← Practical AI in Healthcare
Practical AI in Healthcare
S1, E40 - Jeff Smith — AI Regulation, Transparency & Innovation from the Government Perspective
00:43:40
Loading summary
Transcript55 lines
- [00:02]
A
Welcome to Practical AI in Healthcare, the podcast that cuts through the noise to spotlight real world solutions delivering real world value. From patient care to clinical research, from life sciences to patient engagement, we focus on what truly matters in healthcare today. No hype, no theory, just practical insights where AI is making a true impact. Dr. Steven Lapkoff and Dr. Leanne Rosenblut are your hosts as we explore what's real and moving the needle in this exciting new domain. Welcome aboard and let's get to it. As many of our listeners know, Leon and I work very closely with the DCI Network Division of Clinical Informatics at Beth Israel Deakins Medical center in Boston. This June, the network is hosting Patient powered Digital Health 2026. The conference will bring together patients, innovators, industry leaders, healthcare providers and policymakers to shape the next generation of real world patient centered solutions. The meeting will run from June 22nd to the 24th in Boston at Harvard Medical School. We've arranged for our listeners to get a discount on registration to the meeting. If you Register now before June 15th and use promo code PracticalAI June no spaces, you'll receive 30% off your registration fee. You can learn more at dcinetwork.org patients2026. In addition, we're always looking for sponsors. If you or your company are interested in becoming a sponsor, please reach out to admincinetwork.org see you in Boston.
- [01:29]
B
Hello, welcome to this week's edition of Practical AI in Healthcare. My name is Dr. Stephen Lapkoff and I'm here as I am every week with my colleague, Dr. Leon Rosenblatt. How's it going, Leon?
- [01:39]
C
Oh, I've been better, Steve. I'm dealing with some serious nerve pain and over some meds, so I might be extra wacky today. It's going to be a fun episode.
- [01:45]
B
This week we're going to be engaging in a conversation which we haven't done on the podcast yet. We're going to be talking to a representative of the federal government. One of my former colleagues from my work at AMIA happens to be a gentleman named Jeff Smith. Jeff and I worked together on a number of initiatives over the years and he is now on the staff at the Office of National Coordinator in Washington, formerly known as astp. And Jeff has very kindly agreed to be interviewed for the podcast and we'll ask him about his background. Just a second. But Jeff, welcome to the podcast.
- [02:17]
D
Thank you, Steven. Thank you, Leon. Good to be here. Happy to meet the cast or the pod. Do we call it the pod or the cast or both?
- [02:24]
C
We even Put them together like podcast. But I like where you're going with this.
- [02:30]
D
I think we can play.
- [02:33]
B
So Jeff, our first question that we typically ask our guests is basically tell us your background story. Tell us how you got your superhero cape, How'd you end up at ASTP from amia and what's that story?
- [02:45]
D
So, you know, I'm a meaningfully used baby. Like so many of us in the health IT space, especially of a certain age, I really cut my teeth on health IT policy back in 2010. And right around that time there was about 34 to 37 billion dollars, depending on who you ask, freshly minted, to help incentivize the adoption of electronic health records. Congress in its wisdom, decided they didn't want to just have a grant program. They didn't want to just put a, put a bunch of money out there for people to buy things. They wanted to make sure that hospitals and physicians could demonstrate the meaningful use of electronic health records. And that kicked off a years long and what's turned out to be kind of a decades long process to try and bother establish consistent standards for electronic health records as well as develop policy to incentivize the adoption of those standardized electronic health records. And so that was 2010. I was actually kind of the second policy hire at the College of Healthcare Information Management Executives, otherwise known as chime. And you can always tell who worked at CHIME because they can rattle that off at the top of their head. And this is of course an association of hospital and health system chief information officers and other kind of senior executives within the health IT space on the provider side. And this was a formative time for me, who has a political science background, has a public policy master's, but really kind of being able to work with strategic thinkers like CIOs and the chime Board in a way that intersected with public policy and health. It was really very formative at the time. The organization was of course, very supportive of the funds that were newly made available. And we were able to fashion a relationship with both ONC and the Centers for Medicare and Medicaid Services to help them think through what ultimately we agreed were good things. But maybe we need to think about a different staging or cadence of requirements or a different set of functionalities and tools that CIOs knew their physicians and clinicians needed. So that was, you know, so that was kind of the first five years of my career. And then of course I went to AMIA and was hired on there as the vice president of public policy. And it was at that point that, you know, my policy portfolio went from essentially trying to represent CHIME to Congress and the Hill, but also to ONC and CMS, to trying to represent AMIA to a kaleidoscope of federal agencies and