1upHealth in the News – Where Are We At With FHIR? with Joe Gagnon and Dr. Don Rucker from 1upHealth

John Lynn:

Hi everyone. I’m John Lynn, the Founder and Chief Editor at Healthcare IT Today. We’re excited to bring you another in our series of interviews with top leaders in health IT, and today we have two extraordinary guests. We have Joe Gagnon, CEO at 1upHealth and also Dr. Don Rucker, Chief Strategy Officer at 1upHealth. Welcome guys.

Dr. Don Rucker:

Yeah. Thanks.

Joe Gagnon:

Great to be here, John. Thank you.

John Lynn:

Yeah. So I’m excited for this discussion. I think this discussion will be on FHIR. I think I had to have the mandatory FHIR joke, right? Anyway, before we dive into FHIR, Joe, do you want to tell us a little bit about yourself and 1upHealth?

Joe Gagnon:

Sure. 1upHealth, we call ourselves a disruptor, enabler in the health IT space, bringing a new way of thinking about how data and interoperability are going to transform the health industry. We do that by building out a cloud platform that allows for the use movement, compute analytics against this FHIR data to transform really everything from quality and risk all the way to reimbursement and care. So the next generation data platform for the health industry.

John Lynn:

Yeah. Well, you’ve come a long way. So I’m excited to learn more about and educate everyone else on how far we’ve come, and what’s still left to be done. Don I think most in our community know you, your background at ONC et cetera. But tell us, why did you choose to work at 1upHealth after ONC?

Dr. Don Rucker:

Yeah. As folks know that ONC implementation of the Cures Act really of Congress’s desire to bring a whole new level of engagement and accountability on behalf of patients using modern technology was the center point of what we did at ONC. Almost unanimous Act of Congress, bipartisan support from President Obama, President Trump, and now President Biden.

Dr. Don Rucker:

That all having been said, how do you actually get to do all of this amazing thing, right? How do you get the modern back ends you need to provide all of these consumer frontends, whether that is a frontend driven by analytics from a payer, from a provider, from a startup, an established app company? So all of that requires literally a modern backend if we’re going to do this all in real time. And looking around, that was obviously what I’m most interested in doing. And 1up is unique in having thought through what it takes to have a modern cloud, scalable backend to do all of this. So pretty exciting. And on some level, obvious.

John Lynn:

Yeah. So Don, you talk about this, where are we at in FHIR being that standard and then also health IT vendors being able to provide this modern infrastructure and really a foundation to be able to access the data and make it accessible for people to actually use it. Where are we at on all that?

Dr. Don Rucker:

That? Yeah. Well, that’s a big question, but let me throw out a couple of quick things here for folks to think about. So I’ve been doing health IT since the mid 1980s, arguably earlier, depending on how you count. And this is really the first time that we have an elegant computable standard that comports with the rest of modern computing, right? So all the software stacks that are on our phones frontends back ends. They all use JavaScript object notation. FHIR is the healthcare version of that. So for the first time, we’ve actually gotten health data technically aligned with the rest of computing. And so that is huge. I think your question implies the well, but what about type of thing, which is the data in the electronic medical records is all often bespoke, private formats, right?

Dr. Don Rucker:

They were all built before standards. And so there was no choice other than to do it. What Congress had with application programming interfaces without special effort is really the cherry on top of literally now 10 plus years of work in converting and making healthcare data FHIR enabled, right? So that was really literally the forcing function, but the technology has been there and with the various rules and what vendors have done even independent of the rules. All the major vendors have robust FHIR APIs up and running today. They may not have all turned them on for a variety of reasons until the end of this year for the US Core Data. But they actually have made major steps to take their internal data structures and externalize them as FHIR resources. So that’s the underlying dynamic of all of this, standardized data converted from bespoke, private and accessible data.

John Lynn:

Yeah. Joe, what would you add? And what are the use cases that you see really popping off and people using this, not just, okay, talk of FHIR is great, but what do you see actually being used?

