#244 Using Data to Optimize Costs in Healthcare with Travis Dalton and Jocelyn Jiang President/CEO & VP of Data & Decision Science at MultiPlan

Hi, I'm Maggie, VP of Curriculum at datacamp, and I know firsthand how overwhelming it can be to keep up with the rapid changes in data and AI. But staying ahead doesn't have to be hard. That's why our curriculum team has curated the best learning content to help you and your team stay ahead, covering everything from data and AI literacy for busy executives to cutting edge generative AI for developers, and everything in between. Whether you're upskilling yourself or your team, we make learning practical, engaging and flexible. It just works. Power your data and AI transformation journey on datacamp.com.

You know, I'd say data is like the new oil. You got to find it, refine it and distribute it. And so being able to do that in a reasonable way is going to be really important. And I think platform products that have scale and that offer scale are going to be one of the biggest ways that potentially to control costs in health care over time.

Welcome to Data Framed.

This is Richie. Even though I've lived in the United States for eight years, I find the.

Healthcare pricing system largely incomprehensible.

Turns out I'm not alone. A lot of healthcare pricing is deliberately opaque and this results in individuals and.

Organizations spending too much on healthcare and.

Getting poorer health outcomes. Greater price transparency and predictability has the potential to reduce health care costs and have a substantial impact on health outcomes. And naturally, the key to achieving this is lots of data analysis and machine learning. Today I have two guests from MultiPlan. Travis Dalton is the President and CEO. He's got 20 years of leadership experience in the healthcare space and at this point is one of the leading thinkers on healthcare pricing and previously he was General Manager and Executive VP at Oracle Health. Jocelyn Jang is the Vice President of Data and Decision Science. She's responsible for leading the data and analytics initiatives. Previously, Jocelyn worked in health and welfare consulting and actuarial analysis at Epic Insurance Brokers and Consultants and Aon. So today I want to dig into the details of how machine learning and data are used at Multi Plan and how they can be used to improve healthcare pricing.

Hi Travis and Jocelyn, welcome to the show.

Hey Richie, thanks for having us.

So I'm actually going to start this with a confession. So I've been living in the US for eight years, but healthcare is still very mysterious for me. So just for the benefit of myself and for our audience, can you tell me what happens when someone gets a treatment? So who ends up paying who?

Well, you're not alone. I've been living here my whole life and it's mysterious to me as well. But you know, one of the things we're trying to do at Multiplan is kind of simplify and educate and demystify to some degree how the process works. But really there's, you know, there's essentially two kinds of health insurance in the U.S. so it's government funded, Medicare and Medicaid and then there are commercial insurance products. Typically those are borne by payers that bear the risk and or employers that are self funded that bear that risk. When someone has an issue or needs to see, you know, a provider or someone that's delivering care, they have the choice to go in network. And those network agreements are typically negotiated between providers and payers of health care and employers and sponsors. And then there's a pre negotiated discount or rate. And that's a pretty easy path. I suppose most people are used to that, that have some method of insurance or there's some out of network options that exist and there are some cases where an employer isn't inside of the network or out of the network where care is sought. And that's a very different path with a very different cost structure around it. So I can see where it's confusing. No doubt about it. But as I said, what we're trying to do is demystify the extent we can and also control and contain costs to the extent that we can as well.

Okay, so primarily insurance based. And then you've got this difference between in network where it's like specific providers and then out of network where it's like, oh, you want treatment from someone else and maybe you have to pay a bit more. All right, so it sounds like a fairly elaborate system. What are the problems with this? Jocelyn, do you want to explain the downsides of the system?

Sure. Like Travis mentioned, it's fairly complicated. A few I would say main problems with the US system for the average consumers, number one, I would say the price is not transparent at the current level. It's not very easy to know how much you're facing out of your pocket prior to your getting the care in a lot of the situations. And also secondly, the price can be drastically different from provider to provider even for the same exact procedure that you're seeking. Just simple example like getting an X ray, getting an X ray in the hospital comparing to getting an X ray in the imaging center. That price can differ by 10 times or even more. And then third patient with the insurance, they typically only care about their portion of the out of pocket amount and have no idea of that main risk. Barrier, whether it's the payer, which is the insurance company or. Or their employers, who is actually the risk barrier, are actually paying for the service. So a lot of the times it's hard to educate the consumers or the patients to shop for the service that are shoppable. And then lastly, the patient decision making on the provider selections oftentimes is also less about quality or cost and more about how soon I could get that next appointment and how easy for me to get to that doctor. So that I think that's the area with the recent government requirement of price transparency. We are trying to help filling that gap is to help bring that transparency, helping to bring that cost information, quality information to the consumers, to the payers and providers so they can provide that information to the people for them to make a wiser decision prior or before consuming the care.

