HAL Transcript: Jim Sinkoff on Who Owns Patient Data

This is the full transcript of the HAL: Who Signed Off On This? episode with Jim Sinkoff, Deputy CEO and CFO of Sun River Health. Listen on Libsyn. The show is also on Spotify and YouTube. For the highlights, read Who Owns the Data When AI Touches Patient Records? A Health System CFO on AI Governance.

The transcript follows the published episode. It is lightly edited for readability: repeated words, false starts and short acknowledgements are removed, and names that the transcription software misheard are corrected. Timestamps mark where each turn begins in the episode.

Transcript

Chris Sowa (00:00): Hey everybody, this is HAL: Who Signed Off on This? And I’m Chris Sowa, your host. I’m the CCO over at RedactSure. And today I am joined by Jim Sinkoff, who is the CEO at Sun River Health, and he’s also an in demand AI expert at conferences and somebody that his colleagues turn to for advice on that topic. As you know, this show is about people who make the big decisions in their organizations. And whether that pen is a virtual pen that’s signing off or a digital pen, people like Jim are the people that we bring onto this show. And Jim, thank you very much for joining us today. We very much appreciate your expertise and time. I’d like to give you a minute to introduce yourself, tell us about what you’re what you’ve been up to and a little bit of your background. And then we’ll dive right into how.

Jim Sinkoff (00:57): Great. Thank you, Chris. I’m looking forward to this conversation. I run a very large outpatient primary care system, about 240,000 patients. I’m part of a larger national system of outpatient community-based clinics that serve about 35 million Americans. I also run a for-profit company that Deploys and uses AI as services to a variety of community-based clinics across the country. So I’ve had the opportunity to be sort of in two worlds, both as a seller of solutions and a buyer of solutions. When I first started my own personal learning around AI, I was rather dubious. Brought my I turned myself to school. I went to MIT for a year, prompted by being asked to do a national have a national conversation and lecture on all things AI. And I was quite dreading doing that because I felt I really was not the right person to do so. Turns out that was a pivotal moment in my comprehension and learning about AI. Obviously there’s much more to learn. So really looking forward to dialogue.

Chris Sowa (02:18): And that’s great, Jim. I mean a lot of people, you’re right, are afraid to talk about AI. It’s critical that more people like yourself dive into it so that we can have the conversations about this that we need to have, right? So absolutely very important. Now you may know there’s a couple of questions I like to ask guests before we get into the formal part of the show. Is we and we understand not everybody has a perfect answer to all these questions, but it’s interesting, you know, in your role in charge of such a large organization, if an AI moved $100,000 out of the organization tomorrow, who would you say was responsible?

Jim Sinkoff (02:56): So from a cash management perspective, just in terms of my environment, we deploy no AI tools agents in anything in our financial world. Nothing to do with accounting, procurement. We have digital solutions, but we don’t deploy any AI solutions. But if in fact we got a hack and an AI hack, and $100,000 was moved. The folks that would be directly responsible would be the controller, the senior vice president of finance, the CFO, and our banking partner would be the folks that would be directly responsible for that. The other individuals that would be indirectly responsible for that would be our chief information officer, who has built what we hope will continue to be a very strong system of denying penetrations. So that would be a second layer, an individual that would be very involved in diagnosing how this happened.

Chris Sowa (04:07): Love that answer because what you’re saying is it’s not just the IT guy, right? It’s really the people who are in charge of that subject matter. Right. So if it’s the person who’s the controller, it’s the people who are actually in jaw in ch in charge of those functions. And I’m not sure everybody has is evolved to that level of thinking yet, Jim, but you know, I think that’s really interesting. As a CEO, that’s kind of a that was what appeared in your mind as the list of people that you were going to call and talk to about it first. Next question is about your staff in general, you know, what AI can read off of their screens right now. You know, understanding that surveys have come out that talk about as much as 58% of staff leverage their own AIs at work, whether they’re official or not. So whether they’re logging in through a browser tab or using their phone, you know some staff are doing it, right? At most organizations. So do you do you understand based on talking to your staff right now, what can AI actually read off of their screens?