Joe Gagnon:

I think a place to start John is almost to say every one of us over the best 25 years have experienced how common data standards and interoperability in other industries change our lives. You can book a seat on an airplane using KAYAK or Booking.com and you couldn’t do that 15 years, 20 years ago. You couldn’t buy something online and have someone ship it and track it throughout, by sending pictures to your phone, through text message. All of that experience we have, driven by the consumer, the backend processes for the most part were proprietary. But then all of these industries said, “If we’re going to actually make a difference in people’s lives and give them the capability that they have or would like to have, then you have to get to data standards.” Like, let’s stop arguing about how we move the data or how we store the data, that doesn’t add much value.

Joe Gagnon:

So in health, where we spend $3.8 trillion as an industry, four out of five patients feel they’re not getting served properly. We’re ranked probably 40th in the world and we spend more money than anyone. And the only way to fix the system, it’s not like doctors and nurses are bad. They do great work, not included, right? But what the administrative systems that are implicated in this, by the way, the federal government pays for health and health plans sit in the middle of the whole relationship. The only way to start on the path of doing the same transformation is using this data. And so you start to see use cases that start with the simplest, which is what the ONC said, which is that a patient should be able to request their medical record.

Joe Gagnon:

You should be able to get that data and you shouldn’t have to get it in fax format or get a printout and wait days, weeks, or months. Healthcare, if anything, should be run in real time. It’s nice to have real time information about my black t-shirt that’s going to show up tomorrow, but I would rather know about my health data. And so we started that patient level and then we just continue to expand beyond that. The real win is going to be at the population level. When we can move population data securely of course, with privacy included all of the important elements of that, but then be able to draw insight from that. Apply that both to the cost side of the model and the care side of the model, reduce risk, better quality. These are all things that everyone would like to have. The system hasn’t worked that way because it’s a hundred percent proprietary and we got to break out of that.

John Lynn:

Yeah. And I want to get to the population level eventually, but let’s talk about what data is available today using FHIR and what isn’t. Don talked about USCDI, every EHR vendor has adopted it. It’s amazing what regulations can do to push that forward. How would you look at it, Joe, as far as what data is available and what isn’t?

Joe Gagnon:

Well following along the bouncing bullet here. So far, what’s available is the stuff that the government has asked either a plan or a provider to make available. We haven’t stepped in front of this yet. So right now, the majority of the FHIR data comes from health plans who have to take their Medicare and Medicaid patient data claims data specifically, all the way down to that resource level and make that available for the patient to access. And so that’s the predominance. So we have 31 million patients on our platform where we can get you done to the deepest resource level in FHIR. And that could be for that patient access or to run some analytics. We do also go out and get medical record data. That’s stored in FHIR from the HR vendor.

Joe Gagnon:

And so there are medical record data, it’s just not as readily available. So we have to go out and build that interface to bring it together. So you start to see what happens when claims and clinical data come together. There are more that over time will end up, whether it’s therapeutic data or device data, all of which will end up in FHIR in the same repository. The good thing is that FHIR has been a very well documented over the best 20 years of evolution to articulate a standard data model across all elements. What’s just not happened yet, is all the data has not been converted.

John Lynn:

Yeah. Well it feels like we learned something from HL7 that we’ve now applied to FHIR. Is that a reasonable comment or is that too simplistic?

Dr. Don Rucker:

HL7 actually manages FHIR to be quite clear, right?

John Lynn:

Sure.

Dr. Don Rucker:

So FHIR is an HL7 standard. Yeah. Clearly, the standards construction world that HL7 has employed has been fully used in sorting out the FHIR issues over these last number of years. That’s I think what makes FHIR powerful, in fact, is that it has been through a formal standards vetting process.

John Lynn:

So is the challenge with FHIR, maybe the weakness of FHIR is that it hasn’t been fully adopted or they’re still implementing that as you said. I think it’s a beautiful way to describe it. It’s taking their proprietary data stored in EHR or whatever system it is and changing it to standards base. I think that was a beautiful way to describe it. It is that limited standard, because right now it’s USCDI that’s required. Is that the biggest weakness or are there other challenges that FHIR still faces?

