At the Danville, Pa.-based Geisinger Health System, leaders have been involved in almost continuous clinical transformation efforts for the past decade and more. Geisinger, which encompasses eight hospital campuses, two research centers, a 1,100-member multispecialty physician group practice, more than 21,000 employees, two research centers, and a 467,000-member health plan, serves more than 2.6 million residents across 44 counties in central and northeast Pennsylvania.
Geisinger is justly famous for the development of innovative care models such as ProvenCare® and ProvenHealth Navigator®, which have helped to transform the concept of value in care delivery and payment. Earlier this summer, HCI Editor-in-Chief Mark Hagland spoke with John Kravitz, the health system’s vice president, information technology, and associate CIO, and Alistair Erskine, M.D., its chief clinical informatics officer, to find out some of the latest developments there. This is part one of a two-part interview series. Below are excerpts from the first part of the interview with the two executives.
What’s the latest at Geisinger these days?
Alistair Erskine, M.D.: Let’s start with some of our analytics work. I joined Geisinger nine months ago [in the autumn of 2013]. And I picked Geisinger to come to because of what they’ve been doing around analytics. And they’ve created what is being called a CDIS—a Clinical Data Intelligence System. It’s basically a third-generation data warehouse. It was built around 2007.
Alistair Erskine, M.D.
John Kravitz: I came in late 2009, and it had just gone into production.
Erskine: It pulls in Epic data, through Clarity, Epic’s relational database. And encompasses both inpatient and outpatient data, with Geisinger and non-Geisinger patients across the system. But it also brings in costing data and other data. The CDIS helps us look at quality-of-care gaps, and create bundles, as for ProvenCare. So the care gaps could be something like, my diabetic patients have not had eye exams, for example, or getting a micro albumen, or figuring out why the hemoglobin a1c for particular groups of patients is not where it should be, or any number of things. But what’s unique about it is that it identifies care gaps, and matches the data with a human who calls up the patient and creates an intervention. So closing care gaps is both an automated data piece and also a human-facilitated piece.
Kravitz: And even on the human touch piece, we have predictive dialing; and if someone picks up live, it transfers them to a live person. And it may even be a biology/pathology result that’s malignant, and it helps us use automated tools to intervene or act. And because we have some claims through our health plan, and our population is fairly stable, we’re able to do a lot of analysis on those patients. And anywhere from 30 to 40 percent of our health plan members are seen by our providers, so then we have all our clinical information to do data analytics on them.
So the majority of your health plan members are not seeking care from Geisinger providers?
Erskine: That’s right. It means we’re half-HMO and half-standard health plan. But one of the lessons we’re learning from our HMO is that intelligence can be ported over to our standard health plan.
So you’re learning from the HMO patients, in the most controlled product?
Correct, because we don’t always have access to the clinical data in the looser products. And care managers on the health plan are literally logging into patients’ charts, to be able to understand what the next thing is that they should be doing. And there’s a lot of bilateral effort to make sure the patient who’s in our plan doesn’t get conflicting or confusing information. And in other places, where there’s a disconnected health plan, you just don’t have that level of service and care management. Our CEO, Glenn Steele, M.D., refers to this as the Geisinger ‘sweet spot.’
Is Geisinger in a one of the Medicare Shared Savings Programs for accountable care organizations?
Kravitz: We do have the Keystone Accountable Care Organization, which manages the care of 7,700 Medicare members—it’s pretty small, though growing. More broadly, we have about 80,000 Medicare Advantage members, and 115,000 Medicaid members.
Erskine: And one of the lessons being learned is that a lot of providers are finding out that the level of care management they thought they had wasn’t as good as they had believed.
Kravitz: We should talk about the ProvenHealth Navigator tool a bit. That is comprised of a care management function primarily provided by 167 care managers and 70 health managers at the health plan. We were one of the Beacon Communities in 2010, one of the 17 awarded grants through ONC [the Office of the National Coordinator for Health Information Technology]. That grant was specifically targeted towards treatment of patients with chronic illnesses. Among the criteria involved was to have a functioning health information exchange, so you could connect disparate clinicians to aid in the care management of patients. We focused on CHF [congestive heart failure] and COPD [chronic obstructive pulmonary disease]; we intended to get to diabetes, but it was only a three-year grant. I feel if we had had two more years, we could have gotten to diabetes as well.