Okay. So, yeah, with that lack of transparency, it's very difficult to know how much you're being charged and where you're going to get the best price. I guess there's no, like, price comparison service like you get for shopping sites. No equivalent for healthcare in general.

Okay.

All right.

Travis, did you have anything to add to that? Are there any other problems? You see the system?

Yeah, I mean, it's interesting. I mean, there's, you know, the price for the same service can be different. So if you're going through a Medicare channel, it's a price. If it's in network negotiated, it's a price. If it's out of network, it can be a price. So part of what we're trying to do collectively with all the transparency data and with machine learning and other capabilities, is to essentially bring transparency to the table, but also evaluate those prices against some fair and reasonable basis and determine what an appropriate cost for something might be and bring that to the industry in a way that helps contain costs over time. And so there's a lot of challenges with that. I also serve the provider and patient community for 22 years before I joined the CEO of MultiPlan. And, you know, there's a lot of disparate data out there. Taking that data and normalizing it, making it useful, is something that's really important as it relates to healthcare, particularly in terms of bringing value, but also on the patient provider side as well. We can certainly have a little discussion about that from my past life too, if it's useful to your listeners.

Okay. I love that you mentioned data, and I really want to get into details of what the data is and how it's used. Maybe just agreeing with it at a high level. How can data be used to help solve this problem? With transparency around costs, there's several ways.

So we offer a number of services. So one of the things that we do is we do offer a network discount. So we have 1.4 million providers that have a discount with us and we offer that network as a primary error. Beyond that, we also offer services where we look at claims and we evaluate those claims against some rational basis, like in Medicare as a basis, and then we essentially reprice those claims. So when you look at an out of network claim, sometimes it can be drastically higher than you might see for the same service in network. So we use data, we use information, we use machine learning, we evaluate those claims, we look at them, we peg them to some basis and then we have a market clearing or what we think would be a fair process by which we work between payer and provider to essentially resolve that claim and keep costs down for employers, employees and patients. Furthermore, the team is working on transparency products that take a massive amount of MRF files and data that has been essentially there's been some policy around that making that mandatory that cost data be provided, ingest that, normalize it, use it in a way that you can bring that forward. So you're now getting hopefully cost and quality to the table in a reasonable, meaningful way.

In that case, Justin, do you want to go into a bit more detail about like what are all these different data sets that you have and where do they come from?

Typically those claims data come directly from the insurance company. And then we have a series of ingestion data processing steps to normalize and standardize those data that coming from the insurance company. Oftentimes the way the data is being laid out or sent are by different companies. And then after we normalize them, we will go ahead and enrich those data further. So for example, the data will come in as procedure code or diagnosis code and we enrich them with the description, give them the categorization to help them further analyze their data better. And then furthermore, like Travis mentioned, with the price transparency data that's being public available, although it's very messy and then it's massive that we use a lot of machine learning techniques to sort those out and then clean those data out. So then we can leverage the actual utilization data of the clients coming from the insurance company and then reference that with the public available data from the transparency files to find that balancing point and to compare, you know, like whether services are being paid at a reasonable price point or if there is a Better network the client should be picking based on how their employees are located or how their consumer the cares.

Okay, it sounds like you've got an awful lot of data clean to you then that might be kind of the crux of this. So can you tell me for like what types of data is it? Is this like a lot of just sort of text data that where some doctors scroll some notes somewhere and you're trying to work out what that actually is or what sort of data cleaning do you do?

Mostly the data when it comes from the insurance company, those are post adjudicated claims data. So they are standardized and mostly like they have been processed by the insurance company. So they have the diagnosis information, the procedure information. So you know what has been done to that patient and which provider and then what is the cost, what is the billing amount, what is the discount rates and then you know what is the out of pocket amount for the patient. So those are typically the information we're getting from the insurance company. And then one thing that I didn't mention earlier was after we ingest those data we also apply prediction to it that with the machine learning techniques we can predict someone based on their past 12 month claim experience. What is the likelihood of this person of getting a certain disease onset probability will be like or what is this person's future 12 month spend is likely going to be in a certain range. So with those type of information we can better pinpoint the type of patient that we should focus on or if they have a point solution vendors and oftentimes we can direct them to the right targeted population for them to do early intervention.

