0:00:10
Real world evidence or RWE is a promising way to answer questions about how therapies perform going beyond the controlled confines of randomized clinical trials with narrow patient populations to analyze what happens in the so called real world. But the first generation of RWE hasn't always lived up to its promise. It's often incomplete, inaccurate, or is just not fit for purpose? Hi, everyone. I'm David Williams, president of strategy consulting firm. Health Business Group and host of the Health Business podcast, a weekly show where I interviewed top health care leaders about their lives and careers My guest today is Dan Riskin, CEO of Verantos, which focuses on providing high validity, real world evidence at scale to realize the promise of RWE. If you enjoyed this episode, please press that like button and subscribe. Dan, welcome to the HealthBiz podcast. Well, thank you so much for having me, David. Usually, the first thing I have to do is see if a guest needs to correct any mistakes I made in the interest, so you let me know if I did didn't there, but That was perfect. Sounds good.

0:01:11
Let's start off with how you got to where you are today. Interested in your background, your upbringing, any childhood influences that have stuck with you over time? Well, I think that the decision to go into healthcare at an early age was a big deal for me. For me, I had an equal love of computers and healthcare. And I found that I really wanted a career in both, and there was nothing available to do that. So the formal training for healthcare was certainly the lengthiest. And the decision to go into that gave me a fundamental background that was useful for the rest of my career. I'm not going to say that healthcare is the most advanced technology area in society. It certainly isn't. But it offers an opportunity on the field offers an opportunity to help people individually, to help people globally as a population. And it's been really spectacular.

0:02:14
So what was your education? What what did you go into? And how did you decide and kind of pass it to take along the way? So I I was originally studying biophysics and loved the combination. Had a little bit of a meandering path educated between UCLA and San Diego trying to really experience California and enjoy it, went on to medical school in the turn of the millennium. And love medical school. Just enjoyed it. Enjoy the immediacy of caring for patients. And within medicine, loved surgery because you could really fix something. You could do something, check how it worked and know. And I think that stayed with me through my career.

0:03:12
So I did nine years of residency and fellowship after medical school. Between a number of institutions. My I've completed my residency at UCLA and my fellowships at Stanford. And and and just love the immediacy of it. But as I was going through, it found that I missed the technology work I had done. Even early in my career after my training, I remember a patient who has stabbed in the chest and they had a hole in their heart and I opened their chest and I fixed the hole in the heart and this this kid went on to live the next day I had a conversation and I was I thought it was so powerful to be able to do that. What an amazing thing for society to allow you to do that? And in parallel, my research was technology and artificial intelligence. This is well over a decade ago. And it was the early days, and I was thinking it's also equally cool to be able to influence an entire population to find good data and create evidence and tailor healthcare. So I've had a career that's really parallel path to those Over time, in truth, it's been more and more technology and business because that pulls you in. But the foundation of working with individual patients will never leave me. And I continue to practice what I can to contribute to to individual care. Great. So talk about Health Fidelity.

0:04:47
Health Fidelity was an incredible company. It was named as a combination of Fidelity, which is faithfulness to healthcare, and Fidelity, which is Fidelity of data. It was my second company. It was my last company before Verantos. And the idea was to understand outcomes and risk, effectively understand quality and health care. I started working on a company in two thousand eight, At the time, I had the pleasure of being on the Obama health advisory committee for his campaign, so I saw where national policy was going pre election.

0:05:25
I was working in a venture firm as an entrepreneur in residence, so I knew that I could start a company and have support. And I saw a future of value based health care. The idea that we could use data to understand what was worth it. And I paid for what was worth it. And this this has become a trend in my career, which is using data to understand what's working and what's not working.

0:05:51
Health Fidelity worked in the Medicare Advantage space and then moved on to the rest of value based health care. It turned out that no one wanted to pay for an understanding of quality, but they sure wanted to pay for an understanding of risk because that was how value based contracts were executed. And so there was a market there. And, you know, early more early learnings in my career being CEO. That was the first company of which I was CEO. That you really need to understand the market to grow a business. You can have the the best intentions and the best goals to improve care and benefit patients. But unless there's a business there, you won't be able to have a revenue stream to allow you to scale and I really wanted to do things that could work at scale. Makes sense. How about doc doc? I love the name of that you know, it sounds like I can't decide whether it's like duck duck goose or knock knock so. So I'm I I doubt that's what you had in mind, but what is it?

