This Just In Radio: AI in Action with Chris Paravate

Speaker A: Welcome to this Justin the show, bringing you the latest advancements in healthcare strategy, innovation and public policy. And now for the fastest voice in healthcare, here's your host, Justin Barnes.

Speaker B: Thank you for tuning in. Welcome to this just in. I'm your host, Justin Barnes. In these episodes, I, uh, bring you the latest advancements in healthcare strategy, innovation and leadership. As always, we're broadcasting from the this Justin studios on the Business Radio X network, the Healthcare Now Radio network, the Global Health Podcast network, and our newest broadcast partner, the Health Podcast Library. For this episode I have something special and not just because of the guest. This is episode 400. 400 episodes, 400 conversations with some of the most forward thinking people in healthcare. Leaders who are building, deciding and transforming people who aren't waiting to see what happens next. They're making it happen. When I started this show over 12 years ago, I had a simple belief. The best ideas in healthcare don't live in white papers or research. They live in the heads of people doing the hard work every day. My job is to find them, get them on the air and give them a voice. 400 episodes later, I still believe that. So for this milestone, I wanted to

Speaker C: bring back someone who's been a great

Speaker B: friend of this show through the years. Someone who has earned serious credibility in this industry. Not from a stage, but uh, from the floor of a health system where the stakes are real. I'm talking about Chris Paravate, Chief Digital and Information Officer at Northeast Georgia Health System. Chris has been at ah, Northeast Georgia for over a decade. He has lived through every major technology wave to hit healthcare. EHRs, value based care, interoperability, telehealth, Population health, the cloud, and now AI. He didn't just survive those waves, he helped position his organization to navigate them with purpose. He led the epic Go Live at Northeast Georgia. He built out their first community connect partnership and he holds both the CDO and the CIO hat simultaneously, which is not common. That combination matters because it means Chris is accountable for both the strategic vision and the day to day operational reality. That is a rare kind of leader and exactly the kind of voice you need when the conversation gets real. And it always does, certainly with Chris and I, and it did here. I recently had Chris join me for one of my AI webinars. This session was recorded a few weeks ago and the talk was titled beyond the AI Strategy. What healthcare leaders should do in the next 90 days. And that title was intentional. We are past the point of where having an AI strategy is the achievement. Every major health system has a strategy Deck. Most of them are beautiful. Most of them are sitting in a SharePoint folder. The question is, what are you actually doing and how are you pivoting from pilot to production? How are you scaling? What best practices did you learn? What best practices can you share? Chris and I get into all of that. Why this moment is structurally different from every past technology cycle. How, huh? Epic centric environments are both an asset and a potential blind spot when it comes to net new AI deployment, what real AI governance looks like versus a committee with a name and no teeth, how to handle the workforce anxiety that comes with every major technology shift, and what the two or three concrete moves are that a cio, a cdo, or a CDIO should make right now, this week, before writing any other strategy documents. I'll warn you, this conversation is dense. Chris is not dealing generalities. He deals in what's actually been done, what has worked and what has not. If you are a health system leader trying to figure out where to focus your AI energy right now, this episode is going to give you a framework you can act on. If you are in the vendor space

Speaker C: trying to understand what your health system

Speaker B: customers are actually thinking about, this is required listening. And if you've been a listener for a while, I think episode 400 is the right moment to share this conversation as reflects exactly why I started this show in the first place 12 years ago. Real talk, real leaders, real strategy, real best practices. Without further ado, Chris Paravate from the Northeast Georgia Health System.

Speaker C: Hey, uh, Chris, good to see you again, my good friend.

Speaker B: Thank you.

Speaker D: Yeah, absolutely. Thanks for having me. It's good to see you as well.

Speaker C: You got it. So we'll dive right in and I'm going to cover some of your background. You're a very prolific CIO in the nation. But again, I'm going to pull out a lot of these great accolades and strategies that you, uh, have deployed at the Northeast Georgia Health System. But, uh, you've been doing this over a decade, um, and you've really seen every technology wave come through. How is this AI moment different than past cycles? And why do you think so many health systems have great AI strategies that haven't translated into operational change?

Speaker B: Wow.

Speaker D: You know, I've been in it for 30 years, and I've never seen a piece of technology like this. Uh, and our clinicians and our physicians feel it as well. You know, our cmio, for example, he says, I feel like for the first time, I'm giving my peers things that make their life better, instead of making giving them Things that make their life harder. And uh, I think that's a huge trans transformation. I think that the promise is still there. Right. I mean we, we see this promise AI every day in our lives. I'm planning a anniversary, uh, trip to Ireland and Scotland and I, you know, put together a plan in 30 minutes with, with uh, you know, a tool that's available on the Internet. Right. And, and it's a good plan. And then I can take that plan and I can say what's wrong with it and tell me all the things that are wrong with it.

