Speaker 1: Hello, and welcome to the Proverity Solution Session. Joining me today is Professor Larry Van Horn, President and Founder of Proverity, located in Nashville, Tennessee. Welcome Professor Van Horn.

Speaker 2: It's a pleasure to be here.

Speaker 1: Great.

Speaker 2: In the music city, no less.

Speaker 1: I know. Nash Vegas. Nash Vegas. What a great city. So speaking of the great city and your company, Proverity, where did Proverity come from? What was the brainchild behind it, or what sparked the development of such a company?

Speaker 2: So I've been an academic economist for 25 years, analyzing data, studying patterns of health care delivery, how providers behave. And I came to realize in that space that the data that we have in the health care industry provides tremendous visibility into the practice of medicine. And having worked with PIAA in the past, and MPL, and various carriers, I came to appreciate that they didn't have access to that information. And so I embarked on a journey, probably going back about seven years ago, to create Proverity, where we spent our time analyzing the relationship between the way a doctor practices, their scope of practice, treatment intensity, all those types of things, and malpractice risk. And that gave rise to Proverity in 2016.

Speaker 1: Wonderful. All right. Seven years. That's a long time that you've been investing in this and thinking it through.

Speaker 2: It's been a long road.

Speaker 1: Right. When you embarked on that and found the interest, what was it that you were hoping to achieve? What's the goal, the mission?

Speaker 2: As an economist, I like markets. I like perfect information. I like things priced accurately. And really, the objective was to use the available information and our analytic tools to be able to identify low-risk and separate low-risk from high-risk providers to perfect the market. Because unfortunately, with the limited information that's historically been available to malpractice, insurers, they didn't have the visibility to price malpractice insurance based on the true practice of medicine.

Speaker 1: So Professor Van Horn, can you tell me how your experience with malpractice carriers relates to hospital risk?

Speaker 2: Sure. Well, we're in the capital of health care delivery, Nashville, Tennessee, and we've got lots of great organizations here. And they actually self-insure and manage their own risk through risk retention groups. So they credential physicians, they put them in a captive, and they are exposed to the risk of those providers, not unlike what a malpractice insurer is. Malpractice insurers make different decisions about who to underwrite and who to recruit. Hospitals will make decisions about who to credential on the staff, who to employ. They have other decisions to be made. Do I keep them in my risk retention group? Do I send them out for a commercial paper? In a world of everything else being equal with two physicians, one's higher risk, one's lower risk, who might I choose? So our analytics around risk and visibility into what drives risk are valuable both to the risk retention group captive space as well as the admitted carrier commercial paper space. The solution is a little different. The implementation is a little different. But the analytics in the thought process and the expertise that goes to create that solution is the same.

Speaker 1: Wow. OK. That makes sense. So you're able to serve all industries.

Speaker 2: Well, we're able to serve the $20 billion US of a malpractice industry, both in terms of captives and risk retention groups as well as the traditional admitted carrier space. And we've historically been working in this space. And increasingly, we're being asked to come in and provide risk-based management solutions for hospitals, health systems, and captives.

Speaker 1: Wow. That's wonderful. So not only do you have a breadth of data available to you, the experience of the data scientist and your staff really brings the expertise in order to be able to assimilate and deconstruct that information.

Speaker 2: It's insufficient to be a great data scientist. You actually need to understand the language of health care administrative billing, CPT codes, HCPCS codes, SNOMED, ICD-10, NDC codes, knowing how to organize that information and distill from that features that are impactful for risk, either in terms of forecasting risk or in terms of classifying risk. And so it's the data with the modeling expertise and the deep subject matter expertise of the folks in Prevarity that enable us to do what we do.

Speaker 1: Wow. That makes you very powerful and unique.

Speaker 2: Well, we like to think so.

Speaker 1: Yeah.

Speaker 2: And we're the capital of health care delivery in Nashville. And we're able to leverage our relationships and understanding of the industry. And we feed on all of the great organizations that are here in Nashville to help give us insights in terms of how to understand risk.

Speaker 1: That's wonderful. So what do you think of the future of machine learning and artificial intelligence in this space?

