Building the Data Foundation for Healthcare AI with Jeff Thomas, SVP and Chief Technology Officer at Sentara Health

(upbeat music) - Hello everyone and welcome back to the Beat Podcast recorded live here at Vive in Los Angeles. Today I have the privilege of hosting Jeff Thomas. Jeff is a senior vice president and chief technology officer at Centara Health. Jeff, welcome to the podcast. - Thank you, it's a great pleasure to be with you. It's my pleasure and our pleasure to have you on. How's the conference going for you so far? - It's busy, you take about four or five days worth of meetings and discussions and push it into about three. So it's maybe two and a half. So it's a lot of discussions. And it's interesting is this is my fourth year coming to Vive and I found it really helpful in talking and looking really for some kind of the start up. Some of the newer ideas, people who are looking at the way health care is done differently. And that just means how do we make things different and better for how we deliver the same service or better services. And so it's really, it's a great opportunity to kind of talk with our colleagues, but also really look at some innovative ways of doing things. - Totally, yeah. And Jeff, you guys sit in a very unique position. You're a vertically integrated provider payer. Help the audience understand more about your organization and really how you're using digital health and technology today with that. - Well, it's interesting. It's in Terra Health. We represent the large communities both within Virginia and North Carolina. And that focus really is into making sure that, how to give care, better care every day. It's really delivering it for the individual. And so it doesn't only do us any good to build great technology solutions if it doesn't deliver for that individual user. And so as we look at what we're doing is now, how do we change the way that we come to our consumers? How do we show up for them? And how do we enable them to get the best care? And that means there's really kind of layers there. It's not just great. Do we have the brightest, shiniest object for that front end? Do we use a chatbot and all those things? Those are great concepts, but if it doesn't really drive value and help them get to their care quicker, it's just a lot of noise. And I think we've all experienced that kind of frustration when we interact with something that's not a person even with a person. If they don't have the right information to do that communication, be it an AI chatbot, be it a person on a call, be it a nurse. If they don't have that or a physician, they don't have access to that data that they need right then. It diminishes our interactions and makes us feel like they can't be present because they're searching for the data. And that means that you don't get the care that you need. So to me, I think one of the biggest things we're looking at is how do we create presence for clinicians, for those people who interact with Centera Health Care to get the best care? So how can they be as present as possible because they have the best data, the right data, to help that consumer get to that outcome of, getting the care they need to make their lives better? - Yeah, no, that's really well said, Jeff. And as you think about all the changes that are happening, the pressures that health systems are face-waste, what's top of mind? What's like the number one thing on your agenda? - Oh, I think the number one thing on my agenda as a CTO is how do I deliver those platforms, those structures that provide the strategic advantage we need as a company, but also how do we deliver that care? How do I reduce the cost of what do we do? Every time we turn around, those costs keep going up. And so part of the responsibility I have in my role currently is really how do I make us as an operationally efficient and cost-effective going forward? There's a push to go to cloud, there's a push to go to AI agents. There's a push to change how we interact. And if you look at that, we just keep adding more and more costs on the delivery of that care. Now the question is, what are we doing to reduce those costs but still deliver that care or better care? And so I spend a lot of time looking at solutioning, looking at the way we deliver care from a business standpoint and then say, how are we building those foundations? How are we building those things necessary to get the greatest value out of what we invest? So it's not just enough to say, hey, isn't it great? I've got this new AI agent. I've got more AI coming in than I can shake a stick at. But that AI has to be efficient, effective, and it has to meet the outcomes because if it doesn't meet our outcomes, if it doesn't give us great efficiency, if it doesn't change that, it doesn't solve anything. And sometimes that means we actually have to change how we operate to provide those better outcomes because in technology, we all have a lot of tech debt. And part of my responsibility is to remove the tech debt and to provide the opportunities to really get and use the AI agents in a cost-effective way that gets those outcomes. >> That's great. And while at the same time, as you mentioned, not impacting patient care or even the provider experience, that's where the challenge comes in. That's where it becomes challenging. >> It is. And especially with the pressures of