Dr. Don Rucker:

Yeah, I think there are two challenges and they’re utterly separate, coming from totally different corners of the world, I guess. So one is just any data representation in healthcare is innately complex. Just because every cell in our body is different, the biology of ourselves is vastly unique, and so if you’re going to computationally represent something that’s innately unique, that is, as they used to say in math class, non-trivial, right?

Dr. Don Rucker:

So that having been said, we all use huge abstractions of this. Maybe some of the most problematic ones historically are things like billing codes, quality measures. So there’s some huge abstraction. So we’ve figured out ways to deal with that. The other issue is, we are still in many cases, incenting provider systems to not share data. The brutal reality of the world is that the federal incentives and payment have led to a world where provider systems have a huge incentive to capture high margin procedures.

Dr. Don Rucker:

And they have built inward system IT architectures, to capture referrals, to hire docs and to in fact, not share data under any circumstance. That’s the business model that we have inadvertently funded over the last 50 years of the way we regulate prices.

Dr. Don Rucker:

So what Congress said is, “Okay, let’s try to open this up, right?” That’s what the Cures Acts interoperability provisions are, say, “Hey, if you want to hide it, even maybe ignoring their culpability in the underlying incentives.” They’ll say, “At least we’re going to make this available.” CMS has obviously doubled down on that with a payer to payer API, payer to patient API requirements. And so what you have is biologic complexity that’s often a screen for we can’t share data, it’s too complex. But ultimately, the issues are more around the economics. Now that having been said as patients, of course, our agents are the payers, right?

Dr. Don Rucker:

And the payers are attacking this problem as well, they’re aware of this. One of the things, obviously that 1up is quite involved with is integrating claims of clinical data. So payers can actually be smart about what they’re buying. They’ve never had the ability to be computationally smart about what they’re buying. So all of a sudden, you have the payers side of this whole equation is starting to change materially, right? You have transparency on performance coming from lots of corners, but the ability to integrate claims and clinical data is huge on just what are we getting? What are we paying for? To the extent that providers are not, shall we say, fully engaged in this? What you’ve seen over the last two or three years are massive efforts by payers to actually become providers. The whole payvider model is interoperability by force, right?

Dr. Don Rucker:

If you’re not going to share the data, that’s fine. We’re just going to internalize the whole operation and notice, that what they internalize are the choice points, right? Primary care access, control of resource issues, right? So you have some very interesting dynamics. So when you look at interoperability, there are lots of things that are swirling and FHIR is a little bit of the battleground where these things are happening and where people are actively looking to gain competitive advantage through the smartest use of tools. Just like the rest of the world using software, very interesting, novel and health gap.

John Lynn:

Yeah. Well, it is a challenge, right? it goes back to what Joe said, trillions of dollars spent on healthcare. And so we say, “Okay, well we should be spending less, but I don’t want to hear anyone that wants to get paid less.” Right? Like, no one’s volunteering for that. Well, that’s a problem.

Dr. Don Rucker:

Well, that’s what markets are about.

Joe Gagnon:

Right. Exactly.

Dr. Don Rucker:

You may not want to. Yeah. We all want to earn more money, I suppose, or 99% of people. But 100% of people don’t want to give us more money. So markets balance out needs and wants.

Joe Gagnon:

Yeah. The opportunities will get created in a disruption. And so new businesses get formed that we never saw would be before. And they redefine expectations that the consumer ultimately, so in health, ironically without a patient, there’s no industry.

John Lynn:

Yeah.

Joe Gagnon:

But the patient has limited voice. As they start to get voice, because the way the data can move and the data that’s available, whether it’s on a smart device, like I can go today in order a glucose monitor just online and connect it and have it start to teach me about my consumption of sugar. And then I can go to a company who can then analyze that for me and tell me what care plan I can be on. And they actually are hiring doctors themselves. I don’t even need the system. I might need a little bit of money, but these are very inexpensive ways of taking very expensive patient care, taking it early in its life and actually improving the experience for that patient. These are just like little microcosms of the potential. In that, billions of dollars of opportunity to be created.

