At a time when the broad survival of health information exchanges (HIEs) is in question nationwide, and a number of statewide HIEs are shutting down or in danger of doing so, a small number of such organizations are actually flourishing, among them HealthInfoNet, Maine’s statewide HIE.
The Portland, Me.-based HealthInfoNet has long been known to the readers of Healthcare Informatics; indeed, the magazine named the organization one of its Innovator Awards winning organizations in February 2011, and described its successes up to that date, in this profile.
But much has happened in the past four years, under the leadership of HealthInfoNet’s executive director, Devore (Dev) Culver. Culver and has colleagues have figured out how to keep their organization relevant, even as many other statewide HIEs struggle or even fold. In that context, HCI Editor-in-Chief Mark Hagland interviewed Culver in November for his perspectives, as Hagland prepared one of the Top Ten Tech Trends for the January/February issue of HCI, on HIE sustainability. Below are excerpts from Hagland’s extended interview with Culver.
Would you agree that there is a core sustainability problem for many HIEs nationwide right now, particularly for statewide HIEs?
Yes, I would agree. My best guess, reinforced by some discussions, is that fewer than 10 HIEs will remain within about two years. What I’m talking about are the so-called statewide HIEs. Officially, there were 50, one for each state, based on the HITECH [Health Information Technology for Economic and Clinical Health] Act. And that’s part of the problem; in fact, it’s a major part of the problem. The core issue there is that they never developed a reason to exist. You can do all the nice convening you want, but if you don’t create value, you won’t last.
Was some of what was created not truly needed?
I think part of what leads to success starts with the building out of consensus among diverse stakeholders, so that there is at least the potential for value. And so you know, when I think about our history, the work started in 2004, and we didn’t have an operational health information exchange until late 2008, early 2009. A lot of that time was spent raising seed money, which took a lot more time than I’d expected; and the other part was spent building a consensus on our purpose, including developing a fairly robust economic impact study. We started with some models developed in Oregon and used those models, using Maine data running through them. Maine has had an all-payer claims database since the late 1990s. They had started managing a discharge data set as early as the mid-1990s.
So we asked questions like, what is the potential for reducing redundant testing in the state of Maine? Long story short, in 2008, the analysis coming out of that indicated a $50 million opportunity. So if you’re going to run an exchange costing $6 million, then that potential savings alone would be a huge return on investment. And you know you’re not going to get it all; but you start to build a value proposition, and you go out and get commitment from some of your larger, more data-rich organizations. And we started with a demonstration phase, one that included data from 50 percent of the providers in the state. There’s the core question of critical mass—if you don’t get to critical mass fairly quickly within two or three years, so that you have 70 to 80 percent of provider or healthcare activity in the exchange, it will be very difficult. If I as a provider go to the exchange and don’t find something the first time—maybe I’ll go back a second or third time, but that would be it. So you need critical mass that’s sufficient that people will actually use it. And right now, as of November, 96 percent of the residents of the state of Maine are in the exchange. That’s pretty damned powerful.
And then other things have to start to happen; you have to get experience and demonstrate practical value. Last month, in October, in terms of people actively clicking into the portal from within their EMR, last month was our all-time high for the number of patients whose records were accessed—23,000. That’s almost a 70-percent increase over last year. We went live in February 2009. At the beginning, a good month was a couple of hundred clicks.
And this is another problem with health exchanges. For those that are portal-based and cause you to have to go out to them, as ours is, that feature automatically creates some hesitation, when I have to take extra steps in my workflow to use them. So you’d better be pretty sure that will work. An initial assumption of ours was wrong, BTW, when we thought it would be doctors querying, but it’s actually mid-level practitioners, physician assistants and nurses.
So that is one of the big mistakes that statewide HIEs are making?
They’re making a couple of mistakes. One, because the federal government said it was a good idea, it was assumed to be a good idea. If you haven’t done the work on the ground to build a constituency for this before creating it, you’re probably in trouble. And the other thing is that even ONC [Office of the National Coordinator for Health IT] estimated just how hard it is to put one of these creatures up. It’s hard—it’s hard work, starting from the challenge of becoming a trusted entity, to actually trying to build a business case. It’s hard work.