Speaker B: Mode.

Speaker D: And, and so we see this promise of technology. And so early on what I think we're seeing is there are things that make huge transformational differences. Right. We've seen that in ambient voice recognition. Mhm. And then there's smaller things that remove burden. And I was having uh, dinner with some colleagues last night and I brought this up, I said, well, I can't tell you what the return on investment is to use a generative chart summary for a patient that's been here for 200 days. But I can tell you if you're the hospitalist who's coming out on the first day of a seven day shift and you have to write the discharge summary for that patient who's been here for 200 years, that is a priceless solution. It actually is improving uh, quality and clinical delivery. And I'll tell you in two ways. One is that that summary is far more comprehensive than if that ah, hospitalist had sat there for four hours and wrote it. Mhm. That's number one. Number two is that I now freed up three and a half hours for that hospitalist to attend to patients that actually need the attention of that acute care. And so I think that there are examples like that that we're removing burden. Uh, harder to prove. Roi.

Speaker B: Right.

Speaker D: But I can tell you really, really tangible examples. And, and I think that's, that's kind of the cusp of where we are. And I think we see all this excitement because we know that there's this promise and we also know from let's say the dot com for some of us who are old enough to remember that. Oh yeah, um, is that, you know, now there's a generation that walks around and never experienced what it was like not to have Internet or to use a yellow pages or um, use a corded phone. Right. Or actually have to wait for somebody else to get off the phone so you could make a phone call. I think what we know about this AI, um, and is changing at such a rapid pace, uh, is that the future is going to be so much more promising. Uh, and that's I think, why you see this strategy is far more of uh, here's the things we'd like to do and these are the things we think are important to do to how do you get it done. And I think, ah, that's the, the tempo, the promise, uh, or the opportunity needs to catch up with the reality of delivery still.

Speaker B: Yeah.

Speaker C: And that's why I love to have you as a guest, because you do, you hit the nail on the head there. Not only you're improving much, you know, greater efficiencies, but also access to care and um, also more comprehensive care. And so that's better care and it's compassionate.

Speaker D: I mean if you sat every, I bet you every one of these listeners have sat in a hospital, um, you know, taking care of somebody and what are you doing? You're waiting. And what is waiting? Waiting is like the most painful experience. And I, I would even argue in an outpatient setting, waiting tells me that your time is more important than, than mine and therefore you should wait for me. Right. I, I think waiting rooms are the ultimate patient insult. Um, so I just, my little bit

Speaker C: of crazy love that, love that. So the 90 day timeframe, we actually

Speaker B: made this title to be a bit

Speaker C: more provocative beyond the AI strategy.

Speaker B: What does that mean to you?

Speaker C: And why 90 days is there a window right now that health system leaders need to act within or risk falling

Speaker B: behind in a meaningful way?

Speaker D: Yeah, I think, um, this is coming, right? And I don't care what EMR platform you're on, uh, there's pieces and parts that are coming every single day. And there are, you know, one, I think in the next 90 days you need to really decide. I'm going to really invest some disciplined time in understanding the capabilities. And I'll tell you why one is how ah, do you govern it? And we're going to talk a minute about governance, but also what are these tools doing with your data, how informed they are. And I'll take it back to a practical example. I said, hey, I want to go to Ireland and Scotland and I have 12 days. And it developed a 12 day plan. And when I asked that second question of what's wrong with this plan, it says, well, day four is too compressed and if you had one more day, it would make it far more relaxing and, and break up your travel. I didn't ask that in the first question and I didn't get that answer. So it's all in how we're prompting and answering these questions. And I think that's, this is the uh, the education part is so important is for decades it has gone out and collected requirements. Well, what do you need or what's your problem and how do we solve that and what would you like it to look like? And now at Northeast Georgia we, we are starting and it's still in its infancy stage but resident developers. Our coo, uh, is one of our resident developers. He has a task from the CEO to track some improvement and some savings. He is using a tool that we provided that he built an app, uh, that the entire senior team is tracking what's going on. Now here's this resident application builder that's building largely HTML, which is not sustainable. And we all know that if we're technical. But all the requirements gathering, he's like, this is awesome. This is exactly what I wanted. This is perfect. Guess what? Oh, this is missing here. Let me fix that. And then we work behind the scenes and we build the database integrity and the security and the controls in place. And this is where governance comes important is.

Speaker B: Hm.