acronyms. You know this. Well, both of you members of amia, the, the organization is really bound by this concept of informatics, but at the same time it does have members who look at informatics through a host of different prisms. And that could be through public health or clinical research, it could be biomedical, it could be clinical. And there's, you know, even, even, you know, kind of patient health informatics. And so I'm sure in the years since I've left there's more kinds of different interest groups and affinity groups within amia. But from my perspective on, on the public policy, government affairs side, that meant I could dig into not just ONC and CMS policies but, but I could dig into FDA policies because we had members very conversant and interested in software as a medical device. I could dig into FCC policies because I think at the time we were one of the very first organizations to say that access to broadband is a social determinant of health. So it was at that point in time that I got a secondary education in what I would call health informatics policy. And we really, we tried our hardest to ensconce that concept, health informatics policy, while I was there as a, as a kind of distinct and separate but inclusive policy domain alongside health IT policy, which a lot of folks, you know, today recognize, maybe not at a kind of academic level, but certainly within industry. You talk about health IT policy, people kind of go, yeah, you know what we're talking about. So that's, I'm sorry, so that's how I'm. No, I was going to say that's how I got to, that's how I got to. ONC is from amia.
- [07:18]
B
So the world that you've been in, you've gone from being on the advocacy side to being in the government. The government doesn't sell products. The government doesn't actually make things per se, but it does regulate things. It does make the rules. It does set up the framework through which everything works. You know, you got into the government how many years ago now? Seven or eight years ago.
- [07:44]
D
At this point it's just about six. So I'm at five and a half going on 50.
- [07:50]
B
But in the course of that last five years, you know, you might have been focused in, on, on and you still may be focusing on ehr. But there's been a, you know, just a little tiny new thing that sort of hit the radar in terms of how health it is being practiced. And that's what this conversation is going to be about the rest of the discussion, which is, where is AI from your perspective at the onc, making the most difference? What are you thinking about? What are the issues that we want to go through? And Leon and I are going to go through that with you and sort of say, get an idea of where is the ONC and how is it impacting things? Will there be an AI meaningful use kind of thing down the road? You know, what's coming, you know, before we dive completely.
- [08:37]
D
There
- [08:39]
B
you've been also, you, you did a lot of hims work too, if I'm not mistaken. You've been in and out of hims for a number of meetings. So you, you have the perspective of an advocate. You've got the perspective of someone who's worked with the commercial side of things, and now you're in the government. How have those things sort of shifted in your brain in terms of your own personal focus?
- [09:00]
D
Yeah, that's a great question. Insofar as, you know, I think about this, I do think about this kind of retrospectively when I look back, I kind of think, okay, how did I get here? And to me, it makes sense as a natural progression insofar as I was very lucky, both at CHIME and at amia, to represent members who never had a completely antagonistic relationship with the policymakers we were dealing with. You know, I think oftentimes there were discussions and differences in approach, but not in destination. And I think even on the Hill, we crafted a lot of, and cultivated a lot of relationships with committee staff who, even if they were antagonistic to, maybe the administration, weren't antagonistic to the idea that we need to push forward on interoperability. We need to make these tools better for clinicians. We need to make data more available for patient access and for research. And so I've been very fortunate to be able to come into the government not necessarily with a. With an axe to grind, but with the capability, or at least I think I have the capacity to kind of help craft policy that will then be more amenable to those who have to implement it, or at least be more cognizant of the folks who are either intended to benefit from the policies or have to implement the policies, which, you know, as you know, are not the same. But I think that's. That that has, that has really helped over the Course of the last five years. I mean, I think again, I. I come to the notion of public policy as being an instrument that can be used to do tremendous good at scale. I think that there are even, you know, even when you. When you think about things through that, that line of thinking, then you think, oh, well, more policy means better policy means we should have more policy. That's not always the case. You know, I just think that to be able to. Right. Size the policy and to be able to ensure that the policy is fit for purpose is a really important task. And from my perspective, one of the things that I've really enjoyed about ONC as a culture since I got here, and really a culture since I've engaged with it since 2011, is that it's very open. We get a lot of input from our health IT advisory committees. We used to have two, now we just have one. But we just had our first meeting of this