Latest / Make It Real / Using AI to Protect Healthcare Systems with Kurt Spear | Highmark

Transcript

Using AI to Protect Healthcare Systems with Kurt Spear | Highmark · Make It Real

• Kurt Spear (00:00): AI is two, three years ago, it was more of a differentiator for companies. And you would hear a lot of different vendors in the market say, "Hey, we have AI embedded throughout." Some did, some didn't, but now it's really a business imperative. It's no longer a differentiator. It's the way that we're going to get things accomplished more quickly, more efficiently, more effectively, and allow our people to focus on the most critical and complex areas. It's going to be their tool to help them to operate at the top of their license, if you will. So it's something that we all need to think about leveraging, but do so in a way that's going to be most effective and efficient. (00:44): You're listening to Make It Real, brought to you by EXL. I'm your host, Shirley Macbeth, and on this show we're exploring how artificial intelligence is reshaping workflows, industries, and the way real work gets done. And yes, we're going to make it real. Shirley Macbeth (01:04): Hi, everyone. Welcome to Make It Real. I'm very excited to introduce today's guest. Please welcome Kurt Spear. He's the vice president of financial investigation and provide a review at Highmark. Kurt brings more than a decade of experience at one of the country's most diversified healthcare organizations where he leads teams that are focused on identifying fraud-based abuses, improving payment accuracy and protecting the integrity of the healthcare system. Welcome, Kurt, and thanks for joining us today. Kurt Spear (01:35): Thank you. Shirley, good afternoon. Appreciate the opportunity to be with you. Shirley Macbeth (01:39): Absolutely. Absolutely. Well, very excited to jump right in. Kurt, to set the stage, can you tell us a bit about your role at Highmark and how that kind of vantage point shapes the way that you see fraud and payment integrity in the broader healthcare system today? Kurt Spear (01:56): Sure. So my role is a little bit unique compared to others in the industry. So I specifically have the special investigations unit or the SIU as well as postpay payment integrity. And so it's a little bit unique, like I said, because I do have both of those functions and I think that allows me to see a unique point of view from the perspective of general errors and claim-related issues as well as to the far right side, the true fraud-based schemes that are so prevalent in today's world and allows us to bring both together. And when we find things like claim errors that may eventually lead to more prevalent schemes or even fraud schemes, allows us to use some of those technologies and methodologies to get in front of those schemes and stop them probably a little bit sooner than maybe we would have in the past. Shirley Macbeth (02:53): Amazing. So you've had this unique vantage point. What shifts maybe have you seen over the last few years that have had the biggest impact on payment integrity and compliance over the last few years? Kurt Spear (03:05): At kind of a macro level, I think it's just the overall cost of healthcare, which is kind of an easy thing to say, but I think as you drill down into it, I always look at it as sort of a four-way tug of war. You have the providers of one end and they might implement something like AI technology to help them with billing and coding or ambient listening systems. And on the other end, you have a payer who's pulling on the rope trying to counter some of those activities that maybe we're increasing the cost to healthcare. On the other side, you have self-funded groups who are oftentimes floating the bill for healthcare. And on the other end of them, you have regulators and CMS and a lot of Medicaid organizations on the state side. And everybody's kind of wrestling right now for that healthcare dollar. (04:01): It's a four-way tug of war for that dollar. And it oftentimes puts those entities or those groups a litle bit at odds, but also creates unique opportunities to work together. And I think that's what I'm most excited about as we go forward. Shirley Macbeth (04:16): You mentioned this four-way tug of war and you can feel the tension in all those different areas. Where is AI in all of the mix? Where is that making a big impact on the tug of war and on the system overall for healthcare? Kurt Spear (04:29): Yeah. Each group of entities is either in a journey with AI or are starting to think about it and everybody's kind of at a different piece of that continuum. But I think that the providers, generally speaking, are probably a little bit further ahead than the payers, for example. I mentioned things like ambient listening systems and revenue cycle AI for billing and coding. Those things have been in place for a while. I think from what I've seen going back to about 2022-ish, we started to see a lot more integration of those technologies on the providers. And from the payer side, it started a couple of years ago, maybe two and a half