About the Episode
Unlock the future of clinical research and healthcare with RespondHealth! Join the HealthTech Remedy hosts as they interview founder and CEO, Dr. Vicki Seyfert-Margolis, a visionary who has shaped regulatory science at the FDA and is now transforming real-world evidence and clinical trials.
Discover how RespondHealth is bridging the gap between clinical practice, real-world data, and traditional trials using cutting-edge technology. Learn about their tech-enabled research platform designed to address the inefficiencies of traditional trials and collect real-world data through innovative methods like apps and integrations.
Dr. Seyfert-Margolis shares her journey from early entrepreneurship and foundational work in immunology and clinical trial innovation to spearheading regulatory science and digital health at the FDA. Hear fascinating insights from her time influencing policies around real-world data and the push for comprehensive endpoints in drug development.
Dive into how RespondHealth is built on patient centricity and serving underserved communities often excluded from traditional trials. Get the inside story on their powerful partnership with Harris Computing, providing access to vast amounts of electronic medical record (EMR) data, including unstructured notes, from over 200 million patients.
Explore the transformative power of their platform which integrates AI models directly into this massive EMR dataset. Learn about the groundbreaking use of a knowledge graph to extract, abstract, and connect complex data, enabling large-scale analyses to be performed in minutes instead of months or years.
Dr. Seyfert-Margolis explains how agentic AI allows users to ask sophisticated research questions in natural language, watch the AI write the analysis code, and get results rapidly. See how this technology is poised to become the "next generation EMR," empowering physicians with rapid, evidence-based insights for "smart prescribing" by querying real-world outcomes from representative patient pools.
Hear examples of their work, including a large model built for incretin/GLP-1 research, revealing critical real-world findings in diverse populations. Discover how their strategic partnership with Microsoft Azure is essential for accessing the latest AI capabilities, compute power, and scaling expertise needed to innovate in this space.
This conversation is essential for anyone interested in the intersection of health technology, AI, real-world evidence, clinical trials, and the future of data-driven healthcare.
Don't miss this insightful discussion with Dr. Vicki Seyfert-Margolis!
Transcript
Dr. Timothy Showalter:
[0:00] Good afternoon, guys.
Dr. Paul Gerrard:
[0:01] Good afternoon.
Dr. Timothy Showalter:
[0:02] Trevor, I'm in your home state, man. I'm in North Carolina right now.
Dr. Trevor Royce:
[0:05] Oh, you have picked a tough time of year there. It's probably four to five centimeters of yellow pollen across the.
Dr. Timothy Showalter:
[0:11] Entire state right now. No different in Virginia. Although my biggest allergy right now is the red meat allergy, the dreaded alpha gal. So that's what I'm hating on. Paul, where are you?
Dr. Paul Gerrard:
[0:21] I am up in Maine. Yeah, just rain, rain and cold. Let's do this. Let's get into it.
Dr. Timothy Showalter:
[0:26] That's all you have?
Dr. Paul Gerrard:
[0:28] That's all I have.
Dr. Timothy Showalter:
[0:29] Let's kick it.
Dr. Trevor Royce:
[0:31] Welcome to HealthTech Remedy, the show where three physician leaders in health technology tell the stories of new and established companies and interview leaders from the industry. I'm Trevor Royce, radiation oncologist and researcher with experience in real-world evidence, informatics, and AI diagnostics.
Dr. Paul Gerrard:
[0:45] I'm Paul Gerrard. I started off as a physical medicine and rehabilitation physician before focusing on reimbursement policy, molecular diagnostics, and market access for AI products.
Dr. Timothy Showalter:
[0:53] And I'm Tim Showalter, a radiation oncologist and prior medical device entrepreneur who is now focused on bringing AI advances to cancer patients.
Dr. Trevor Royce:
[1:01] This week, we're discussing RespondHealth, a company transforming clinical research and health care by bridging the gap between clinical trials, real-world practice, real-world evidence through advanced analytics and AI solutions. So basically tech solutions for clinical research and real-world data. And we'll talk in more detail about what that means.
Dr. Paul Gerrard:
[1:18] We'll be speaking with Dr. Vicki Seyfert-Margolis, founder and CEO of RespondHealth, about how their platform is enhancing patient care and clinical efficiency.
