Dr. Eric Topol is founder and director of the Scripps Research Translational Institute, and co-chair of Doximity’s PeerCheck initiative alongside Dr. Regina Benjamin, the 18th Surgeon General of the United States.

...

Alexis Kayser

By Alexis Kayser

Healthcare Editor

0

Share

Newsweek is a Trust Project member

See more of our trusted coverage when you search.Prefer Newsweek on Googleto see more of our trusted coverage when you search.


This is a preview of the February 12 edition of Access Health—Tap here to get this newsletter delivered straight to your inbox on Thursday mornings.


Good morning. Yesterday, we got news from the Assistant Secretary for Technology Policy/Office of the National Coordinator for Health IT (ASTP/ONC), which is currently hosting its 2026 Annual Meeting at the Omni Shoreham in Washington, D.C. It’s been fairly productive, if productivity is measured by the number of bullet points in my email briefing. (I kid. There are, indeed, some notable updates.)

Read More on Health Care

Let’s zoom in on the one about “fighting information blocking.” HHS, through the ASTP/ONC, has initiated its first oversight actions to “seriously address information blocking,” per the Office.

“Specifically, ASTP/ONC is in the process of issuing ‘notices of potential non-conformity' to certain developers of certified health IT,” the Office said in its news release, “requesting explanations related to potential nonconformance with Certification Program requirements, including the Information Blocking Condition of Certification and API Maintenance of Certification requirements.”

Other than a nod to section 45 CFR 170.580 in the Code of Federal Regulations, it didn’t get more specific. When I spoke with Sean Sullivan—a partner in the health care group at the Atlanta-based law firm Alston & Bird—a few hours after the announcement, we still hadn’t received additional details on which developers the agency would be targeting. He did share some context, though.

“As an attorney, this [information blocking] is something that I've heard a lot about in the industry, especially about EHR vendors that have been very protective with their data—which really isn’t their data, [but] the data of their hospital/health system customers,” Sullivan said.

Many EHR vendors claim that they’re being protective of their intellectual property and guarding the security of their data, he continued. But other vendors claim they need access to those records to do the jobs that the shared health system customer hired them to do.

We’ve seen this tension play out in recent lawsuits involving Epic, like the Health Gorilla case (in which Epic is aligned with health systems as a plaintiff, alleging misuse of patient data), and the Particle Health case (in which Epic is accused as the defendant, accused of anticompetitive behavior for its “stranglehold” on patient records).

“In many cases, we've heard stories about EHR companies that blocked other vendors from being able to access their top customer’s information,” Sullivan said, speaking generally (he did not point to specific cases). He added that you do have to examine these on a case-by-case basis, “but in many cases, that does look a lot like information blocking.”

For health systems (and smaller tech companies that feel stifled by the industry data giants), yesterday’s announcement will likely be a welcome signal that the government is finally ready to intervene when data access appears to be unfairly restricted.

But—and this will sound familiar if you’ve ever found yourself in the cybersecurity village at a health tech conference, or even just chatted with your IT lead—every additional integration, API or third-party connection also expands your potential attack surface. More pathways to data can mean more opportunity for misuse if controls aren’t airtight (and recent breaches tell us that in today’s cybersea, there’s no such thing as an unsinkable ship).

Here’s the question I’ll be looking to answer as we watch this all play out: at what point does data protection become data restriction, in the government’s eyes? And which is the bigger risk for a health tech company: being sued for monopoly or for a medical records leak?

As always, I’d love to hear from you. Send me an email at a.kayser@newsweek.com and let me know your thoughts.


In Other News

Major health care headlines from the week


Pulse Check

Executive perspectives on key industry issues

Dr. Eric Topol is founder and director of the Scripps Research Translational Institute, and co-chair of Doximity’s PeerCheck initiative alongside Dr. Regina Benjamin, the 18th Surgeon General of the United States.

...

Last week, I sat down with a man who needs little introduction—at least for those who are following the health care-AI discourse.

Dr. Eric Topol is one of the most prominent voices on safe, responsible AI deployment in health care. He’s the founder and director of the Scripps Research Translational Institute, and most recently made headlines when he signed on to co-chair Doximity’s PeerCheck initiative.

Click here for my full writeup on the details of PeerCheck (it’ll be useful background, if you aren’t familiar), and read on for a look into the second half of my conversation with Topol. We discussed how much automation risk health systems should accept, and how challenging it is to bring clinicians into the AI fold. Spoiler alert: it might be easier than you think, if PeerCheck is any indication.

Editor’s Note: Responses have been lightly edited for length and clarity.

When I speak with health system chief innovation, digital and technology officers that describe themselves as “bullish” on AI, they frequently tell me that humans make mistakes, too—that we don’t need a zero-percent error rate to accept an AI tool as better than the baseline. What's an acceptable margin of error for AI tools, from your perspective? What threshold must it cross to become trustworthy?

I'm glad you brought that up, Alexis, because that all sounds really good—you know, just get a net benefit.

The problem with that is that it works at a population level, but not for the patient in front of you. If you're looking at it from a societal point of view, or [considering] the practice of medicine, [you can’t say], “Oh, humans make mistakes, AI makes mistakes, but they cancel out, so this is a net benefit. Let's just adopt it.”