years ago, using things like machine learning and natural language processing to really make our jobs more efficient and more effective in a lot of different ways. Shirley Macbeth (05:28): Maybe you could say a little bit more on that. You said that the providers may be a little bit ahead of the payers, but what are some of the types of things that AI is helping you spot maybe a little more specific for listeners to understand the types of maybe fraud or abuse or other things that you're able to spot? Kurt Spear (05:45): Sure. So if I think about maybe going all the way from what I consider the left-hand side of claims errors all the way to the far right of a true fraud scheme, I think AI is kind of used in each part of that from what I've seen from things like on the far left-hand side, I mentioned Ammy and listening systems and coding systems using AI where these technologies are allowing providers to now capture member encounters, allow the clinicians to spend more face-to-face time with the patients, spend more time in an exam room versus charting their notes and their medical records. So creating a lot of efficiency I think from that perspective all the way through from a payment integrity perspective, some of the ways that we've leveraged it over the past couple years are things like natural language processing, which allows us to upload a medical record that maybe is a thousand pages in length and really highlight and call out some of the key areas that we need to focus on as part of our review. (06:51): So instead of having a clinician or a coder pulling up a claim and a medical record and holding those things together and trying to go back and forth and figure out, hey, does the record match what's being put on a claim and where do I go into medical record to find that information? Technology's allowing us to do that. So it's allowing us to look at more claims. It's allowing us to perfect our hit rates and make those hit rates higher, meaning looking at fewer claims but finding more errors. And so it's allowing us to be more effective and more efficient. On the far right-hand side all the way to the true fraud schemes, unfortunately, where we're seeing issues there are things like fabricated medical records, fabricated claims, because it's so easy now to use technology that in most cases is available to everyone, it's free or low cost. (07:44): It knows healthcare so the perpetrators don't need to understand healthcare. They don't need to understand billing and coding or clinical care pathways. All of that stuff can come very easily through these solutions and a lot of times they're free. So what we're seeing there is it allows the bad actors to perpetrate fraud more quickly with higher velocity overall and much more sophisticated and done so at a lower cost. Shirley Macbeth (08:13): It's crazy. On the one hand, you talked about allowing the clinicians to spend more time and more easily have that interaction and have it automatically coded and it is amazing. And then on the other side, there's always going to be a bad actor in finding ways to leverage technology in these incredible ways that you wouldn't have thought possible and how sophisticated they've become so quickly. Think about the balance of technology and the benefit versus the risks and you kind of play on both sides of that. How do you think about that moving forward? Kurt Spear (08:46): Yeah, the way I try to think about it is really what's the good and the bad, meaning how is it going to be used against us from the bad actors perpetrating fraud, as well as how can we leverage it for good? So I gave a couple of examples of how we're using it for good, but the thing that I always have in the back of my mind and trying to balance it all out is leveraging the skills and the experience of our people I think is most important because the technology's going to continue to get more robust. It's going to get smarter. I mean, anytime you go into LinkedIn or anything, I mean the volume of AI articles and how sophisticated the technology is, I mean, it changes every day, every hour, which can really help us. But to me, it's about keeping our people really front and center and being able to leverage their knowledge and expertise to figure out how can we use AI to make us more effective, more efficient, but also what are some ways that it could be used in a nefarious way potentially against us and against our members and trying to be able to keep that investigative kind of mindset and the people really at the center and keeping that human at the helm, I think is the way that I really try to focus on it and think about AI as another tool in our toolbox, right? (10:07): Super important one, very sophisticated, can really help us in many different ways. But at the end of the day, I try to think of it as a tool and that we need to arm our people with it, teach them how to use it and think about how it could potentially be used against us as well. Shirley Macbeth (10:24): Exactly. Well, when you think about guardrails for AI, and I think people listening could take what you've