Dr. Timothy Showalter:
[1:26] Maybe I can give a little bit of context how I came to know Vicki Seyfert-Margolis and RespondHealth. So we're actually, I've had the experience of working with RespondHealth and I've been really impressed about what they're up to. And so I've learned a little bit about the company. I'm excited to learn more from Vicki when we talk to her later. And I think this is going to be one where we're going to actually have to bring out our hardcore, hard-hitting investigative journalism type approaches here, guys. And we're going to have to really go straight to the source and really probe Vicki to get more context about RespondHealth.
Dr. Trevor Royce:
[1:59] That sounds great, Tim. I'm looking forward to the interview with her because, you know, in a prior life, have worked on some projects with her and she is very impressive. So I would never bet against Vicki. I look forward to hearing what she has to say.
Dr. Timothy Showalter:
[2:09] Fantastic. Well, I can give a little bit of like information that I was able to find through the power of the interwebs. So what I do know about RespondHealth is they were founded in 2019.
Dr. Timothy Showalter:
[2:21] Vicki is the is the CEO and co-founder. And their idea is to really focus on tech enabled research. They've got an interesting model for registry studies and to marry that with real world data to really set up an improved focus on data access. My understanding is they have a real strong focus on medical device companies. And, you know, like for us, for my project I'm working on with her for diagnostic companies to maybe deliver data from clinical venues that are not otherwise well served by, you know, companies that might focus on, you know, pharma facing projects for metastatic cancer patients or other scenarios. So their idea is really to fill that market need and bridge that overall gap. They have developed some proprietary AI solutions to focus on delivering real world market intelligence and data sources. I've met some of the key leadership team, of course, Vicki Seyfert-Margolis, Tom DeLay, who is the chief business officer. I've also met Sammy Mentis, who's director of operations, and then I've not met, but see from the company website. There's also chief scientific officer, Dr. Ed Kim, and Rich Fody, who's director of epidemiology and analytics. So I think they've got a small but experienced team.
Dr. Trevor Royce:
[3:41] Yeah, we could take a deeper dive into the problem that they're trying to address. And the sort of known truths are that, you know, clinical trials are very expensive and inefficiently run. There's a lot of sort of outdated methods of data collection and capture, like using paper forms. And so there's this whole ecosystem coming on the board where you have these tech-enabled clinical research solutions. They're doing like interesting technology things like building apps for data acquisition. They are doing things like leveraging data that's collected during the routine course of life like real world data sets to collect vitals through technology collect you know glucose monitoring through technology and so forth so basically you know a technology-based platform to help facilitate the execution of clinical trials and I think to your point we'll have to hear more from Vicki but they're sort of heavily indexed towards trial design and execution medical devices and.
Dr. Paul Gerrard:
[4:32] It seems like their solution really focuses on the technology and integrating and aggregating information from different sources. I mean, we've all dealt with the issue that data is a big deal in healthcare. You need data to move studies forward, but getting access to data is tough. It often has to be manually abstracted from existing clinical charts, or you have to set up and run an expensive clinical trial. And it seems like Respond has created a platform that can be deployed across settings, and they are combining that with AI to try to enable integration and collection of data across these different settings.
Dr. Timothy Showalter:
[5:14] You know, as I think through what their broader solution is and what their platform is, I'm kind of interested to know what's transportable across project types and what are the solutions that they're building that scale? So I hope we get into that with Vicki. I can certainly say, you know, from my own experience, I think, you know, really building cloud or internet native registries is something that I've seen them do quite well and to be flexible with respect to surveys or data coming in from EMR integrations. And so I think they're generally well positioned to build around whatever the company needs are. But I want to know like what what happens beyond just project specific efforts and really, really take a deep dive into to what their platform can do. So, Paul, do you have thoughts about this that you want to share about potential revenue sources and this sort of business model where it's real world, but it's also, you know, some clinical research organization type type partnership services?