administration last week over a health IT advisory committee. And this is such an important group of folks who are nominated by different quadrants of Congress as well as the Government Accountability Office. And these individuals, these are experts from the field who give us windows into what's going on. You mentioned hims. There are a ton of industry activities, industry conferences. And I think this year I was somewhat aghast that I was working on at least my 11th or 12th hymns, if not something higher. And I decided I didn't need to really go back and think about it all that too hard. But certainly the industry engagement that we have is actually really important because the job of a policymaker on the inside is really to try and find that middle ground, find that compromise that will help both sides, that will help push forward the policy objective, probably in ways that are not directly amenable or at least directly preferred by either industry or by providers or by patients, but like, you know, really trying to craft that policy in a way that says, okay, given all the input that we have heard, these are the steps that we need to take from a regulatory perspective, or this is a set of issues that we need to elevate from a priority perspective. And so, you know, I think that the culture here, again, is just one of openness and wind up having a lot of meetings with a lot of different interested parties or stakeholders. And those are, you know, those are all kind of, you know, some, in some ways, very directly, in other ways, you know, public comment periods and other ways, you know, just kind of more informative, but they always kind of come in at the end of the day. To help us think through where's the direction we want to go, big picture. And then what are the policies that we want to take to get there? And then what is the negotiation that has to occur in order for folks to actually implement those policies? Because at the end of the day we have a pretty solid policy tango that we have with cms. I don't think maybe tango isn't the right dance that we've got. I don't know, maybe it's a two step. But you know, we have a voluntary program. There's not a single EHR in the country that has to, that has to certify to this, to our program. But they do it because then their clients will get reimbursed 100 cents on the dollars. And so that's a dance. And it takes a lot of, a lot of people to figure out how to do it properly.
- [14:14]
C
Well, Jeff, first, delighted to have you with us. Miss working with you from the amia, your AMIA days. I've always found you so sensible, pragmatic and knowledgeable. I'm glad to have someone like you working in the federal level and shaping federal policy, whatever the dance steps required.
- [14:33]
D
Right.
- [14:34]
C
Whether, you know, probably a waltz is in there somewhere.
- [14:37]
D
Yeah, absolutely.
- [14:38]
C
And also thank you for making us introspect. Whether we're a POD or cast. I'm sure Steve and I will have to have several high level strategy meetings. But what I want to come back to is that there's something you said in a prep call that really struck as a leverage point. The EHRs are both the wellspring of data for AI and the mechanism for getting AI outputs into clinical care. Right. So it's there to pipe out and then there's a pipe in. But they can also be the bottleneck. Walk us through what the tension looks like from where you sit at onc.
- [15:11]
D
Yeah, yeah. So I mean, I think I'm glad that that resonates because I think it's, for me it's really helpful to help people understand the centrality of the electronic health record, even though it's not necessarily seen, I think universally as this central point in terms of we think about things in a couple of different ways. Technical standards and capabilities, behavioral and kind of business practices by the developers who deploy the or who develop those technical standards and capabilities. And, and then we've got a broader surround of information blocking and so to try and kind of frame this for you a little bit and to bring you into kind of current day onc, we have over the last several years pushed to try and advance a nationwide kind of infrastructure, one that is predicated on FHIR, the HL7 standard, and one that's predicated on APIs. And I think we've had a certain amount of success in this regard, especially as it relates to patient access. I think that the value proposition hasn't quite developed yet on the patient access side. But more specifically on the kind of provider to provider exchange side, I mentioned three things, so technical standards, business and behavioral, and then information blocking. And I would kind of think about, we regulate the technology, we regulate the developer of the technology, and then we exist in this statutory environment that Congress established in the 21st Century Cures act of 2016. And essentially what this does is it makes information blocking illegal. Just point blank. Okay, well what does that mean? That means that information blocking is essentially defined as a practice that is likely to interfere with access and exchange and use of electronic health record, or, sorry, electronic health information. And there's a caveat there, like the actor who is supposedly engaging in information blocking, if the actor knows or should know, that's when they're, you know, if they're kind of actively getting in the way, that's when they're in violation of this statute. And ONC was charged by Congress to establish various exceptions to this. And you know, a good one to kind of, to keep in mind is like privacy. If