learned as far as some core principles that maybe they should be thinking about in healthcare around AI. Do you have any advice about guardrails and how to implement them in your organization? Kurt Spear (10:44): Yeah, I get the question a lot, especially from some of the smaller health plans and thinking about where do we start with AI and how do we implement it and how do we use it? And it's always a great question, especially for an SIU that might have three people in it or four people and one, not having the people power to launch maybe as robust as a department would like, but also from a funding perspective, how do we leverage our dollars in the best way possible? And I always think about it in two ways. One is start with something that you're almost certain you can be successful with. So come up with a use case that you've worked across the enterprise to develop because it's probably going to take people from your IT functions, from your governance teams to make sure you're applying it in an ethical way and according to all your state requirements, et cetera. (11:42): So really leverage the core team of your people across the enterprise and come up with a scenario that you feel really good and confident about. The second is try not to get tied down to one specific technology because again, the technology changes, like we just said so frequently, so quickly that you want to leave yourself open that as things evolve, as your use case changes or expands, that you can continue to apply it to those different technologies. I know the one way that I've always thought about it is from a member encounter perspective. So we sign up for a health plan policy, we go to our doctor, we have a visit, we have follow-up care, we get a prescription, we get a EOB in the mail, we pay our bill. The way that I've thought about it is kind of just laying out all those key activities and what's the way that we might be able to embed AI into it to make things just as efficient and effective as possible and again, add it as another tool to the toolbox to allow your team to be successful. Shirley Macbeth (12:47): Yeah. So you're talking about a journey and thinking where along that journey of a good medical experience or interaction, if you will, where that can speed up and improve the process, improve quality of care and the patient experience. Another topic we've talked about before is around data and so switching gears slightly, but data to have that experience and have that billing experience after the fact and all of these different things, data often boils down to the data and you've talked about data quality and unstructured data as key to this, but it's still a major issue across the industry. Can you comment on the role of data and how that plays into taking advantage of the benefits of AI? Kurt Spear (13:33): Sure. I think we've been talking about data for 25 years and it's still the same. Shirley Macbeth (13:38): Never perfect. Yep. Kurt Spear (13:39): No, how do we figure this out? And it's still the same situation that we faced 20 years ago, which is the data is in a bunch of different systems that use different technology and it's unstructured, like you mentioned. And can we even get to it? Are we getting to the right data, et cetera. But yeah, having access to the data and having it in a way that you can actually leverage it and use it and be comfortable and confident with the results that you're getting is always a challenge. And to be honest, I don't have a silver bullet, I wish I did for how to manage that. From my experience, what's been the most successful way to go about it is, again, leverage the people within your enterprise across a lot of different functions, including your data, your governance teams, your IT team, to really make sure, hey, whatever we leverage for these solutions, it is the most appropriate set of data. (14:37): It's as structured as it can be. It's something that our AI systems can leverage, can learn from and can use. And it's something that's scalable and repeatable. We are going to continue to get data from the same source. It's going to continue to be a source of truth even as we acquire new health systems or continue to expand our businesses, that's going to be the core center of our data universe, if you will, I think is super important. Shirley Macbeth (15:04): You started to talk a litle bit about people in that as well, talking to your teams and the data that's important to do their piece of the function. Let's step back a little bit when you've been implementing AI and continue to be doing it in more sophisticated ways. And healthcare in particular, being so highly regulated, there's a lot of probably built in, we just talked about guardrails, but skepticism and you talked a bit about picking a use piece to be successful. What can we learn from change management and any advice that you have around bringing people along and bringing them from sort of skepticism to implementation and really driving those benefits forward? Kurt Spear (15:46): Yeah, it's interesting. At Highmark, we implemented a tool