It’s one of the most overused buzzwords in the industry, but when it comes to understanding what ‘big data’ actually entails, Kirk Kirksey believes there’s a gap the size of Texas. “If you want to see big data, come here,” says the CIO, who shares his thoughts on the ‘holy trinity’ of high-level analytics, the model UT Southwestern is developing for academic medical centers, and why he believes we’ve just scratched the surface of big data. In this interview, Kirksey also discusses what his team is doing to provide IT support for the organization’s three missions — education, research, and patient care; being an early Epic customer; what he believes we can learn from history; and why “culture eats strategy for breakfast.”
- Building an analytics model for academics
- Clearing up misconceptions — “If you want to see big data, come here.”
- The big data holy trinity
- Bench to bedside
- IT’s role in new hospital construction
- Inspiration from the Art of War
- Why IT projects fail
LISTEN NOW USING THE PLAYER BELOW OR CLICK HERE TO SUBSCRIBE TO OUR iTUNES PODCAST FEED
I call it the Holy Trinity, because with big data, you need three things. You need computers to process it, you need mass storage to store it, and you need big network pipes to move it around.
If you look at the plans of the hospital, you can see where the teaching and research missions are sort of instantiated in the physical footprint of the hospital.
Everything in this hospital is going to have an IP address, even the light switches. Everything is going to be on the network. Managing that is going to be a challenge, just in terms of the infrastructure and the speed to ensure that we deliver the information when and where it’s needed.
Once you take ownership and people start working with the traffic flow goes and rehearsing how to get a patient from ER to surgery, then you’re going to want to make changes in your IT systems and your workflow, and you want your IT systems to reflect that.
They continue to fail or be suboptimal at about the same rate as they have been for many, many years. I don’t think that we have cracked the nut. We really haven’t discovered the pathology of IT project failure in my view.
Gamble: Just based on the nature of what your organization does, are you in a fairly advanced place in terms of data analytics and things like that?
Kirksey: I think we’re fairly advanced on the clinical side. We have an initiative right now to expand the data analytics to a more institutional data warehouse. The analytics on the clinical side, I think, are pretty well established in terms of data model capability — a lot of institutions have those. There are some leaders there. What we hope to find for academic medical centers is a data model and an analytics solution that covers the entire institution. That’s what we’re in the process of building now.
If you think about it, you have a set of providers — some physicians and some people like executives who are interested in measuring productivity. On the clinical side, that may be pretty straightforward. How many patients do you see? How many appointments get cancelled? Things like that. But you have the same problem with faculty who do research — measuring productivity in terms of how many publications do you have, how many grants have you won, things like that. It’s a different problem that has expanded out from what traditional healthcare analytics I think is doing.
Gamble: That’s one of the things we’re hearing so much about. One CIO told me that big data is the most annoying buzzword since Meaningful Use, especially since there seem to be some misconceptions about what big data actually does entail. Did you find that to be the case?
Kirksey: Whoever said that is exactly right. I would agree with that. What people think of as big data is not really big data. If you want to see big data, come to places like this, where you’ll see genetic sequencers that will generate 10 terabytes in one run. These genetic sequencers are getting so cheap they’re all over the place. I call it the Holy Trinity, because with big data, you need three things. You need computers to process it, you need mass storage to store it, and you need big network pipes to move it around. If you’re going to manage it, you need really all three of those things in place. We’ve had researchers in the past before we built these pipes literally put big data on big disc drives on a truck and ship them to San Diego or to Austin for processing. It was faster than moving over a traditional network line.
Gamble: It’s interesting.
Kirksey: It’s very interesting to me because it is really a whole new way of managing data and managing information. There are always issues, but on the commodity side that we’ve been doing for the past 30 years, these problems of big data and managing are much, much different and very, very interesting. We’re just sort of touching the surface of it.
Gamble: There’s just a long way to go to get there and to achieve what’s really possible. It’s interesting to me because you’re talking about organizations like yours and then we have community hospitals and there’s no getting around the fact that there is going to be a wide gulf in terms of what organizations can do. It’s not going to be, ‘everybody do big data now.’
Kirksey: Well, no. Our mission has really been bench to bedside. What we want to do with our science here — even starting with basic science that may not have anything to do directly with patients — is to develop technologies, medicines, equipment, and algorithms in the lab and then go through the process and eventually move to the bedside where you first do clinical research to validate the methods, and then turn them in to real treatment. That, to us, is a lot of what big data is about.
Gamble: Yeah. So you talked a little bit about the new hospital that’s set to open in 2014. This is something that seems pretty interesting to me. This is being built from the group up?
Kirksey: Yes, it is.
Gamble: In your career, had you been part of something like that before?
Kirksey: Yes I have, at Perot Systems. When I worked at Zale Lipshy in Dallas, that was literally a hole in the ground or close to a hole in the ground. So I was here from almost the beginning of that. With the new University Hospital, going back to the mission integration, if you look at the plans of the hospital, you can see where the teaching and research missions are sort of instantiated in the physical footprint of the hospital. There are places to teach. There’s going to be a lot of video conferencing — not only between teachers and students, but patients and physicians, research areas, things like that. So really this is going to be a very, very unique facility in that it physically is going to integrate the visions of an academic medical center.
Gamble: That’s very interesting. With something like this, it’s almost like a clean slate. You’re starting from scratch so I can imagine there are a lot of possibilities, but are there challenges as well from that standpoint?