No. If you're the patient who's [on the receiving end of a] serious mistake of their treatment or their diagnosis, it's dreadful. I don't think you ever want to accept serious medical errors. There are plenty of those with humans today, but that doesn't mean that we should accept them, and we shouldn't just blindly accept AI errors without trying to reduce them. That's what the mission should be, is trying to bring these down to the absolute minimum.

Is zero the target, or is there another “absolute minimum” that could be acceptable?

Zero should be the target. We'll kind of never get there.

Having experts in these layers of review to go over these [AI] responses, it's certainly going to help, but we'll see what the significant error rate comes in at, and then hopefully it keeps coming down over time. That's the goal. We should never be happy with errors, but if we get down [to] approaching zero, that would be phenomenal.

Change management is another kind of risk that health systems take on when adopting AI. Thinking back on your work with PeerCheck, what is the most challenging part of bringing physicians into the AI development and governance process?

I would have thought that getting doctors involved would be the most challenging [part of the project]. That turned out to be the least challenging [part].

I guess because AI is so big right now, there's tremendous enthusiasm within the medical community for academics and leaders to be part of a solution, and not just be complacent with part of the problem. And I didn't anticipate that. I have to say, I thought it might be difficult.

Getting 10,000 physicians on board this quickly who are more than happy to do many reviews... I mean, wow. I didn't expect that. I thought this would be much, much more difficult.

Why did you think that it would be so difficult?

Well, because we're so busy and we're getting pulled in so many directions. I think it shows a dedication to making medicine better in a new world, a world that's moving quickly, maybe too quickly. I think most people know that AI will have a vital place in the future, but just don't accept it on [its] face. We can do better than just taking AI-generated responses. This [physician involvement] is, I think, the way to do it and pave the way to accurate information that's needed in the management of patients.

Not all AI is built equally. It’s still rare to see so many MDs included in a tool’s review. What would you recommend that hospital and health systems look for when assessing AI tools that could earn physicians’ trust?

Transparency is a big part of it—being completely upfront about how this is being done, what the results [are].

We do lack transparency in the AI medical world. A lot of algorithms get FDA approved or cleared or used without anyone ever seeing how it works, its performance.

There are too many AI devices (the FDA calls them devices. I don't know why they call them devices. They're algorithms) that we've never seen the data [for], and most of them don't even get [that data] published, ever. And if they do get published, they [the AI developers] call it “proprietary.” So transparency is something the medical community really likes, and health systems, when they buy into these sorts of tools, should demand transparency.


C-Suite Shuffles

Where health care leaders are coming and going


Executive Edge

How health care execs can manage their own health

That rural physician outreach story—the one I flagged in the “news” section—was difficult for me to write. I came home from that plane ride with so many stories from the team’s years of working together. It felt impossible to capture such camaraderie in a single magazine feature.

I ultimately had to bend these parallel plots into a single angle: the clinicians who make sacrifices for their patients, and the clinicians who make sacrifices for each other. The former is easy. Some variation of it can be found in health system mission statements all across the country.

The latter, I learned during my visit, is equally important. It helps to know who you’re facing the long, hard days with—and it helps to trust your colleagues in work that demands both emotional regulation and practical precision.

This isn’t a novel idea. In fact, recent research published in the journal Organizational Science found that tightly-knit surgical teams who have experience working together are significantly more likely to provide positive patient outcomes. According to the study, patients with experienced surgical teams spent three less days in the hospital, and nearly two days less in the ICU post-surgery than those who had surgical teams that weren’t well-acquainted with one another.

Here’s a section from the feature on the power of camaraderie in health care. If you’ve been keeping up with this column, you know that shared burdens are lighter—and that community amongst health care workers is as vital as it’s ever been.

Dr. Michael McNamara (front, far right), a maternal-fetal medicine physician at Sanford Health, and his outreach team aboard the plane they fly weekly to care for rural patients. This photo was taken in December 2022, when a winter storm left the team stranded 320 miles from home after an outreach clinic.

...

McNamara fondly recalls a December clinic in 2022—the last one before the holidays—when snow stranded them overnight in Rapid City. The sonographers bought matching Christmas pajamas before crashing at a hotel. When inclement weather continued into the next morning, they went shopping for essentials (and, McNamara said, “several items to upgrade my bland wardrobe”) and joked that if the blizzard continued, they could all get pedicures. They got the all-clear before they were able to check out and rushed to catch a plane to Marshall, Minnesota, and begin the drive to Sioux Falls.

“In the end, I did miss out on that team-building pedicure, and the once-promising upgrade to my bland wardrobe was left stranded in that abandoned shopping cart,” McNamara said in an email, reflecting on the experience. “But with those five people with me and getting back home only a day later, I call it a win.”

Click here to read the full article: In rural South Dakota, the doctor is a plane ride away. Luckily, he flies.


CEO Circle

Insights from health care thought leaders around the world

“At 3 a.m., an ICU tells the truth,” said Dr. Der-Yang Cho, a neurosurgeon and the longtime superintendent of China Medical University Hospital, in his latest thought leadership piece for Newsweek. When every second matters, every delay can be deadly.

Cho—who is also a member of Newsweek’s CEO Circle—shared how his hospital is reducing risk by modernizing critical care. Before you go, give it a read.


This is a preview of the February 12 edition of Access Health—Tap here to get this newsletter delivered straight to your inbox on Thursday mornings.

View 0 Comments

Request Reprint & Licensing

Submit Correction

View Editorial & AI Guidelines