In 2010, Randy Gaboriault left an industry that thrived on innovation to answer the call of another that “had not incurred disruption in decades.” Healthcare, he believed, offered him an opportunity to “reshape it from my backyard,” and he accepted. Five years later, Gaboriault is leading a top team at Christiana Care that is focused on creating a ‘true community health record’ and harnessing the power of predictive analytics to improve outcomes. In this interview, he talks about the major IT plans on his plate, what he believes are the core competences of health care, what leaders should mean to their teams, and what surprised him most about the CIO role. He also discusses the innovation challenge that was issued to his team, and the trend that CIOs must work to reverse.
- Reporting/BI vs. true analytics
- CMMI grant to reduce readmissions
- “Healthcare is really underutilizing big data.”
- The flawed approach to analytics
- Embedded teams that “live, breathe and understand the problem.”
- Fostering innovation
- Time management challenge — “How do we flip that?”
LISTEN NOW USING THE PLAYER BELOW OR CLICK HERE TO SUBSCRIBE TO OUR iTUNES PODCAST FEED
As human beings, we’re not really very good at seeing things or patterns that we’re not looking for. So we’ve been focused for the last three years on developing a specific capability to effectively be able to do that
Let’s understand it. Let’s be able to explain it. Let’s actually begin to predict who, and then begin to move upstream and actually modify things that are happening to prevent the readmission.
For the most part, all industries are really just at the opening chapter for this. We look at these tiny slivers of information. We’re not looking really big and wide at an ecosystem of information.
We’ve taken on an innovation challenge to think about how we actually begin to solve those issues and create a much higher fidelity, quality experience and quality outcome by designing novel solutions to problems, and we do that by bringing together the people that sort of experience the problem from different perspectives.
As an organization, you spend a tremendous amount of time just keeping and managing the present, and very little time thinking about the future. And so our objective, is how do we flip that? How do we actually spend much more of our energy focused on where we need to go? That’s the challenge.
Gamble: When you talk about everything you have about going in a large organization, I can imagine that the whole issue of managing the data is a big priority. So I wanted to talk about your strategy there and some of the work that you’re doing or plan to do with business intelligence and analytics.
Gaboriault: Sure, in fact, it’s an area we’ve been actually focused very heavily for a lot of different reasons, obviously including the changing healthcare commerce framework as well. I think that like a lot of folks now, it’s helping our constituents understand the difference between reporting and BI and analytics. To me, when I think about BI or the specific reporting component in healthcare, I start to think about units per thousand — how many falls per thousand, how many readmissions, how many this or that per thousand. And getting a diagnostic on how are we performing looking at that information; having designated teams of individuals looking at that and determining how to execute performance improvement.
The analytics is a really different piece. For us, it’s about will the person with brown eyes that had a stent placed on Tuesday with a BMI of 25 be readmitted? And by the way, we actually want to be able to identify the risk of that person being readmitted at the time of their admission for the intervention, so we can start determining what things we have to actually put in place to modify risk for the intervention.
We’ve been very focused on doing that and actually building a lot of machine learning capability and recognizing that as human beings, we’re not really very good at seeing things or patterns that we’re not looking for. So we’ve been focused for the last three years on developing a specific capability to effectively be able to do that, and we’re doing that with a particular patient population that we started on. We actually launched this with a CMMI (Center for Medicare and Medicaid Innovation) Grant, an IT-generated grant that focused specifically on the ischemic heart disease population. And what we saw was that after the intervention, if the person’s experiencing an event — you can have an open heart, valve, or bypass or have a PCI or a stent placed or you can have a medical therapy — the question is why are people being readmitted, understanding that a large population are readmitted within 30 days and a significant population, a vast majority of those people, are readmitted within 365 days, and being able to pose that question, which is why? Let’s understand it. Let’s be able to explain it. Let’s actually begin to predict who, and then begin to move upstream and actually modify things that are happening to prevent the readmission. So from our perspective, we’ve made a lot of investments in the space and we’ll continue to take that capability and scale it across chronic disease conditions, and then cystic fibrosis and cancer as well.
Gamble: That’s really interesting work and such a great use for the data. It’s a matter of being able to leverage the data to really be able to change outcomes.
Gaboriault: It’s true. One of the challenges I think is healthcare is really underutilizing big data. Actually, for the most part, all industries are really just at the opening chapter for this. We look at these tiny slivers of information. We’re not looking really big and wide at an ecosystem of information. The way we would traditionally do this type of stuff, we would think we have to go and pick and choose what attributes we’re going to put into a model, and then out of those few attributes, some biostatistician and clinician are looking at that and saying, ‘Okay, what are we going to include?’ And then they’re going to use that to try to predict from that limited set of information of who’s going to be readmitted. It’s flawed in the fact that upfront, we historically have to choose the factors that we actually want to consider and then build some sort of model around it, and then use that model to extrapolate to all of the actual real experiences we have with patients we’ve seen in the past, and run some sort of regression to determine that. It’s effectively flawed. Our approach is that we’re not actually going to pick a model. We’re going to start with a question and leverage the data and actually find the patterns that we’re not looking for. And then we can get to the point where we can actually optimize action for each unique person.