Dr. Paul Gerrard:
[6:15] They offer their solutions through a subscription-based model, and... In that, they are providing access to a combination of the data as well as AI to enable analytics of large amounts of data. I am assuming that some of this is going to be unstructured as well, which further raises the need for tools to create useful information out of this. And presumably, that's where the AI comes into this. They have notably gotten some important collaborators, including Microsoft Azure, to enhance their capabilities. And as you just mentioned, Tim, it seems like they're really well positioned to develop tailored solutions to particular research projects as needed and can be flexible in terms of providing something relative to a particular research project.
Dr. Timothy Showalter:
[7:05] They do a good job of thinking from a regulatory lens and from an efficiency standpoint for getting research done, like how to really stand up registries and other sort of evidence generation exercises really well. I want to understand more about their data integrations and what they've built to sort of reach in to health systems and collect insights for their research partnerships. So I'm curious to learn more from Vicki about what her thinking is on that.
Dr. Trevor Royce:
[7:33] You know, one clear thing that is very impressive about RespondHealth is the experience of their leadership team. I mean, Vicki has seen it all from the regulatory side and seen all sorts of different trials. She's a very knowledgeable and experienced about how to navigate the design and execution of clinical trials. And then when you layer on this technology component, you have the opportunity to really expedite how these trials are executed while still meeting the regulatory burdens and requirements. I'm going to sort of riff a little bit. You know, I was, when we were thinking about this interview and preparing for it, I went and looked at one of the keynote talks that she did and she made the analogy of, and flying where you go from a paper boarding pass to a mobile boarding pass. And like, you know, people, you know, my kids probably never even know what a paper boarding pass is. And I think that's like her vision for clinical trials is like, let's not see paper consent forms. Let's not see paper, I'm putting words in her mouth, but like, you know, this very clear sort of common sense move to a technology enabled execution for a very complicated clinical trial. It just makes sense. And they seem very well positioned to do it. There are some, you know, use cases that might help highlight kind of how they see this. And we should ask Vicki about this, but I'll just highlight one very quick example where they were tracking for community health workers in Texas, blood sugar control amongst the population there. And I think they had over like 20,000 participants on a platform that they clearly like kind of custom built for this exact purpose. So I think they're pretty flexible and can tailor needs to like whatever that clinical research question is that's trying to be addressed. Paul, any thoughts on your end?
Dr. Paul Gerrard:
[8:59] Yeah, well, I mean, you know, one of the things that has gotten the recent news from them is this collaboration with Microsoft, which could significantly enhance their AI capabilities. And they also have this thing that I think sounds like an interesting concept. I'd be interested to know more about it is this idea of a knowledge graph that sounds like it tries to help connect different data elements to get higher level conceptual relationships.
Dr. Timothy Showalter:
[9:24] Yeah, I want to ask her what a knowledge graph is. I will tell you, it's fascinating to me. It sounds like really can improve discovery and leverage the power of AI. I've checked it out with the AI experts I know, and it is definitely a thing. It's real. And I want to understand more about how that approach can be used to accelerate the work they're doing.
Dr. Trevor Royce:
[9:45] That was a great overview and discussion of RespondHealth. We think they exemplify innovation in health technology. Stay tuned for our in-depth interview with Vicki Seyfert-Margolis, where we'll explore how RespondHealth is shaping the future of healthcare.
Dr. Timothy Showalter:
[10:01] Dr. Seyfert-Margolis, welcome to HealthTech Remedy.
Dr. Vicki Seyfert-Margolis:
[10:04] Thanks, Tim. Good to be here.
Dr. Timothy Showalter:
[10:06] We are very happy to welcome you and to learn more about your company as well as your own career.
Dr. Vicki Seyfert-Margolis:
[10:11] I'm excited to talk to you about it.
Dr. Trevor Royce:
[10:14] We are excited to hear about what you guys are up to and what you guys are building. You know, we certainly have had the pleasure of working you before in some projects, and we came away very impressed with sort of the technical capabilities that you guys have. So we look forward to hearing a little bit more from your perspective about how, you know, your story led to that and the tools you've developed. Before hopping into the discussion on, you know, what you guys have built and are building and the projects you're executing on, we'd love to hear a little bit about what brought you in life to this position and a little bit about your background and how you came about to founding this company.
Dr. Vicki Seyfert-Margolis:
[10:43] Oh my, so how far you want to go back?
Dr. Trevor Royce:
[10:46] As far back as it'll take us.