Kirksey: From an IT point of view, we’re entering with faster networks and 4G and LTE. We’re entering an age where we’re going to have an internet of devices and not humans. And so although it’s an exaggeration, I like to say everything in this hospital is going to have an IP address, even the light switches. Everything is going to be on the network. Managing that is going to be a challenge, just in terms of the infrastructure and the speed to ensure that we deliver the information when and where it’s needed, and on time. That’s a new challenge. I think that the delivery of information via cellular signal is a challenge. I think a lot of buildings are in need of these distributed antenna systems inside. You’re building essentially a cell tower so you can deliver a cellular signal to every nook and cranny and you’re not dependent on outside power surrounding your building. That’s important to us. We think that’s going to be a big area of where we can exploit the technology. There are a lot of new things you have to think about and manage with a hospital like this.
Gamble: But then at the same time I’m sure it’s a real positive to be able to, like you said, put in these technologies so that the new building really does reflect what the organization is doing in terms of research and helping to meet those needs.
Kirksey: It’s very exciting. It’s really a lot of fun to come to work. We talked about it a long time; a lot of planning has gone into this. I remember when I first saw the blueprints and saw that there were really going to be spaces to do this — things as mundane as cable trays to hold the weight of the cable that’s going to be the fiber optics that’s going to be needed in this hospital. When you really saw instantiated, you knew it was going to be real. It’s one thing to sit around and talk about it and go to committees, but when you see it in a blueprint, you know it’s going to be real.
Gamble: Do you think it could happen where once this hospital is up and running, you find that the other hospitals want to make more upgrades, just because you have a new hospital that’s build from the ground-up and is using all the newest technologies?
Kirksey: Well, one of our hospitals is going to be a neurosurgery center. One of our big centers of excellence is neurosurgery/neurology, so Zale Lipshy will be turned into that kind of hospital. I’m not sure that we have decided exactly what to do with the other hospital, but we are using Zale Lipshy, which is a relatively new hospital, as sort of the test bed for some of these technologies that are going to go into the new Clements Hospital. Depending on the age of the building and a lot of other factors, it would be very hard to go back and retrofit something that looks like this hospital that we’re building.
Gamble: Maybe it’s just more of a case of envy sometimes, like ‘Wow, the new hospital has X, X and X.’
Kirksey: There’s plenty of need for hospitals that don’t look like this — community hospitals and urban hospitals. This is going to be a very, very specialized and the leading edge facility, but we have missions that may be different from a lot of other places.
Gamble: I’m sure this is a major part of the focus for the next you say 18 months or so.
Kirksey: Yes. In my experience, the IT has got to be in place about nine months before you actually admit a patient, because you rehearse and make sure the geography is in place. We’ve just been able to start walking through the facility very recently. My experience has been that once you take ownership and people start working with the traffic flow goes and rehearsing how to get a patient from ER to surgery, then you’re going to want to make changes in your IT systems and your workflow, and you want your IT systems to reflect that. We’re shooting to have sort of this hospital instantiated in software about nine months before the opening.
Gamble: That’s a really exciting thing. I hope that down the road I can touch base with you and see how everything goes once you get closer to the date.
Kirksey: Sure, we’d love to. It’s a real privilege to get to work on something like this.
Gamble: Okay, I want to switch gears a little bit and talk about your blog. It’s a very interesting thing to me as someone who has a writing background. So first I want to talk about what inspired you to do this — to get your thoughts down, and how the experience has been for you.
Kirksey: The inspiration, if you call it that, came from something that was done in the 80s and 90s. I don’t know if you remember, but in business schools, there was this sort of belief that the Japanese economy was on the rise. So there was a study of these ancient texts and how they related to Japanese business. There was the Legend of the Five Rings, and one of those books was The Art of War by a general named Sun Tzu. He lived about 2500 years ago. So academicians and scholars were dissecting these books, and then the Japanese economy kind of tanked and they figured out there was not much mystical about it at all; there was not a lot of voodoo behind it.
But I got interested in this book, The Art of War. Not so much what it said, but I would read it several times — it’s a little short book with 13 chapters — and what I wanted to figure out, and what sort of struck me, is why would a man go into a cave and write this book? I think he wrote this book at the end of his career and it was sort of a collection of his insights. I’ve been in this a long time. I’m on the south side of my career. And so I started a blog — it’s strictly a personal thing — and it was an exercise to see if I had learned any insights in my career in practicing my craft. My rule on the blog is I don’t look up any data. I don’t back up anything with any data, which is heresy, and I try to keep each entry to less than 300 words, which I don’t do all the time. It’s really an attempt by me to see if I’ve really learned any original insight or if I had any original insights that are of any value as a result of doing this for about 30 years. That’s about as complicated as it gets, so it’s pretty straightforward, I think.
Gamble: I like that concept. I know that one of the blogs focuses on why IT projects fail. This subject attracts so much interest. We’ve got conferences, events, blogs, everything — there’s so much talk about this, and we haven’t necessarily come up with the solution yet.
Kirksey: No, they continue to fail. If you believe Gartner and Forrester and your own experience, they continue to fail or be suboptimal at about the same rate as they have been for many, many years. I don’t think that we have cracked the nut. As I like to say, we really haven’t discovered the pathology of IT project failure in my view.
Gamble: But it will be continued to be dissected in the same way. I think that that’s one of the concepts that you touched on as far as looking at the pathology of a problem.
Kirksey: In my arrogance, I guess, I have my own opinions about why projects fail. In my view, it doesn’t have anything to do with sort of the common talk that people usually talk about — your PMO methodology. I think there are much deeper causes of failure than what most people look at today.