Grahame Grieve is a principal with Health Intersections of Melbourne, Australia and was the architect-developer of HL7’s Fast Healthcare Interoperability Resources (FHIR, pronounced “fire”) specification that allows EHRs to exchange information.
Was it weird to see FHIR as the only universal topic of HIMSS19?
Not so much weird. Obviously it was gratifying for us to see the community investment that so many people have made becoming justified. It’s definitely worth saying that we really value HIMSS’s active participation in driving the conference in that direction. There was an organicness to the fact that FHIR became the big issue given the way the industry overall is, but HIMSS definitely actively drove that and that was an important part of the picture. I thank Hal for pushing that.
I thought there was maturity at the HIMSS meeting this year. You and I talked about bad FHIR puns and expected to see them all over the place, but we didn’t actually see anything like that. We saw instead quite a lot of maturity around the discourse and the challenges of sharing data. I thought that was really good.
I always call you the father of FHIR without asking you if you accept that title. Is it fair or unfair to call you that?
I did initially draft it and propose it and I’ve curated the passionate community input over the years. If that makes someone the father, then I guess I am. The community is the real father. I get undue attention as if it was some magic that I achieved, where actually it’s just thousands of people passionately contributing to the common values that we hold.
I’ll repeat a question that I asked you last time we spoke. Are you worried that non-experts mistakenly believe that we’ve figured out interoperability because they keep hearing about FHIR and APIs?
That definitely happens. People assume that since FHIR is now the designated answer, it is the answer to all of the problems faced. But it’s not.
As HL7, we can only take on and mandate solutions that everybody completely agrees to. This is healthcare, so there’s a very limited set of things that everyone completely agrees to. Additional agreements are required. The scope and scale of the agreements required are beyond any single organization. We’re spending increasing amounts of our time investing in collaboration with other organizations to get a seamless process around scaling up agreements and consistency across multiple organizations like IHE that are helping out with the problem.
As long as hospitals buy an EHR and then spend a $100 million customizing it to their workflows, then interoperability is going to be a challenge. On the other hand, the fact that hospitals make those kinds of investments indicates the complexity of healthcare. There is no easy win. There is no easy victory to get interoperability with some kind of tick mark against it. It’s an ongoing process that we’ll be going through for a long time yet.
It seems like every EHR vendor has at least some customers exchanging information with the EHRs of other vendors and now CommonWell and Carequality are connected. Can product shortcomings still be given as an excuse for lack of interoperability?
The vendors work really hard, and from my perspective, the vendors are committed to making patient data as easy to move as possible. On the other hand, the vendors basically fight with their old legacy code bases that are extensively customized and very had to work with. That’s the nature of any mature software product. I think that if I was a consumer, I would be unhappy with where they are, rather than if I’m an engineer looking at their problem. It’s kind of a challenge for the vendors.
But increasingly, as I observe the space, the challenges are with the providers. To what degree do the providers want to share information? To what degree are the providers prepared to standardize their record-keeping practices and their clinical practices to make it easier to exchange data and to transfer patients seamlessly? Not many of the colleges really understand that problem. I would like to call out the American College of Obstetricians and Gynecologists, which understands the problem very well and has very standardized record-keeping practices on paper. That puts them in good stead to get interoperable.
A lot of doctors I talk to think about this as a technology problem, but it’s not a technology problem. It’s an information problem, and so technology can’t solve it. It needs clinicians to make clinical agreements in order to get clinical interoperability.
There’s one more thing I’ll say, which is that interoperability is not a binary thing. We get a degree of interoperability. We can routinely exchange patient summary information. But seamless transfer of care will require a deeper agreement. We’re not there yet. We’re working on those as a community. But I believe that increasingly the load will move away from the IT side or the technical side to the clinical side as time progresses.
Efficiently accepting patient information from an outside source requires placing it into the receiving clinician’s workflow and being willing to use information that was entered elsewhere. Not that we need another interoperability frontier, but is figuring those issues out the next one?
The trust issue is really important. I’m glad you brought it up, because increasingly as I look at projects around the world, the question, is who trusts who and why? A lot of the complaints I hear from patients about poor record-keeping actually comes down to no established trust framework. If the patient provides you with a written statement concerning their medical history and you read that, are you liable for not asking them verbally? Can you rely on that written statement? If you get a written statement from another institution, can you rely on that? The interplay between trust and liability is something that we’ll have to revisit as a community and make that a fundamental part of our interoperability considerations.
What could I do as Provider A if I find that I’m regularly receiving incorrect or unreliable patient information from Provider B?
