hospitals at over 2,000 patient census and we had some algorithms and some software surrounding that so that sort of launched my work in this space on the other side of the fence now in industry. Thank you so much. My name is Abhinav Sharma. I'm an assistant professor in the Division of Cardiology at McGill University in Montreal, so just north of the border there. I am a heart failure cardiologist by background, trained in both Canada and the U.S. and a lot of my research anchors around the triangulation between digital health, artificial intelligence, and cardiometabolic diseases and a lot of the work we do right now is running digital essentially clinical trials in trying to evaluate a lot of technologies on screening and identification and stratification using digital tools to identify chronic diseases like type 2 diabetes and heart failure. So these are studies that have required you know a lot of complex interactions with patients, with sponsors, and also we've had some really interesting interactions which you can maybe discuss a bit later with other partners as well to see how this can all evolve and serve our patients. So that's a little bit about me. So thank you so much. Good afternoon everyone. My name is Jean-Philippe Coudert. I'm the CEO of a company called VPG Medical which is located in upstate New York, Rochester, New York, and we develop with this technology a new way of monitoring cardiac activities through video cameras. So we extract actually the pulse from the face of the patients and we do that in a very passive way. So it's running on your computers, laptops, and smartphones. So while you're interacting with your computers we detect your pulse and we can detect irregularity of the pulse which might help detecting atrial fibrillation. So my background is biomedical engineering. I started my career as an academic person at University of Rochester where I did my postdoc and then become professor of medicine and professor of electrical engineering and started developing technologies specifically around body surface electrocardiogram. And while I was interacting with this type of technology and patients we quickly realized that we needed other solution to monitor the heart of patients that was the less cumbersome I would say that was also not limited in time. It's why we developed this video technology that works on your smartphone all the time and doesn't have any limitation. The time someone gets to a heart failure specialist or an electrophysiologist, by definition, they're already a patient of some type. But Liz, Jean-Philippe, you are working at a point where many of the people that are gonna use these products are just consumers. They are well people and you may be using the technology to screen them, to help identify them as someone who may need further help. So can you talk a little bit about how you think about that spectrum of the person who's not yet a defined patient and all the way to helping these same tools manage someone with a defined disease state? So maybe Liz, I'll start with you and then I have Jean-Philippe put comment. Sure, we look at it both from the screening and then diagnostics. So screening oftentimes, even for home testing, we do things like figuring out someone has colon cancer through a fit kit, which is you put in a little sample and then you mail it back. The separate other idea is right now, we're doing a lot of screening to diagnostics and then ultimately testing to treat. So being able to, let's say, say an A1C, we are now doing it where we can actually put someone on metformin through something called LDTs. It's a way that the FDA has approved where we are allowed to actually start to treat off of home tests. Just like we did with Paxlovid with COVID, when you have a positive COVID test, you don't necessarily need to have a physician monitor and review that. We're doing that now with even putting people on levothyroxine if we see like a thyroid issue. What we're seeing now, especially from the idea of screening and prevention is that we don't wanna pan screen. Sometimes we can risk stratify now. So we work with a lot of payers who are now reimbursing for these tests, home testing and biomarkers saying, we know that these people are high risk. For example, even those with A1Cs that are pretty high and also have high blood pressure, we now are screening people for kidney disease. Chronic kidney disease is about 30 million, more than 37 million people in America. 90% don't know about it. So what we're trying to do is do more screening for those that are higher risk. And then ultimately we can diagnose and then treat. So that's been the goal for a lot of our home testing. I will say though that Apple and some of the other more commercialized products, like you said, consumer testing, when they've looked at things like AFib and others, they're not necessarily trying to figure out exactly if the person has AFib because they're not medical grade, but they're able to at least say, we're detecting something interesting. Can you add the followup? Because even a lot of these studies, we've known that in the academic world, if they don't have the right patient population, for Apple study, their AFib was most people that were under 40. That kind of data isn't as accurate as those that are like, you know they have AFib, you want to detect it, you want to know when they have an episode. Jean-Philippe. Yes, thank you for the questions. Very good question. It's obviously two different beasts to tackle, in a sense, right? When you think about developing a technology that would help screening or creating a diagnosis for already, I would say, diagnosed patients. For us, our experience has been to really start at the lower levels, so really targeting patients that have already a diagnosis of AF, that are paroxysmal AF, that have an increased risk of having a recurrence of the arrhythmias, and propose actually an app, which is a prescribed app by the physician, to not only give peace of mind to the patients, because the technology monitor them all the time, but also a technology that really provide to the physician all the information they need to treat optimally the diseases, in a sense. So why as a small startup you do