Earlier this month, HCI Editor-in-Chief Mark Hagland wrote a blog analyzing two articles about health information exchange (HIE) that had appeared in the July issue of Health Affairs. The two articles are entitled “The Number of Health Information Exchange Efforts Is Declining, Leaving The Viability Of Broad Clinical Data Exchange Uncertain,” authored by Julia Adler-Milstein, Sunny C. Lin, and Ashish K. Jha; and “Engagement In Hospital Health Information Exchange Is Associated With Vendor Marketplace Dominance,” by Jordan Everson and Julia Adler-Milstein.
As referenced in the blog, the way in which HIE is described in the first article is as a fairly fragile phenomenon in the wake of the loss of most federal and state funding in the past two years—something that has been widely known—with a strong need for operating HIEs to prove strong market value (though that term is not used) in order to survive. Meanwhile, the second article looks at the issue of intense market dominance by a few large electronic health record (EHR) vendors, and its strong impact on how HIE is pursued in markets in which a single vendor dominates. The authors of the second article did make it clear that the impact of market dominance on the part of the Verona, Wis.-based Epic Systems Corporation is quite different from that of market dominance on the part of any of the other largest EHR vendors—the Kansas City-based Cerner Corporation, the Chicago-based Allscripts, or the Westwood, Mass.-based Meditech.
As a follow-up to the publication of the blog, Julia Adler-Milstein, Ph.D., the co-author of both articles, spoke recently with Hagland recently, to discuss some of the implications of the two articles at this moment in the ongoing evolution of the HIE phenomenon. Dr. Milstein is an associate professor at the University of Michigan, with appointments in both the School of Information and the School of Public Health. in Below are excerpts from that recent interview.
Reading both Health Affairs articles carefully, I came to the unmistakable conclusion that there is no ideal “silver bullet”-type solution for fixing all the challenges facing the HIE phenomenon. Whatever we do, we’re trying to make the best of what was not an optimal seedbed for HIE development to begin with, correct?
Yes, I definitely agree. Now, if I put on my policy hat, from a policy perspective, I would ask, what was the market likely to under-invest in? And it was pretty clear from the start that HIE was not something that the market felt it needed to invest in. So it seems like that is the place where we should have started doing the most policymaking around. We started doing the most policymaking around EHR, but it really should have been the opposite. I think it should have been much more on interoperability and not as much around EHR adoption. We put so many eggs in the EHR adoption basket, and so few into interoperability, so now we’re having to catch up.
And because of the diversity of HIEs, it makes it that much more difficult to give good policy guidance?
Yes. I think running an HIE is one of the most difficult things to do. The governance, the technology, the business processes, and the workflows, are all too diverse. And the problem is not well-defined. Are we really envisioning that every piece of data can be shared seamlessly with every other piece of data? We’ve never defined the endpoint. I don’t think HIEs were ever going to get to that seamless interoperability. But there remains that fundamental problem of what we want to get to at the end of the day. We let ourselves think that we’d get there by having everyone share everything with everyone. But that was never feasible. But we never had the discipline to say, OK, what is feasible, and where can we reach consensus?
If you were the federal HIE fairy, and you could just wave a wand, what ideal things would you make possible, from a policy standpoint?
What I struggle with most is what lever we press. Do we just hold providers accountable for certain things, require some basics, and pay for what we care about and pay for that? We haven’t sorted it out. So maybe we should essentially hold providers accountable for cost and quality measures that they cannot do if they don’t share information across the care continuum. That holds appeal, because we’re paying for what we care about, and providers would put pressure on vendors. If you said, we won’t pay for readmissions, period, that would force hospitals and long-term care facilities to coordinate in a way they’d never done before, and that would force interoperability. That’s not feasible, but is tempting. The other alternative is to put providers and vendors on the hook for whether data is moving. And we could do a lot more to make sure providers are sharing data. Decertify systems, take money away from providers in the absence of interoperability and data-sharing. I truly believe that if Epic and Cerner believed they wouldn’t be in business in a year, that they’d figure this out.
Might it be feasible or worthwhile for HHS to force EHR vendors to take very specific steps?
We don’t have a fully specified approach [on a healthcare policy level]. A lot of it comes down to, when you’re implementing standards, to interpret those. Right now, vendors have no incentive to work together to implement standards. So in concept, yes, you could say that they would be much more motivated to work in lockstep with an implementation guide. So I think that could happen. I think in industries were standards adherence is essential to staying in business, they’ve figured out how to make it happen.
And Epic has been motivated to share information among Epic customers. But why have they not been motivated to share with non-Epic customers? We have not set up requirements for them to share with non-Epic customers. And we shouldn’t criticize them for sharing among their own customers. And the question is not, how do we bring them down, but how do we bring others up to their level?
So I understood from your article that Epic being dominant in certain markets is discouraging non-Epic customers from creating health information exchange. Did I understand that correctly?