That sounds really cool. The fact that you can make predictions about how much people are going to have to spend over the next year, things like that just help people budget. Are you able to share any details on what those machine learning techniques are that you use?

Yeah. So before we were purchased by Multiplan, the company was founded out of MIT spin out project led by Dr. Bert Smith, that's a pronounced professor in the optimization techniques. So we mainly uses the decision tree methodologies to finding that optimized decision for the individual based on their past care pathways that we could see in the claims and then finding those detectable signals in their claims and then determine okay, based on all these signals, combination of the signals, how much spend this person is likely going to spend or what type of disease this person's likely to have and then it's not widely any disease we can predict and oftentimes is not those ones that have certain patterns that can be traceable or trackable. Those are the ones that can be detected better.

That's fascinating. And as well, the use of decision trees means that it's quite easy to explain what's going on, which goes back to your idea of having transparency around like how everything works.

Correct. And I'm glad you mentioned interpretable because I think that's one main differentiator in our methodologies. Comparing to other companies that are applying AI or machine readable methodology to their data. Oftentimes it's a black box, like you can't understand what's going on before the results being spin out. But with a decision tree like you can actually interpret the results or prediction and have a better understanding and then like guide the model if needed to better fit the prediction or results that you're seeking for.

Just going back to like the main use case of this, which seems like it's trying to save money for people so they're not spending so much on healthcare. So I guess if you are paying for insurance, like what's the sort of step one into like making use of all this technology and data analysis in order to save money on your healthcare? Travis, do you want to go with this?

Yeah, happy to. I actually want to make a little comment on what Jocelyn said too, if that's okay. As an industry, my view is that the BST acquisition was made because we actually believe that's the future of healthcare is data. You know, I'd say data is like the new oil. You got to find it, refine it and distribute it. And so being able to do that in a reasonable way is going to be really important. And I think platform products that have scale and that offer scale are going to be one of the biggest ways that potentially to control costs in healthcare over time. And so quality and cost are two of the biggest issues that are facing healthcare. And I think the use of technology is going to be the thing that ultimately helps quality go up, access go up and costs go down. And I think that will be scale technologies. And so you'll see that from major, major technology players that are interested in the space because of that. And also, you know, as we move to the cloud and other capabilities, I think it's really important for care. You know, there's lots and lots of point solutions out there. There's lots of things to do. I mean, as an individual, you have to just be proactive in your own search and your own career. If you are part of an employer plan, it's likely that someone like us or another Technology provider provides capabilities where you can actually search, you can do some search where you can look at some price, quality and availability data or it can serve up someone that's in your network that is at a particular price and quality that's emerging and getting better and better over time. As the data and the data science improves and as the transparency data, which we wildly support as a policy initiative becomes available, we can use that and we can start to map, as I mentioned, cost quality to availability and that makes the process more simple. So you can start to have choices. But it still feels a lot like hunting and pecking or luck or referral network or I just know somebody, I think you have to advocate on behalf and don't underestimate the tools that you may have through your employer or through government websites that are actually starting to develop and do better here, of which, you know, we offer some of them, but others do as well. I think that's somewhat the state we're in.

I suppose the other part of this is things like the provider side of things. So like the, the doctors and the hospitals. Does having better data affect their costs and. Yeah. Is it going to affect the, the monetary situation for these providers as well?

Yeah, I mean, I put on my, my healthcare hat and in total, yeah, I think the thing that's going to be really fascinating I think with the provider and the patient side, not to take us too far off track here, but there's a couple really interesting developments. One is the ability to use ambient technologies and listening and other techniques to where you're now collecting data in a way that's very different than the past, which is essentially the keyboard. So we always used to say, you know, in prior life that the most dangerous thing in healthcare is the pen because you were writing script. Now it's, oh, the most dangerous thing in healthcare is the keyboard. Now it's a voice to voice conversation where you can actually collect information and data in an ambient way. I don't know if people like that word or not, but I've heard an actionable way, that's what I would prefer to call it, that is done real time. And also you can run rules and you can use other technologies behind that to where you're taking a lot of what would be administrative work off and burden off of providers and caregivers. And now they're able to focus on the patient and the actual conversation that's happening and then you can run prediction behind that. So there's fascinating things going on in healthcare. I encourage your listeners to really look at it as a cool space to go work. I mean prediction models using data around. I've seen really interesting models around suicide prevention, looking at social determinants combined with medical data to get early detection or early prevention technologies to people or capabilities to people opiate other abuse situations where you can look at state data and you can look at other data and you can determine a full view of someone's healthcare and what's happening. So you can intervene in a way versus wait for someone to present. There's some really fascinating things that are happening with that. And also around precision medicine where you're looking at the genotype and you're figuring out the best way to actually prescribe in a precise way certain therapies or procedures or protocols that yield the best result. The next 20 years is going to be incredible. I Wish I had 30 years left in my career. I probably don't. Jocelyn probably does. But I mean it's going to be like the best 30 years because it's going to be the most exciting time because this is all available now because a lot of it's been digitized in the US and across the globe actually. So sorry, I got a little excited there. Richie.