that? Because there are a couple of different constraints when you develop for already diagnostic or for screening. And the main one is the FDA submission, obviously. If you go for a screening tool, it's a very different level of requirements than if you do just a digital health diagnostic tool. And in a sense, for us, it made sense to, and I think all the other companies have adopted the same strategy, to really target a population that represents some kind of niche, in a sense, and from there, when you have your clearance of your technology, try to push forward for other populations. So there is, it's two different, obviously, objective, but there are two different paths with very different level of difficulties when you are doing that. a big issue we have in just getting. Did they get access to this test somewhat? in your field, what will it take to get this right? Yeah. I mean, that's a fantastic and a complex question that by itself could take the entire session. And so I do think that the issue of access is certainly tremendous. There's both not only access to the physician, but access to the technology. But as we've seen in very recent, you know, analyses, you know, across multiple socioeconomic strata, the availability of digital health tools is certainly increasing. So in a way, there is a little bit more of a democratization and access to those types of tools. How those tools link into clinical care is still a big unanswered questions. But one of the ideas is can we at least create very simple closed loop circuits, for example. So, you know, I love Liz's comment on, you know, can you stratify, diagnose someone with diabetes and initiate a treatment all without necessarily having that person needing to come to a doctor to speak to them in a clinic for someone to hand a prescription and then they go and fill it out. Like those types of simple closed loop circuit systems can be done at the first layer, which can address then the huge sort of, you know, sort of population that does not have access. So I think that's where a lot of things are going forward. Increased automation is going to be absolutely essential. Of course, we'll always require regulatory approval and oversight, but these may be ways in which we can accelerate the clinical care without overburdening the healthcare system already. So Dan, you know, Murat talked about infrastructure. you write? Is it possible? How did you do it? Is it possible to teach large systems new tricks or do we just have to think about this totally differently? And maybe you can share with the audience a little bit of what you went through in that telemetry experience and what you learned. Sure. Yeah. No, thanks, Sunil. And I do agree with all your comments. I think that was very well said. Just to start with the last part of your question. So at the Cleveland Clinic, we created a bunker, a command center, not just for telecritical care, but for continuous monitoring of non-ICU patients. And our big success was getting out front of cardiac arrest. So we were able to deploy our emergency response teams directly to the bedside as those patients were deteriorating. And we published these results in JAMA in 2016. We got a 93% return of spontaneous circulation, which for US hospitals, the non-ICU ROS crates are around 25% or so. So we were able to scale that up to, like I said, main campus plus 10 hospitals over 2000 patients continuously. And to answer the first part of your question, I think that that's the model. Rendering care effectively at a distance. Murat made a very important comment earlier about staffing shortages of physicians. It's not just physicians, but it's also nurses. And it's also the level of experience of the nurses that we do have. You know, as we all know, the pandemic drove a lot of really experienced nursing out of practice. And so the next generation that's rising up to fill those ranks don't have decades of experience. And so I think as we're talking about sensors and we're talking about these configurations, this digital ecosystem, it has to be insight oriented. It can't be, you know, this world of what we've had in telemetry for the last 20, 30 years of just dumb alarms of heart rate high, heart rate low, blood pressure high, SpO2 low, SpO2 off. It has to be insight oriented. Whether it's a clinician, a nurse, or a physician, or even a consumer or a patient, we have to deliver insight to them of something like a disease state that we're detecting, whether that's sepsis, bleeding, decompensation of heart failure, because those are all things we know how to treat those. We don't have to innovate the therapeutic part about it. We know how to intervene and manage those conditions. It's about getting there earlier within that window where the intervention a step back. I mean, a part of this session was to talk a little bit about where we are now and, you know, where we're headed in the future. So maybe I could ask you all, you know, as we sit here in 2023 or, you know, 2025, as we were asked, what digital biomarkers have we perfected? Like, what's out there that you say that, like, it's, you know, in the bank and it's just a matter of getting to patients, whether, you know, it's on an outpatient, you know, consumer thing or wherever? You know, each of you has a different perspective on this. I would like to hear it. What are we doing great at this point? Murad, do you want to start? Well, you know, I have maybe the most extreme forms because we have invasive monitoring on heart failure patients. And, you know, in my world of advanced heart failure, people have invasive devices and we have pressures from inside the heart. So more and more people will wear devices that measure pressures from inside the heart, but we will soon have variables on top that will be validated against intracardiac pressure measurements. That's our gold standard. So I think we can figure out how to automate that data and create feedback loops. That would be awesome. I think that will be a revolution to where, you know, every heart failure patient early in their career of heart failure, so to say, gets a device applied or implanted, and now it always lives with them and feeds data back. So I think that is, we are on a verge