Looking around the world, I routinely hear that more than half of patient records contain wrong information about the treatment history. Some of those are really, really wrong, and you can easily find examples of that in the media. Surveys that I’ve seen show that it’s more than half the records contains something wrong, and yet we make those available to the patient without any consideration for what a patient should do if they look at it and say, that’s not right, I’m a guy, so I don’t think that a pregnancy test was actually performed. Life’s a bit more complicated than that, but what do they do?
You asked the same question about providers with each other. There’s one organization working on the policies and technologies associated with this, which is Carin Health. But we should start moving towards a culture where it’s a professional obligation that if you share your records with somebody, you have an obligation to have some sort of error detection and correction process running so that your records can be corrected. But in today’s environment, we’re a long way away from thinking like that.
What has been the impact of Apple exchanging information with EHRs using FHIR?
There’s certainly discussion happening around Apple in particular, but more generally patient access to information and what kind of difference that will make. Obviously that was a subject of the keynote at HIMSS. Apple brings a particular sharpness to that debate because of its global consumer reach, the style of its consumer reach, and the potential for Apple to disrupt health in the way they’ve disrupted other industries. I certainly hear discussion about that. Some people are wildly in favor of any disruption. Other people are very much not in favor of any disruption. Some people are concerned about what a consumer company like Apple might do.
My perspective is that getting patients their data doesn’t really make much difference to patient satisfaction or behavior, because it’s all historical data. What makes a difference to a patient is the services that you provide. You need data to support the services, but it’s the services that matter. As long as healthcare services are fundamentally delivered in the flesh in the physical world, there’s a limited degree to which the consumer electronics companies can disrupt health.
In order to provide substantial healthcare services, you have to put people on the ground. That raises all of the classic “how do you manage healthcare” problems, for which I don’t think there’s any magic bullet. I think that their impact will be significant, but ultimately limited by real-world constraints.
Joe Biden and Seema Verma have recently expressed disgust that they, even as high-ranking government officials at the time, were unable to get the medical records of their relatives, and Verma in particular seems outraged. Do you think the government sees its role now differently than it did originally?
It has become more clear across the industry that what we have is not a technology problem. We have a business and an information problem. The government laid down a whole lot of money as far as stimulus, partly to spend money — which it did effectively — and partly to move past the technology barrier to the information barrier. Aneesh Chopra has told me that what happened was relatively predictable. We’ve now solved the technology problem. We can focus on the business and the information problems, and here we are doing that. The NPRM focuses on cleaning out the technology problems and moving the business and information problems to front and center.
But as the government, the levers that you can pull have limited effectiveness. That’s even true in autocratic countries. I was in one country where they showed me that certain things were happening in a particular way. The next day, I would meet with the programmers. They would say, “This is how we do it, but don’t tell the bosses, because they’re not allowed to know.” The levers that you can pull as a policymaker and a money-spender are a lot more limited than people believe. At least the US government is acutely aware of that, much more so than other governments I deal with.
Nothing leaps out at me as any quick solution here, so since the NPRM is marginal improvements being made over time, we can look for improvements. I’m particularly hopeful that we can solve the access to healthcare records problem through thoughtful change. I already saw that happen with vendors. When I started dealing with health information exchange, vendors were suspicious about exchanging patient data. They saw that as a business threat. Now when I deal with C-level people at the vendors, they’re all like, “Well, why wouldn’t we do that? We can’t not do that. It’s part of our business. It’s a business opportunity.” Whereas if I talk to providers, I see providers very much being, “Why would we do that? Why would we spend money doing that? Isn’t that a business threat?” It’s about making that same cultural adaptation to their thinking.
That’s the key thing that we need to chase — the understanding that exchanging patient data with the patients is a business opportunity, not a business threat. But it’s a cultural transition that needs to be bedded deeply through the provider before the provider is ready to see healthcare as a different kind of business model. There are a number of institutions around the USA that are pushing that as hard as they can. Hospital in the home, seeing the hospital as part of a wider network, the whole ACO thing is pushing that. There’s a bunch of things happening, so I’m not particularly pessimistic about it.
Does anybody still care about Blue Button?
There’s a really active community around the new Blue Button work that CMS is doing with FHIR. The FHIR community is picking up and processing the data. There certainly is interest in cross-correlating data from Blue Button with data from the Argonaut interfaces that patients can get, and creating a market in that space. The White House is interested in that. It makes a lot of sense to try and leverage some efficiencies out of the system by cross-correlating payment data and payment efficiency data with individual healthcare data.
That’s the logical place to look for where you, as a funder, could seek to provide more efficiencies in the healthcare system. In the end, the providers of healthcare are not motivated to perform systemic repair to healthcare. It’s the funders who are motivated to perform systemic repair. That’s an important part of the overall picture.