Geisinger's Alistair Erskine: How we're using FHIR to create EHR-agnostic applications

When it comes to the widespread use of the Fast Healthcare Interoperability Resources (FHIR) standard, Geisinger Health System Chief Clinical Informatics Officer Alistair Erskine calls himself impatient but optimistic. Erskine has helped Geisinger to develop a number of applications that take advantage of FHIR and public application programming interfaces to be able to exchange information with multiple electronic health record systems, but he'd like to see more.

"I think we need some real world examples, both of an implementation within a particular organization, but also of an app store for healthcare that would be vendor neutral and agnostic," Erskine told FierceHealthIT at last week's College of Healthcare Information Management Executives fall forum in Orlando. "The latter would have a mechanism to curate these apps, make sure they do what they say they'll do and be able to distribute that to the people who want to be able to consume them."

In an exclusive interview, Erskine discussed the process behind the development of such applications, as well as the implications of his work for the industry at large.

FierceHealthIT: What went into Geisinger's decision to build EHR applications?

Erskine: We became interested in writing electronic health record apps, even though we've had Epic installed for almost 20 years at Geisinger, because in our process of reengineering various workflows within our system for the whole continuum of care, there are some places that Epic hasn't gone yet. They probably will at some point in the future, and frankly, if they do, we'd probably be inclined to use whatever they come up with because it's integrated within our system. But in the meantime, we got in the business of saying, "if something doesn't exist yet, or it's not on their roadmap, let's go ahead and try to contribute to the community," the community being anyone with an EHR system.

That's where FHIR became interesting to us. Rather than being a proprietary data service out of Epic, FHIR is designed to be a restful, Web-based interface that works with anybody who supports the FHIR-based standards in a non-proprietary way.

FHIT: Geisinger has commercialized a rheumatology application via xG Health Solutions. Are there plans to do the same with other applications?

Erskine: There are a number of different products that exist that are in the pipeline that do a whole series of different things. Rheumatology was really just the first concept. There are roughly 5,000 rheumatologists in the country. xG Health Solutions isn't going to make it big on rheumatology; once that's established, though, we'll make it big on the next things that we have in the pipeline.

One of them is called SuperNote, which defrags some of the information in EHRs in terms of the problem list and what you're writing into your note, and reorganizes that information in a way that's more intuitive. It identifies some of the care gaps that a patient may have and puts them face up so that the clinician has to do less work gathering the information and can focus on writing the key components of the note. That information then gets sent back into the medical record as a clinical note.

FHIT: Have you encountered any hiccups in the creation of such applications?

Erskine: Several. The first one was some reluctance on some of the vendors to prioritize developing production-based FHIR receiving amongst a series of other things that they have to do. If I'm trying to get Meaningful Use done and ICD-10 done and other things like this, and we have one person that says "hey, we'd love for you to put production FHIR in place," it may not be the top thing on their list. Trying to convince them that this was actually an important and pivotal moment for the industry was one thing.

The other issue was the business models that needed to be built around the app store concept. If I was a client of that particular vendor, I could develop an app, but if I wasn't; could I still put my app in the app store? And if I put my app into your app store, then do you get access to all that IP and can you put it into your application? Or do I get some protection that actually continues to be mine and I could sell it to you or I could sell it to somebody else? It's kind of a willingness to do FHIR resources, as well as the business model piece.

Additionally, there is a discipline to this. We take care of patients. We're not necessarily an HIT development shop that has 7,000 employees where this is all we do with product management and application development.

FHIT: Have you talked to/worked with other organizations on these efforts?

Erskine: One thing that was useful is, in Project Argonaut, which we participate in, they assembled people that were early adopters and early implementers, and therefore we could exchange information amongst each other. One person that comes to mind is Ricky Bloomfield at Duke, who's been involved pretty heavily; he's written some apps.

Another person is Mark Braunstein at Georgia Tech, who has tasked his engineering students to write their applications against a series of different FHIR resources. Even though it's not a medical school or a health system, he believes that there's value in saying to these engineers, "use these data services, write whatever apps you're going to write and then create future FHIR-enabled application developers that can go out to the community."

FHIT: What are the implications for care that these apps are EHR agnostic? And what are the implications for other organizations also interested in creating their own apps through similar methods?

Erskine: One of the things I like about the use of FHIR is the fact that it can democratize functionality across any number of EHR vendors. I don't have to pick Cerner or athenahealth or Allscripts or Meditech or Epic because it has this one thing in it that I want to be able to use; I can get my base system of record and then pick this one app that works on any of them.

A second benefit is that in the patient activation and patient engagement space, there is less effort on the EHR vendors' side. There are portals and a number of other things, but there's a lot of white space in terms of how to better engage with patients. Something the health systems are probably better set up to do compared to an HIT vendor that doesn't have access to patients.

In terms of the formative evaluations and design and development and implementation of those things, it puts us at the advantage because we have the patient relationship and we have the data about those patients. So in terms of patient outcomes, making sure that we think through what the patient and their family proxies are as we design apps is going to be an important part of how we develop these apps.

Editor's Note: This interview has been condensed for clarity and content.