Ep. 14: Steve Yaskin - AI Agents, Clinical Data, and the Future of Digital Health
Bradley Bostic
Healthcare Technology Investor, Founder, CEO, & Board Leader | Host of Boombostic Health Podcast
"I think the major revolution is happening right now, and it's not just about AI in healthcare or training models. It's about executing automated workflows, reducing administrative burdens, and looking holistically at clinical data to address gaps in care." – Steve Yaskin, Founder & CEO of Health Gorilla
Welcome to the Boombostic Health Podcast newsletter!
Get weekly insights and candid discussions with experts who are shaping the future of healthcare. Explore trends, technologies, and challenges in an accessible way for all. Click subscribe for our newsletter. Join the conversation by listening to our latest podcast episode here.
Episode Spotlight: From Theoretical Physics to Healthcare Innovation
Episode Summary:
In this episode of Boombostic Health, we welcome Steve Yaskin , a visionary healthcare entrepreneur and founder of Health Gorilla. Steve shares insights from his journey in healthcare, starting from his background in theoretical physics to becoming a leader in clinical data exchange. He discusses the evolution of Health Gorilla and its role in the Trusted Exchange Framework Common Agreement (TEFCA), a significant milestone in healthcare data interoperability.
Steve also introduces his latest venture, BeHuman.Health, a startup currently in stealth mode, and shares his excitement about its potential impact on the healthcare industry. The conversation delves into the importance of clinical data, the challenges of data utilization, and the transformative role of AI agents in healthcare workflows. Steve emphasizes the shift towards preventative care and the need for comprehensive, longitudinal patient data to improve health outcomes.
Key Topics:
A New Venture in Stealth Mode
BRADLEY: Hey everybody and welcome back to Boombostic Health. We are in episode 14 here and have an outstanding guest, Steve Yaskin, who he and I go back, I don't know Steve, you think it's been 10 years, 12 years? I don't know, but yeah probably like let's say 12, so it's like 84 web years because we have to measure everything and you know like dog years as we've, when he and I first met, we were just crawling right out of the nursery and getting things up and going.
But Steve is a fantastic health care entrepreneur, fits the mold perfectly for our theme around bold innovators that have found that model to have the economics to support their ideas and I think for Steve you'll hear it wasn't a straight line, a little bit of evolution, maybe a couple pivots here and there. He's joining us from his home down in Florida.
Steve, just to give you a quick introduction on who we're talking to here, founder and board member of Health Gorilla, which is an incredibly powerful clinical data information exchange platform and also is with DHVP as a founder, I'm sorry, a partner where he is involved with making investments into digital health and data-oriented businesses.
But I think maybe most exciting is that Steve has that glimmer in his eye again because he's got a new startup that is still in stealth mode. It is called BeHuman.Health. I think I may be the first to actually report that and or at least one of the few. Am I the first?
STEVE: Yes.
BRADLEY: You heard it here first, so Steve, when that is, when you cross the billion-dollar enterprise value mark, we'll look back and say this was that moment we first shared with the world that you had your new idea underway. The business is in stealth mode, so there's no website up yet even, but I would suggest keeping track of where Steve goes with the business BeHuman.Health.
So Steve, I'll turn it to you. Would you mind just giving your own introduction and your own words? And I would love to have you start with why did you get into healthcare to begin with? Lots of ways to apply technology. How'd you find yourself in this space?
Health Gorilla: Shattering Data Exchange Barriers
STEVE: Right. Well, thank you for having me, personal, friendly, and thanks for the warm introduction. We do go back probably about 10 years, right? And my brief journey started, first of all, my training is in theoretical physics. For some reason, I was always fascinated with that. A couple of degrees later, I got into data science, largely because numerical physics these days is mostly large data, massive data processing, various models, and things of that nature and that ultimately brought me to the world of healthcare, where we produce petabytes of data. Healthcare is one of the last frontiers where technology is still in the world, but very, very slowly proliferating in healthcare, and largely because of the regulatory space and because of the large amount of data. Frankly, because the stakes are very high in making a mistake. So, that combination brought me into the healthcare space.
I was also the founder of one of the largest primary care delivery networks in the Bay Area, which started probably about 15 years ago. We grew to close to 50,000 patients under management. It was a very traditional primary care delivery model, and powered by technology. This is where we started bringing tech into the primary care delivery model. It was acquired later on. Into the primary care delivery model. It was acquired later on as we scaled and that experience, I used to call it fighting the war in the front trenches because primary care is where it happens, right?