Speaker 2: I mean, if you read the NPL magazine and whatnot, everybody's talking about predictive analytics, big data, and the future of AI. And I think there is, as data becomes more replete and we have greater technology, more powerful computers, we can do lots of things. The issue is that that has to be connected with subject matter expertise. You can throw a lot of data at a machine and let it whirl, and it will come out with stuff that makes absolutely no sense, that is not actionable.

Speaker 1: So the black box really isn't effective.

Speaker 2: You can't do this in a black box without having context and industry expertise to guide, shape, and constrain the optimization process. And that was one of our early learnings, is that you just can't throw a lot of data at something and have the machine solve the problem in a way that makes any sense to any of the folks who live in the space. It's got to be curated and constrained to be able to come up with solutions that have tractability, relatability, and the ability to communicate to somebody why this provider is higher risk or lower risk.

Speaker 1: Right, right. Really rock solid information there and how you derived it.

Speaker 2: So I mean, what's great is that our head of product, Matt Kerlin, is going to do kind of a deep dive on what we do, how we do it in terms of the data, how we analyze things, and how we translate that into actual information. And so he's going to follow us now and do a little bit of that conversation.

Speaker 1: That sounds great. I'm looking forward to learning more.

Speaker 3: Hello, everyone. My name is Matt Kerlin. I'm chief product officer here at Proverity, and I'm here to talk to you a little bit about our solutions. At Proverity, we consider that we're creating the next generation of malpractice risk management. What do I mean by that? Probably the best way to start is by giving you a little bit of an analogy. So Progressive Insurance today has a product called Snapshot. Now what Snapshot does is Progressive will put a little device in your car, and it will transmit information about the way you drive back to Progressive. So as an example, it will record how fast you drive, how fast you turn, how fast you stop. And based on this information, Progressive is able to model the types of risks you're taking when you're driving your vehicle. And based on those risks, it's able to change the way the amount of premium that they charge you. Now what Proverity has done is we've taken that same model, and we're bringing it into the medical malpractice industry. How do we do that? Well, to understand that, the first thing you need to understand is a little bit about how the medical industry works. So if you think about that for a second, a patient will walk into a doctor's office, and they'll receive treatment. Maybe they were injured, maybe they're sick, maybe they're just going for a checkup. But what they do is they walk in not with any cash for that transaction, but they walk in with their insurance card. And they give that over to the doctor, and the doctor performs the services. Now what you often don't see is that once the provider performs the services, they record that in what's known as an electronic transaction called the 837. Now that 837 contains a lot of information about what the doctor did to you. What was done is encoded in something called a CPT code. So that's basically the services that they provided. Why it was done is encoded in something called an ICD-10 or an ICD-9 code. Those are the diagnosis that the doctor came up with when the patient was in front of them. Where it was done is the place of service code. So were you treated in a hospital, were you treated in an outpatient setting, or perhaps a doctor's office, urgent care center. And then of course, if you go to a pharmacy and pick up a prescription, they record what kind of medicine you picked up and what quantity and what dosage. Now all that is recorded electronically and sent to a medical clearinghouse. And the job of the clearinghouse primarily is to take that information and distribute it to the payers. So these are the Blue Cross Blue Shields, the Aetnas, the Medicares of the world. And once that information is distributed to the payers, the payers will process that and return payment back to the physician. But the other thing about this transaction is because all that is recorded electronically, companies like Prevarity can receive a feed of that information. Now this feed is de-identified, so there's no patient information in there except high-level demographics. But what Prevarity is doing, and what other companies can do, is they can extract a lot of information about what the provider's doing, why they're doing it, to whom they're doing it. And we can use that information to create not just a profile about the physician, but also a risk. What types of risks are the physicians taking in the way that they practice business? And then that way, it's exactly analogous to what Progressive is doing with their Snapshot product. We're doing the same thing, but the difference is that the practice and the providers have to change nothing about the way they're doing business today. As a result of that, what Prevarity has is we have already assembled risk profiles in detail for about 80% of the provider population across the United States. And this is what's very different about how we act. So if you look at how those risk profiles actually behave in the market, here's an example of what we do. When you look at internal medicine doctors in the state of Florida, what we've done is we've taken them all for the year 2013, and we've separated them into quintiles, 20% of the providers in each bucket according to our risk profile. So the lowest risk providers are in quintile one, and the highest risk providers are in quintile five. And when we do that, and we overlay the event rates that occur in each quintile, you see that there's quite a bit of difference. In particular, the providers