Gamble: Absolutely, it’s the end goal. It kind of dovetails to when people talk about how innovation doesn’t necessarily mean we’ve invented this super cool technology, but sometimes it’s just thinking things in a different way and using different methodologies. I wanted to just get your thoughts on what innovation really means and should mean for advancing a health system and advancing patient care forward.
Gaboriault: Think about innovation from the standpoint that we’re going to have algorithms that are actually going to be part of the care team in the future. There’s a private equity and venture capital firm called Deep Knowledge Ventures. They’ve actually given a voting membership seat to an algorithm as they assess capital investments about where to place their capital. So teeing off of that innovation, I see that we’ll have lots of analytic-driven innovations taking place. But the broader picture of innovation for us comes at the point where you begin to have intersecting disciplines that each have a different sort of lens on a particular problem.
We’ve done a lot of innovation around particular problems that we found, but we’ve chartered our organization to think about problem sets from a standpoint that 80 percent of serious medical errors come basically from communication issues. So we’ve taken on an innovation challenge to think about how we actually begin to solve those types of issues and create a much higher fidelity, quality experience and quality outcome by designing novel solutions to problems, and we do that by bringing together the people that sort of experience the problem from different perspectives.
The other piece is that from an IT perspective, there’s not one goal, not one question, not one challenge, not one initiative that’s not going to be enabled though IT, and so you begin to pair the ability to drive very different levels of innovation, because some of the technologies that exist where we can do things that would’ve been cost prohibitive to do even five years ago. From that perspective, we’re very much focused on how can we transform care by creating this intersection; creating the weather conditions for people to feel safe to actually challenge what’s been built from an existing perspective, and encouraging them to try ideas and fail fast against trying different things out, and how can mock up ideas very quickly.
Gamble: Is that something you do by assigning specific roles or having committees or meetups — anything specific you do to try to encourage this or foster this type of innovation?
Gaboriault: Yes. A lot of organizations have done things like create a formalized innovation center. We have an innovation team of people, so what we do is obviously there are human beings that have the ability to move to where the problem is and embed themselves. The first piece for us is creating an ability for our folks to be invited to witness the problem. Going back to that innovative team, a great example is our folks in emergency. We run one of the busiest emergency centers on the East Coast. We push upwards of 200,000 ED visits on an annual basis, and so we’ve had some major challenges in terms of how do you operate at that level of scale?
Our folks in emergency leadership were facing a couple of particular problems. There’s a comfort established where they can reach out and say, ‘we’ve got this particular problem,’ and what we do is we actually embed our innovation team inside emergency to basically live, breathe, and understand the problem from multiple dimensions. And then we begin to iterate what might be potential ways we can help solve the problem through a combination of technology and process in conjunction with them. We’ve done that in 10 different ventures. Internally, we’ve been able to do that with our folks in oncology around helping providers better understand pain, symptom presence and intensity for patients. We scaled that solution to heart failure and into behavioral health. We’ve been able to do that in areas around pressure ulcers. We’ve been able to do that in areas around actually capturing handoffs — not just clinical handoffs, but sediment handoff from one provider to another at shift change. We’ve been very effective at leveraging embedded virtual teams inside the problem.
Gamble: It’s a really interesting because there is so much potential, and I’m sure I can almost get at overwhelming that there is so many things that you do want to address and so many outcomes you do want to approve. But in the meantime, there are all these other priorities that the organization has, so I can imagine that that’s where the challenge comes in to play.
Gaboriault: We challenge ourselves — I do along with my organization — from the standpoint of, how much of our time is spent managing the current state? How much of our time do we spend really shedding the legacy processes tools and capabilities? And then the third piece is, how much of our time and how much of our capacity are we actually spending on designing the future? When you start to look at that question, you find that as an organization, you spend a tremendous amount of time just keeping and managing the present, and very little time thinking about the future. And so our objective, is how do we flip that? How do we actually spend much more of our energy focused on where we need to go? That’s the challenge. And we find that when we do that out in our service lines and with our folks, problems emerge, and from a governance perspective, you have to be able to allocate that capacity. And that’s where, as we look at all the innovation ideas, we do use some decision-making process to figure out which ones we think are going to provide a return. And that return can be in the form of efficiency, safety, quality performance — those types of dimensions.