We all watched the Super Bowl this weekend and the primary care doctor is like a quarterback. You coordinate, you direct, you manage, you oversee the whole field, and you need to know where the puck is going to go in specific treatment for that patient. It's a very interesting analogy there.
From that experience, I moved into the world of clinical data, founded Health Gorilla. This was a pure tech way, and we always had this vision of making entire longitudinal chart history of every patient available on demand in real time to providers treating that patient.
Interestingly enough, this bold vision coincided with regulatory space quite nicely. Back in the day, on the heels of something called 21st Century Cures Act, President Obama actually passed that act, and it had good intentions, but at the time, it had been just sitting on a post, not seeing a lot of execution, until much later when Health Gorilla was formed.
The timing of the company's vision coincided with Congress coming down to this Division of Human Health Services called ONC at the time, they got recently renamed, and Office of National Coordinator for Health, and the basic question was, hey, we have this act passed by the Congress, what have we done? And there was not much to show, so they started actually implementing that in a very different way from how government traditionally passes acts into the healthcare field, and this was formed as a partnership between the public and the private sector.
Health Gorilla fit very nicely into that mold, and we started being the partnership between the public and private sector, building technology in accordance with the regulations, and that culminated in something we call TEFCA today, Trusted Exchange Framework Common Agreement, and Health Gorilla is an essential part of TEFCA framework in the United States, it's one of the qualified health information networks, since December of 2023.
To make that long history short, I was very excited to be able to be part of that journey, build that national infrastructure, and now it's time, I think, to execute and actually build on top of that infrastructure, which is partially, I wear two hats:
BRADLEY: That's very cool. So, I just have to circle back to one thing, so I, total tangent, but I was at the Super Bowl, and so I didn't see all of the ads, I, you know, they put some of them on the Jumbotron, but you're people watching, and was there one that stood out that was the most interesting? And even, to your point, also, anything healthcare, you know, any ads that jumped off the page for you?
STEVE: I wasn't watching too many ads, to be honest with you.
BRADLEY: You're probably working on theoretical physics during the downtime, okay, all right, so you didn't really, and the game was a blowout, but what about the Halftime Show, did you watch that?
STEVE: Yeah, I did, I was actually, I liked it.
BRADLEY: Did you? You liked the Kendrick Lamar show?
STEVE: Yeah, yeah, I actually enjoyed that.
BRADLEY: Yeah, I thought the dancing was pretty cool, awesome, well, hey, with the whole range of things that you're doing, which, by the way, I missed the, all the details on how the primary care business grew up, I'd love to talk more at some point about that, but one of the areas that is of particular interest for you, as I understand it, is artificial intelligence, and in particular, these AI agent kinds of constructs that are doing some really innovative things to how you can transform the way people are cared for and how healthcare operates, and I'd like to talk about that, but as we build the foundation for that, I think a common passion you and I have is the data, and not just data generally, which historically in healthcare people would refer to claims, which are just this sort of artifact related to what was billed for, but in reality, there's this massive untapped opportunity with clinical information that goes down to the specificity of what's actually happening with the patient's story. story. And that's what I know you're passionate about and I share that passion.
So maybe we just start with that foundation of the data, how important it is, how you think about that from an investor and an entrepreneur perspective. And then maybe hit on some of what's that regulatory world like, which is an area where you're an expert. Then I think ultimately I'd love to get to, let's talk about this whole idea of AI agents and how powerful that is. But I'll turn it to you, Steve. What are your thoughts on just healthcare data, where we are today and what that opportunity looks like?
The 3% Dilemma: Unlocking Healthcare's Untapped Data Potential
STEVE: Yeah, no, sounds good. First of all, let me just say, we have a very interesting investment criteria, investment thesis at my investment fund called DHVP, Digital Health Venture Partners. It's based out of Washington, DC. And one of the postulates, if you wish, how we look at companies coming through the deal flow is very simple.If a founder comes in and says: “We build this great go-to-market strategy and this is a great idea, but we don't really need to know anything about the patient. We don't care about clinical data history or anything like that.”