a hospital thinks that somebody is trying to get access or use information that they don't, that the hospital doesn't think the patient, the person should have access to, then you know, they have an exception, a privacy exception. Well, over the course of last several months and years, again we get a lot of input from industry and the notion is how can we do we need to modify those exceptions, do we need to modify the definitions? And in a recent regulation that we put out as a proposed rule called HTI 5, Health Data Technology and Interoperability version 5, we proposed a number of updates to our information blocking kind of regulatory regime that were really meant to prevent exceptions and contracting leverage from becoming kind of de facto interoperability gates, especially in the context of care delivery that increasingly relies on API driven, automated and AI mediated workflows. I'm happy to talk through a couple of the finer points there, but I think what I want to try and convey here is HTI5 is largely known as a deregulatory rule, and it is, the industry knows that we're proposing to remove a really high percentage of the existing certification program certification criteria. And we're looking to scope down Some of the conditions and maintenance of certification requirements and these are kind of the policies that regulate the developers, not the technology. But we did have a section of HTI 5 very much focused on trying to ensure that the regime around information blocking was inadvertently inhibiting the proliferation of, of agentic workflows and wasn't used as a contractual level lever to create kind of discriminatory access.
- [20:03]
C
Yeah. So Jeff, I think that the consequence or the practical consequence of what you may be reacting to as relevant to AI is that AI solutions developed by EHR companies themselves are much more easily deployable than third party or in house solution.
- [20:21]
D
Right.
- [20:21]
C
There's this asymmetry and the concern is that EHR vendors who control the platform might be creating artificial moats of various kinds.
- [20:29]
D
Right.
- [20:29]
C
Some intentionally, some unintentionally against third party AI. The Hi5 is the framework. Can you share with us how you think about separating legitimate technical and safety concerns from genuinely anti competitive behavior?
- [20:45]
D
Yeah, so I mean, I guess I'll go in just one layer deeper on some of the proposals in HTI 5 to help tease that out. So you know, the information blocking regulations, which we kind of affectionately refer to internally as those at 45 CFR 171 are.
- [21:07]
C
I can sense the affection, as you say it.
- [21:09]
D
Yeah, yeah, yeah. The effects that he has. Actually I don't have as much affection for them. All my regulations are in 170. So that's, that's where I light up. 171 though is the information blocking section. And these regulations are essentially are meant to be a playbook for industry, particularly lawyers, to try and discern when was an action legitimate and for safety reasons or for privacy reasons or for, you know, a small list of additional reasons. And when was this action nefarious, for lack of a better phrase? And you know, I think a couple things. So one of the things that we've tried to do is I mentioned that information blocking was essentially defined as a practice that's likely to interfere with access, exchange and use. One of the things that HTI 5 does is it clarifies in regulation that access and use includes automated technologies, including robotic process automation and agentic or autonomous AI systems. So we're basically saying it doesn't matter if it's a human trying to access an exchange and use or a robotic process automation or agent. Those, those things should still be able to, you know, should be able to engage with the electronic health information. And so that, that trying to, that kind of sets like a, that clarifies the floor that Everybody should be standing on now in terms of trying to walk that balance between legitimate use and, you know, information blocking. There's a couple other provisions in there and they, they really do try and articulate things that we're seeing in the market. And that is the, the blocking of write back capabilities for some when it's not being done for others. And there doesn't appear to be any meaningful difference between the some that are giving, getting right back access and there's some that aren't. And you know, a lot of these, a lot of the information blocking stuff is very, very legal insofar as you have to have a very discern, you have to have a very particular fact pattern. So it's hard to explain exactly in the, in the abstract, but what we're trying to do is create clarity around those scenarios where an, where a, an actor, I'll just have to use that, that that blanket phrase cannot claim an exception because what they're doing is clearly preferencing one party or one mode of access over another. And they have no other category, reason, categorical reason to be doing that. So, you know, I know there's going to be, there, there are, there are a ton of nuances in this space and I will say for your audience and, and to get me out of this, I would say please send all inquiries to healthit.gov portal or health IT inquiry portal.
- [24:20]
B
Yeah, so Jeff, I know that the ONC has put out an RFI relating to AI and I know that there's going to be folks listening to this podcast who may be interested in responding to it. Can you maybe give us a quick overview what it's calling for and what guidance you could provide to those who are interested in responding?