Dr. Timothy Showalter:
[10:47] To little, to little tiny Vicki, you know.
Dr. Vicki Seyfert-Margolis:
[10:51] Oh, well, little tiny Vicki is from Redding, Pennsylvania. Right. You know, I grew up very, uh, very modest. Um, I would wager on the verge of, uh, poor. But I think it's really interesting. I always had this like sort of entrepreneurial spirit. So I had two businesses when I was young. I had a business that was selling art. I was actually doing album cover, redesign album covers of Grateful Dead album covers in pastels. And they were actually pretty hot sellers.
Dr. Trevor Royce:
[11:27] Just real quick, we might want to stop the interview here because We've got the coolest guest by far we've had so far.
Dr. Vicki Seyfert-Margolis:
[11:32] I mean, that's fantastic.
Dr. Timothy Showalter:
[11:33] You win for sure, for sure.
Dr. Vicki Seyfert-Margolis:
[11:35] Well, it gets kind of cooler. So then a friend of mine from high school and I founded a silkscreen. I learned how to silkscreen, you know, make T and we decided to found a T-shirt business. And so I started doing T-shirts for concerts, but also for like older, older bands. So our first, our biggest seller actually was this T-shirt that was just like the Doors logo on the front. And then on the back, it said, people are strange. I swear to God, people love that T-shirt. But we would go to like, you know, parking lots at concerts and stuff. And we would we would make different T-shirts for different shows and we would sell them. And it was actually a pretty a pretty good business for a high school.
Dr. Timothy Showalter:
[12:15] And there was a lot of wisdom there. People really are strange.
Dr. Vicki Seyfert-Margolis:
[12:18] They are. I know. So so I guess the Doors knew the Doors figured that out a long time ago. They were strange, too. So, yeah, I kind of always had this entrepreneurial spirit. And I also had like I had to have a lot of drive because I put myself through school. So, you know, I went to a college that is not nobody heard of. It was called the Philadelphia College of Pharmacy and Science. And now it was just recently merged with St. Joseph's University. But I went there because I got a scholarship. I mean, we just didn't have any money. So I got I got paid and and I always liked science. Probably one of the biggest inspirations of my life was when I was about 12, my mom decided to go back to college and get a degree in med tech. And I helped her study. You know, I went to undergraduate and I was a biochem major and I really loved biochemistry. I then got, I don't know, I took an immunology class and I got really hooked on immunology. And I think the thing I like about immunology is it's the systems thinking.
Dr. Vicki Seyfert-Margolis:
[13:10] And what I also think is really interesting about immunology is I almost feel like it's a predecessor to how we think now about data, because everything that that you learn when you say the immune system is sort of this principles of homeostasis and one thing impacts another and, you know, complex data, complex systems. So I wound up getting in, you know, a couple of graduate schools. I went to Penn and I did my Ph.D. There in immunology. You know, I would say that was really a formative time because we were merging, at the time, we were merging molecular biology into immunology research. And that was really new at the time. You know, there was T cell receptor gene had been cloned by Mark Davis. And there was a lot of action going on, beginning to move from old school immunology into newer technologies and how to study the immune system. And then, you know, I went to Harvard and did a postdoc in a basic yeast lab because I was really interested in transcription. So I had this like weird career of trying to fuse different disciplines. So after that, I wound up the NCI and then I decided to leave the lab and I went to work at NIAID. And I actually started out with a grand portfolio, but I quickly developed this sort of reputation for pushing the limit.
Dr. Timothy Showalter:
[14:23] In what way did you push the limit?
Dr. Vicki Seyfert-Margolis:
[14:25] At the time, you know, HIV was really an issue and there were big new models being developed for how to do clinical research. If you think about it in terms of what it did, it was transformational in how we think about drug development. So I was there as these big clinical trial networks were forming and as the first reagents were being developed for what turned out to be the first surrogate marker, which was T-cell counts. So this was the first time in drug development that we were actually on this fast track and where we were looking for ways to use indirect measures for understanding whether a drug was having efficacy or not. So I wound up setting up an office of innovation and I basically went to the leadership of the institute and I said, look, I think we need to start fusing engineering, molecular biology, and immunology as sort of cross-disciplinary approaches for how we look at basic immunology research, but more importantly, how we think about clinical trials, that we need to really think about testing cells in the immune system and doing novel technologies to really understand what's going on when we apply, you know, different therapies and clinical trials. And so I wound up getting this Office of Innovation. It's like one thing led to another to another. And interestingly, from there.