I walk away right away. That's the deal breaker for me. I don't want to hear anything else after that. And then there are founders that do say, we have all the data in the world, right? I also, that's another turnoff for me. Nobody has enough data. There's never enough data on the patient. There's always more sources to discover to fully complete the history of that patient. And then there are founders that look at the data as a commodity, meaning it's the same data as your financial data. As long as you encrypt the data, you're in the game. You encrypt it in storage, you encrypt it in transit, and people just change the data from that perspective. And that's also a non-starter because it's a heavily regulated data and you need to know what you're doing at the very early stage of the company. Hopefully, it's going to scale and you want to be compliant with all the regulations. Once you know how to do that, you just execute on that, but you need to be serially aware of that.
So going back to the premise of utilizing clinical data in healthcare, I think I started this conversation by saying we, in healthcare, produce petabytes of data. It's staggering the amount of data systems automatically generate, providers produce. There's a whole huge industry of medical devices now automatically spitting out data around the patient. Unfortunately, as an industry, we utilize close to no more than 3% of that data is actually actionable and it's actually utilized. And then this is, I think, the last frontier in this country, the place to learn and recognize the value of utilizing clinical data.
BRADLEY: Hey, Steve, real quick, on that note about the 3%, so I hadn't heard that stat before. So unpack that a little bit. So we're saying that there, I don't know, I love the comparative analytics around how the amount of data being generated from healthcare compares to all the information in the Library of Congress or whatever. It's some massive amount, but where's the 3% coming from? What does that mean?
STEVE: Yeah, I wanna make it very different. Clinical data is very different from claims data, right, like you mentioned already. Let's utilize claims data and there is a massive industry that is consuming claims data, producing and consuming claims data and we learned how to analyze that data and maximize that data. But this is revenue cycle management, this is extracting more dollars, savings and things like that. It has very little to do with actual clinical data and actual treatment outcomes when it comes to that. This is where we are only using 3% of that data and the way we are using that, if you look at the entire history, everybody is looking at a very narrow element within that data.
BRADLEY: Like a point in time based on some acute symptoms or that kind of thing?
STEVE: Point in time, if you are a provider, if you are a lab, you are contributing that blood work, if you are radiology, you are contributing an icon image with a CAT scan or ultrasound. Some providers know how to look at the last 30-60 days of the clinical data and make diagnosing decisions for example, but it is all bits and pieces.
I think the major catalyst in turning the point where we start recognizing the value of the clinical data, surprisingly it does not come from providers, payers or value-based care, but it comes from technology and this is where AI agents come in. This is how I connect the two industries. I think the major revolution is happening right now and it is not pure AI in healthcare, it is not training your models, it is actually executing automated workflows, reducing this administrative burden and actually looking holistically at the clinical data to execute very narrow workflows, whether it is onboarding, recognizing gaps in care, reducing spam, preventing readmissions.
There are a lot of clinical workflows that actually today are being automated. Nodes capturing ambient AI agents that create nodes and reduce administrative burdens. AI agents are how we will make that massive amount of data we produce in healthcare actually actionable.
AI Agents: The Silent Revolution in Healthcare
BRADLEY: So, you say AI agents. You can think about a humanoid walking down the hallway. We had a podcast guest that was talking about the need to create the sterile environment and how there were robots basically that could do that kind of thing and that was a whole different layer of atoms versus bits. How much of what you are talking about is an agent in a physical construct versus something that more fits into a software flow and maybe specifically, you talked about onboarding. Does this make it so finally we do not have to fill out the same admission information 50 times?
STEVE: Exactly.
BRADLEY: Talk to us about that.
STEVE: It is not as exciting as a humanoid roaming the hallways of a hospital, even though that has been done. I have seen these tiny robots bringing medications and changing sheets and hospital beds. This is happening in parallel, but this is a very different area. I am talking about AI agents that actually sit in the background. Sometimes they would surface through voice interactions with the patients. Sometimes you would not even know they are deeply embedded into the clinical workflow when you just freely, openly talk to your provider and there is an ambient listener capturing everything you are saying, transcribing that into a clinical note and summarizing that later to identify improvements in the outcomes like gaps in care and missing tests and things of that nature.
These are silent agents that are working in the background and I think one of the biggest hurdles in healthcare today is the administrative stuff. This is where most of the waste in healthcare comes in this mundane administrative burden, creating burnout with providers because they are inundated with the amount of paperwork they have to produce. Perfect job for an AI agent. It is impossible today. It is really hard to navigate simple tasks like creating appointments all the way to that ubiquitous clipboard that you get handed and you fill out. I don't know how many times each visit to the provider organization. Perfect jobs, perfect workflows. To the provider organization. So perfect jobs, perfect workflows for AI agents to tackle over the next couple of years. We'll see more and more of that.
