Why Geisinger Is Moving its EHR to the Cloud Now

The Pennsylvania-based health system is partnering with Amazon Web Services to move its electronic health record platform and other services into the cloud.

Geisinger recently announced it has chosen Amazon Web Services (AWS) as its strategic cloud provider. The project, one of the largest migrations of electronic health record systems to AWS, will save the Pennsylvania health system several million dollars a year, which executives say they'll use to make investments in improving the health of its patients.

Geisinger Chief Information Officer John Kravitz, MHA, CHCIO, spoke to HealthLeaders about this partnership and how it will impact innovation at Geisinger. This interview has been lightly edited for brevity and clarity.

HealthLeaders: Your move to the cloud -- why now? Why not earlier or later?

John Kravitz: For us, as an organization, the time was appropriate to start. We looked at the cloud for the last two years, did an RFP process with the three major public cloud providers, and we decided we wanted to get to the cloud, so we're doing some things like setting up an API management platform to provide better integration with APIs. The ability to move to the cloud, to be agile, to be scalable, to move our applications quickly, to deploy applications quickly, that makes sense to move our strategy for the organization forward.

In the meantime, we're looking to simplify, as much as possible, our environment, as we're starting to migrate to the cloud. We're in early stages. We've just done the design and some other things for it, and started to collaborate on some small applications. It will take us a while to get to Epic, probably 18 months to 24 months before we get to an Epic production cloud. It’s going to be a three- to four-year journey before we get there.

John Kravitz, chief information officer, Geisinger. Photo courtesy Geisinger.

There are some cost savings associated with it. But that wasn't our main driver. It was agility, disaster recovery, business recovery, all those components of it. We had two data centers, but they're just in the same region, too close together and not a good scenario. We need to be diverse, and we need to have the capability to recover our business and keep business moving forward.

HL: Regarding the 400 applications mentioned in the announcement, is that typical of an organization your size?

Kravitz: Actually, it's atypical. We have about 1500 applications. We're rationalizing about 450 applications (and) eliminating them, as we're going to move to the cloud. Our goal (is that) we'll move about 90% of our applications that we can into the cloud., There'll be some things like PACS imaging, you know, for radiology, cardiology, other things like that, that really have to sit here. The middleware to connect analyzers for laboratory and other things have to sit here and connect to devices; those obviously stay on premises in our primary data center, and then we will integrate that data right into our cloud environment.

So in actuality, it's probably going to be closer to 900 applications that will migrate to the cloud. That would be about 90% of just over 1000 when we get there. Our goal is to continue to drive down specialized applications for every little thing, really driving closer to more vendor-centric, deep relationships with our EHR vendor and some other core vendors that we utilize, to utilize their stack as much as possible and not have best-of-breed systems.

HL: Going to the cloud is really the time if you're ever going to do that.

Kravitz: Exactly. You want to clean it up before you move it, because it'll cost you a lot more to move a lot of inefficiently developed applications to the cloud. When you get to the cloud with SaaS (software as a service), or infrastructure as a service, you’ve got to be able to utilize resources appropriately, and not over-allocate, or you're paying money for wasted resources. We're looking at a complete migration, moving applications to either PaaS (platform as a service) or infrastructure as a service where possible, or rewriting applications to make sure that they're efficiently used in the cloud.

HL: How do you integrate these different services once they're in the Amazon cloud?  What kind of challenges does that represent?

Kravitz: We're working through those right now, using APIs where possible, or HL7 integration to make those things work, or even EDI if they're business-related transactions. Because our organization isn't just a clinical enterprise. We have a large clinical enterprise, we have a health plan, we have a medical school and a graduate education school. I've got different platforms for like EDI transactions that work for the health plan. And that has to be on an EDI engine, Tibco in this case. We can go across the backplanes, to the clouds, to connect things back and forth without having it go out to the public cloud and back in. We're going to be doing it with that environment, working with the architects from Deloitte, as well as Amazon architects. We've already started scaling some of that.

HL: What role is Deloitte playing?

Kravitz: They are working with us now in the initial education and training for our staff so that we are able to do this move on our own. We have a commitment with them for a short period of time, and then the knowledge transfer is there. And we will then transition over our own people to do all this work.

HL: What does that education entail? What does the average Geisinger employee need to know or do differently?

Kravitz: It’s not the average. It would be an infrastructure person from Geisinger, from the IT team. They're learning how you do segmentation within the cloud, because there are certain things that have to be segmented away, like the health plans, and the health system has to be protected. The information has to be kept separate and distinct. They're learning that, they're learning the firewall ruling and everything else that works with it: data loss prevention, you know, CASB, some of the security tools.

HL: So many people these days think the cloud is a real inflection point for innovation.

Kravitz: Cloud providers like Amazon have done really innovative work in their own environments. It provides us an opportunity to use AI and some other tools that can leverage the data that we have that's protected through nondisclosure agreements and other business associate agreements. To be able to have access to that data to assist us in doing deeper evaluations of the data and looking at things that we can do to improve the way we provide health care to our customer or to our patients.

HL: Is it true that if you have a cloud infrastructure, you don't necessarily have to ship data around as much as you did before, which inherently can provide issues of privacy and security? Do you adhere to that frame of mind about this?

Kravitz: Conceptually, yes, that is going to be the case. In practice, we hadn't had that experience just yet. But what we've learned in discussions with Deloitte, with Amazon, that is certainly the intent as we move forward.

HL: Albert Einstein once said, 'In theory, theory and practice are the same. In practice, they are not.'

Kravitz: Well, we'll find that too. I've done it a long time. I've seen a lot of things that don't work out as planned. But this does make sense that you would have the capability to use those analytics tools, and be able to effectuate change with those. We do a lot with analytics now on premise, with big data lakes, and Hadoop environments and other things like that. But we're looking at newer tools that are really cloud enabled. An example could be Snowflake, and other tools that are available for analytics purposes.