Dr. Vicki Seyfert-Margolis:
[15:44] I was recruited by a bunch of transplant surgeons and a guy named Jeff Bluestone, who decided that there was this really interesting issue going on where they were having great success with immunosuppression and transplant, but they were seeing, particularly in children, that the long-term consequences of long-term immunosuppression was development of lymphomas and cancers.
Dr. Vicki Seyfert-Margolis:
[16:06] And they were starting to look for how can we modulate or change the immune system so that we can use different types of approaches, but not sort of gross immunosuppression that leads to sort of a sentence of cancer later in life. And again, interesting time because genome was just starting and monoclonal antibody therapies were really starting to hit as the beginnings of gene therapy. So they formed this network called the Immune Tolerance Network, and it was, I think, probably still is one of the largest contracts that was ever awarded by NIH, $150 million. It was like a startup. And I was the third person recruited to basically build a different model for how to do phase two and phase three clinical trials. And it still lives. This network is still going. It started in 1999. My job was to think about a totally different way of doing trials that incorporated in mechanism of action studies. So we developed ways for integrating into protocols, serial sample collection, and then I built a variety of cutting-edge assay technologies into...
Dr. Vicki Seyfert-Margolis:
[17:15] Testing these samples to see if we could correlate changes in immune system dynamic. We were the first group to use the AFI matrix gene arrays in a clinical setting and to really try to understand what was changing in the biology of people and correlate that with clinical data. And so this was probably the most formative thing for me because it was like we had great funding and we were partnering with all of the biotech and pharmaceutical companies that were involved in monoclonal antibody development in either transplant, allergy asthma, or autoimmune disease. We had this really great purview, really innovative clinical trials that we were doing with this huge biomarker discovery effort that I ran and led.
Dr. Vicki Seyfert-Margolis:
[17:56] And it just blossomed into this incredible sort of juggernaut of research that still is going. And so after that, in 08, when the Obama administration came in, I sort of had this reputation of being like the biomarker maven And the FDA was really interested in thinking about novel approaches for clinical trials. And I was recruited by the administration to go to the FDA. And I wound up working in the commissioner's office for eight years. All kinds of great stories from there we don't have time to go into. My job was really to spearhead regulatory science as a discipline, which was how do we do science in a way that's like we were doing at ITN, that's sort of more rooted in reproducibility, moving academic type approaches into a framework that allowed things to be done at FDA grade as opposed to just research grade. And so that was really my primary job, but I wound up doing a whole bunch of other really interesting stuff. Another really interesting set of things happening, healthcare reform was rolling out.
Dr. Vicki Seyfert-Margolis:
[18:58] Digital came, you know, like I was involved in writing the first guidances for digital health technologies. And that really got me interested in real-world evidence because I had done a lot of work at ITN on large data. And we had started building linear models and advanced analytics and machine learnings back in the 2000s. And then I saw, wow, we're going to start with FDA. We're, approvals of drugs, but we're going to have to start thinking about what the performance characteristics look like once these products, medical products are out on the market. What actually happens? So this idea of efficacy to effectiveness. And I think I was a little bit early, naively maybe, I thought things would move quicker.
Dr. Timothy Showalter:
[19:39] I'm curious, when you were at the FDA, were you, did you work in efforts to sort of make new policies for using real-world data?
Dr. Vicki Seyfert-Margolis:
[19:48] Yeah, I mean, there was so much going on with healthcare reform, we were talking about things like value-based contracting and, you know, the FDA really didn't have a role in pricing, but, you know, we would get together. I was in the commissioner's office. I mean, we were meeting with like Bio and all the CEOs of the major companies and we were sort of saying, hey.