BRADLEY: So as you're thinking about the future of your focus and you talked about building on top of this clinical data that's now being consolidated, I think as I hear you say that, I reflect back on this journey that we've been on in healthcare, not just in the United States, but in particular in the United States with the high-tech acts to establish just a foundation of electronic health records and making that information digital. When that happened, there was this, I think misperception or the alignment of expectations around what that was going to get. All of a sudden, because everything was electronic, we would have universal access to comprehensive longitudinal records and clinicians would be able to much more efficiently do their jobs. It seems as though what's actually happened is in some respects, we now have a lot of information that is stored, but not really utilized other than when the physician shows up in the office.
I think about lab data with our company hc1 and Health Gorilla has got a major component that's lab related. Typically a lab result is looked at in a point in time because a patient shows up with some symptoms and then it's trying to figure out what medication to put them on or diagnose them with something, but it misses that bigger perspective on what's going on.
You would have thought, wouldn't have electronic health records solve that, but there's this layer of intelligence that is the next frontier, I think, that starts to look at that data in a holistic way. Like we talk about with hc1 strategy, we're all about unifying the patient journey across the provider world and all the way into clinical research in the event that people fail some kind of a conventional therapy.
And that requires a view that is comprehensive and longitudinal. And that requires technology that can identify those constellations of signals. Is this in line with how you're thinking about this next leg of the journey, Steve?
The Shift to Preventative Care: A New Era
STEVE: Yeah, absolutely. As a founder of several companies and a CEO and operator, I always have to know where the puck is going. And there's a lot of things going on on the Hill with the new administration coming in. So it's largely unknown on the small details yet, but it also is very much known and visible that the market is changing towards prevention away from acute care.
And I think we're gonna start seeing that over the next decade where United States healthcare will finally start moving towards preventative chronic condition management away from fee-for-service models that are designed to, you know, this vicious circle. We create the problem, whether through food supply and many other things that are targeting healthcare in this country, all the way to the acute treatment where we spend more than 80% of the budget treating small percent of populations with chronic conditions, right? We never really focused on preventative measures and I think part of the problem was lack of this longitudinal data set. We could not look until it was too late into the data and that comprises a lot of many factors, right?
If you look at the history in the United States, regulatory history and how market responded to them, we've done some great things. For example, meaningful use, the HITECH Act, which moved providers away from paper chart. For those cabinets with those manila charts where the term pulling the chart comes from, actually somebody had to go and pull the chart off the shelf. We moved to computers and this was a great thing. This was the first step in creating this framework infrastructure that we call FHAS today. But it also created drawbacks. It created a lot of negative because markets reacted by creating a lot of silos in how we store digital data. We had over 700 electronic medical records, software vendors selling it to the providers.
Yeah, sure, there's half a dozen large acute EMRs, but there's a long tail of second tier EMRs and labs and radiology imaging centers. Of second tier EMRs and labs and radiology imaging centers and PACS systems and lab systems that actually continuously silo that data and break it down. And they build walls in regulatory environment. I'm not gonna share it because of regulations. I have to adhere to them, right? And that created a monster.
Now we are at the second stage where we spent decades creating an exchange framework where these walls are, you know, in some arguments they are being removed, but in the same argument they are being strengthened by improved regulations like what QHINs are doing in TAFCA. We narrowed down definitions of the permitted purposes: “Who are you?” “What is the reason that you are trying to access this data?”
We defined treatment, we defined payment operations, we even defined consumer access, a boring government name called IAS, individual access services, but this is a basic right of every patient, every consumer in this country to gain full access to that data. It's still early, it's still probably a couple of years away from being ubiquitous throughout the nation, but it's coming. Utilizing, making that data actionable.
HIEs: From Obscurity to Central Stage
BRADLEY: So Steve, health information exchanges or HIEs, that was a concept that, I don't know, 15 years ago started to become in vogue as the idea I think was to make the information more portable and allow for patients, for example, that go to a doctor or some other place to have access to their diagnostic results. And then Health Guerrilla comes along and did what exactly, and how does it relate to those HIEs that were more regional and almost like non-profits? Because I think Health Guerrilla is a private, like for-profit enterprise. Is that right?
STEVE: Yeah, it's a private, for-profit enterprise and we did work with a lot of HIEs and HIEs served a purpose, right? There were pretty hot commodities when the movement started, again, some years ago when they started aggregating data within the state, county, basically regionally. In concept, there were no different than EMRs that were building this walled gardens around the big elements, but at a larger scope, sometimes within a state, sometimes within one or two counties within a state.