Jim Sinkoff (05:21): So I think we need to break it down between the administrative processes. So administratively, the AI tools, whether it’s co-pilot embedded in Microsoft tools or folks using Gemini or other layered agents, they can pretty much read anything once they’re launched. They are reading the f the full dossier. And we know they’re reading the full dossier because everybody wants the notes post. Meetings from an administrative point of view. So we know that while it looks like they’re just, giving us the transcript, they are able to provide a great deal more information if we prompt it. So we’re aware that they can read everything. That’s not the case in the clinical environment. The clinical environment is much more rigid. The embedding of an AI tool within the clinical environment. And I don’t want to go on too long, but our core application, our core electronic health record, does not have any embedded AI in it itself. It has third-party vendors that have maybe ambient listening tools. And then there’s an API layer that is built in order to effectuate some of the workflows and some of the dynamics. That API layer is obviously the most sensitive point of penetration where AI can begin to spread, if you will, in ways that are unintended. So I would say that as far as I understand. And as far as I’ve observed, in the clinical environment, it’s very restricted and very limited.

Chris Sowa (07:13): Well thank you for that. I know I’ve never spoken to somebody who was a CEO that was in charge of a large health system like you and heard that there was not pressure from a cost perspective, right? And this constants need to do more with the people that you have or with fewer people, right?

Jim Sinkoff (07:37): Let me tell you how we’ve sort of gone about that. So if you look at the arc of AI, everything from you know neural networks to even more advanced forms of AI today, all the way to robotics, right? Whether it’s, you know, the robotics of a robot running down a track or robots consummating transactions, where we have deployed for the purposes of cost savings, is really a great deal of robotics. So in those robotic processes, and this will make sense intuitively to everyone. So if you’re billing a medical claim, it’s very routinely routinized. There are a number of steps that you need to go through in order to go from beginning to end. And that can be done robotically. And so we have deployed a great deal of robots. To actually consummate those transactions. The way the environment is secure is the robots don’t sit natively within the EHR, they sit outside the EHR, and the instruction is then fed to the robot. The robot completes the work, and then there is a transaction back into the native system. So, and that has helped us from a cost perspective, just to finish it off, in not having to recruit and hire. Those human beings that once used to do that.

Chris Sowa (09:02): No, that makes a lot of sense. And I there was an article in the I think it’s a Wall Street Journal today that talked about how health costs are going up because of this AI battle between the hospitals and healthcare providers and the insurance companies. And I had actually done some analysis prior to that article myself and I learned that well the this is universal, but in many cases the insurance companies are winning. Insofar as that the number of times a first payment is requested by a hospital is denied, actually went up. So I think it was like 76% or something like that. And this is national stuff, obviously not specific to any one regional organization. Was approved, was coded correctly for payment. That number was pushed down to 67%. And I know from my work when I was a consultant and Censure, you know, a lot of insurance companies were getting involved with AI of reviewing payment information and just to check that. So I know a lot of that is being done by AI in the background, and a lot of those initial denials are most definitely related to AI. From a cash flow perspective, many of those may still be paid up to the seventy-six amount, but that cash flow is still, money that the insurance companies is that are using that is your money, essentially. Maybe not in your specific organization. When I say you, I mean the collective group of people that are providing.

Jim Sinkoff (10:43): Have you seen this within your own organization, the these trends there are a couple of interesting points that you may that are prompting me. So w one of the issues in healthcare that is going to slow the deploy of AI or limit the sort of deploy of AI. Is going to be in all of the sort of regulatory environment that really does look at fraud and abuse. So in your specific example, it won’t maybe raise to that, but I’ve got a different idea around that problem. But where it does show up is as the country moves more towards accountable care and risk-adjusted coding, which yields savings. CMS has the power of limiting the and capping the rise in upcoding. And by imposing a regulatory scheme, no matter how good the AI tool is, the payer, and in this case Medicare, can basically limit, let’s say the AO tool is really great and the code should be. You know, an upcode of five, you know, 1.4, but that’s going to be a 40% rise. They’re going to limit that code. And there’s a great deal of historical regulation. Some of it is now vestigial and needs to be updated in contemplation of these tools. But I think it’s important to remember that there’s still a regulatory environment out there that helps to circuit break. Some of this. To the particular example that you raised, there’s still a lot of market forces. So what we’ve seen in healthcare is a tremendous amount of con consolidation, both on the payer and on the provider side. Lots of roll-ups, into Gallagher for general liability and so on and so forth. And in the hospital system, we have now mega hospital systems out there. Where those issues, I think, get negotiated. Debated is what was originally construed in the payer contracts. So both sides can start implementing penalties Yes. For bad behavior, if you will. The deploy of the AI tool can certainly help the insurance company perfect their denial process. So their prior denial process was maybe a bit digital and a bit analog. Now they’ve deployed AI to basically perfect that system. Likewise, the provider side can use it to actually understand all the reasons for the denials and use AI to obviate the original reason for the denial. So I think it plays out both ways in the marketplace. And then when there’s bad behavior, there’s just bad behavior and it has to be dealt with in a different way.