0:06:51
So as I was completing my work at Health Fidelity, I was invited to work to get a national policy. This is now in two thousand and fourteen. I was invited to testify on an initiative called twenty first Century Cures. And Congress, this time, invited me, it wasn't the administration. And I got to talk about the future of healthcare in the US. At the time, I had a crisis, a personal crisis to say, well, wow, I just, you know, I'm transitioning into the next thing. It's hard. It's hard.

0:07:25
I actually a second company of Fidelity was very successful and, you know, ultimately sold very well, and and I was so proud of it. And yet, that company grew from being an adolescent to being an adult. It was continuing to grow on its own. I was moving on I was pulled into national policy and I said, god, I wanna work globally. I wanna have an impact globally. So I took a position as an advisor at Apple. I took a position as an as an advisor at Google. I got to work on on everything from first party apps to to trials that these groups were running.

0:08:08
And I wanted to explore building a next company, and I tried a number of things. I had friends who had sold a company and we tried and I supported them. I had a friend from business school who had built a company called Dock Dock, and his name is Cole Surchak. He was a successful investor in Asia to Monza and he was building this company in Asia that could measure quality and improve care for patients. And I thought that was spectacular. So I started helping him in those days back when I was test twenty first century cures and building my next company, Verantos, and helping friends. And I've I've stayed with it I've stayed with helping him that company has been successful in Asia, and there's nothing like a global perspective.

0:08:57
It may be that that the US has uniquely suited data and technology based on all the efforts that's gone into it. But the population and the needs through the world, particularly in areas that are developing or growing rapidly, it's spectacular. I think one of the things that U. S. Healthcare lacks is actually that global perspective. Certainly, the healthcare systems are very different in in different parts of the world. A lot of it has to do with sort of historical and cultural reasons, so they can be hard to compare. At the same time, it's you know, it's much less expensive everywhere else. And although everybody elsewhere complains about, you know, quality and and cost as well. Somehow they're doing it for half or less, so it's at least probably worth understanding what's going on.

0:09:47
So if you say, well, let's say if even if we didn't reduce what we spend. If you took away the constraint of cost and applied some of the other things that are going on, like, what should we be able to achieve with what we are spending? Yeah. What we found in our internal data, and now this is between Verontos and Health Fidelity and Dock dock. What we've been finding is in the US, there are some infrastructure challenges. The payer system, the way that the provider system is set up. There are some limitations. But the data is spectacular, the technology, particularly with the advent the compute power of the cloud, the advent of artificial intelligence that really is usable in healthcare. And the the newly available data based on the high-tech act over the last decade going to more than ninety percent EHR adoption.

0:10:39
We've got unique capabilities in the US. If I look globally, some of the power is the scale of the population. There's high variability in quality. You know, someone in, let's say, Indonesia goes to a orthopedists for a knee operation. In the US, there might be a half a percent or one percent chance will never walk again. And the delta between the best provider in the U. S. And the worst provider in the U. S. Or top quartile and bottom quartiles better said. Might only be a two x difference. Right? That's meaningful for a patient. Yeah. Two times the chance I'll never be able to walk again.

0:11:20
But if you look at other areas, more developing areas, it can be a five x or greater difference between the most trained doctors and the least experienced doctors in the space. And so you do see a real need to identify quality and to send people to the to the doctors who are best suited. That is what doftop does, by the way. And I think it's a real need. These startup companies are hard, hard building a startup company, hard capitalizing it, hard growing it, But if you pick areas where you really benefit patients and where there's a business model, I think the entrepreneurs stay with it, they do very well. Even through the ups and downs of the economy. Let's talk about Verontos now and start with the unmet need. Clearly, something that you'd been interested in all the way through from education to clinical practice to companies that you founded and worked with an advised So coming to Verintas, you must have identified something pretty important to go after. How did you think about that? I think so.