It’s a complicated thing, but when Epic is not the dominant vendor, the greater the dominance of that vendor, the better off everyone is in the market in terms of developing HIE. If I’m in a Cerner-dominated market, it’s easier for me to set up an HIE when the majority of the market is any one vendor, except for Epic. And my guess is that that has to do with the fact that Epic is not playing well with others. And so when they’re the biggest, whether at 30 or 80 percent, it is having an effect.
Is Epic having a deleterious effect nationwide, on non-private HIE development, because of its dominance in some healthcare markets?
It seems to be so, though perhaps there are other explanations. That is what you hear. It appears to discourage organizations. We’re even seeing it here at U of Michigan Health System. They’re having to fight the fight every day. It’s almost as though CareEverywhere puts up a contrast to others who would have to set something up on their own. You’re going to have to pay for something that you have to set up separately. There are so many pressures that lead people to say, I can’t justify paying to participate in a third-party HIE. That is real. But you only see those when you’re in the decision-making meetings of healthcare organizations…
Would you agree that if it were true, that that offers a policy conundrum?
I agree with you, yes, it would. At the end of the day, I’d like a provider to be able to share patient data appropriately. There are several ways to go about this. The question is, do you come down harder on the providers or vendors, or both? And you think about CommonWell and others… they see the problem. But CommonWell has been around for a while, but what’s actually changed on the ground today, is not obvious. So how do we move these issues up the priority list? I don’t know the answers, at the end of the day. But if you do pressure vendors, it will have a bigger impact. The provider market is much more fragmented. So if you could figure out a way to make the six biggest vendors share information with each other, you’d have a huge impact on the market. And it would also give leading vendors an advantage. And ONC [the federal Office of the National Coordinator for Health IT] is very hesitant to do that; they don’t want to pick winners and losers. You could clearly make a difference by going to a select set of vendors, but that would be problematic in that sense.
Meanwhile, we haven’t done a good job of making sure that data flow is patient-centric. And here in Ann Arbor, the big University of Michigan Health System has Epic, and the smaller system down the road has Cerner, and patients are going between them a lot. But if we could do a better job of measuring the problems inherent in markets where the two big, dominant health systems are not sharing data, that might be helpful. So I think we also have to decide which kinds of problems we want to solve. It’s very different, too, to get smaller practices connected into the major health systems in their communities, versus getting the “big dogs” to play together. Those sound like different solutions to different policy problems.
What do you think will happen in the next few years?
That’s a good question. I’m not involved in FHIR [the Fast Healthcare Interoperability Resources standard], but those involved in it do believe it will be a game-changer. I hope that that is true. I don’t feel that I really understand enough about it. I never think that the standards are the hard part, I think it’s the implementation and the governance; and if you reduce the technology barriers enough, that could possibly move the market. Part of me feels like things won’t be that different three years from now, but part of me believes it will be better, once we’re able to extract data from EHRs and move it around. I really don’t know.
What do you think of the vision of the future some have put forward involved with the sharing of very small pieces of discrete data, rather than the sharing of entire CCDs [continuity of care documents]?
It seems like we do need to move forward on atomic data-sharing. The CCD sharing has clearly not been working very well. And moving forward in more of a use case-based way, which FHIR supports—I think we need to try that. There is a risk of perhaps going to extremes. And I haven’t pondered that scenario that much. But clearly, some of the information in CCDs is not being used. So I would say we need to try that approach; I think we should go in that direction. I have not heard anything that suggests that we shouldn’t go in that direction.
What should CIOs and CMIOs be thinking about now, and knowing right now?
One of the things I’ve tried to bring clarity to is the distinction between perverse and weak incentives. A perverse incentive is the idea that perhaps Epic and other vendors might feel better off by not sharing information. That’s very different from a weak incentive, where I’m interested in sharing data, but progress is slow because it’s a hard problem. So being able to be clear about where stakeholders stand on whether specific incentives are perverse or only weak; and then creating clarity on priorities. There are so many challenges in healthcare today, there’s so much to do, and it makes it hard to make meaningful progress. We see it manifested in so many different areas, in the struggle to create and sustain health information organizations; but you make fast progress when it’s at the top of your priority list.
So I would say that, for your audience, a key message is that they need to decide whether this [HIE development] is high on their priority list. CIOs and CMIOs may believe that this is an important priority, but is it one of their top priorities? They need to know whether it is a top priority for their organization. And to go to Epic and say, we need X next year, it takes strong organizational muscle. So it’s not so much a message to CIOs, but it would be a helpful thing. I think it’s hard to pursue the right priority actions, without knowing what the most important points of friction are for CIOs and CMIOs. And the reason we haven’t made as much progress as we could, is that people aren’t openly saying that this hasn’t been a priority. I think, in general, being able to discuss and raise these issues, is important. And forums in which there can be more communication and collaboration around this, will be helpful. There are so many people trying to independently pursue this, but it’s a problem that by definition requires a lot of multi-stakeholder participation. Everyone keeps talking about the same problems, but why aren’t we getting together to discuss these issues?