the big beautiful bill and other things that are coming in, we're being asked to cut and become a lot more efficient. And so efficiency means I have to make the right decisions on technologies. If I'm in public cloud, how do I reduce those costs with public cloud? Do I look at innovative solutions that deal with reducing those platform costs? So I use products. I look for those kind of products in my world. Can I leverage them fully? Can I build off of them? And do they really have that economies of scale I need to really utilize things like public cloud, to utilize things like agent to AI? How can I do that in a cost-effective manner? And if I'm not looking at that, if I don't run it like a business, the IT side, there's no way I can deliver the right solutions that help the business to run their business better. >> Yeah, I love that, Jeff. And I love your focus on, at the end of the day, we're delivering care. Do you mean how do we do that better? Reducing tech debt, number one item on Gartner's agenda for tech leaders. Let's get rid of it. AI is also up there. So talk to us about your target areas for AI. And what adjustments are you making to deliver the desired results? >> Well, when you talk about AI, we keep forgetting about the data. Because just like any other AI, the basis of it is I'm going to take a whole bunch of data. I'm going to connect it together in interesting ways and I'm going to tell you why it's important to you. Why is it more relevant? Well, the problem is, is that data is locked? That data is not accessible. And when we start putting the strain as we're going to say, "Hey, we're going to go and run all these "agentic workloads against it." How do I get that data to be usable? How do I get that data to be accessible the way we want to with the real time? So if anybody uses things like chat GPT or Clawed in that, they'll notice that there's a number associated with them 5.2, 5.2. Well, they don't tell you as each of those are actually using a older version of data that they're having to update those models to connect new data. So every time that next version, they basically have new data in there that they've pulled in. Well, when we start talking about how we can help care work with it, we've got a whole lot of data, but it doesn't do us any good to look at data that's three years old, that's a year old, that's maybe even a day old. Some of the stuff we may need to do to get those interactions for AI is we need to get that data in a way that's usable, that inference. And so we talk about inference. It's how do we have our structure or infrastructure, our data layer in a way that's accessible for an inference. So all of a sudden, now you start saying, okay, I'm connecting back to legacy systems, older ways of doing things sometimes, they may not handle that new way of connecting. And so how do I present out that data? I use tools, there's tools that I use like silk in the public space that really helps me in that way. It allows me to do AI inferencing and present that data into my data warehouses into my, so my agents can see it and it doesn't impact my production. So I'm able to see that real time view of my data in a way that I can then start using agent AI in really more meaningful ways. And so how do I do that in a cost effective manner without creating tons of data vision. I've got data sitting here and data here and I'm going to go make replicated copies that I'm now having to manage at high costs, which is a problem in public cloud. So how do I do that? That's one of the key things that we do is, is we look at how do we utilize the tools we have and then how do I look at other tools? How do I use tools that I've bought like Databricks or any of those others and make them really cost effective. And so it's not just that I have new technology, I implement interesting things that's then how do I financially manage it? So the thin-up side of this is huge. So how do I keep those costs down while I provide really innovative solutions? And if I can't do that, then I'm probably in the wrong business. And I'll say Jeff, and you know, you have sort of the two sides to think about. You're thinking about the provider business. this, but you're also thinking about the payer business. Talk to us about that. You know, I'm always curious because you know, one job is there, but you have both sides to look at. It's interesting and this is the interesting thing about health care, both payer and provider, the technology underpinnings are not that different. Now, they have some different systems, but really what you're looking at doing is bringing those two together in the sense that I can connect the data. I can get that data so that they everybody can see them and then we can utilize them appropriately within their appropriate regulations. We have that means you still have the segregation of the data. So you're not breaking hippo, you're not violating any of that customer confidentiality, the PHI. You keep those separate, but you also provide a platform that does that. And so really, I think, and this is one of the things about moving into health care. And this is coming from the outside. The complexity is getting rid of the tech debt. The complexity is not building the new capability. The new capabilities are going to really be driven by getting to the right solution set and start thinking about