As an early adopter of electronic medical records, Geisinger Health System is well-equipped to use health data in new and innovative ways. Because the Danville, Pennsylvania-based provider has had a system-wide EMR since the mid-1990s, both clinicians and researchers are very confident when navigating mounds of patient information and using it to improve patient care processes, Chief Data Officer Nicholas Marko says.
"We've got around 20 years' worth of data on the back end of our information stack," Marko tells FierceHealthIT. "We've been in an environment with a lot of information and a lot of data for quite some time, and we're fairly comfortable transacting in it and molding processes around it."
In the second part of FierceHealthIT's exclusive interview with Marko, he talks about some of Geisinger's current programs built around the use of data for better care. He also discusses educational and training opportunities for future potential healthcare chief data officers. Read part 1
FHIT: How is Geisinger using big data to improve patient care?
Marko: One of the interesting things about Geisinger is that we were an early adopter of the idea of a system-wide electronic medical record. Some of our earliest processes that were built in a data-driven fashion were a variety of evidenced-based care pathways that Geisinger called ProvenCare. They essentially were disease- or condition-specific pathways that provided optimized roots of care for patients to make sure they were getting the best possible available treatment options in a consistent fashion. Those pathways also generated data on the back end so that we could always monitor the outcomes of those patients and the process of care delivery so we were sure that we were actually functioning the way that we envisioned functioning.
We also have a fairly extensive program that we call "care gaps" that's designed to find patients who might slip through the cracks, somehow. For example, a diabetic with a series of test results that are a bit off that may not be getting appropriately noticed for some other reason. We tie the back end data infrastructure to a series of rules and predictive algorithms to try to identify patients who need attention refocused on the need to get access to some sort of special type of treatment, and we work on connecting them and closing those care gaps.
There's also the "Geisinger in Motion" program, which focuses on the fact that data now is coming in from a variety of different places and that patients are interacting with their healthcare environment in a variety of different ways. It's helping us to improve communication with patients while extending the delivery of care.
We're also using natural language processing techniques to look at a lot of plain text; a lot of healthcare information gets entered into notes by doctors or reports that are dictated, so we're using NLP and other analytics on top of it to help find information that is clinically relevant to patient care. For instance, incidental findings on radiology reports that merit follow-up, but aren't the primary reason that a doctor ordered that study may get lost in the shuffle.
FHIT: Are there any measurable results that you can talk about?
Marko: The "care gaps" program has been up and running for a couple of years now. There are hundreds of patients who have been identified through those processes as having some part of their care that can be better optimized, and then we've communicated back to primary care physicians that have shown measurable improvements in a variety of metrics that we track.
In the ProvenCare systems, we've shown that a variety of different ProvenCare systems patients have better long-term outcomes.
With the natural language processing efforts, we've identified 20 or 30 patients in the last year that required treatment for abdominal aortic aneurysms that might otherwise have been overlooked. We've prevented problems and gotten patients more optimal care earlier on in the course of their treatment so we can focus on prevention rather than reaction.
FHIT: You previously talked a lot about the need for more chief data officers in healthcare; are there any educational efforts geared specifically toward molding hospital CDOs?
Marko: We're relatively new in the process of chief data officers, in general, as a position. The first ones in any sector were named about eight to 10 years ago, and the position has been growing since. In healthcare, we're much newer in that space. While there have been one or two chief data officers in healthcare for seven or eight years, it's really only over the last two or three years that you're starting to see the position take hold in a variety of healthcare operations.
There's certainly not a formalized program for training somebody who would want to have that role as of yet. Most of the people I know found themselves in that role doing something else first, and got connected with and very interested in how the information was managed. They then grew into that role as they became their organization's go-to person for that sort of thing.
As the position continues to take off, there will be a role for some additional formalized training in that. I tend to think that there's some value to having domain expertise when you're a chief data officer, although it's not strictly necessary because many of the challenges faced in healthcare and in other places are 90 percent similar, regardless of what the data content is.
One thing I think might be a little bit unique about healthcare is that the product that we deliver is very different than the product or the culture around delivery of other products in other sectors. In healthcare, I think there's an inherent advantage to having somebody who's your chief data officer be somebody who has a background in the delivery of patient care.
The way that we train people now is essentially through healthcare informatics fellowship programs. These are for physicians and there are just a few of them that have popped up over the last couple of years; I believe there are five or six nationwide now. Geisinger has just started a health informatics fellowship here, also; and we're accepting our first class of students now.
Editor's Note: This interview has been condensed for clarity and content.