BRADLEY: Sometimes there wasn't a lot of technology either. I remember being very underwhelmed when I went and met with the chief technology officer of an HIE and I think it basically was a call center where you could call in and get access to a medical report that was static. I just didn't understand what there was. I guess it was a, instead of having to call the doctor's office to have them fax you something you would call the HIE and they could fax it.
STEVE: So yes, there's been a lot of mixed success with HIEs. Some became really good, actionable HIEs, some fizzled out and closed due to lack of local funding. But I think going forward, HIEs do have a role in this federal network of data exchange.
BRADLEY: And then how does Health Guerrilla relate to that? How does Health Guerrilla relate to an HIE?
STEVE: A lot of HIEs started connecting to the federal backbone while still, you know, playing their role in the local community. They are actually building that bridge to the national federal highway that's going across all 50 states. And they're kind of on ramps into that information highway. So they play a very central role in regions where they're active. They are slowly but steadily joining the federal program.
BRADLEY: Okay, so like you said, similar to EHR where you have to take one step at a time. First the data has to be digital, then the data has to be regionally accessible. Then you create this layer, this network that then can unify all, you know, the entire country, all of the HIEs across a common fabric or network. And that's where the Health Guerrilla capability helps to create more accessibility, more portability, more longitudinal information. Is that fair?
STEVE: That's absolutely fair. And as this federal program was largely successful and we've been at it for the past four years, I think now is the time to build vertical solutions on top of that infrastructure. And that's where my head is these days.
BRADLEY: Got it. So there is a massive consideration as it relates to regulations and information privacy and I couldn't agree more that so many of these regulations have been misconstrued and misdirected to become the enemy of progress. Have been misconstrued and misdirected to become the enemy of progress, which is not what they were intended to be. Even just HIPAA, for example, is something that I think many people, even in the provider world, don't even understand what it was for.
And they're bringing up HIPAA when it isn't even relevant in the conversation. It can be as a means to say, no, we're gonna block all the access to everything because we're afraid of HIPAA. When in reality, that's not the intention. You're missing out on the chance to actually prevent illness, to make people healthier. An example that's a really cool one, I would give the data science team at hc1 actually constructed a model using tens of millions of laboratory test result values over a lifetime, multiple lifetimes, to analyze who were the people that, or what were the indicators that would say somebody who's going to develop cognitive decline.
So you talk about the example of being preventive. Well, with the new drugs that are very promising for cognitive decline and Alzheimer's, they can't reverse that problem. They can slow it down though.
If you could see these things proactively in that constellation of signals over time, you could actually manage that like a chronic disease much more effectively, keep people independent, which you want to talk about a huge drag on the health system and the costs as well as obviously the quality of life issue for everybody involved.
In order to do that, you have to have a really unique access to information, you have to utilize it in a de-identified way, then ultimately you have to start integrating those models and those capabilities into the actual delivery of care. There are a whole bunch of steps that are involved.
You talk about building on top of now what's been assembled on this national network, but is that all de-identified? For the entrepreneurs that are listening to this, they're like, hey, Steve, that sounds awesome. It sounds like you're of an abundance mindset where rather than being proprietary about this, you're out there evangelizing like, hey, let's go plant a thousand flowers and see which ones are going to bloom. What are the constraints? What can you do and what can't you do? Are you actually able to identify an individual patient with a solution in this kind of a network? Talk to me a little bit more. I'm not really clear.
TEFCA: The Game-Changer in Health Data Sharing
STEVE: It's no secret there's a huge industry out there that trades in unidentified data, sometimes with bad consequences, sometimes with really breakthroughs, powering research, life sciences, and many other industries. The current infrastructure that I'm talking about, powered by TEFCA, which stands for Trusted Exchange Framework Common Agreement, is actually the ability to exchange identified data for one of the six permitted purposes. It's part technology, part regulatory.
Technology powered by several laws that were passed recently, like data blocking laws and many other things, prompting providers to actually kind of forcing them to change data, but also regulations that say you need to identify yourself as a member of the TEFCA slash QHIN organization. For each query, for identified data sets, you have to specify what type of organization are you. You do it one time as you join QHIN, but you have to specify what is the actual reason for your query.