0:12:26
So when I was working on policy in two thousand and seven and two thousand and eight on the Obama health advisory committee, we had a we had health IT sub committee and we're discussing what the future might look like. And one of the discussions was around the decades of health data reform. We knew health payment reform was contentious and difficult and unclear or future. Health data reform, people are very supportive, everyone on a better data in healthcare. If you look at the big health data acts you know, the macro and twenty first century cures in twenty fifteen and twenty sixteen, they were bipartisan. So the discussions back in those days, two thousand seven, two thousand eight, was envisioning a first decade of health data reform that just got information in the system. At that point, we were sixteen percent EHR adoption that was minimal data get the data in the system, allow us to implement an existing standard of care. That was called population health disease management. I'm not going to say we did it perfect in that decade, but we got a lot better. The second decade twenty twenty to two thousand and thirty was supposed to be a lot more powerful. Data in the system better compute power, better technology, we could not just implement an existing standard of care defined by randomized trials, we could create a better standard of care. We could start to tailor therapy. We could figure out what worked best in which subgroups. And we're really just embarking on that journey.

0:13:51
The fear when we were working on twenty first century cures in twenty fourteen and twenty fifteen was passed in twenty six seen. The fear was that this area of using routinely collected data would be done badly. We had decades of experience of using routine data badly. Yeah. Hormone Replacement Therapy where a small set of real world evidence used to define care for a large population, and many people got treatment that didn't benefit them and increased the risk of cancer. We saw even later in COVID, early information was so confusing. New England Journal and Lancet needed to retract early publications because they said the data wasn't good enough. And so the fear in two thousand fourteen and two thousand fifteen was that we would allow all this data to be used, but the data wouldn't be good and we would figure out the wrong answers and we would treat patients in the wrong ways and we would hurt people.

0:14:48
In twenty first century cures, we included a clause that said, the FDA will start to use real world evidence and said that the FDA will create an evidentiary standard, will put out a framework and guidance. On what that should look like. The idea being that we would have an entity that was so good, an agency that had created evidentiary standards in other areas, they would say this is what good real world evidence look like. My fear in those days in twenty fifteen and on was that even if we created the evidentiary standard, no one would be able to do it. Yeah. Meaning, we had a lot of legacy firms that had used low quality data, minimal technology. And I was worried that there wouldn't be anyone that would say, yes, I can pull the torch and really create high quality data and high validity evidence. And so I started a company to make sure someone would be doing it. And that was called Monta. That sounds it sounds good. You know, I I've seen this term high validity real world evidence, and I can think I understand from your description of the need about why it's needed. But, you know, why isn't that the norm? Like, why why do you even have to say high validity as opposed to just you know, RWE. Because it's hard. Yeah. It's harder to do, which means it's expensive to do. Yeah. On If you look at the standard, so Verontos has become an FDA demonstration project to help in defining guidance for real world the draft guidance is out. The draft guidance is used our ideas and and ideas from other demonstration projects.

0:16:25
And the standard is is increasingly becoming accuracy completeness and traceability. Accuracy is tough because If you want the most accurate information, you don't go to billing information. Billing information wasn't intended to be detailed, granular accurate. So if you want the most accurate information, you go to the information intended to convey clinical issues, which is doctor Meredith describing care. Hard to work with. So accuracy, you have to go into the narratives, you have to use technology to pull out information, and you have to check how good your technology was and say what level is good enough. You know, we use eighty percent. The industry standards closer to fifty percent. Yeah.

0:17:11
If if you want completeness, you need to have multiple lenses into the patient. Lenses mean you have the clinical care, which is the electronic health record, you have what drugs they were given, which is pharmacy claims, you have what outcomes happen, which is often in billing claims because that would include if someone had a heart attack in the other side of the country, It's national and a death registry. And so completeness requires linking all these different data sets and traceability just says that you're not gonna mess around. If you say that a patient has a certain disease, might bring with aura or severe asthma, you'll be able to track it down to the initial information that made you believe that. That is hard and expensive. You need a bunch of different partners and technologies. It's not the norm. It's far, far, far from the norm. So if you think about all those challenges that are out there and what she would do about it, I mean, as you say they're expensive. It's it's difficult.

0:18:11
Others know about these things because they've partly they've read what you've, you know, they've read what you've written or how you've testified and they've had their own experience using something that turned out to be it showed you one thing and it leads you off in a wrong direction. Hopefully, not down like the HRT path, but just even in trying to go and validate something. And so what do you do differently? What, you know, what makes it so that ranches can be successful? Is it that the data is better? Do you have better technology to make sense of the unstructured data? Is it that you've got, you know, epidemiologists and bio statisticians that are just more that are just deeper in their therapeutic area. Like, what what is it that makes Verantos be able to go beyond what you could anybody could sort of copy onto the website, right, and say, high validity.