And a lot of it was under-funded, right? About $640 million was bestowed through grants. And then they further complicated things—they pivoted and decided that the DIRECT solution was going to be a requirement of every HIE was funded. You can’t make a living through DIRECT, I’m sorry; it’s secure e-mail. We were lucky because we got our application into ONC very early and were already a live exchange. DIRECT hadn’t even been thought of yet when we were set up. We were the fifth or sixth operating plan approved by ONC, and DIRECT hadn’t even been thought of yet as a concept. When that happened, that put most of the state exchanges into a difficult place; it forced them into an operating model that couldn’t be financially sustained. And the value of being a post office? Meh. And frankly, the value is not about me delivering you a lab result, it’s about the data, about making that lab result truly interoperable so you can do something with that result; and that lab result in conjunction with other data around the patient, that is important. So simply moving data points from point A to point B, is just a necessity. But not sufficient. And from the get-go, your value proposition was endangered.
And the handful of exchanges that are truly succeeding now, they figured that out early. You can count on one and a half hands the number that truly are doing well and will survive. And if you look at the so-called private exchanges—and I take exception to that term, because we are a privately run organization—but those created by providers—they have a different starting point from the concepts of value, they become an expense line. They’re not trying to create operating margin by themselves.
And everyone says, “Oh, that’s the better model.” Well, no, because inherent in both of those, both the payer-driven and provider-driven models—if the provider controls it, it automatically makes other providers uncomfortable. And the same thing is true of the payer-driven model—it makes providers uncomfortable. And that’s why you set up a neutral-convener structure as we have, a neutral convener that has no stake in the use of the data. And frankly, it took us a few years to get there. But the epiphany came when we realized that the statewide exchange was a business line and BTW, a loss leader, and that HealthInfoNet was a business organization operating multiple business lines, and that was very liberating.
So the exchanges that are succeeding are either heavily funded by state money, as in Vermont, or have figured out the business value. Indiana is a great example of the latter. They started where the genesis was of delivering value by reducing costs. A group of hospitals that were competitors joined together to save money. And they were part of Regenstrief, which was a research group; you had multiple dimensions at play. But that’s a successful exchange. For us, realizing we weren’t the exchange; the exchange was a business line—and that opened up multiple thought lines. And also the concept of lazy assets—something you have to do in order to do your business, but that others will pay for—and one of the things we do is manage patient identity. And lo and behold, there’s a marketplace out there for provider organizations with multiple record numbers looking to get to a single record number. So, hello! That’s what I do. And people pay us, sending us multiple record numbers, and we send them back a master number per patient.
So the key is that we’re doing some really amazing stuff around analytics. We have a set of tools in the field that is dynamic—the data being used is a day old, driven by clinical and event data, and is predicting things like who will show up in the emergency department. Who’s going to get readmitted in the next 30 days? That runs while the patient is in the bed.
And we’re starting to see impact. One of our clients has been using this to manage their discharge functions, particularly around self-pay and around Medicaid patients—it’s economically more advantageous to keep those patients out of the ED or a hospital bed, because they lose money on them anyway. And any patient running more than 30 percent likelihood of being readmitted, they’re put into primary care coordination right away.
What will happen in the next two years in terms of the evolution of HIEs overall?
I think a couple of things are going to happen. I think the marketplace for exchange will become driven by things like analytics, sort of turning things on their head. Everybody thought exchange was the answer—and it is the answer—but in a different sense. We’re taking data from disparate sources and standardizing it through LOINC, SNOMED, RxNorm, NDC (those two are medication standards), ICD-9, ICD-10—all of those are standards. You map as much clinical data as you can to those standards. I call it a CPC and you call it a CPC with differential.
Here’s a classic example: there are two private reference labs in Maine, one in Bangor, one in Portland. Each has an order catalog of tests built locally, so the same test will have different names. They also have local numbers, with no standardization. So when you’re mapping in the catalog for purposes of taking data into the exchange, LOINC allows you to equalize what is the same test, regardless of what it’s called locally—same body part, same type of test (fluid, etc.), and that allows you to equalize. And whether wisdom, genius, or stupid luck, but the people creating the first exchange in Maine deemed it necessary to standardize language. Most exchanges around the country haven’t yet been standardizing data, well, hello? That’s the whole point.
So what is market-driven will lead to survival, right?
Yes, exactly. Ask simple questions around complicated issues. What are your pinpoints, and what is critical now, and how can you help? If an organization is managing a patient’s life under an ACO structure and is responsible for that, there is a pain point that needs to be addressed, and that an HIE can address. The other thing that makes the exchange valuable is the data. Most claims data is 60, 90, 120 days old. So you’re predicting events based on old data. My predictors are based on new clinical data coming in every night and recalculated every night. And we’ve got a very complex, sophisticated engine, with the ED analyzer looking at 40,000 variables. And you can predict core activity based on one-day-old data. And that’s incredibly valuable.