable when someone needs to send data. Chopra said he’s recently convinced companies to ship patches in a timely manner when his team has contacted deficient EHRs about problems and is inviting more people to do this.
Neighborhood Watch co-supporter Brendan Keeler from HTD Health told me that there was a ton of immediate interest from vendors after the announcement, with many EHRs asking where they currently stand. Chopra told me that he wants to get the idea off the ground before any eventual new ASTP/ONC head is announced so that they might be able to run with such an idea.
Trump’s chilling effect on AI research
Already, researchers studying the performance of clinical AI are feeling a chilling effect as a result of the Trump administration’s early executive actions. By default, much of that work analyzes algorithmic bias — but bias, and several other terms that map to statistical concepts in machine learning, are among those targeted for exclusion from federal communications and government-funded research because of their association with diversity, equity and inclusion efforts.
Uncertainty over the interpretation and execution of those orders is catching clinical AI research in the dragnet, leading to delays in publications and young scientists redirecting their research focus, Katie Palmer reports. This isn’t just a problem for academics: Device makers and health systems facing regulatory hurdles and compliance deadlines can be impacted too.
Ex-Cerebral CEO’s new mental health startup
After three years at the helm of Cerebral, former CEO David Mou is teaming up with long-time mentor Tom Insel, the former director of the National Institute of Mental Health, to launch a new startup they hope will help buttress high-quality mental health solutions. Called AdvocateMH, the company will use licensed clinicians to triage people to the most appropriate service to meet their needs, be it a meditation app, an online therapy provider, or maybe even something in person. The clinician will serve as a kind of case manager who can track a client’s progress. Ultimately, using their visibility into a number of services, the company hopes to help raise the bar for for mental health care in the digital health realm.
“What are the standards that make a difference? How do we define safety and quality in this ecosystem, in this field?” Insel told me. “I think it’s something that we’ve been missing, and we’re hopeful that really, within the next couple of years, by doing this, we can reshape the industry, and give people a really good sense of what to shoot for.”
The founders say they are motivated by their feeling that, so far, venture-backed mental health startups as a whole haven’t delivered on their promise because they’ve been chasing scale at all costs. (Nevermind that Cerebral, which has raised $460 million, settled with the U.S. Attorney’s Office over allegations tied to an aggressive growth strategy that led it to improperly prescribe stimulants like Adderall.)
To help point incentives the right way, Mou and Insel structured AdvocateMH as a public benefit corporation with the objective “to transform the behavioral health ecosystem to ensure best outcomes for anyone needing care.” They’ve still raised money though: An “oversubscribed” round of an undisclosed amount led by Morningside Group. The company is very early in development, but it plans to start out by offering its services in Massachusetts.
HIMSS announcements and other news
- The Coalition for Health AI announced the launch of a “model card registry” where algorithm developers can all put details about their products in hopes of making it easier for buyers, like health systems, to vet them. I was told “The goal is to have the complete model card disclosed publicly,” which would be a win for transparency. CHAI’s model card template is designed to be compliant with ONC transparency requirements for some algorithms baked into EHR software.
- The Peterson Health Technology Institute, which produces evaluations of whether health tech solutions deliver outcomes and are worth the money, announced it will vet digital health solutions for opioid use disorder. It should be a juicy read when it comes out. It’s also interesting because of a little backstory. PHTI was very much born out of the work of the Institute for Clinical and Economic Review, which years ago did a notorious evaluation of OUD digital therapeutics. DTx pioneer Pear Therapeutics was never able to escape the tepid assessment of its evidence. It’s likely that the new owner of Pear’s OUD treatment, PursueCare, will be evaluated as part of this new PHTI report.
- There were bunches of clinical documentation and AI workflow announcements timed to HIMSS. *Deep breath*: Microsoft announced Dragon Copilot, a new virtual assistant for clinical workflows; Rush University announced an expanded deal with Suki for “Enterprise-Wide Ambient AI Rollout”; Abridge struck a deal to bring its clinical documentation solution to Athenahealth’s EHR software; Memorial Sloan Kettering announced it will use Abridge for clinical documentation; SalesForce announced Agentforce for Health; Freed raised $30 million, including from Sequoia, for an AI scribe that’s gained traction with small and medium-sized practices.
What we’re reading
- Even critics of the scientific establishment give Trump’s drama-filled weeks mixed reviews, STAT
- CMS rescinds Medicaid’s health-related social needs guidance, Fierce Healthcare
- Amazon’s One Medical CEO stepping down after less than two years at helm, CNBC