A lot has been going on lately at the University of Pittsburgh Medical Center (UPMC) health system in western Pennsylvania. Last October, the UPMC system, a $10 billion annual integrated health enterprise based in Pittsburgh, which encompasses more than 20 hospitals and 400 outpatient care sites, employs more than 3,200 physicians, and counts 55,000 employees, announced a five-year, $100 million enterprise healthcare analytics initiative to foster personalized medicine and other core strategic healthcare IT goals for the organization, partnering with several prominent vendors—Oracle, IBM, Informatica, and dbMotion.
As a press release made public on Oct. 1 noted, the leaders at UPMC are determined to harness the data found in various sources in their organization to create a best-in-class data warehouse to drive massive advanced analytics and predictive modeling capabilities. The initiative also accords with an explicit strategy on the part of UPMC leaders to partner as much as possible with IT vendors and other healthcare business organizations with compatible visions for the future of healthcare, a strategy that has already led to breakthroughs in such areas as digital pathology and semantic interoperability.
The scope of the big-data initiative announced in October is virtually unprecedented in private healthcare in the United States. But it accords perfectly with the ambitious overall agenda for corporate development and performance improvement that enjoys consensus support across the UPMC organization. Among the senior leaders in the vast integrated healthcare organization helping to move this agenda forward is, of course, Dan Drawbaugh, UPMC’s senior vice president and CIO. Drawbaugh welcomed HCI Editor-in-Chief Mark Hagland to sit down with him in his offices at the top of the US Steel Tower in downtown Pittsburgh for an exclusive interview, at a crucial moment in UPMC’s ongoing evolution. Below are excerpts from that interview.
What are your high-level thoughts on where IT fits into the broadly ambitious overall corporate agenda for UPMC as an organization, and the journey forward towards the new healthcare, towards a healthcare system in this country that will be more transparent, accountable, reliable, cost-effective, higher-quality, and consumer- and payer-responsive?
A couple of things. One is relates to the role of IT, as we move into the era of big data and big science, something that Dr. Steve [Steven D.] Shapiro [M.D.], our chief medical and scientific officer talks about; and that is that intensively leveraging IT for overall organizational performance improvement is becoming the norm. So while healthcare was always was a data-intensive industry, the amount of data before electronic health records hit, was manageable. But now, if you look at the growth curve in the electronic health record world, the data is doubling in volume every 18 months. But that’s just the tip of the iceberg. What happens is that, with the science side also advancing, you’re going to get to exponential growth. And I like to put ourselves in a place where we’re ready for things, almost as though something might happen tomorrow, so that even if you know something is three years out, I like to ask people here, what if something in particular happens tomorrow? Are we ready for it?
So from a strategy perspective, I see things like systems engineering being valuable, as some are calling it. And in the storage world, we’re moving from disc to computer storage right on the system, with flash technologies. And Oracle and IBM are working on that area.
And accessing the data, making it available, providing predictive analytics, are a part of that. It’s almost like a tapestry of different elements is involved, and it’s becoming clearer, and you’re starting to see the pieces of what you’ll need in place. I look at our strategy here at UPMC around the electronic health record, then interoperability, then semantic interoperability, and into predictive analytics, and our strategy has to keep evolving very quickly, as the technology evolves. Our focus is on execution. And it’s easier when you’ve seen someone else do something. But we’re pushing the envelope, so in some ways, that requires trial and error.
Meanwhile, as knowledge gleaned from academic research gets transferred to practicing physicians and as they become knowledge about the technologies involved in genomic medicine, I can see where that will also involve an iterative process feeding back into the electronic health record and into patient care.
With information from both genomic medicine and from outcomes research emerging, do you see a continuous cycle of analytics and performance improvement for medicine?
Yes, with an added element in which you’re actually reaching the age of artificial intelligence, so that data feeds back into that workflow, and what will happen is that the processing of these vast amounts of data will allow for the clinician to evaluate the data available to him or her. The system will say, ‘This is what we analyze in terms of Mark’s condition, and we have a 90-percent confidence level in this analysis; do you feel this is right, Doctor?’ In other words, rather than requiring the physician to collect the data points, the data points will come to the physician, who will still make the decision. But it means real-time predictive analytics brought to the point of care. And we believe it will turn into where the physician or clinician, sitting in a diagnostic cockpit, if you will, with the electronic health record, the analytics, the imaging information, molecular diagnostics, the genetic information, will all feed into that cockpit. So a lot of components will automatically be populated, and this power will be at a level higher than where we are today.