0:18:53
The same way that in health fidelity in the early days, we had early visibility as work in the Obama health advisory committee in the campaign and we really tried to work on Medicare Advantage as it was becoming a growing area. And the same way that in twenty in twenty first century cures, we had visibility into a framework that would be developed to define high reliability or high quality data. Verontos has been doing this now since twenty fifteen. So we've been doing this for almost eight years. Actually getting to high quality data and high validity evidence. We've not only helped define the framework. We've known for all those years what the framework be. So we were some of the first groups to measure accuracy and completeness and traceability, and we figured out what would be needed. The contracts with the health systems, the technology to pull out of unstructured data, the methodology to measure accuracy. So I think we've been successful because, well, the team has done it ever since early days from early in our careers. And we've had a lot of time in front of us to figure this out now. We spent three years building technology we commercially launched in twenty eighteen. Yeah. We built some tough stuff and then we refined it. We distilled it to practice over the years.

0:20:21
You talked about it. I I think I've seen on the website discussing the use of AI as part of what you're doing with RWE. And AI is the sort of thing that, you know, it's it's exciting. Seven or eight years ago, people are talking about it a lot to the point where you start to rule your eyes when you hear about AI. And now all of a sudden, in the past few months, you know, with chat GPT and other generative AI that's that's come out. Now all of a sudden people are kind of interested and not just that, but saying, gee, their whole world has changed now. And in six months or years, it's gonna look totally different. And I even think that some companies that have been funded more recently are saying, oh, wait a minute. If this was around, you know, it's just a general purpose thing better than all the stuff that I've that I've built. How does what you do with AI differ from, you know, what a generated AI like Chegg T would do? And how do you or will you incorporate these new available technologies? Yeah. I think we're well positioned to incorporate new technologies I think that part of the challenge, there's a lot of buzz in our industry.

0:21:19
There's quite frankly buzz in the generative AI, but there's also buzz even in let's say, in real world evidence, people will say, oh, yeah, I've got access to tons of data and some narrative. Well, when you dig into some narrative, it's, you know, tiny. And then they're pulling pieces out, but they're not measuring accuracy. And then they're saying that natural language processing. But natural language processing, particularly traditional, NLP, is limited. It's limited to the sentence foundry. I'll give an example of what I see happening. So in real world evidence, NLP natural language processing is considered state of the art. In my mind, it's necessary but not sufficient.

0:22:00
Let's say we're doing a study on migraine. And let's say the doctor wrote in a patient's narrative, patient with MA. And NLP is limited well, traditional NLP is limited by the sentence boundary. It doesn't really take into account a ton of other content. And so it says this could either be migraine with aura, Medicare Advantage, Mass, or medical assistance. Yeah. And, you know, that's where you get into accuracy problems. Who knows? When we do the work, we go beyond an LP to use AI based inference where we're taking into account the rest of the record. So if there are if there are medications that are appropriate for cancer and there are tumors and other things that's that MA is probably a mass. If on the other hand, we're seeing headaches and photophobia and medication for migraine, that MAs probably might have been with aura. If we have a study on migraine with aura, that's very important to us.

0:23:01
Similarly, we see we see groups working with claims data and heart failure where they're saying, okay, the subtype of heart failure we're checking is heart failure where the heart doesn't pump strongly, heart failure with reduced ejection fraction. And the doctor doesn't ever write heart failure with reduced ejection fraction won't take them too long. And so you need to pull out pieces. You pull out a piece that says this is heart failure or that suggests heart another piece that shows an ejection fraction that's low on an echocardiogram. This is far beyond natural language processing. This is really inferring It's AI based word. And it's something that chat, GPT, or other things, they're not well suited to look at many different pieces in a medical record and put it together and infer and maintain accuracy, completeness, and traceability. It's it's specialized work.

0:23:56
Speaking of specialized work, a lot of RWE companies I'm familiar with tend to focus in a specific therapeutic area, most often oncology, I think, because that's where a lot of the money is and the opportunity. From what I can tell your broader base, is that right? And and how do you how do you think about that? We feel strongly about it. We we we think that high validity, real world evidence really started in oncology. And we think it's very good.