called Sidekick. I want to say it was probably two years ago now and it's basically our internal ChatGPT type tool and it's secured, it's within the walls of Highmark. We've trained it, et cetera. So we're comfortable with it from a governance perspective, it's given us the right information. But we first rolled it out to my team, I had a lot of people hesitant like, "I don't want to use this. I don't know what it's giving me. Is this really going to save me time?" And now as I walk by my team's laptops, I see it open all the time and they're like, "Oh, I can do this with Sidekick to create this letter or whatever. It's going to be so much faster." I think as people just continue to get more comfortable, they're going to be our best asset of figuring out ways to launch it in other areas. (16:32): To me, that's always the hardest part as owner of a department is saying, "Hey, there's so many capabilities, it has so many benefits and opportunities, but where's the best place to plug it in and drive the most value?" That's something we're thinking about now with Agentic AI is we've built some solutions to really automate some things that require clinical expertise, coding techniques to kind of pull it all together and do so in a sequential order. But to me, it's also the little areas like where can we plug these things in some routine areas to make, again, people more effective, more efficient? And a lot of that is going to come from our people saying, "Hey, I do this manually today and it takes a lot of time, it's prone to error. And if we could automate this in some way and help it make decisions and have our people at the back end approving it, that would be really valuable from a departmental perspective and from an overall enterprise perspective. (17:34): So that's something we're trying to focus on as well. Shirley Macbeth (17:37): We've had past guests on the podcast talk about their AI agents or the automation as suddenly the sidekick that you mentioned maybe as an example, it almost becomes a member of their team, if you will, as an added tool, an added benefit and an added edge. Is that the case for your teams or how would they describe this new capability? Kurt Spear (17:57): Oh, absolutely. Yeah. And if we were to say, Hey, we're going to get rid of this tool today, or some of the other AI technologies like the natural language processing or other solutions we have, it would be mutiny. I think they'd be like, there's no way because we get so much more done through this. But keeping that human at the helm, they're the ones ultimately that are helping to train these systems. And if something doesn't give the right feedback, they're the one that's going to catch it and help train the system to make sure it's better going forward. I mean, these systems aren't perfect. When they start, it's going to have some level of misinformation and it takes the knowledge of our people to say, yep, that's not accurate. We need to change this or add different data or something to make sure it's learning in the appropriate way to give the right outcomes. Shirley Macbeth (18:44): Absolutely. Well, and as you move more and more into Agentic, we've talked about how still human judgment is still a very important piece of that. It went from human in the loop to human on the loop, if you will. And I would imagine that's a big part of where you're seeing the evolution as well sort of upping the importance of that human judgment as things go faster and you're dealing with more claims, et cetera. But talk a little bit more about that human judgment element. Kurt Spear (19:12): Yeah. And again, I think that's where I look at our people as the most important cog in the AI wheel of making sure the system's giving the right output. And one of the things I fear the most is our people just relying on the output and not second guessing it or thinking about it, just taking what it gives you and running with it. And we really try to stress to our teams that you're the expert, this is your tool, it's going to help you, but you have to make sure you're comfortable with the results. And I think on the fraud side, what we're hearing and seeing a lot from the regulators and law enforcement is, hey, you have to be really confident that you can explain how the system came up with an answer because when you get into testify, you're at a deposition, et cetera, somebody's going to ask you, how did you come up with this finding, this observation or conclusion? (20:04): And it can't be, "Hey, the system spit this out and gave it to me. " You've got to be able to back into it and say, "Here's exactly what it looked at and how it came up with that conclusion. And by the way, I can repeat that. " So I've gone in and I looked at information A, & C to make sure that the conclusion it gave me was accurate, but if we can't do that, that's really going to be a problem for some of these fraud cases. You've got to be able to demonstrate that. Shirley Macbeth (20:30): I think one of the interesting things when we've talked before is in no other industry as much as when you and I