No, no, I love the excitement and that's kind of interesting that a lot of the stuff you mentioned about. Well, okay, if you want to improve costs in healthcare, you just got to intervene early and detect things before they become a big problem. I can certainly see how that's the case with opiates. You want to, you want to identify the problem before people have messed up their bodies. So are there any other areas where you think there is room for like some big improvements in healthcare there? So like are there any particular disease types or any particular conditions where you think there's room for a lot of use of data to improve the situation?

Yeah, I think that there are quite some areas where especially I think the shoppable services where people can have the time to evaluate multiple options. That's where we have the better chance to provide using information to provide more actionable insights before someone making that decision. So I would say things like orthopedic surgeries or trying to find the best provider to treat kind of a long term disease conditions. That's when I think using data and having that well rounded information, both cost and quality to help make that optimal decision for the patients. I don't know Travis, if you have some more to add, I'll answer a.

Question a little differently than a specific condition or area. What I would say is that for Us a really important thing is figuring out what to work on. The world of things to work on is enormous, but the resources to do it are finite. And so having product life cycle capabilities that allow us to understand the market, to understand the needs of clients, and then ultimately to make things that matter is really important. Or you end up doing quote unquote science projects that don't have scale benefits. And so companies, I think, have to think deeply about how they actually create product life cycle so they can make organic capabilities that matter most to the markets they serve. And then you start to surface the things that can have the biggest impact, some of which have been mentioned here today. And so there's just many, many, many things you could look at from a morbidity and comorbidities perspective and insights. But picking them and having impact and getting the market is also important as you run the business.

Since Justin, you mentioned the idea of like long term diseases and Travis, you talk about which areas have the biggest impact. So it feels like with healthcare there are some things which are kind of cheap but happen a lot. So something like flu happens to millions and millions of people, but also it doesn't have a big impact. And then you have some things which are rare and kind of devastating. So cancer is less common but has a tremendous impact. So is there like one area or the other where you think data is most important?

Data as a usable commodity? Where I said earlier is, you know, the new oil is going to be interesting in the sense that it's not all in healthcare usable because of lots of rules, lots of regulations, lots of phi, lots of other things like that. So if you could truly figure out the best way to look at social determinants, to bring in search information to combine that with the medical record, you really could start to predict with much greater accuracy certain things like obesity in certain areas, looking at diabetes, instances of diabetes, certain things like that could have predictive capabilities or qualities. And that way you could start to target programs around wellness and other things in targeted geographies and areas, or you could spend social dollars in a way that's meaningful in a community to help prevent the spread of some of these things over time through education and otherwise. And so I think that most things are predictable in some way if you can get the right amount of data. But there's a large debate in healthcare around what's acceptable, what's usable. Is it phi, meaning health information that should be protected? Should it be usable, should it be ingested, Is it cyber secured? All of those things are just A massive part of the data infrastructure that have to continue to be answered so that we're able to use predictive models in a way that makes sense and most meaningful. Probably didn't answer your question, but I thought I said something.

No, that was really interesting.

Yeah.

So the idea that there are some sort of limits on what you can do ethically is very interesting because I suppose you can sort of try and ingest everything about everyone's lives sort of Facebook style, and then get more information for better predictions. So can you talk me through, like what are the limits on what data you can get in order to make predictions?

I think when we analyze those insurance claims data, some other data that we can combine to help making some more accurate predictions are things like social vulnerability index, just as an example. So you can use a public available data in combination with the client or the patient's specific data together to bridge the gaps. So you are not exactly using some of the specific phi information, but using some general information. Knowing this person lives in this zip code and then by knowing by public information what is the social index look like in that area, you can make certain prediction around, okay, four people live in a certain geographic area. What do they like, like what their average educations are, what their dominant ethics in that geographic regions are. So things like that, when you combine the specific data with more general data together you can still make some pretty useful predictions around healthcare.