Chris Sowa (13:57): No, I think that’s perfect. I mean that and that’s something that the discussions I’ve been having with some of the other health providers are is exactly that. It’s like, okay, we wanna understand which codes are being denied and why and which codes we actually should be applying. So there’s a little bit of market force balance that I think will eventually come into play.

Jim Sinkoff (14:19): The debate that we’re having internally around data governance, and I don’t know that we’ll touch on it in our time today, but I do encourage in my conversations with my colleagues and others to really think about data governance from this very core perspective who owns the data? And when organizations really understand who owns the data, I think they can begin to craft policies, procedures, procurement policies, testing verification over their data. And the reason why I say this, Chris, is very simple. If there’s a cyber breach, the covered entity is the entity that’s sued. And therefore, if that’s true, then if you can be sued for the breach of that data, that means you own that data. And therefore, you should be doing everything from a risk management perspective, including AI, to not just secure that data, but make sure that all of your contracts make it very clear with all the third parties you’re working with that it’s not their data.

Chris Sowa (15:38): That’s a great perspective. I’ve never heard anybody frame it exactly that way. And you’re right, if you could be held liable for it, then there’s an argument that it’s your dators. So that there’s there I guess there’s a silver lining in there somewhere as well.

Jim Sinkoff (15:54): I think so. And this is where I think the power of AI can really shine. So let’s say you work in an organization and I don’t know, you use Oracle on your HR side, you use some other application on your finance side, you use a third application on your clinical side, and now you’re deploying a number of AI tools. When you deploy those AI tools, Oftentimes there’ll be provisions for subrogation, there’ll be provisions for data use agreements, and a variety of important documents that frame and define the relationship between the parties. Often what happens is organizations don’t ask for bi-directional feeds to make sure that they’re getting all of their data back into their central repository. It’s an outbound transaction. And then that company is using that data for research and at other things to perfect products and so on and so forth. And when you want the data back, they often say, well, it’ll cost you X number of dollars for us to build an API for you to get it back. Okay, with that as the background, why is that interesting? It’s only interesting that if you want to do your own large language model research, you need large data sets. And those large data sets, if it’s only an outbound to third-party vendors, diminishes the power of your data set for you to begin to use AI internally for strategic purposes, financial purposes, cost purposes, clinical improvement purposes. And you can go on and on because the power of AI is that it is almost endless. It is limited by the curiosity of the asker. And now we know that it’s unlimited by the curiosity. And I’ll introduce a new theme here, which is I believe we have now as human beings literally created a new species. And standing up next to humanoids are a new species called artificial intelligence. I had a colleague when I asked them sociologically, did they have a f sense of what we were doing? And she said, I said, do you know what A AI is? And sociologically, and she said, artificial insemination. And everybody chuckled. And I said, you know, you’re not wrong. That we have artificially created a new species, and it’s standing up next to us.

Chris Sowa (18:40): And I’m not sure that everybody fully sees that yet. And you’re a very interesting person because on the one hand, you have become very knowledgeable and expert in AI, right? On the other hand, you know, when we were coming into this conversation, you mentioned to me you prefer actually prefer a lot of the analog, absolutely, is a part of you of your own life and what you talk about with your family and maybe some of your peers. If you could maybe just expand a little bit. On this concept of you know why you think that’s important for people.