0:24:25
There are a number of firms that do this. They do it manually. They do small studies, you know, twenty, fifty, a hundred patients They pull pieces out of records with two annotators on every record pulling little pieces out. It's very good. You can create high accuracy. You may not get completeness. You may not have linkage. You may not have traceability very easily. You can create high accuracy. It was the first generation of high quality data in real world evidence.

0:24:54
So when we looked at it, we said there were two problems. One, they can't leave oncology. Everything they've developed has used oncology data. That's how they negotiated their contracts. They they pull out the pieces in oncology, that's who their annotators are. The second piece is they can't get scale. If you're not using technology, you know, you may have five hundred to a thousand annotators in your company pulling out pieces. Yeah. You if you wanna have patient sets that are ten or a hundred times as large as you would need for heart failure or neuroscience or other areas you just can't hire a hundred thousand annotators. Yeah. So, yeah, we thought there was a real need and we're fulfilling the need in those areas that are beyond oncology. We're comfortable working on oncology also. It's just the biggest need is those areas that are unserved that require scale and there's just no one who can achieve accuracy, completeness, and traceability at that scale.

0:25:59
When I think about health systems, you know, and other data sources, they they start to understand, gee, there's value in data, at least they realize we're generating data, there should be value here and it may be more thoughtful even more thoughtful or just saying, gee, if somebody wants to come and and leverage I don't know about about that. Maybe I wanna, you know, get more value for myself. How how does the RWE ecosystem and maybe Verantos specifically encourage these data producers to allow, you know, data to be used for the kind of benefits that you're that you're describing Yeah. So patient first, we want to be rigorous in privacy and security. And that's our first standpoint. We're high profile from testimony before Congress to FDA demonstration projects to working with multiple top ten life sciences firms funded by NSF, NIH, or high profile. We can't afford to be on the front page of the New York Times harming patients any more than they can. We wanna be on the front page of the New York Times saying that we are doing the most advanced research in the world to benefit patients the most strongly. And so that's the starting point of a discussion with the health system.

0:27:17
The next point is we wanna be good actors. We don't wanna make money from health systems. That's not quite frankly, they're they don't have that much money to work with for this kind of effort. They were never the ones who funded the randomized trials and RWE is similar and that it's creating evidence. Any one health system doesn't have that much incentive. Compared to let's say life sciences where they have all the incentives. They were the ones to fund most of the RCTs. They're the ones to fund most of the RWE. So we don't want to make our money from health systems, and we want to benefit them, meaning if they can optimize care regardless of which drugs they're optimizing for, we want them to have the data. So if we have the ability to work with their data set, pull out the pieces and understand what's working and not working. We wanna feed that information back. We've had a great success with our network. Feeding the information back and allowing them to identify areas where they can improve their care and they have. And it delights us to know that that's another avenue for benefiting patients. Great.

0:28:27
The RWE is a business that has gotten a lot of tension over time, and there's been a lot of spending on it both from the investor side. And then also, I would say from the customer side too, largely on the life sciences side, but there are not a lot of companies that are very profitable in the space. Some of it has to do with the Army or, you know, I don't know what kind of a force you would describe hundreds of thousands of people doing annotation or whatever whatever the number is. It's high cost is that you described. It's expensive to actually get data that's that's good.

0:28:56
How do you think about it kind of innovation on the business model side? I mean, there are sometimes these sort of short term projects can be done. There's sometimes subscriptions. There are things that are done at the, you know, at the individual product level and somewhere at the enterprise level. How do you see what have you tried to do and what do you see as being the way to success there?

0:29:19
So the first segmentation is REO Pharma Services firm or Pharma tech firm. I don't believe you can be both. The contract research organizations are pharma services. You can identify it at typically not massive growth. It's typically not large margin. It's typically project based. The pharmatec firms are our higher growth, higher margin, some portion of our annual recurring revenue. I think that you need to know what you are There's nothing wrong with either of those segments in the market.

0:29:50
If you are a pharma tech firm, you're tech based. What you're building as a platform and applications on it, and you can be efficient. And you can, yes, provide services on top of it. But largely the bulk of what you're doing is using your platform and application to deliver value. And so we started from that perspective. If you look at firms that are largely manual, that's a hard transition to make from a business model where you have an army of annotators and you're fulfilling projects to being a a tech company.