Is it for treatment purposes, and I'll talk about that a little bit in a second, or is it the consumer asking through one of the APIs for his or her longitudinal data set? Is it the payer, or is it the health care operations company, or is it one of the government agencies like CDC or VA, and they have different purposes - Eligibility verification or actual CMS is now playing a bigger role in health care data exchange
So everybody has their reason, but they have to specify, and it has to be within those six doors, and I think one of the main reasons TEFCA is becoming prolific and actually growing so fast is that these doors are now very clearly defined, and not all of them are open yet. We're still working on several doors that...
BRADLEY: So once you go through one of these doors and you get information...
STEVE: Yeah, treatment is the one that's working.
BRADLEY: Okay, treatment.
STEVE: There's been a lot of national networks like Care Quality, Commonwealth, and others that pioneered data exchange for treatment purposes, and TEFCA kind of consolidates regulatory platform, brings technology into that, and makes it really federal, ubiquitous, and treatment is the first door that went live more than a year ago now, and there is this growing exchange for treatment purposes.
BRADLEY: So once a... So specifically, a patient presents for treatment, the clinician is able to make this request against the network. Does that request go through their electronic health record, or is there a separate way that they make the request?
STEVE: Yes, most likely it will go through their source of record, which is EMR, EHR. There are APIs that are published by QHINs and maintained and audited by QHINs, and it's pretty much easy, it's ubiquitous, and most providers don't even know they're participating in them because data starts to come in through their EMR, and kind of then breaches their own records on that patient from external sources nationally.
BRADLEY: Okay, okay, and then once that information is retrieved, does it have to, is it stored then by that endpoint that made the request to, okay, so once you've made it, then you're able to retain that information as part of their health record, and it was made for treatment reasons, which is a valid reason, and, all right, okay.
STEVE: Don't forget, one important thing here is that treatment purpose on the TEFCA is bi-directional. So if you are able to pull the data, you must submit data back. So it's bi-directional exchange, and this is kind of a network of growth for that purpose.
BRADLEY: Okay, so what would you give back then? Once you make that request for that information, what's then flowing back into the exchange?
Regulations vs. Innovation: The Ongoing Battle
STEVE: It's in the definition of that exchange. It's for treatment, right? So you are a provider treating that patient. You generate your own progress report, your own data and plan of treatment, and you do your own diagnosis, and you prescribe medication so the entire clinical record that you produce goes back automatically into the TEFCA so that the next provider who is treating that same patient alongside with you or in the future has the same access, and this is the network effect of growth.
And when I say provider, this has been a long debate, right? It's not like invite. If it's a digital app that, for example, goes after a heart condition, there is a doctor at the end of that app that has created the clinical algorithm that allows to catch a chronic condition related to heart. Is that considered providing treatment or not, right? So there's been a long debate, and finally TEFCA kind of streamlined and defined very succinctly and precisely what constitutes treatment.
And obviously, like you mentioned, HIPAA was created more than 25 years ago before we had internet, right? So TEFCA actually looks at the current conditions, current technologies, and there's been all kinds of very interesting debates, and leading all the way to one of the bills introduced recently is can AI write prescriptions, right? I think this is kind of where the industry is going. This is how we're thinking right now. Not necessarily replacing providers, but can AI be trusted to write prescriptions? And of course, pushback is no, doctor has to verify. Fine, it's a co-pilot mode, but the bill actually says, well, we should probably look at that because it's proven that AI could be more effective than providers.
BRADLEY: That's interesting. Yeah, I'm a big advocate for RLHF, reinforcement learning through human feedback, and incorporating that kind of a virtuous cycle making the providers bionic, or whether that provider is a pharmacist or primary care doc or whatever type of specialist. It's not that we're going to all of a sudden leap to using AI as a substitute for that human's judgment because clearly there's a lot that goes into health and goes into how a patient is doing that goes well beyond just the basics of what does a specific test result say that humans need to be informing.
DecisionRx, Inc. I think you and I touched on this a little bit. It's one of my portfolio companies that I co-founded. It's interesting you raised that prescribing example. In that business, we're actually using the technology layer to analyze at a level of granularity all of the information from a diagnostic and prescribing perspective, identifying those high-risk people, and then surfacing that information through an easy-to-use interface. It's almost game-like for pharmacists who then are able to say, well, why is this drug red and how do we make it green or yellow? Because it could be doing more harm than good. People generally take for granted the fact that you're always getting the right prescription or that the prescriptions that a doctor would give to you would be good for help you and it's not intentional in general to hurt you with medications, but it's incredibly common. I think that is an area where what you started out with on the AI agent front, you've got this ability to power up the pharmacists who want to engage to optimize care for patients.