that data as an asset. And how do I expose that asset to be utilized correctly? Addition through subtraction. That's right. And why? Because we all know from from basic things, the more variance, it's more costly. If I can reduce my variance, then I'm going to be able to exploit the value of what I have a lot better. I've been hearing a lot about the value of archiving and decommissioning to to achieve that. Yeah. Oh, yeah. And so it's so when I came to St. Tara eight years ago, we had 13, 1500, 1500 systems. We're running between about five and 600. And so that kind of rationalization, that's eight years. And so we are. So we sit, we sit about 85, 90% in cloud. There's a few, there's a few areas that I'm still looking to get something that works really well for some of the imaging space is still kind of behind in their tech, big, big technology that they still haven't made that leap. But we're giving these great economies of scale in cloud in our, in for all of our other systems. And so we're able to reduce costs. We're able to operate. And so with a good practice of architecture, of Finops and doing that full life cycle and managing out what we're building, we're able to kind of keep those infrastructure costs on a flat and actually get economies of scales and we reduce actual daily costs while we grow. So we can flatline it. And that means that we have the ability to really innovate because now it got capacity and have got money to go innovate against. Jeff, I love that. I love your approach. Congrats on the, on the huge progress you've made doing the team of made there. You know, as the CTO, what key PIs do you use to measure your success? As a CTO, a lot of my success is how do I deliver every day to the customer? So we run, you know, we deal with the end users. We deal with how do we help the clinician to do their job better? How do I simplify how I do it in a cost effective way? Do we make those rationalized decisions that may not be initially very popular, but really give us the foundation to move forward? So, you know, eight years ago when I came to Centaria, the whole purpose was to go to cloud. Everybody's like, well, why do you want to go to cloud? And then 2020 hit. And because we were already 60% in cloud at that point, I was able to pivot my organization to allow users to work from home very ineffective, very cost-effectively. Let them use their own desktop for a while, route things through it. And they were able to function very quickly through that pandemic. And then as we start moving through that, it's the same thing that we're doing with now. Our costs are going down. Our costs are, our costs need to go down. We have to operate more effectively, but we have to deliver more technology, more capability. So my measure is how effectively do I deliver services at a lower cost? So when I think about it as a business, what do those things? So as we're growing, every, every healthcare out there has probably 20 to 30 million dollars of new technology growth they're trying to do, which comes to that bottom line. The question is, can I take out a portion of that and reduce that infrastructure cost so that we can add more to that growth? Well, you're certainly touching on a lot of the hot topics that leaders are here to solve. Jeff, what advice would you close us with on how leaders should be tackling these technology challenges, but also making the most out of the opportunities? I think as technology leaders, we miss the change management piece within our own organizations to make technology effective. So we are great to go get a vendor. We're great to build something new to drive to what is going to be that future land that we all want to be in, but we never talk about what the changes we have to make in our organizations, that change management both with our users and with our internal staff to really get there. And so technology is moving quickly. A lot of people have a lot of fear around it. And part of our responsibility on the technology as a CTO, as a CIO is to help the organization get through that change and do it in a proactive way, in a productive way and not just do it out of fear. And so really when we talk about it, all the technology in the world will not solve those things if we don't get the people to change and we don't help them through that change. That means we're standing beside them as they move forward. What a great way to close Jeff was with that note. And if people want to reach out to you and the team at St. Tyra to learn more about what you're up to, where can they, where can they grind out? We, my personal contact, LinkedIn is probably one of the best ways to get to us, but reach out. We love to talk about what we're doing. There's a lot of opportunities out there and what we're doing and we're loving to share it. And just not just the technology, but how to operate, how to think about things differently. The leadership piece. That's right. That's fantastic. Well, ladies and gentlemen, thanks for tuning into the beat podcast here at five in Los Angeles with Jeff Thomas, Senior Vice President and Chief Technology Officer of St. Tyra Help. Thank you all for tuning in. Check out the show notes where you'll see shortcuts on what we discussed today with Jeff, as well as the best ways to get in touch with him to learn more. Jeff, thanks for being with us. Thank you. [Music]