that appear in quintile one have about a half percent or three quarter percent risk of seeing a claim in the next year, whereas the providers who are in quintile five see roughly 16 times that risk. So as you can see, there's a huge differential between what we can do within a single specialty, within a single state, by looking and examining the activities and applying our risk models. So that is fundamentally what Proverity does as a company. Now is this an anomaly? Can we only do this in internal medicine? Well we've had this information and our models validated independently by actuaries, several actuaries in fact, and all of them come back and said exactly the same thing. Our models for individual providers provide additional lift over and beyond the traditional actuarial models that are doing the same thing. So there's a lot of value to be had by using our models. How does that value manifest itself for risk retention groups? Well here are some ideas that we've had in talking with some customers that we're currently piloting with. The first model is very simple. When you're considering a new provider to bring them on board into your risk retention group, there's a hiring process that they need to go to. They need to interview with different doctors, you may need to fly people around to meet people across the country, and there's a cost associated with that. Very simply, when we provide you a risk profile for an individual, before you start flying people around and having them meet your doctors and taking them off of the hospital floor for example to meet them, you can examine our risk profile and you can say, hey is this doctor risky, not risky? Is he someone that we'd like to have in our portfolio of providers from a risk perspective? If the answer to that is no, then you can screen them out before you incur any of that cost. Of course if the answer to that is yes, then absolutely go ahead with the hiring process. So that's a very simple example. Here's a little bit more complex example. So if you're looking at our product or our risk profiles as a way to evaluate the portfolio of risk that you have, there are a lot of things you can do with that. So first is if you bring a single provider on, how does that impact your overall risk portfolio? Does it increase your risk? Does it decrease your risk? And if you're doing it for one, you can do it for many. So in the case of an acquisition, what if you're bringing on 20 or 30 or 50 providers? It's the same sort of analysis just repeated over time. The other thing you could do is you could look at your current portfolio of risk and decide for example, do you want to take a set of providers and move them into the market as opposed to taking them into your group? And if you do that and you make these decisions, what impact does that have on the overall reserve that you're holding aside to handle malpractice risk? Ultimately, the goal is to minimize the reserve that you have on hand so that you can move that money into different higher value purposes. And our models help you do that a lot differently than you're doing today. The next example is something that we call the financial view. So when we talk to our customers and we understand, we try to understand how they're allocating the cost of malpractice risk across different locations or different groups. And today that approach is somewhat rudimentary. So perhaps there's a relativity value that they assigned to different specialties and otherwise it's just a count of doctors and you take your current expense and you divide it up by that count and you move it out and you allocate it to the individual groups. We're using our risk score, you can get a lot more sophisticated with that and you can actually charge different amounts depending on the amount of risk you're taking by employing a particular doctor. So in that case, if a particular practice is hiring especially risky doctors, then they can be charged relatively more. In this case, as we're showing behind you, that change could be very substantial. And if you adopt this approach and you have your local risk managers making different decisions based on the risk of the doctors, effectively what you're doing is you're allocating the responsibility for lowering the risk down to the sites and you have a financial check on their ability to do that. Another view that we have is the operational compliance view. So if you think about the information that we're collecting, we're collecting data about the types of procedures that providers have done over time. So thinking to the application process, so when you have a new provider applying for credentials into your risk retention group or perhaps they're applying for employment, one of the questions is generally what types of procedures do you perform and in what quantity? Well, we already know a lot of that information and we can provide that to you prior to the application. So from a process perspective on an operational perspective, that's a much more streamlined process and a lot less information you're asking for your potential hires. The other thing that we can do since our data is updated on a monthly basis is we can monitor that. So if you're hiring somebody with the assumption that they're undertaking certain procedures and therefore a certain risk class, we can make sure on a month in and month out basis that they're adhering to your normal assumptions and therefore validating the model that they're operating under. So when you combine this with other monitoring that we do in the market, such as things like sanctions that they may receive from government bodies and what have you, we can provide you a good updated monthly list of different things that may be impacting your provider risk model. So hopefully these examples provide a good basis for how you might get some value out of Proverity Services. If you're interested in learning more, please do contact me at matt.krillin at proverity.com or visit our website at www.proverity.com. Thank you.