And I think what this does actually using that agent model, where there's this copilot, as you said, that gives you this chance to make it so that people aren't falling between the cracks.
And those people that fall between the cracks.
I'm sure the ones who end up costing the healthcare system the most and having these terrible outcomes. Is that a decent way to look at it?
The Future of Healthcare: AI Agent Equation
STEVE: Brilliant description of how to make data actionable. Yes, I am familiar with that company. I think it's a brilliant idea to actually make data actionable using existing infrastructure. This is a very exciting example of how you build vertical solutions using the infrastructure, the new regulations, and the AI agents. This is where the industry is going to be focusing, I think, for the next decade. This is a big frontier Rubicon that we have to cross right now.
AI agents are only as good as the data that they have to cope with. I did not say the ML training needs to get all kinds of data. What I'm saying is agents executing specific workflow for that given patient have to work with the longitudinal history for that patient. It's not anything big, national. It's actually looking at the individual patient and automating that workflow looking for the gaps in care.
The second part of that equation is making AI agents equitable, meaning we spend the majority of our spend in these budgets on underserved population in treating acute care. What I think will be changing also is creating actionable preventative measures on this population. If you look at the federal agency called HRSA, more than 40% of US population is underserved. They have various criteria to determine that, but one of them is access to primary care. People have to drive long distance to see a doctor. Doctors are not available, there's no schedule, there's insurance gaps, and there's a lot of factors here. This is where we spend most of the money once this population gets sick and gets admitted. Then we spend billions and trillions of dollars, more than 70% of GDP now, and 80% of that spend is treating chronic conditions that we missed to begin with.
I think this is where AI agents could play a crucial role in approaching the underserved population, making it really easy to access. Of course, telehealth is one of the platforms or delivery mechanisms for crossing that chasm and reaching out to that population, whether it's preventative screening, or kids, blood work, testing.
BRADLEY: Well, and especially, Steve, as we have a shortage of actual clinical professionals. The problem keeps getting harder and worse. That's interesting. I did talk to your pseudo co-founder, probably like three years ago. We're talking about some of that exchange of information that when you actually draw down from the network, you have to then reciprocate with that. What is the scale of that network now, of the Health Gorilla Network, in terms of number of humans, number of US patients that it's touching?
STEVE: Well, I'll say one thing, just so you understand the level of adoption. The first year was kind of slow, until Epic announced that they're shutting down all legacy exchanges in lieu of joining TEFCA, so all their hospitals went to TEFCA. Epic is the second, Health Gorilla was the first, Epic was second. Epic announced abandoning, shutting down all legacy networks in lieu of the new TEFCA. This is where it accelerated at a tremendous rate. By end of this year, most of the exchange in this country will be done on TEFCA.
BRADLEY: Okay, very cool. So, TEFCA basically showed up at the right time to align with what you were working to build and then that created this...
STEVE: It created the right time, right? It actually forced partially through private sector adoption and recognizing the need to exchange, powered by the regulations like data block. So TEFCA did not show up at the right time, it created an opportunity, it created the right time.
By the way, the third QHIN that I want to mention briefly to you, where you asked me about HIEs, e-Health Exchange is the third QHIN that is a federal number. E-Health Exchange is the third QHIN that is a federal nonprofit, but it is largely working with the state and regional HIEs to bring them into the federal program. And those are the three things that are driving TEFCA right now.
How Gorilla does payers, consumers, treatment, digital health apps, value-based care, Epic serves majority of the acute hospital, not majority, but more than 50%. E-Health Exchange is really prolific with the states, with HIEs, and those three lacks that same stool is how TEFCA became ubiquitous, becoming ubiquitous in the country.
BRADLEY: You and I should find a time to get on a whiteboard here soon. I'm incredibly excited about all these possibilities, and I think it's super cool how you persevered with your idea at Health Gorilla. I was there for some of that journey, and that's a whole other conversation. But now that there's this platform there, as you say, this is that next big leap. How do we make that together?
So thank you so much for being on the podcast. You're an amazing innovator in the healthcare space, and I kind of get the sense you're just getting started, which is exciting. I definitely ascribe to the Amazon coined day one mentality, and I think it seems like you do as well. You have to get up every day like it's day one and go make it happen. I'm really excited to see what comes out of your next major innovation at Be Human.