Speaker D: We think about governance historically about either prioritization or uh, efficacy or uh, you know, managing risks or issues or challenges. The governance in this domain is also looking at how are you controlling the environment and the tools you're introducing and do you know what those tools are doing? Just like we have employees and we have policies and we have m, you know, things to protect the organization, we need to build that for our AI tools. And so I would say in the next 90 days, the second aspect is really governance. And governance not in a traditional way that you think about that. So one aspect of governance is similar to what we do with cyber security, right? We have a, um, security operations center or soc. We're looking at all of our AI, uh, our cyber security tools and one pane of glass and we know what's going on. Well, we need to have the same thing for our AI, uh, we need a single pane of glass, one visualization to really understand what's going on and to take interventions. Right. And that's the technical side of this that we really still are very much in the infancy stage on.

Speaker C: Yeah, I mean on the governance side, this came up at ah, one of the conferences recently. We're hosting a panel there and one of the speakers who does manage the AI for the organization said, you know, you have to get your hands on governance much sooner than you realize because you're going to have these AI, um deployments and you might have one to start off with, and you may have one or two or three. Well then next thing you know, you're going to have five different AI governance models based on those five plays that you have or strategies going on. You need to have one overarching one because you can't because soon you're have

Speaker B: 15 and then you have 15 different

Speaker C: governance models and that's not going to work. So I mean, what do you think to that and how would you manage that?

Speaker D: I agree. I think, um, ah, what we have aligned, at least at Northeast Georgia is our, our IT Executive Steering committee has responsibility for AI governance. Now we have a methodology or a strategy that we've applied that says, um, we're going to be an early adopter. We're probably not going to be a development shop with some limited use cases. Uh, we're going to uh, work with established partners, um, that have a strong security posture. And then we're going to put controls in place to measure not only are these tools performing, um, but are they doing it in a controlled, safe manner. And then when they're no longer performing, take them out, get them out of the environment. And we're starting to do that. As we've implemented some stuff, we've turned a few things off that we said, gosh, we thought that was going to be valuable. Maybe not as much as we thought, let's take it out of the environment. Um, and you'll know because just like at home when your, your kids are on Chat GPT doing something, of course my kids are a little bit older and they were picking house colors. You realize if they don't use it, it's not valuable. And so you can look at utilization very quickly, um, and, and discover that, that it needs to come out.

Speaker B: Well, talk about picking house colors.

Speaker C: A little off topic, but I, I was building a house up off Lake Lanier about 15 years ago and I picked pink by accident. It was supposed to be like a beige. But I wish I had Chat GPT to help me at the time. I had to repaint the whole house. It's a brand new house. And my builder called me and said, your house is pink. And I'm like no, and 10 miles

Speaker B: from where you are right now.

Speaker C: So anyhow, I digress, but I wish I had Chat DBT back then. Um, real quickly though, you carry the title Chief Digital um, and Chief Information Officer hat you know, which is not, you know, universal in most settings. How does holding both roles shape the way you approach? We talk about governance, but Also investment decisions in general. You know, what does that mean organizationally for you guys?

Speaker D: Yeah, for us, you know, my role is really making sure that we've got the, uh, digital tools and architecture for delivery and we can really provide continuity of care, um, that has a direct impact on patient experience. And so we have three big overarching goals, um, that we're trying to focus on. One is to be the best run health system in Georgia. The second is to be the best place to, to get care. And the third is the best place to give care. Right. And so as you think about that, it touches every aspect of our universe. Right. And digital as well. And so I think it's a more, uh, comprehensive role that really kind of expresses that we are a digital company and we are driving, um, our business solutions. Our durable value is our patient care. Our patients have different needs and there is a digital need in that domain. And so that then cascades through our whole organization and the way that we deliver.

Speaker C: That's excellent. We see that. Um, so kind of on the workforce side now this comes up on top. It came up earlier today, even on this webinar, um, AI and healthcare always, you know, services, concerns about workforce displacement from clinicians to IT teams from everyone in between.

Speaker B: How are you handling the culture and

Speaker C: the change inside of Northeast Georgia health system? What advice do you have for leaders who are having the resistance?

Speaker D: You know, I would tell you, one is look at where some of the software companies are going. And I've had some conversations with some leaders on the software side and often they're saying, you know, it's not that I don't have the money to hire another person, but I'm challenging my team to build agents that report to them. I thought, wow, that's kind of, that's kind of interesting, right? So then I was thinking somewhere last night, um, you know, we do a lot of build in our emr. We do departments and M, we change units and we open new clinics. This, it's repetitive work, but it's unique with what we're building. And I'm like, and I thought there's an individual who works for us that predominantly leads at work. And so the conversation I'm going to have with him is, I want you to hire three agents. I want you to build three agents. And he's going to look at me probably a little crazy, but that's where I want to start is just how do I then change his mindset to say, gosh, I don't need another employee, I need an agent. And I need to maybe have a team of agents. I would predict that in the next two years, it's. It's incredibly plausible that 20% of my workforce will be agencic. And that's, that's. I. I know that's, that sounds crazy, but, um, it's not. It's. It's not. And, and we're not only that. We're seeing AI tools start to identify connections and data that are not existing today. Right. So, for example, um, we had a, ah, pa. We were looking at referral data, and we said, how many referrals, uh, did this primary care physician send to this cardiology clinic? It was like four. And then we said, well, how many patients on this primary care panel was seen by a cardiologist in this clinic? It was like 10. It's like, oh, look how, uh, this attribution and this assignment, I think that speaks to how the AI tools will start to evolve into, starting to make these connections very much the way that we do when we're analyzing work.