spoke, did I think about AI as you're really having to be ahead of it, to be thinking about what those bad actors are going to do, you really have to have an offense strategy around that. I think that's just what you said. Yo have to be tracking and understanding and knowing the root of where answers are coming. Talk a little bit about that unique role that you play with having to be more on the offense to outthink the bad actors. Kurt Spear (21:00): Yeah. I mean, to me it's fun. We're always learning something new. There's always some new technology and that's how we try to stay ahead of things, knowing that it's changing so fast and how do you stay on top of it. But I try to tell people to go out and play with these tools and you're not going to be an expert. You don't have to be, but just understand that there's so much unique capabilities of each one of these and some platforms are better than others, but as you play with the different technologies, you're going to learn a little bit more about, hey, this solution can do this better than that. For example, I remember three years ago when AI was in our day and age, fairly new with what it can do and available to the masses. And if you went in and tried to create a hand, it would give you a hand of six or seven fingers and you're like, "Okay, AI curdled that. (21:53): " Now it's perfect. I mean, it's really hard to tell, which is good and bad, but just having that knowledge of knowing how to do things, it's going to help our investigators to understand and think, okay, is this really legitimate? Could this be fabricated? I probably need to go a step further and check, do some old school investigative techniques, go onto Google Maps like, "Hey, does this facility really exist? Is it located there? Can I do more boots on the ground? Do I need to do a drive-by of this facility and see if it's really there?" So it's just having that trust but verify skillset that we've all been taught over the years, probably even more so now with the capabilities of AI. Shirley Macbeth (22:38): Well, Kurt, amazing. Thank you so much. And I wanted to just give a macro view. Maybe if you could see where all of this, give us your advice on where this is all headed with AI and payment integrity, where does that look in two, three, five years? Kurt Spear (22:51): It's a great question. I think health plans need to make sure they have an AI strategy in place, knowing that it's going to have a lot of twists and turns on the road, but knowing at least where you want to get to and where you want to leverage it. I think of AI is two, three years ago, it was more of a differentiator for companies and you would hear a lot of different vendors in the market say, "Hey, we have AI embedded throughout." Some did, some didn't, but now it's really a business imperative. It's no longer a differentiator. It's the way that we're going to get things accomplished more quickly, more efficiently, more effectively, and allow our people to focus on the most critical and complex areas. It's going to be their tool to help them to operate at the top of their license, if you will. (23:41): So it's something that we all need to think about leveraging, but do so in a way that's going to be most effective and efficient. Shirley Macbeth (23:48): Well, that's great. And just thinking back over our conversation, I think there are about three or four things that really stuck out to me. So you'll have to see how I do, Kurt, but you just said that AI is no longer necessarily differentiator. You just need to do it and it has to be part of what you do. And as a way to get started, you said, pick a use case, pick a use case that sets yourself up for success to get started. You've also said don't get too tied to the tech. I mean, the tech changes literally minute by minute, but don't get too tied for the tech, but move forward and experiment and get in there. You also talked about people, a lot about people and you said people are literally the most important part of the AI strategy and the AI journey. (24:28): So bringing people along, having them drive forward, experiment, et cetera, was another key takeaway. And then the final thing that you said was have an offensive as well as defensive AI strategy. And I think all the more important in the world that you live in for payment integrity. So I think this has been an amazing conversation. How did I do for key takeaways? I Kurt Spear (24:50): Think it's perfect. It's perfect. I appreciate the opportunity to be with you today. I look forward to hearing some other podcasts and learning from others, so I appreciate your time. Shirley Macbeth (24:58): Awesome. Well, great conversation. Thank you so much, Kurt, and we appreciate the time today. Thank you. (25:04): Thank you. Thanks for listening to Make It Real. We hope today's conversation gave you ideas, insights, and inspiration to help bring AI to life in your organization. Remember, big ideas don't drive change, people do. Keep learning, keep experimenting, and keep embedding AI where it matters most. Follow along so you never miss an episode.