Jim Sinkoff (19:16): I you know, s some of it is very I don’t know the right word, but what I’ll say is the value of speaking to you in this subject is what I call analog. You get to read my body language, you get to, you know, you get to apprehend me not just by my language. You get to get a feel for me. And that feel is can lead to. What I think is the most important thing is trust. So you asked me early on about who would be responsible. Well, if I told you it was a series of robots that were responsible for it, I have no way to manage from a place of trusting my colleagues’ ethics. Trusting my colleagues’ integrity. I have no way of imbuing ethical discernment in the work that we’re doing. Now, I work in the healthcare space, and I fundamentally believe that healthcare is a human endeavor. Regardless of how much AI can tell us about cancer cells or too much calcium buildup in plaque, all good, all great. Stuff that should be done. But at the end of the day, when somebody is losing a loved one or is frightened about a potential outcome or is not understanding their medication, that human touch is of so much value and it creates the kind of trust that I think is so important. The other piece of the analog world that I also think is important. Is I write longhand. And I also type my own emails. And I don’t ask AI to do it. Now I’m slower, I’m not as quick. But if I’m making grammatical errors, I need to reread my work. The simple act of rereading, editing, thinking about the reader, thinking about what I want to say. AI will come up with their version of what they think is a better conveyance. But once I become so wedded to it, I basically have given over to it, and I have now taken my own learning and my own skill and now subrogated it. And so I really feel it’s important, and especially for new generations. You know, you could call me an old salty dog if you would. I’m not sure that I would disagree with that generationally, but I would say that it is so important that people do longhand math so that they understand the theory and they learn to think Mm-hmm. And not just ask the machine to tell them the answer. There is so much lost when we just ask for the answer.

Chris Sowa (22:25): Well, and I had a another guest on earlier and you know she was she told me she’s a grandmother and she said that you know used to be what grandma said was the answer, and now they don’t believe her unless ChatGPT tells the same story. Right. And you know, for myself, I do some analysis on different things and I was my son was looking to go get his master’s, which is he decided what he was gonna do. But we know to gather information quickly, we use the AI. And it was recommending a particular school, and I’m not going to mention the schools. And I was surprised by it. It was a good school, but it wasn’t when I expected. And what I when I challenged the AI, and I think this is something that also is important, is we need to learn to challenge the answer of the AI. We need some critical thinking. And I and because I know a little bit about how the AIs work, I said. You know, are you looking at Reddit and the number of people that are responding and using that larger data set as a reason that you favored this particular institution? Because I knew they were much bigger than the smaller institutions that we were we had in the West.

Jim Sinkoff (23:38): Such a critical point that needs to be underscored constantly. I wish I had said it. I it is so important what you’re saying is that critical thinking is so vital in the deployment, the procurement, the use of AI. And beyond the prompt and the response. If you start from a native place, you start from the place of what is the problem that is trying to be solved. And the way AI development works is not classic code writing anymore. It actually works in a reverse engineering process. It doesn’t code from the start to the end. It goes in reverse engineering. The idea of reverse engineering, maybe for some of your guests, it’ll be like, yeah, that’s obvious. I understand exactly what reverse engineering is. But for a lot of folks, that concept is foreign. It what do you mean reverse engineering? So I really appreciate what you said. It’s so important to be said over and over again.

Chris Sowa (24:45): Now, you know, you’re somebody that is on the AI healthcare transformation task force. Is that right?

Jim Sinkoff (24:53): And It was started by Governor Levitt, who was the Secretary of health and human Services under Bush One.

Chris Sowa (25:01): And are you finding this people are thinking about transformation that AI is on the agenda from kind of a value perspective?

Jim Sinkoff (25:12): There is no question that you can’t find yourself in any room. I in any industry sector I think at this point, and healthcare’s no different, where folks are actively engaged in how to solve the quality and the cost issues in our healthcare delivery system, without talking about what AI tools might be available. To help with that transformation. There’s no question. It’s here to stay. It’s part of the landscape. It’s accelerating. And in that acceleration is also a great deal of tripwires. Our approach in my organization is to be very holistic about our procurement. We don’t really like to do one-off projects. We like to try to do projects that are integrating into a population health perspective because we’re trying to manage populations, not individuals and their individual diseases. Of course, we’re trying to do that as well, but we’re trying to really do it in a holistic frame. So when we look at the reason why people are overusing the hospitals, for instance. We’re trying to understand their sociology, their family dynamic, their transportation barriers, and a variety of issues in addition to their clinical issues. We don’t need an AI tool that can solve just one of those problems. But yes, within the healthcare space it is. And the amount of private equity that’s out there is some of it’s just stupid money. I mean, people are just throwing money at it and thinking something’s gonna stick.