0:30:27
So I can't speak to the pharma services side what their path to profitability would be. I could say on the I mean, on the pharma services side with the path you have for the building. But on the pharma tech side, I can say that that if if you do it right, you can stay cash efficient, you can be largely breakeven. You can have your path to profitability. You can be, in my mind, a good business with good growth and good margin. And I think our customers like that about us -- Yeah. -- which is we don't feel they don't feel that we're subject to the whims of the market. Yeah. You know, the last the last year has been rough in the market and firms that are bleeding money have been penalized. Yeah. Clearly. Yeah. Then we're we're not as subject to that. We're more privately held. We're more rapid growth. We're not losing a lot of money. And I think that makes for a good business, which has been a priority of mine for all of my businesses that I've grown.

0:31:32
Dan, clearly, from, you know, as you were just even getting into your educational mindset and then parts of your you've always been looking forward and you you were talking in terms of decades and describe what was gonna be this sort of twenty twenty to twenty thirty decade. Help paint a picture a little bit looking ahead for Verint just five years, ten years. You know, what's the vision for where you for where you get to over whatever use sort of long term time frame makes sense to think about? Fifty years. Okay. Fifty years is the time frame. So there's been a lot of research on industry disruptions that come in fifty year cycles. Most industries have them.

0:32:10
In healthcare, it's very obvious what the enabling factors are in the major shifts. If you look at the eighteen seventies and public health, it was journ theory and pasture and the biggest drop in mortality, and preventable mortality ever in healthcare. Fast forward fifty years, nineteen twenties, discovery of Penicillin, and everything that went into that, medicines that really were Yeah. The nineteen seventies, evidence based medicine using randomized trials to create medicine as a science rather than an art everything that had gone before with speech to report has syncing declaration, all of the enabling factors. Well, now it's very easy to see what the enabling factors are for the next fifty year cycle.

0:32:57
We have data availability. We have compute power. We have technology in the form of AI. You can see where it's going. Yeah. You can see it in other industries. You can see it in Google Maps. It's changed all of our lives. In Spotify, in in any of the systems, Amazon, that's that's that knows us. Yeah. Right.

0:33:19
In healthcare, of course, it should be that way we should learn from every patient interaction. It's insane that we don't. So a health system where we can learn from every patient interaction, where instead of giving everyone with high blood pressure the same drug because that's the one we use last out of our twenty choices and it seems good. Yeah. One, that's that's not right. That's not right. Of course, we should take advantage of the technology data and and compute power and learn and tailor that's the future. And, Rontellus wants to be a meaningful part of the future to help make sure we do it right. The scariest thing would be to do it wrong. Yeah. I'm we're learning, but learn the wrong lessons on bad data and and hurt people. But if we do this right and create the right frameworks and have companies that are really working hard to believable data and credible evidence, I think the future is extremely bright in healthcare. Great.

0:34:16
Well, the last question I have for you, Dan, is when I ask everybody, which is whether you've read any good books lately and anything you would recommend either from things that you've read lately or from longer ago? You know, I've it won't surprise you to learn that I love science fiction. Yeah. I have a geek in heart. And I recently reread Neil Stephenson's snow crash. And I look at his view of AI from decades ago, and it's just powerful. It's it's lovely to see what's come to fruition and what's still what's still pending. I won't say it didn't happen because it might happen soon. Yeah. Well, then then, I I just love what's happening in technology and how it's revolutionizing so many industries. And I enjoy that in my books too. It tells you that I probably don't stop working that much. Yeah. But it's good. I love this stuff. That's okay. Well, Dan Ryskin's CEO of Verantos, sci fi fan, futurist, Thank you very much for joining me today on the health of this podcast.

0:35:24
David, it's such a pleasure. Thank you and thank you to the listeners. Appreciate your time and hope that com, we all can make a better world together. You've been listening to the Health Biz podcast with me, David Williams, president of Health Business Group. I conduct in-depth interviews with leaders in healthcare business and policy. If you like what you hear, go ahead and subscribe on your favorite service. While you're at it, go ahead and subscribe on your second and third favorite services as well. There's more good stuff to come and you won't want to miss an episode. If your organization is seeking strategy consulting services in healthcare, check out our website health business group dot com.