- [24:38]
D
Yeah, so absolutely. I think one of the things to kind of think about is that like HHS is always trying to make sure that our policies and our incentives and those incentives could be regulatory, that could be grant making, they could be cooperative agreements and other types of those kinds of levers. But we always try to make sure that these are kind of aligned, even though it is difficult to ensure alignment across a department as vast as hhs. And so this administration put out an AI request for information, really at the same time as the, or around the same time as the HTI 5 request for public comment on our notice of proposed rulemaking. And really the questions landed in three buckets. It was where are the research opportunities that HHS needs to be strategic about? What are the regulatory levers either regulatory levers that need to be strengthened or what are the regulatory levers that we need to kind of revisit and maybe loosen and then what are the reimbursement pieces that are hindering the proliferation of AI? And I think this is really one of the first times that HHS has been very specific in thinking about the healthcare sector or even the health sector broadly. But I think really the RFI did try and focus on, are basically trying to help facilitate the proliferation of AI in the health sector in a way that's both responsible and fair. But again, this is coming from the Secretary's office. It's intended to focus on what can we do as an entire department to ensure that reimbursement is supportive of the new paradigms that we're seeing in terms of agentic workflows? And how is it that we want to, how do we want to promote this in a way that helps relieve burden and create efficiencies, but doesn't actually then end up in a scenario where the doctor no longer gets paid for it? But we're interested in areas of research. Certainly there's a lot in terms of trying to create common data platforms and research platforms that could be useful and reused. And then on the regulatory side, I think we are thinking how to align regulations, especially those that are appearing to be out of step, and how do we ensure that the regulations that are there are fit for purpose, especially for, again, a technology that continues to evolve and change after it's deployed. And I think within the context of software as a medical device, there has been a very long standing conversation going back several years focused on treating software as a medical device different from a medical device, which of course doesn't change once it's gone through an FDA review process and once it's, you know, once it's given the green light to get marketed Software as a medical device, AI driven, decision support, these kinds of things continue to evolve and you know, for the most part very intentionally. And so the question is, can we design a regulatory regiment that will help facilitate market access to those tools? And then I think there is a very interesting regulatory question around the emergence of LLM based and generative tools that are increasingly being used not just for note taking, but for a whole bunch of things emanating from that AI scribe and offering billing codes, offering orders and drafting other types of things, kind of getting more integrated into a clinical workflow and trying to think about how, how would, how would we go about as a, as a community? You know, one of the things we think about is like, what are the levers of government. But then what. What are those levers that. That are just outside of government? The levers that government will never be able to have or at least won't have until Congress gives us those levers. But. And that necessitates a distributed accountability approach that, you know, I think there is a lot of thought going into and, you know, hopefully over the course of the next several weeks and months start to see some more kind of, I guess, kind of strategic documents coming out and then from there, hopefully some concrete policies that are, you know, aligned with what we heard from industry.
- [29:21]
B
So, you know, I think you're hitting on something that we've seen both at our work at Harvard and in the virtue in the virtues of this podcast, which is that this field is moving at a pace that is outstripping regulators and lawyers, for that matter. They have pointed out that, you know, everything is moving at, like, warp speed and the government is able to move at, you know, snail speed, maybe, if we're lucky. And those two are incompatible. And it leaves this gap of how do we move in a way that is both deliberate and actionable that can help us actually address some of these. And folks are looking for guidance right now, I think, is a real challenge. As you bring. You talked about AICDs, and that's one of the areas that I've been focused on in my work at Harvard, which
- [30:11]
D
is
- [30:13]
B
do we regulate AICDs? Do we have rules? Do we have to have it certified? Do the AI models have to have some kind of testing to make sure that these are safe and effective? I mean, all these are questions that are out there. And I don't know if the ONC is the right place or the right body to, you know, regulate this or comment on this or have a. But I'm looking to you as someone who's been in the field from all angles, and I don't know if you can comment on that or not, but it feels like this is a, you know, folks are looking for what to do next, and the lawyers are telling us we don't have answers yet. We don't have case law. The government's moving on things at a similar pace. So, you know, if you can't say so because of where you happen to be, and I get that. But do you have a thought on that?