Dr. Vicki Seyfert-Margolis:
[20:08] You know, as you're designing your phase two and phase three programs, you should start really thinking about broadening what you look at because just a single endpoint may not be sufficient to demonstrate the actual value of your product. And I think it really came to a head with type 2 diabetes drugs, which ironically now are like the hottest drugs in the universe, you know, and sort of the beginning days of where that came from. And I think this is a great story is that we were really focused on driving the companies to doing long-term cardiovascular outcome studies, as opposed to just looking at reduction of A1C. So if you looked at all the diabetes drugs, even until recently, before Novo and Lilly went for these different measures, everybody had the same indication. It was better control of A1C with diet and exercise, no matter what the mechanism of the drug was, it SGLT2, you know, a GLP-1 or, you know, some metformin, some of the older drugs. And so we kept saying, hey, you know, it's great to regulate A1C, but the real problem with type 2 diabetes has long-term consequences. And so there's this big argument going on with industry about, we don't want to do these long-term, you know, cardiovascular event trials are really expensive. Why are you making us do this? You know, well, it turns out if I look now, I'm like, that was probably the.
Dr. Vicki Seyfert-Margolis:
[21:27] One of the things that's sort of an untold story in that look at where the value of the type 2 diabetes drugs are now and where the value is made is in those longer term endpoints. Reduction of renal complications, reduction in heart failure, reduction in risk of cardiovascular events and weight loss. Right.
Dr. Timothy Showalter:
[21:47] Yeah, it's super interesting.
Dr. Vicki Seyfert-Margolis:
[21:48] Those were the primary endpoint. And all of those were things that we were talking industry about, you know, back in 2010, 2011. But we've come full circle now. And I look at drug development and it's much more comprehensive and much more complex in terms of the different number of variables and endpoints that we have to collect. And now I think we're seeing where real world evidence really needs to play an important role in understanding the longer term outcomes because there's value there for the for the companies in terms of expanding their education. And there's also value in understanding what long term consequences of drugs are.
Dr. Timothy Showalter:
[22:24] That's like the perfect transition. I'd love to maybe if we can if we can step into a little bit of like how all those great experiences that you had before are translating into what you're doing with RespondHealth. So maybe for our listeners, could you give a little bit of overview for what what's the specific need that you're that you're filling with RespondHealth? And, you know, how are you taking all those great experiences and really helping move the field forward?
Dr. Trevor Royce:
[22:50] I'll just quickly add, I think we heard how you were there when this digital era came into play, how real-world evidence, you know, was recognized as valuable and how to use it. And so tying that into RespondHealth, what you guys, what's sort of that next big entry inflection point and how you sit in there?
Dr. Vicki Seyfert-Margolis:
[23:06] Yeah, so the next step was about patient centricity and real-world evidence that incorporated the patient perspective, right? And if you look again at FDA, what was happening was it was the first time we acknowledged that patient-centeredness needed to be part of drug development. That was part of the user fee negotiations that went on something called Fidesia when I was there. So the first thing the company did was to say, look, we need to understand more fulsomely what is going on in products, but not just in healthcare settings that are indicative of where clinical trials are performed, but healthcare settings that are like real world. And so our first foray was to develop digital technologies that supported what I call community-based studies, studies in places where trials weren't typically done in populations that weren't typically represented in clinical trials. So, you know, low income settings, federally qualified health centers.
Dr. Vicki Seyfert-Margolis:
[23:59] Underserved communities, etc. Because we were seeing, again, if you take type 2 diabetes, we were seeing there was this huge burden happening in those communities. And the representation in phase 2 and phase 3 trials that industry was putting out was like overwhelmingly Caucasian. So we started with that and we created a very successful digital platform that we use mostly with academia. And I had this, like, theory that we had to really prove ourselves anything we were going to do, and it probably comes from being at FDA, that we had to work with academia. I didn't think it was going to be the best market to start with. You know, it certainly wasn't the best thing to do to raise investment capital, I got to tell you. But I felt like we had to be really rock solid and have a foundation where technology was being used and delivered in the right way. And I think that that has proven to be important. And I told my original investors, and I'm the biggest one, honestly, in my own company, that it's going to take a long time. That developing good software and developing good real-world evidence tools was going to take a long time. It wasn't going to be a short term. It wasn't going to be like tech. It was going to be software that was more like a drug development path or a medical device development path. It was going to take time. So what we do now, we still have that. We still have this what we call intelligent engagement platform, but we really knew the, the gold, was going to be getting our hands on electronic medical record information.