And we’ll see convergence; in fact, we actually have a project called Convergence, and you can see elements of that being demonstrated at our Technology Development Center [which displays prototypes of emerging projects within the organization]. And look at Watson, and what Larry Ellison and his colleagues are doing with Exadata. I think it’s moving much faster than what we realize. And staying out in front will be important.
Might not a digital divide between the haves and have-nots be emerging in this area, with some healthcare organizations having the resources to “do big data,” and some not having those resources?
I think you’re going to see a lot more consolidation to deliver on these healthcare reform and regulatory requirements. I just don’t see the one-two doctor practices continuing; I see large groups or large health systems achieving this. I think you’re accurate when you say that, that there will be a digital divide between those who can move to those technologies and accomplish a convergence and an integration into the workflow, with analytics capability. I think it will be a requirement of doing business, of delivering care. So fortunately or unfortunately, this is the way it’s moving.
And you see it with healthcare reform and the ACOs; you look at these groups of providers and insurers together… and we’re already there. But there are organizations that will stand out in terms of their IT capabilities.
And that will be a market differentiator, right?
Yes. And you see situations where people might say, the CIO should be doing that, but the science area now is all about technology, but the reality is that almost everything we do now has clinical elements intertwined, so it requires collaboration with senior clinical leaders, such as Dr. Shapiro and his colleagues, here at UPMC. And in the past, we used departmental information systems to accomplish specific tasks; but now it’s moved to where IT is the solution overall.
And you see that necessary coming together of non-clinicians and clinicians in IT, right?
Yes, I do see that as a team coming together. And I see the CIO, the CMIO, and the CTO all sitting together, across different types of reporting structures. The reality is that the reporting structures don’t matter; but those responsibilities have to be approached in an interactive way. And I look at Dan Martich [G. Daniel Martich, M.D., UPMC’s CMIO] and Vivek Reddy [chief medical information officer for UPMC Physician and Hospital Services Vivek Reddy, M.D.], and I don’t feel as though my responsibilities are “higher” or anything; instead, we’re all bouncing things off each other. The same thing is true with the CTO role, which is executed here by a few different individuals.
And let’s just take the data network or let’s say the server environment, the virtualization environment—if any of that isn’t right, you’ve pretty much shut everything down. And staying ahead of that, it is about skating to where the puck is going, as Wayne Gretzky said. And we’ve been fortunate in the culture of the organization, which is very future-focused. We had a situation here just last week, when Adrian Lee, Ph.D. [a researcher working on applying IT solutions to the task of translating ongoing research advances in genomics to breast cancer care at the University of Pittsburgh] made a presentation to our IT board of directors meeting that articulated how sophisticated IT is helping to improve breast cancer care for our patients, and Mark J. Laskow, the vice chair of the IT board of visitors and a director on the UPMC board of directors [and CEO of Greycourt and Co.], said, ‘This is exactly why we invested the billion dollars in IT, to improve the care for the individual, to predict and address illnesses upfront, and find cures for illnesses, this is what it’s all about.’ And think about that statement, it’s pretty cool, it’s pretty exciting.
Are there any lessons learned so far on this big-data journey at UPMC that you could share with your healthcare IT leader peers nationwide?
One thing I reflect on is that I look at the challenges when we started out; and when we first discussed analytics, it was a challenge. It’s nice to use buzzwords, but setting up to where you can execute, that’s a challenge every day. Every day, you’re learning, educating, communicating, and finding out every day who can help you execute in the organization. And you have to go outside of the circle where you’re comfortable in. And I remember two and a half years ago, going to meet Ramayya Krishnan [Ph.D., the Dean of the Heinz School at Carnegie Mellon University], and learning about big data sets and such; and now we’re at the point where we’re training individuals in big data. Two years ago, we were focused on beginning the journey; but as you execute, it’s going to be challenging, and persistence around your effort is essential; don’t give up. It took us 18 months to where everyone believed we should fund an analytics initiative. And now everyone sees that.
I think that was a good signal to the industry.
And you were one of the first to really publish about that. And was interesting was the phone calls and the energy around that, after your story appeared. And at that time, only a small percent of the industry had achieved stage 1 of meaningful use. And you were ahead of the curve [at Healthcare Informatics] in educating the industry on interoperability, on the next steps that will happen. This is why we made the billion-dollar investment, as that vice chair said.