Speaker B: I love it.

Speaker C: I love it. So this kind of brings me to my last question here. I know we have, um, a hard stop at 4pm Eastern, so kind of measuring success, and you kind of went into it a little bit there. Um, a year from now, how will you know if your AI investments actually worked? What's the metric? Is it, is it the number that you just mentioned or what's the moment that tells you that Northeast Georgia Health System made the right calls?

Speaker D: Yeah, I think, um, you know, I put them in two buckets. Right. I, I think of, um, removing burden. Right. And, and maybe I don't need to have it hard or I on that. And, and maybe the cost is not tremendous. And then there are those things that are transformational. You know, ambient is a good example. I could tell you physicians are happier, patients are happier, we're seeing more patients. Reimbursement is positive. That's, uh, transformational. Right. And so that is going to have far more structure into. How are we capturing the roi? Do we know what the implementation measures are? What are the metrics? How are we measuring those things? All of those things will be in place and it'll go through a higher level of rigor. Uh, and so we bucket those kind of in those two areas, uh, and those things that remove burden, I have a lot of latitude to just say, okay, we're just going to do it. We're just going to turn it on and we're going to see what it produces. And then we still capture an ROI on it or try to capture what that benefit is. But in some cases there's not a direct attribution. I can't tell you, uh, an example of the hospitalist that they saw five more patients or that their care was, was better. But I know intuitively that that's the case.

Speaker A: Mhm.

Speaker D: It's not. I'm not going to go create a science project to try to prove it.

Speaker A: Right.

Speaker B: No.

Speaker C: Excellent. Um, take one question here from the audience. I see, you know, are you seeing a push from patients for faster adoption of AI in their care digital tools that they interact with or conversely being met with resistance or fear from the unknown, from the patient perspective?

Speaker D: So, um, I think you're seeing adoption in AI and some interesting use cases. Um, for example, we have a goal to, um, most health systems have a goal to follow up on patients after they're discharged, um, creating a AI bot that could call that patient and ask them a few questions and then connect them to a care coordinator. I think we're seeing faster adoption of that and patients being super responsive to that and feeling connected.

Speaker E: Absolutely.

Speaker D: Like, you know, you, you close the loop. I think there's other aspects of AI that the patients don't really see. Uh, you know, patients. My wife's taught me more about patient care than anybody. You know, my wife said, um, you know, I, you got a real problem over there. And I go, what do you mean? She says, well, I can't get an appointment. That's quality issue.

Speaker E: Yes.

Speaker D: Well, you know what, that's an interesting perspective. Right. But um, your patients don't expect you're going to hurt them. They don't expect you're going to, you're going to, you're going to, you're going to heal them, you're going to make them feel better, um, you're going to help them live a healthy life. They just expect that. Right. So if you're using AI to improve those things, I don't know that patients are going to be wowed by that as much as they're going to expect that that clinical and care evidence is incorporated into your flow. And I think, um, you know, those things will, as those things mature into the environment, you'll see those pop. Right? You'll see them in quality scores or in mortality, uh, rates or in medication utilization. Uh, those things that are more, um, of that, uh, traditional quality as we think of quality, but not necessarily quality, is what consumers consider quality.

Speaker E: Right.

Speaker C: Excellent insights, A lot of wisdom there, Chris. I appreciate your time, my friend, and thank you to everybody for joining us today.

Speaker E: Please tune in as this show broadcasts weekdays at 2:30pm Eastern, 11:30am Pacific. As always, you can track me on LinkedIn @justintbarns or track me on Twitter @hitadvisor and use the hashtag this Justin Radios and respond to your comments from the show. If you missed any of this episode or want to hear more, all my shows are posted at YouTube, Spotify, Apple, iTunes, Amazon Music, Megaphone, SoundCloud, iHeartRadio and TuneIn and you check out the latest content and thought leadership we just Posted@Justin Barnes.com Appreciate everybody's time. Thank you so much. Have great rest your day.

Speaker D: Sam.