Dr. Vicki Seyfert-Margolis:
[25:20] And for a long time, I was pursuing the partner that we now have to get access to large amounts of electronic medical record information and not the claims data and not the structured data so much. We had a bunch of that stuff at FDA and we did a lot of large-scale analyses and we saw a lot of junk, honestly, in the data because we would get a lot of false positives. As you know, coding for billing is very different than the story that's actually going on with the patient. So we finally have our hands and have a partnership with a company called Harris Computing that's acquired lots of electronic medical records, about 36, and they have about 200 to 250 million patients worth of data. Their biggest acquisition was Allscripts. And so we have a partnership whereby we can build AI models directly into this data. And this is like my dream, like my dream has arrived. So now we're able to use the power of AI to mine all of it. Every pathology report, radiology, all the unstructured notes, all the structured data, and put that data extracted and abstracted into entities that we then put into a graph database or a knowledge graph. And once you have it in a knowledge graph, the power of the analyses that you can do is just unprecedented.
Dr. Vicki Seyfert-Margolis:
[26:42] I mean, the speed and the dimensions of the data that you can look at are extraordinary. So we can take long, large-scale epidemiology-type studies or real-world evidence studies that would take months, years to do because people would have to read through notes, you know, and at a much smaller scale. And we can scale that and do it in honestly. Once the knowledge graph is assembled, you can do any analysis you want within minutes to hours across tens to hundreds of thousands of patients worth of data. It's absolutely incredible.
Dr. Vicki Seyfert-Margolis:
[27:12] So we've just built this platform that allows for the integration of agentic AI, which gives even another dimension because you can ask these questions in natural language, very sophisticated analyses, and then you can watch the agentic AI write the code in real time for you to do the analysis. And we provide full transparency of the data, the data provenance and all the code so that as the code's being written for a sophisticated user, they can see the code, they can alter the code if they feel like there's a problem with it. And then you just run it and you'll get like your bar charts or linear regression models, whatever you want within minutes across 25,000 patients worth of five years of visits, all that data. And so it's like incredibly cool.
Dr. Trevor Royce:
[28:01] Are there any specific like use cases just for the listeners to visualize that?
Dr. Vicki Seyfert-Margolis:
[28:06] So our first big model that we've built is actually in incretins or GLP-1s. And we're learning really incredible stuff about we're getting ready to publish. We're learning a lot of interesting stuff about, for example, one of the questions I wanted to know was what does it look like in terms of the weight loss benefit when you're starting A1Cs or high? Because all the trials were done with an average of about eight, eight and a half. And, you know, I had been working in studies in Texas with an underserved community down in Brownsville, Texas, where they have the highest rate of diabetes. It's primarily Hispanic, Latino.
Dr. Vicki Seyfert-Margolis:
[28:43] And the average A1C there is 10. And so my question was, you know, what's that look like? Do you have the same weight loss benefit on the same timeframe? And the answer looks like, no, you don't. There's some really interesting things in the trajectories of change over time between the different incretins and the rates within things occur, depending on what starting A1C is, for example. And we're getting ready to publish that stuff. We're just starting to scratch the surface. We've just built a Parkinson's model, and we're out selling into life sciences companies, and we're working on, you know, some contracts there that will really benefit different therapeutic areas or different programs. For the use of these models and also through the use of being able to assemble, you know, cohorts and registries like really rapidly. Because not only do we have the data, but we have access to the electronic medical record network so we can access the physicians and the associated patients through the network. So it's a very powerful system.
Dr. Timothy Showalter:
[29:42] In terms of the technology side and like the integrations, are there current challenges that you're facing now or do you feel like your team has sort of built the integrations to leverage all that EMR data really grow.
Dr. Vicki Seyfert-Margolis:
[29:54] I should also mention, we had got a contract from NIH and the whole idea for this actually came out of a solicitation that NIH issued in 2023 asking if knowledge graphs could be built. For disease models. And the guy I had been working with on advanced analytics for some large-scale COVID studies we were doing, and I said, hey, I think we can do this. And I think if we got this data, if we get this data deal done, we're going to be able to do this. We went through the contracting process. And that was really informative because all the back and forth with the reviewers turned into a business plan for a whole new area. And we got the contract. And then that really set the stage for, okay, let's get started at small scale. But we realized pretty quickly, as talented as my AI guy is, and he is super talented, and my epidemiologist, and I think this, again, confluence of epidemiology and AI is critical for being able to build what we want to build. I'm a firm believer in cross-discipline drives innovation.