- [31:00]
D
I think I'll, I'll observe a couple of things. One is that, you know, the RFI that we, that I just mentioned did come from, from onc, but ONC is, I would, you know, it's got a regulatory purview, and that's through certification of EHRs. It's. And another health it and it's got a regulatory purview on establishing exceptions for information blocking. But I would say where we're looking to make an impact here is more on the coordination side of things, because one of the things that you learn early on once you get to ONC is that we actually can't do all of the things that we want to do. We have to work through our partners at the FDA and through the CDC and through cms. And if you look at meaningful use, we certify the EHRs, but that's voluntary. CMS is the one that incentivizes and regulates the use of this certified technology and regulates the exact use. We can make this certified technology be capable of handling all of the USCDI data elements, but we can't make providers share that information. We can't make providers collect that information. We can't make the providers send that information to specific places. And so there's, there is a very finite role, but on the coordinating side, you know, I am, I'm quite optimistic and that the, the answer to the, you know, to the question on, you know, how should we be thinking about this? What should we be thinking about? This is going to be something that I'm, I'm hopeful. What, what ends up coming out of the, out of the AirFi is, is a, is a sense of strategic direction that helps people understand where they fit in the puzzle. And when you think about the number of interested parties engaged in the task of ensuring that a single patient can benefit from the wonders of this technology, that is, it's a huge cast. It's not just clinicians, it's not just organizations. It's not just industry. It's, you know, honestly, there's lawyers, there's academia, there is a galaxy of, there's a universe of folks who are involved in trying to ensure that this technology is used to its optimal benefit and that its harms are reduced to its lowest level. So I think. Yeah, sorry, I was going to just
- [33:32]
B
say so given that what you're describing around this rfi, what we didn't ask you, which is when is it, what are the parameters around responding?
- [33:40]
D
So actually I just checked and the deadline for comment has already passed. However, that just means that the task of actually making something real out of the RFI and then having people kind of weigh in on that is, is nearer than, than it would be otherwise.
- [33:59]
B
So there's a synthetic time of this exercise I would guess then, yeah, but
- [34:04]
D
I think what comes out of this is going to. I hopefully, again, I hope what comes out of this is something that is. You know, I hate to say this because it's such a cliche. It's a, It'll be a living document. But, but I do think the, the intention is that. I think the intention will be that whatever, whatever kind of products, whatever kind of resources come out of the, this exercise, they, it will help move the community, the universe, the galaxy of different professions and professionals. Try and understand, okay, we don't have the roadmap exactly, but we're starting to build out a blueprint that can help people, you know, kind of help people understand where they're. Their role in building out the scaffolding is.
- [34:52]
C
So, Jeff, I just gotta get to like a really important question about the HTA 5, which is a redefinition of what a user is, that includes automated processes. That's a significant conceptual move. And I just want our audience to understand what that implies.
- [35:11]
D
Right.
- [35:11]
C
So the implication here is if an EHR blocks an authorized AI agent from accessing data, that's information blocking. It's the same as they block the patient or a human clinician from accessing data. How are the HR vendors responding to this framing?
- [35:30]
D
You would have to go to their publicly available responses to, to check to know that. I, I don't, I don't feel comfortable characterizing exactly because I haven't read through them all. But I think, you know, I think everybody. One of the things I really enjoy about health care is that generally speaking, everybody has a sense of urgency and caution at the same time. And it depends on who you are, whether, you know, and it also depends on whether you have urgent caution or you're urgently cautious or you preference urgency over caution. But, you know, I think the tug and pull is real. And, you know, I can't even say exactly when we might finalize HTI 5, but I will say that some EHRs are kind of looking at this and going, yeah, okay, this makes sense. And I think there's probably some out there that are less enthused about this, this notion. I will say, I will say that this is, you know, this is one of those classic government situations where we tried to make clearer in earlier communications. I think there's an FAQ that we put out a while ago that tried to make this, this, you know, kind of this, this notion that agents are. Agents can be people within the context of user and information blocking. But we, we felt that it was Important to do this through a regulatory process so that we could get notice and comment so that we could get some comments on it. And you know, I would, I would be out over my skis if I said how exactly we're going to land on this. But it is something to keep your eye on.