Joe Gagnon:

So yes, every legacy company, General Motors in 1978 had 52% market share. They now are down at around 15%. Why? Because they stopped paying attention to the customer need and it’s going to happen the same way just because these guys don’t want to share data and providers have a monopoly does not mean that, that endures. And so they will either have to evolve or they will be replaced. There is no they’ll be marginal. Right? And so the only question for all of us anymore is, what’s the time horizon not, is there a different outcome at the end of this past?

John Lynn:

Interesting. Well, I’m sure we could talk economics of healthcare for a long time, but let’s shift gears back to something you mentioned, Joe, around population health level approach. I think when most people hear FHIR, they think about individual patient data access, right? FHIR into the provider to pull it down to my iPhone or whatever record I’m using to do it, right? Or maybe a doctor requesting your record as part of a transfer or something like that. So is FHIR really going to play a role at the population health level? And how will that play out? Maybe you can start Joe.

Joe Gagnon:

Yeah. And then Don can certainly comment on, that the bulk FHIR API might have been the most significant part of the act that was really going to allow for the bulk movement of data. And so, yes, you’re right. There’s this API that allows for patient access, but under treatment payment operations, you can make a request to get bulk access roster of patients. And then in that you start to have access to data that you never had before. And so as we start to see that happen and the same can happen on a payer to payer request. It doesn’t just have to be for the patient. It could be for a roster of patients.

Joe Gagnon:

So then you have all this data and now you can start to run analytics. So even something as simple as we announced a sequel on FHIR user interface, that would allow you to take all the data you and start to run very simple queries that have never been able to run before, because you didn’t have a common data set. And so we’re in a walk stage, but really getting moved to a run stage because of this ability to get bulk data through the FHIR API.

Dr. Don Rucker:

Yeah. I think we have historically looked at interoperability from a very narrow lens of getting one person’s medical record to one other provider or maybe one other provider system, right? That’s an extraordinarily narrow view of data movement of computing. And it certainly doesn’t allow anything to do with major things like population health, social determinants of health, configuration of delivery systems, measurement of quality or measurement of performance.

Dr. Don Rucker:

We took the opportunity since people had done the heavy lifting of converting their proprietary data elements into FHIR data elements is, and this is work with kudos to Ken Mandl at Boston Children’s. We put together the bulk FHIR standard, that is a requirement at the end of this year for the US Core Data for Interoperability. So med list, problem list, some of the notes, things like that to expose that API for an entire population of patients.

Dr. Don Rucker:

So all of a sudden, you will, if the two parties agree have for the first time ever in American medical history, robust, computational, comparability, and accountability for care. We have never had a computable way of doing that. Now that having been said, these are, as Joe mentioned, under the classic treatment payment operations, the way payers and providers exchange data today. So in the past providers have said, “Well my data’s too idiosyncratic. I can’t put it out.” There have been lots of these very bespoke data fields for the narrowest of purposes at huge cost.

Dr. Don Rucker:

Now, if both parties agree, they can get it in a standardized way. And obviously, I think, some providers will say, “Hey, I’m not sharing that data, but it won’t be for want of a technical standard.” And my guess and assumption and hope is that these things will actually have win-win contracting, right? Because these are contracts between counterparties that have ongoing business relationship in the Roger Fisher getting to yes, school of negotiation.

Dr. Don Rucker:

So I think you’re going to see joint efforts with payers and providers to use this data and to really figure out what works, what doesn’t work and change incentives of all of the parties involved. This is now possible, really for the first time ever in a robust way. This scales in terms of impact, orders of magnitude, more than classic narrow gauge, quality of measures, which has been our prior attempt to purchase value in healthcare over the last 20 years.

John Lynn:

Yeah. It is such a huge change in mindset. I’ve often said to people, the problem with interoperability, the challenge of interoperability in healthcare is that we’re asking Pizza Hut to share their customer lists with Domino’s across the street. And yet that’s what we want from healthcare organizations, right? Because it’s the right thing to do for the patient. But as you said, will they agree to it? I think that’s the golden question now.