Dr. Vicki Seyfert-Margolis:
[30:53] So the key, though, was we had to go after big tech. And so we went after a partnership with Microsoft. And that really has been transformative. And I think what we have now in the partnership is access to all the latest stuff coming out of OpenAI. We have access to compute that we need. We have access to know-how in scaling software products. And I think all of that is critical. I don't know how you can play in this field without some sort of a partnership with tech.
Dr. Timothy Showalter:
[31:24] We're big on clinicians and healthcare specialists getting along with technologists. I think that's a great theme. While we, I know we need to wrap up soon, but I'm curious, like, so the concept of a knowledge graph is new for me and it may be new for many of our listeners. And if you're a researcher or a physician or you're just interested in the space, can you just explain from your perspective, like, why should we care about a knowledge graph and like, how does it help so much?
Dr. Vicki Seyfert-Margolis:
[31:52] You know, all of our healthcare system, including EMR data, so for physicians, is sitting on ontologies, ICD-10 codes, and something called SNOMED, which is a painstaking hierarchical classification of diseases and all the subdiseases, right? What the knowledge graph does is it sort of builds these relationships between different entities using mathematical probabilities. And so you can basically begin to see new things in terms of the relationships between different concepts. But I think what's going to really, what I really think the knowledge graphs are going to do is provide the next generation electronic medical records. And we're working on that. So the idea that a physician can have a patient who comes in the doors, you know, a 53-year-old woman with, I'll stick with diabetes, type 2 diabetes, who's been on metformin and what's the best option for her and the ability to then go and interrogate a knowledge graph with a representative pool of 30 or 50 or 100,000 patients in and look for others like her to see what the outcomes are associated with drug A versus drug B. Wow, that's game changing in terms of smart prescribing. And that's what these knowledge graphs are going to bring, is the ability to interrogate data at speed and flexibility that we've never seen before.
Dr. Vicki Seyfert-Margolis:
[33:11] Using agentic AI that's going to help you do it by just asking a question. I mean, it's really remarkable. And we do it all the time. We'll be like, hey, can you tell me X, Y, Z? And the model will, you know, will show you how it's planning it. The agentic AI shows you how it's planning it. Then it does the analysis. It gives you the answer. And then, by the way, if you wanted a report or you want a draft manuscript, we have a Gentic AI that does that for you, too. Or we can give you a treatment plan. We can give you a treatment plan. You know, I mean, that stuff is just game-changing. We now can really get into evidence-based medicine.
Dr. Trevor Royce:
[33:43] Those dots you connected to that picture you painted, I mean, it makes perfect sense how that partnership with Microsoft can really just help you guys take off.
Dr. Vicki Seyfert-Margolis:
[33:49] Yeah. Essential, I think. Essential. We're very thankful for that partnership. It's really fun.
Dr. Trevor Royce:
[33:56] It's awesome.
Dr. Vicki Seyfert-Margolis:
[33:56] See all the cool stuff coming out. Yeah.
Dr. Timothy Showalter:
[33:59] Vicki, thanks so much for your time. And I know we ran substantially longer, but I mean, we both could talk, we could talk to you all day.
Dr. Trevor Royce:
[34:07] We just get warmed up.
Dr. Timothy Showalter:
[34:08] I want to go back. Yeah, we need to hire you to still create some shirts for HealthTech Remedy too.
Dr. Vicki Seyfert-Margolis:
[34:14] Yeah, of course.
Dr. Timothy Showalter:
[34:16] Your slogan still holds true, believe me. So particularly around us. Well, it's been a real honor for us. Thank you so much to Dr. Vicki Seyfert-Margolis from RespondHealth. That's it for episode. and we really appreciate your time.
Dr. Vicki Seyfert-Margolis:
[34:30] Thanks, Tim. Thanks, Trevor. Great to be on your show.
