Skip to content Skip to navigation

Providing Value Through Analytics

March 5, 2013
by Rajiv Leventhal
| Reprints
UPMC's director of provider analytics shares his thoughts on the system's big data journey

At the University of Pittsburgh Medical Center (UPMC) health system in western Pennsylvania, predictive analytics is the next big step in its big data journey. The UPMC system, which encompasses more than 20 hospitals and 400 outpatient care sites, employs more than 3,200 physicians, and counts 55,000 employees, has been moving rapidly to aggregate and harmonize all of its data.

On March 5th at HIMSS13 in New Orleans, Assistant Editor Rajiv Leventhal had a chance to catch up with Oscar Marroquin, M.D., director of provider analytics and director of the Center for Interventional Cardiology Research at UPMC, to talk more in depth about how UPMC is moving forward with its big data initiative. Below are excerpts from thats interview. 

What are you hearing and seeing at HIMSS?

The main expectation was to see if there were other systems doing similar things to what we are trying to do in terms of their analytics efforts. And I’m hearing that a lot of organizations are talking about what they want to do moving forward, but they don’t have a very detailed plan as how to go about it. So far, there hasn’t been a lot of breakthrough that makes us say, “this is something we have to do.”

How are you using data analytics to improve patient care?

Specifically within cardiology, we are doing several things. First, we are trying to manage our data to the point where we can get information out of it by applying analytics on top of it. The first big hurdle we had to face, which a lot of organizations face, is structuring data at the point of care. We think we have successfully achieved that by changing the culture within our physicians. We have a robust informatics team, and have now started to bring reports to physicians so they can see what are the patterns of their patient care, how they are doing with their outcomes, and recently, we have followed metrics on the appropriateness of their procedures. We have also been developing clinical pathways to help guide physicians to the appropriate steps that need to be taken to deliver better patient care. Currently, we’re in the process of digitizing some of these pathways so they can be brought into the workflow. From that learning, we can refine those pathways and processes.

Is data becoming more value based rather than volume based?

Absolutely. In the past, when everything was around volume, the physicians were less interested in understanding the impact of their outcomes. Lately, we’ve seen a significant change in the culture and an acceptance of more information. There is now a more willingness to say, “Would you be able to share some of that [data], because I think it will be helpful?”

In terms of doing analytics, what are some lessons you have learned or challenges faced along the way?

When it comes to data, it is a rigorous process to make sure you understand the data and really grasp what it ‘s telling us. Also, it’s important to know how the clinicians are receiving the data. There have been times where we have sent data to clinicians and we didn’t get the response we were looking for because there is doubt about where the data is coming from and questions about the validity of it. Some physicians have told us that their patients are sicker than the ones we are generating the data from, so this information doesn’t apply. We need to have answers to those questions in advance so the data can be used in the most impactful way.

What are your visions going forward and what opportunities do you see with analytics at UPMC?

Well there are two big things, and they’re not necessarily separated. First, we are doing the analytics in a retrospective way. We’re asking how did we take care of the patients over the last week or month? Then, we share that information with the clinicians. And that is new but clinicians will learn from it. The idea is to introduce decision support tools that incorporate analytics into patient workflow at the point of care. Similarly, the other big part is getting patient involvement in the decision making process. But for that, you need the data to show the patient, and you need them to be educated so they can, along with the physician, make the most appropriate decision.

Are patients ready and willing to be more active in their own health?

They’re as ready as they’ve ever been. But as physicians, we need to get them more actively involved. Some physicians still refer to the notion that “the doctor knows best,” but in this era of information technology, it’s our responsibility to get patients more involved in the decision making process.