Dr. Don Rucker:

In market, instead of functioning, people have to share data, you can move your money between bank accounts and brokerages. Joe mentioned the travel we can decide whether we want to use English system or metric wrenches, but ultimately there are two sets of wrenches. In functioning markets, consumers enforce interoperability of goods. It is a bizarre measure of healthcare that enforcement is very weak and is really only historically been about internal provider efficiency.

Dr. Don Rucker:

We’re thinking, if we go back 25 years, it was actually Congress with HIPAA, which is now a privacy law, but it was actually the HIPAA Act that required payers and providers to do X12 electronic data exchange for claims, right? Before then, there was not electronic data exchange in any material way. So the government giveth the government take it away. There’s a funny history to all of this, but the excitement is that modern technology is leapfrogging some of these things. And we’ve seen this in almost every industry in the rest of our lives. And I think healthcare is soon to follow.

John Lynn:

Yeah. Joe, what were you going to add to that?

Joe Gagnon:

I think that when you have something of this scale that operates in a very integrated and in an embedded way, that does feel like it’s hard to change. And yeah it sounds like no one wants to share, but it goes back to, that the consumer is going to have more choice. Even COVID provided an opportunity for remote diagnostic care that we weren’t able to accept for probably the next 10 years. And so we are changing how the delivery system works. We’re changing the way that you get access to care. We can even change the cost model around it. And yeah, there are big systems that will somewhat resist, but like every adoption curve over the history of technology early adopter, early majority, late majority in the laggards. Laggards will stay where they are and they’ll become less relevant over time.

Joe Gagnon:

And that the early adopters will start to continue to get new patients. They’re going to get employer related plans instead of just normal health plans as a place to get different kinds of care. We have different health systems that are coming up that operate in a much more localized fashion. And so I think that it’s all pointing in the good part. So in the past, it would’ve been very expensive to try to articulate how you could use a cloud platform. Well, that’s already there for us. It would be hard to articulate a data standard. Well, now we have FHIR that can do that for us. we have a security standard with HIPAA, like all the piece parts are already there. So the time to change isn’t that long, because you don’t actually have to do anything special.

Joe Gagnon:

We’ve mastered the conversion of this data into FHIR. Take the legacy data, map it to FHIR, ETL it onto a cloud platform. We can do that for any size plan in just months, a couple of months and get that done. This used to take a year to be able to do something of that magnitude. We can get it down to under a month to be able to convert tens of millions of members if it was even a plan of that size. So as you start to see that, and then you have the tools that make that data accessible, you start to see the innovators say, “Wow, we really couldn’t do that easily before. So why don’t we go do something about it?” And in that, opportunity is created on both sides of the equation.

John Lynn:

Yeah. So as we wrap up, at healthcare IT Today, we love to talk about successes and highlight successes, right? So maybe Joe, you could start and then Don, what is the best real world success that you’re seeing with FHIR? What do you see? When you see it happening, you’re like, wow, that’s an exciting case study example of what can be done using FHIR?

Joe Gagnon:

I think we’re seeing interesting dynamics. One of our customers ended up using the FHIR data over the past couple of years to track COVID vaccine rates that were happening, that they couldn’t see across their data set. We never thought anyone was going to do that. We have another customer who ran a query that sounds like the hello world, first programming thing you could ever do. They ran a query of smokers versus non-smokers across their entire Medicare, Medicaid population. They were never able to do that.

Joe Gagnon:

We’re starting to see the health plans request of their largest provider. That clinical data to combine so they can start to do a better value case assessment. So we’re seeing the analytics happen faster now that they’re getting this data. So they started, because they had to do compliance to meet the government regulation for patient access. And they’re like, “That’s not enough.” And they’re running with it. And I think that that’s the early green shoots of success that we’re starting to see.

John Lynn:

I love it. That it’s things that they couldn’t have done previously. That’s [inaudible 00:30:28].

Joe Gagnon:

Its sounds so hard, right? These are such simple, basic level insights that you can get from data that they haven’t been able to. So yeah, it’s amazing. Don?