And I suppose related to this, there's an awful lot of data privacy regulations around healthcare data. So you've got like in the US there's the HIPAA regulations. So how do you go about ensuring compliance with all these regulations?

We have very high standard on securing the data that we collect from the various sources. So not only we're HIPAA compliant, we're also, you know, hitrust certified. Just as an example, like certain techniques we use to protect those information is even when the data comes in to hit our server, the very first step we do is to separate the phi information and the claim information of that individual into two separate repositories and using a common person ID to link the two information together. So the phi is stored in a completely separate storage place compared to the claims information. That way, as another layer, we're trying to further protect the information become identical in the unlikely events of things like certain data being breached.

One thing that you've both mentioned a couple of times is the idea of quality and cost being separate things. So I think a lot of the focus on this has been let's try and reduce the health care costs for people. But can you also use data in order to improve the quality, quality of outcomes as well?

Absolutely. I think that's one of the major themes of the industry as a whole right now is figuring out how to best use data in a way that's meaningful to improve outcomes. Our focus as a company really has been, I would say, primarily around providing cost containment, waste and abuse solutions, network capabilities, and now we've added the decision science piece. But we're also, you know, very much want to facilitate anything that we can do to help the providers and clinicians and otherwise figure out ways to use data in a meaningful way. I mean, I came from the Mr. Background, so I don't, I don't want to talk too much about where I was, but you know, there's an incredible amount of data that's generated and there's an insights and technologies and rules and capabilities that are there that ultimately allow for much better care and capability and early detection of things that never existed in the past. And the thing that's most interesting is it brings protocols forward in a way that in the past someone would have to cognitively process all this, but now they actually can think about things in a very different way because it's being brought forward than having to go find it. So it's kind of changing the way work is done over time. And I'll just say this, Richie. We, we want to serve the whole ecosystem of healthcare. So we actually, we want to work with rural America to look at the idea of maybe helping them figure out ways to create better access to care in rural parts of the country to help keep costs down and quality up. We want to bring those transparency products to the full continuum of care. Actually, as we go forward to kind of help with some of these problems we're just discussing, I like the idea.

Of just helping out rural Americans. Have you got any other success stories where some projects have worked and it's.

Improved people's healthcare as a data platform where we have all these prediction powers. One example is using the claims data and then we predict those high likelihood of people of getting MSK surgery in the next 12 months. That way when the clients have a MSK Point solution vendor that specifically support the members and help them guide and navigate them through the whole journey of their surgery events by knowing who those, you know, like high, high probability members are getting those surgeries are extremely helpful and they improve that outcome tremendously because rather than trying to, you know, like wait for people to contact them, they actually know who Those people are so they can proactively as a point solution vendor, contact those members that are having the conditions and then understand which stage of the disease they're in and then help them find the optimal care pathway for their MSK conditions. Some may not even need their surgery. Some may need the surgery. So you need help finding the right provider, combining cost and the quality information. So having the data and the predict and then trying to early intervent and using the data further to find the right combination and the cost and then the quality providers. I think that's the type of problems using data we're trying to solve together.

I think something Jocelyn, we talked a little bit about it. I think she hit it really well. I mean, that's a great example. And I think what I believe is with the future of healthcare and data will really be moving from sick care, which today it's like, I'm sick, I better find someone that can help me or maybe I should show up somewhere if I'm really not feeling well, to more of a wellness point of view where you can look at data, look at information. You can have early detection or prediction of the conditions that may present themselves and then you can target certain groups of individuals or certain conditions and then you can do things around that to create wellness. And you see that with registries which get putting together registries based on data that you can actually reach out to people versus waiting for them to show up when they're at their worst. And that's a big change that I think can happen over time in healthcare that really would make a big difference for the world's population in terms of health and wellness and benefit.

That is really interesting because yeah, now I think about it, don't even go to the doctor when there's a real problem happening. And if you can intervene earlier, then it's probably going to be cheaper as well as a better experience for whoever's sick. Now this all sounds very exciting and I'm sure there are a lot of people who are interested in working in this field. So if you want to get involved in like data science or data analysis in the healthcare field, what sort of technical skills do you need?

Well, I'll start and you can say what really is needed, Jocelyn. So I mean, you're talking to someone who learned to code cobol. Okay, that's my expertise is COBOL coding. But nonetheless, I mean, look, you know, I think it's again, fascinating area as opposed to some other interesting areas of technology. I personally believe you can put a mission to it. Which is cool. Serving health and wellness and otherwise around the world is an important thing to do. So you can get behind that. I think healthcare informatics will be just a huge need for that, particularly in the United States, but also NHS and other parts of the world. Having worked in those health systems myself across the world, data science, analytics, machine learning, AI, the attributes of intellect and problem solving. And I would say curiosity will be important. So I think employers myself look for attributes as much as skills. And how you show up in a way that indicates that you'll grind out when things aren't going well, but you'll also seek new capabilities and skills over time are really important. And Jocelyn could probably speak much more eloquently to the technical capabilities needed.