Thanks so much, Steve. I appreciate you, and thanks for being a friend of the podcast.
STEVE: Thanks for having me, Bradley.
BRADLEY: Have a great day. All right. Thanks, everybody, for joining the Boombostic Health Podcast to hear from Steve Yaskin. Next up, we'll be getting into the verdict with Emily to talk through some of these regulations in a more granular level. Take care.
The Verdict with Emily: Legal and Regulatory Insights
BRADLEY: Hey, everybody. Welcome to the verdict with Emily, where we have just had a great conversation with Steve Yaskin, who's a true innovator, who has built Health Gorilla, which is an information sharing network across the HIEs in connection with the TEFCA regulations.
As we get into this legal and regulatory analysis, I think, Emily, you have a point of view about the promise of this kind of capability to make information about patients more portable and also some things that we really need to look out for that could be big pitfalls.
EMILY: Yeah. Yeah. So the idea of an integrated health information exchange is a really powerful concept that can no doubt improve patient care. The thought that my provider here can share my information, and if something happens when I'm on vacation in another state, my information is readily available. They know my allergy information, any pre-existing conditions. There's significant benefit that can derive from this type of information sharing when done correctly.
There is also the concern, if not done correctly, right, especially when you have a basic name like mine, that the information will be combined together in a manner that could cause my information to merge with another Emily Johnson's out there, which obviously would heighten risk to patients.
The other thing that we hear a lot about right now is how information is being used and shared with folks who might have access to this type of health information exchange. I think from a consumer perspective, patients don't really understand how their information is being put into the exchange, what rights they have with respect to limiting use and disclosure of that information. And that is a very real problem for the providers too, right, making sure that they are only using and sharing PHI in accordance with what's permitted under HIPAA as well as state privacy regulations.
Like all things privacy, right, this evolves on like a near hourly basis. So we're definitely going to see this evolve as well over time. But when you think of like the regulations that were, was it December when the Biden administration changed the security guidelines for how providers had to secure information, I think we're going to see a lot more of that because the last thing we want to have happen is, you know, we established this great network of patient information and then another Change Health like incident happens where people's information is breached, you know, millions and millions of people's information is breached.
BRADLEY: Yeah, it's a tricky issue because there's no question patients are best served when caregivers have comprehensive access to their record. Even more so, it's even better when you've got this capability to use technology to power up better care by analyzing that data.
How do you recommend or what are your thoughts on that balance between these regulations being used as a justification for blocking access? Because everybody's so afraid of all the risk and on the other side, or those people who are trying to block access say they're afraid of the risk on the other side, patient care can't be improved or optimized if you don't have access to that full picture. Like we sort of have to figure this out, right?
EMILY: Yeah, it's definitely a problem that needs to be resolved. I do think though, I would challenge you on the aspect of people blocking access to information. I don't think that's necessarily the intent of the regulations is to block access. I think instead, it's to secure the information that is being shared, right?
I mean, if you look at the information blocking provisions that were passed in 2020, that basically give patients full rights to their records, lab results, things like that, without even a provider reviewing it. So there is certainly a change, I think, based on the concept of personalized medicine to give patients access to their information quickly.
This situation to me is different because it's not the patients that are getting the access to the information, it's the healthcare providers. But we've talked about this before, as the practice of medicine continues to become more and more electronic, no doubt this has to happen, right?
It's just making sure, and I think that is the intent of the regulators, is to make sure that the information that is shared is done so securely.
BRADLEY: Absolutely. Well, really appreciate the perspective, Emily. Always love to double click a bit on what we hear from our guests. And certainly Steve had a lot to offer about the promise of building capabilities on top of this platform. I look forward to continuing the conversation as we move forward and seeing where this can go.
And thanks everybody for being part of the Boombostic Health community. We'll look forward to seeing you next week.
Get Involved:
Are you passionate about transforming healthcare? Join the conversation by listening to our latest podcast episode here. Share your thoughts and insights in the comments below or reach out to us directly. Let's work together to drive meaningful change in healthcare!
Stay tuned for our next edition, where we'll feature more expert insights and innovative ideas shaping the future of healthcare. Don't forget to subscribe to Boombostic Health for regular updates!
Healthcare Technology Investor, Founder, CEO, & Board Leader | Host of Boombostic Health Podcast
4moNobody has more experience with making healthcare data actionable at scale than Steve Yaskin . Check this out to hear where Steve sees healthcare innovation going in the future! 🚀💡