John Lynn:

Yeah Don.

Dr. Don Rucker:

Yeah. To me, the real dynamic here is the ability to do stuff real time, right? We’ve built up this vast business with quality analysis and prior auth that are not real time because that’s what the technology was. When you look today at apps and some early apps are using the 1up platform, those APIs are all available on our website. If you’re a developer and want to see what they look like, so you can start doing that. But the ability to power all of this real time in platforms where you have atomic FHIR resources, the ability to do any analytics you want and put it out to an API, that’s the essence of cloud computing. Every app, we expect to be real time on our phone. We’ve never had that opportunity in healthcare. And now with this type of a platform, we’re bringing that opportunity. That to me is extraordinarily exciting and extraordinarily powerful for us as consumers paying for healthcare and us as patients who have personal medical issues.

John Lynn:

I’m surprised you don’t think the 30 day retrospective claims data is not going to change healthcare. Jokes aside.

Joe Gagnon:

We can wait a little longer than that if you want.

John Lynn:

No, it was what it was right? As Don pointed out. It was the technology of the time. So Don, before we finish this. Is something coming after FHIR or is FHIR the standard for the next decade? We had X12, we had HL7 v1, v2, we had CCDAs somewhere in there, right?

Dr. Don Rucker:

Yeah.

John Lynn:

Which is an overlap of some, we have FHIR, is something else coming or can we bank on FHIR for a while?

Dr. Don Rucker:

Well, actually, I chuckled when I saw that question because I obviously had to ask myself that and we had to ask ourselves that at ONC, when we were putting out a rule, right? If you’re even vaguely responsible public servants and ONC is a great team, we had to ask ourselves long and hard. I think the answer to that is pretty straightforward, FHIR is it. And it’s not that it’s FHIR, it’s actually JavaScript object notation. It is having something that is essentially as minimal as possible as a data structure. It’s human readable. Right? So we’re talking about how many brackets, braces, commas, curly cues do you need in order to represent something in computer science and JSON is pretty much a reduction to the absolute minimum needed to do that. FHIR bills on that.

Dr. Don Rucker:

So I think that is it. Whereas, their work to be done, obviously, each of those things to be FHIR enabled, the standards community has to figure out what are the details of putting in the JavaScript object notation? And that work is moving quickly, much of healthcare. This is already done. We have the entire US Core Data for interoperability, but there’s other fields, other data things where that is still a work in progress in that shift from proprietary to overall standard space. The other dynamic here, you mentioned CCDAs and that brings up the point that a side dynamic here is a speech recognition and text, become extraordinarily powerful. In my practice side, I was involved in the building the first dragon based EMR back in 1988.

Dr. Don Rucker:

Now, without putting a plugin for that vendor, that’s now owned by a big company which starts with an M, speech recognition is absolutely amazing. You just walk up and it’s extraordinarily accurate. What does that mean for your audience? It means that the speed of adoption of NLP tools, the same AIML tools that you can use on the 1up platform for other types of analytics, they’re also going to be able to process these notes.

Dr. Don Rucker:

So there’s a very interesting dynamic because FHIR will be turbo charged with text. And it’s part of that where are you doing your analytics? How are you thinking about modern data science? How are you doing that real time? It all blends. So I think the advances will actually come in the use of the platform and generating insights. I don’t think there’ll be any need for changing the underlining data representation. That’s my two seconds. Yeah.

John Lynn:

Well, sounds like Joe’s made a good investment and 1upHealth in a good place, then you’ll be happy. Excellent. Well, this was a great insight into what’s happening with the FHIR data standard and obviously the work you do at 1upHealth. So thanks so much and thanks everyone for watching and listening.

John Lynn:

If you want to find more great healthcare IT content like this, be sure to check it out at ealthcareittoday.com or search for Healthcare IT Today on your favorite podcast an application. Thanks guys. Appreciate your insights.

Joe Gagnon:

John. Thanks so much.

Dr. Don Rucker:

Thanks John.