Sure, yeah. So it was really interesting that it's all about attributes as much as skills. But Jocelyn, do you have anything to add on technical skills that you need for working in this field?

I think on top of that, just maybe a little bit of basic understanding of the US healthcare system, like a little bit understanding of some basic coding or whether it's like the diagnosis coding system, the procedure coding system, just so you know, like how things are being categorized or being coded or being interpreted to better apply the right technologies onto the right data. I think that two combination together will really power up the way that we leverage data and then make it super actionable and predictable.

Okay, yeah. So just understanding what these disease codes actually mean and, and how they relate to like a real condition. Okay. Now we've talked a lot about the US healthcare system. I am curious as to whether any of these sort of approaches you're taking are applicable to other medical systems around the world, like within Europe, for example. The structure of the payments is very different from the us so how globally applicable is all this?

I had the opportunity to cover global healthcare for Oracle, which I'm sure many folks know as a big technology provider. So global healthcare operates differently in terms of how the payment system works. A lot of times it's a, it's more of a social approach to it based on taxation, it's government driven with some, some private option typically available at times. But all the same problems still exist in many ways. Quality, wait times, high cost, readmission issues and otherwise. So I think a number of companies are working on that for us. You know, we're primarily focused right now on the US and the work we're doing here. I do think over time, as we mature as a company, as we grow and as we develop that a lot of the work that we do particularly direct to employer or through other channels to employer where we use our solutions for reference based pricing. We're looking at claims, we're looking at costs, we're looking at prediction. I think those are very relevant for employers around the globe. There's just last thing I'll say about that Richie, is that there's a massive sometimes barrier to entry for companies to do work in other parts of the world. Almost every region has their own set of quality standards, cybersecurity standards, data rights and usage restrictions. It's high cost and sometimes it's a big barrier to entry actually for innovation. So I think it's something interesting to think about. That policy really could help move forward the industries across the world if we had ways to better to reduce the cost of entry for technology companies and businesses. Different topic, probably for a different day, but something to think about too.

Yeah, I can certainly imagine how the data cleaning would become dramatically harder if you've got different diagnosis codes for different countries and things like that and different data formats as well. All right, so maybe a plan for a different company in a different country then.

Yeah, you probably won't invite us back, but if you do, we'll talk about that too.

Super. So just to wrap up, what are you most excited about in the world of healthcare and data at the moment?

I think I'm super excited about these, I think GPT like artificial intelligence technologies and what can really lead us, just like what Travis mentioned earlier about the doctors and providers these days are using these ambient technologies to really rather than type, hand type those EMR information when they talk with patients now is using technology to translate those patient conversations into a digitalized standardized data sets where you could really leverage that, normalize the data rather than messy unstructured data to further predict and better prevent certain condition from worsening or happening. So I think that's what I'm super excited about in this industry.

Yeah, I think we talked about a little bit. This idea that you can start to focus a little more on wellness than just sickness I think is important, really giving individuals the ability to have some autonomy over their own health and wellness and choice is important. So I think you know that access autonomy, understanding your choices through transparency, frankly can be a comforting thing when you're in a, in a moment where, in a confusing process where someone you know or love isn't feeling their best is disconcerting. And so I think there's an emotional element to it as well over time where, where predictive models can help with wellness. But we're also bringing cost control, we're bringing other transparency solutions to the table that just make it eat. Just. Just makes it freaking easier. It's hard, man. It's hard figuring out this discussion today, you know, I mean, we live our lives doing this every day. I can only imagine how the average consumer feels sometimes. So kind of demystifying it, making it simpler, making it easier, making it less cognitive burden to have to figure out where to go and what to do. I'm excited that I think that the next decade in healthcare is going to be the decade of true and meaningful innovation, and it's going to really come to the market and it'll be through technology. So that's a pretty cool thing to be a part of.

All right.

Yeah. Just empowering people to make decisions about their own healthcare. That sounds like a wonderful goal. Excellent. All right, thank you so much for your time. Jocelyn and Travis. Yeah, lots of insights there.

So it was great. Thank you.

Yeah, thank you very much.