When Rick Skinner took on the role of CIO at UVA Health System last year, he came with a specific goal: to accelerate the academic medical center’s use of analytics. But for Skinner, who had spent four years at Cancer Care Ontario, it was also an opportunity to “return to the front line” in supporting IT. In this interview, he talks about what it was like to go from a single-payer, government-sponsored system in Canada to an academic organization in Virginia, his team’s goals in starting an ACO, and his experience with an integrated record system in the Army 20 years ago. Skinner also discusses UVA Health’s support model with Epic, his patient portal strategy, and his thoughts on population health management.
- Big plans for big data
- “It’s starting to bear fruit”
- Predictive analytics in the NICU
- UVA’s Data Sciences Institute
- Starting an ACO to “transform the way we operate across the board”
- Acquiring Culpeper Regional
- ConnectVA & Epic Care Everywhere
LISTEN NOW USING THE PLAYER BELOW OR CLICK HERE TO SUBSCRIBE TO OUR iTUNES PODCAST FEED
The last time I talked to our chief quality officer, she reminded me that she has 493 quality metrics that she has to report on at least a monthly basis. So just doing the routine stuff that isn’t really even analytics is a major task.
Our intent was not so much to go after the potential savings, but rather to learn about population health and accountable care, and to get an early start in transforming the way we operate across the board.
Deciding what to change when and what impact it’s going to have on the volume of patients you see, the reimbursement you receive, the setting you see them in — all of those are variables that are going to change over a short period of time, and change pretty dramatically.
The Department of Health has mandated that we report public health statistics like syndrome surveillance through this HIE. Obviously the hope is that that will foster adoption of the HIE by more facilities. We’ll see whether that happens or not.
Gamble: Being an organization like yours, I can imagine the amount of data you deal with, and I just wanted to talk a little bit about your data management strategy and what you’re doing with analytics, or what you plan to do.
Skinner: Sure. As a matter of fact, that’s a particular interest of mine. I moved down here from Toronto about a year and a half ago, and had been responsible for a fairly large analytics effort for the Province of Ontario. That was one of my major objectives; actually, one of my tasks coming here was to accelerate our use of analytics. We’ve built an analytics team led by data scientists who came to us from outside of health care and have not only the data stores from what I’ll call our transactional systems like Epic, PeopleSoft, and others, but an enterprise data warehouse and a set of analytical tools.
We’re just now getting to the point where we’re pretty regular in terms of being able to effectively and efficiently generate all the routine reporting. As a matter of fact, the last time I talked to our chief quality officer, she reminded me that she has 493 quality metrics that she has to report on at least a monthly basis. So just doing the routine stuff that isn’t really even analytics is a major task for a health care organization, in particular an academic one. Then you add to that all the internal reporting — how many people did what last night and those kinds of things. That’s another huge layer, especially in an organization like this one that is focused very intently on performance improvement across quality, safety and other dimensions. So you can imagine the input or the information needed to support that, before you even get to the fun stuff like predictive analytics having to do with monitoring newborns in the NICU, which is actually something we’re doing where we’re able to, with a high degree of accuracy, predict which cardiac-challenged newborns are going to worsen over the next 24 hours and direct people to that bassinet.
All of that taken together has been a major focus here. It’s starting to bear fruit.
One other aspect of this is that we’re a brand new Medicare Shared Savings Plan ACO. Of course, population health analytics is a big piece of being successful there, and we’ve created a small analytical team to support the ACO to do those kinds of analytics.
Gamble: I can imagine the amount of legwork that has to happen before you start getting those really cool uses out of the data and doing things like the predictive analytics. It’s really interesting.
Skinner: I should mention one other thing that, at least so far, has been fun. Being part of the University of Virginia, the academic side of the university has spawned something called the Data Sciences Institute, which is the university’s effort to harness the power of the buzzword ‘big data’ to explore advanced computational methodologies for using large datasets. We’ve partnered with the university to use some healthcare datasets in those efforts, and that’s very interesting. It’s very early on, but still very interesting.
Gamble: It’s one of those times where I bet it’s really beneficial being part of a university like that.
Skinner: Exactly. They have access to a whole faculty of PhD mathematicians and data scientists and so forth.
Gamble: As far as the ACO, can you talk a little bit about that—who is involved and how that got off the ground?
Skinner: Early last year, we decided to take an opportunity of CMS’ call for an expansion of the ACO. The year before had been the first year of ACO applications. We applied during the second year, which was last year, for a Medicare Shared Savings Plan ACO, which in essence, CMS decides which individuals in our geographic area — which Medicare-eligible beneficiaries—had received the majority of their primary care at our institution over the last three years. They have a very complicated algorithm to figure that out that, one, I don’t remember and, two, I won’t bore you with, but in any case, we applied. We were accepted.
What that means for us is that there are approximately 20,000 Medicare beneficiaries in our geographical area for whom we provided the majority of their primary care. Going forward, we are responsible for 33 different quality measures for that population of Medicare beneficiaries, and in future years, are eligible to share in any cost savings that we generate over the traditional fee-for-service Medicare for that population. Our intent in getting into this was not so much to go after the potential savings, because this is a fairly small proportion of our total number of patients, but rather to learn about population health and accountable care, and to get an early start in transforming the way we operate across the board not just for Medicare beneficiaries in this very emerging and new world of accountable care.
Gamble: It’s interesting and it seems like it’s something that you really have to have a very deliberate approach to because this is such a drastic change in the way things have been done for so long.
Skinner: It is, and in particular for academic health organizations, it’s almost diametrically opposed to our current practice. And so deciding what to change when and what impact that’s going to have on the volume of patients you see, the reimbursement that you receive for seeing those patients, the setting you see them in — all of those are variables that are going to change over a short period of time, and change pretty dramatically. Not getting too far ahead of and behind that wave is critically important.
Gamble: You mentioned earlier in the interview about acquiring Culpeper Regional Hospital. Was that something where previously or currently you had a partnership with them where you shared IT expertise and things like that?
Skinner: No, not really. Several years back, we had obtained a 49 percent ownership stake in Culpeper Regional Hospital, but Culpeper has always operated independently of UVA. They were a standalone regional hospital, and although we owned 49 percent of them, other than having a couple of seats on the board, we had no stake in or influence over the way they operated that facility.
And then at some point, their management and governance decided that they would offer the remaining ownership to University of Virginia. We accepted that offer, and came up with an agreement for how to do that. Exactly what that’s going to look like in terms of who does what is yet to be determined.
Gamble: So it’s still in the early stages? Now, as far as data exchange with other organizations, are you involved in any statewide or regional HIEs at this point?
Skinner: We are. We are a member of ConnectVirginia, a Virginia-based HIE which is fairly new; it’s been around for a year, maybe a year and a half. We are one of less than a dozen members at the moment, but the interesting part about this particular HIE is that the Virginia Department of Health has mandated that we report public health statistics like syndrome surveillance and other kinds of things through this HIE. Obviously the hope is that that will foster adoption of the HIE by more facilities. We’ll see whether that happens or not. But the other piece is that we are, through Epic, connected to all of the other institutions in the state that use Epic. That’s been very beneficial as well, since as an academic health center, they are all part of our referral network.