Chris Sowa (26:56): So knowing what you know about artificial intelligence and obviously the you know the transformation aspects, if you were giving counsel to a peer and you know or to the board as a group, and you said, you know, if there are two key things that we need to decide before the end of the year as it relates to AI, or we’re gonna regret it not talking about those two things. And if there’s only one, that’s okay. If there’s three, that’s okay too. But what would be you know that key decision that they ought to be making as a team and as an organization before the end of the year?

Jim Sinkoff (27:38): It’s a longer list than two, and I think you get the flavor, it’s hard for me to be succinct. I keep working on it, but I’m it escapes me.

Chris Sowa (27:47): Doing a great job. I w I’m loving it.

Jim Sinkoff (27:49): Let me start from a philosophical perspective or p policy perspective. What I have said to my board is that we are all in on AE AI resources and AI solutions. As an organizational strategic imperative, we will continue to innovate, explore, and deploy AI tools as quickly and as fast as we can with this proviso. So long as it’s done in a way that we believe meets our policy, which we’ve crafted. And is secure. And last but not least, is that we are a co-creator in the development of that AI. We very rarely take an off-the-shelf product and deploy it Mm-hmm. Because we it’s very hard to verify. It’s very hard to understand as you just talked about your Reddit example. What are the data sets that they’re pulling from? Are they relevant to us? Not relevant? They’re just a number of things that we interrogate before we do the work. The other business piece, and this may surprise you, Chris, and I told you I run another company. We also will tell a potential co-creator organization that we want, especially if they’re in a startup phase, we want equity in their product as they think about going to market. That creates a very different dynamic in and it’s go and why do I do that? Why do I insist on that? Because it’s our data. Your learning and building your business model on our data. I think what we can all conclude, your listeners, and well beyond this, is that the new oil is data. And it is being monetized all the time. And there is no reason for us to, and I say this to our board, for us to not absolutely insist that. We monetize these opportunities in protection of our patient care. So I it was a longer-winded question. I don’t think I answered it one, two, three, but I think it’s an approach to the work. We want to engage, we want to find these great solutions, but we’re not just opening the vault and saying, go forth and so on and so forth. There’s also a huge training aspect behind it. I’ve over two thousand employees.

Chris Sowa (30:39): That’s huge. I mean, there’s so much that they need to know. So unf I mean we I think that you and I could talk for hours. I think that we might we me. We may ask again to have you on the program, you know, as this environment continues to change, you’re obviously a great resource, so thank you. I want to give you an opportunity to give a little plug for your organizations here at the end. I’ll give you the last word, Jim.

Jim Sinkoff (31:05): Sun River Health, I am very proud of the organization. We work with almost entirely underserved communities. You know, in 50 different clinical sites throughout Greater Metro, New York. You were extremely proud of the impacts that we’re having on communities in terms of their health and wellness. And then on my other company, solutions for community Health. Which is a for-profit tech company. We’re equally excited about how we’ve been helping a lot of clients deploy and think about the deployment of AI. And it’s been a lot of fun watching the industry slowly but steadily begin to adopt different solutions that are helping their communities. So it’s been very satisfying. And so I encourage Anyone who’s looking to earn a good living and do good in the world, look at community health. I think it’s a great place to learn and expand and live well.

Chris Sowa (32:07): Fantastic. We really appreciate you taking the time today. This is HAL: Who Signed Off on This? And we bring the key decision makers, the people who are actively in the moat and doing it them s doing it right now. And we’re very lucky to have had Jim Sinkoff on this program here today. Thank you very much, Jim.

Jim Sinkoff (32:28): Thank you, Chris. Bye now.

About the show

HAL: Who Signed Off On This? is a show of short conversations with the people who have to sign for AI: how they decide what to approve, where they insist a human still looks, and what they would need to see before they trusted a machine with more. HAL is the computer from 2001: A Space Odyssey. The show is about the people who make sure that never happens on their watch.

About the host

Chris Sowa is a founder of RedactSure and a former CEO of an AI company; he started it years before ChatGPT existed. He previously led AI at Accenture, served as Global VP of Strategy & Innovation at Schneider Electric, was CCO of Sovos, and spent more than a decade at Oracle, with earlier roles at SAP and IBM.