- [37:09]
C
We're going to try to get you to answer just a couple of more questions before we start to wrap. But there's one more thing I need to get to about the changes in direction. So right now the paradigm is about getting data out of EHRs, right? But you're asking a slightly different question in your work. Is it information blocking to prevent data from being written back in? And walk us through that tension. I mean, FHIR supports CRUD operations.
- [37:38]
D
Right.
- [37:38]
C
You can update and delete and create. But HR certification always traditionally required read and search. However, that's just like tradition. And so there's a gap between what's technically possible, what's actually required. How are you thinking about that as a regulator?
- [37:54]
D
Okay, yeah. Now this, now you're, now you've jumped back over into 45 CFR Part 17 0. Now I'm, now I'm on a little bit firmer ground and the infection is
- [38:05]
C
back in your voice, I can tell.
- [38:06]
D
Yeah, I can hear. So I think a couple things. I mean, you're right. The current regulations right now as they stand, are very much focused on the capability to read and search. But I would also point out that we've got a new set of regulations focused on facilitating prior authorization. And for those wonks out there, it's 45 CFR 170.315g, 31, 32 and 33. Essentially, across those three new certification criteria, which we finalized, I think In August of 2025, we established for the very first time requirements for health IT modules that certify to those criteria to be able to support specific subscriptions and specific CDS hooks. And you know, I think to your question, Leon, I think one of the things that we're excited about is this notion of pondering a world where create, update, delete, read and search are all part of what a modern electronic health record can support. And of course we know that they can support this already. I mean, like, we know that this is supported in various contexts. But you know, I think that there's a whole galaxy of, of use cases out there, from wearables to updating of kind of writing back for clinical notes and the Argonauts. I don't know if you're that familiar with The Argonauts, this is a group that was, you know, instrumental and industry group that was instrumental in trying to get that first kind of US core data for interoperability set off the ground and really kind of helped push fire as, as a, as a core forward. They've been working pretty diligently for a number of months, if not years on trying to define these broad use cases where you could essentially establish, write back capabilities in a consistent way. And this could be again, for things like continuous glucose monitors, or it could be for medical device data, or it could be just a common way of helping patients understand when there's been updates to their clinical notes, not just to their lab results. So, you know, I think there's a lot of really interesting work being done in the standards development space. We are not regulating that today, but we're certainly interested in it and we're certainly interested in, you know, following along where the industry goes.
- [40:44]
C
So, Jeff, it's been phenomenal to hear your perspective and, you know, such a service to the industry. So thank you for, for joining us. You're sitting at an intersection of policy, technology and clinical care in a way that very few people have to and get to. And there's one thing, if there's one thing you want our audience of health system leaders and life sciences executives to understand about where health IT regulation is heading. What is it in 60 seconds?
- [41:12]
D
I want to say that the target that we're always aiming for, but I think particularly now, it's fit for purpose and it's also trying to be cognizant of where the future is, where the future is coming at us most rapidly. These are pretty incredible times to be in a position like this because we watch the, we watch the technology advance so incredibly fast in the, in the industry, especially on the consumer side. And what we try to do is we try and look at that and say, okay, where are there connections and technologies and capabilities taking place that we think are good and should be celebrated everywhere? Because right now, as is often the case with technology and the future, it's now, but it's not evenly distributed. And the same is very true of care and high quality care delivery. And we have a say in it. High quality care delivery is going to be through the use and access and exchange of data. And so we look at this and say, there's really cool stuff going on over here, here and here through our levers of certification. How can we make sure that that's happening everywhere?
- [42:18]
C
So, Jeff, it was wonderful to have you, you know, thank you for joining us. You've where can people follow the ONC's work and engage with these policies and conversations if they want to learn more?
- [42:31]
D
Sure you can go to healthit.gov we just rebranded the website to recognize that we're back at ONC no longer ASTP. So healthit.gov of course that domain actually never changed but you can find us there and we put out every and we operate in the public goods so we operate out of the public domain.
- [42:56]
C
We will all like and subscribe this particular department of the federal government. I just want to thank Steve and Jeff and our audience for a great conversation today and hope everyone else everyone can join us again next week for another exciting episode of Practical AI in Healthcare.
- [43:21]
A
Thank you for joining us this week on Practical AI in Healthcare. If you're ready to go beyond buzzwords and hype and explore how AI is truly transforming healthcare, stay tuned for more conversations that get us to what works. Until next time, stay practical.