of making this a mass. So let me just push you on that. So for you, that biomarker is some sort of intracardiac pressure as opposed to a natriuretic peptide or some other, you know, serologic biomarker. Yeah. So I think, so this speaks to an important point. Cardiologists are very simple, not electrophysiologists. You guys are very smart. So heart failure cardiologists are very, very simple people. We know pressures very well. You know, we knew it for a hundred years. You give me a biomarker that now I can get it on a higher frequency that gives me variability and gives me feedback when to act on it. Well, we understand it. So I don't have to learn. If you give me a biomarker, I don't have to learn. We love it. If you give me now a sensor that is complex, is unitless, and I've never heard of it before, now it gets a little more complicated for the average cardiologist to deal with that. So I guess I'm going to have Liz comment. I'm going to just put my personal perspective, you know, having been in medicine like 25 years, I think we all, you know, recognized the times where our patients would finger stick themselves, you know, four times if they were diabetic for glucose. And now I see all my friends, you know, with a free Libra or a Dexcom device and seamlessly meant, and I guess in my mind, like it's the perfect, you know, biomarker, it's patient friendly. The patient is, you know, completely engaged in the process. The physician is pretty much out, like it solves all of these issues. You know, I'm sure Liz, some of the things you were talking about earlier, you know, are also designed to be, you know, somewhat, you know, patient friendly on their own. So how are you thinking about this? Yeah, I'll mention that. I think from a generalist standpoint, I feel like there's some really smart people here. I didn't go through the cardiology route, but I think for me, I look at it from the future of care, like where are we going? So affordability, volume, the masses are, there's like blood pressure and all that. So one thing I think about is anything that's detected, whether it's like heart rate, blood pressure, all these things, it has to come with the return on the investment. So for payers and insurance companies to want to reimburse for it in their homes and whatnot, it has to have a reason. So it's a hospital discharging early. So someone came in with congestive heart failure, and now they want to discharge, but they just need to monitor their blood pressure, or they want to make sure that they just went through surgery, and they want to make sure that they don't have an infection. So they want to monitor temperature. There are things like that, that I think from the payer perspective, if you can get people out of a hospital, say 5,000 overnight earlier than you would normally do, it's both good for the hospital, it's good for the patient, less infection risk. I think there's a lot of interest in that. I think we have Massimo, and like there's some great opportunities out there where if you can reduce the total medical costs, there are ways where you align with the payer and the provider and the patient, and ultimately the future pioneers, the industry people. I also think of it with another, we just talked about the four or five Ps with policymakers. The other area is that CMS is now reimbursing for CPT codes for home remote patient monitoring. So you have to track blood pressure 16 days out of the month in order for you to get reimbursed from a provider. They're reimbursing up to about $1,200, almost $100 a month. So for a provider, like we said, it has to have an intervention. They have to monitor it, but they also will get reimbursed for it. So we're seeing the government changing. We're seeing also devices coming out. We're seeing our providers getting aligned. But I would say the last thing that's important is that it doesn't, not just, it has to be medical grade because Apple's and the other Amazons may not be able to do it, but the other thing is that it has to also, like you said, have an actionable insight. So I think about also like the five Ps. So it's preventative, prescriptive, which is really important, personalized, predictive, and ultimately there's also a lot of very interesting places where we're thinking about customized and also driven by the patient to actually want to engage because a lot of patients don't want to be monitored and they are a little bit lazy. We all are sometimes. So making sure that patients are engaged. Liz, before I turn it to Dan, I'm just going to, you mentioned this issue of billing code. I'm just going to ask your perspective of an issue that I feel like a big pain point for me is, yeah, it is true that for many of these things, CMS has a defined coverage policy. But what I find really frustrating is that often commercial carriers do not. They don't follow the same rule. And it's very hard as clinicians to have two pathways of care, you know, one for a CMS patient and one for a commercial carrier. And then within commercial carriers, who knows, you know, which commercial carrier you're dealing with. And I think from where you sit today in your non-medical hat, I'm sure it's something that you grapple with. I mean, what will it take, you know, to get it more uniform in the U.S. as opposed to such fragmented coverage? So unfortunately we live in a multi-payer system, fortunately or unfortunately, some people believe it's a good thing for competition. I think the reality is we, a lot of the commercial side, so fully insured and self-insured, they follow what CMS ends up doing and the guidelines, Medicaid and Medicare. Medicaid is more separate, as you know, it's like state by state. So sometimes there's a certain governor that's really excited about mental health or really excited about remote patient monitoring and that gets put out in the RFPs. What's also interesting though, Medicare, yes, we're starting to reimburse for it, which is really exciting. Commercial plans will start to look into it once we see, again, going back to like the return on the investment. So outcomes-based, reducing medical costs. What I think commercial payers always look for, and we don't always talk about, is they're looking for the 12-month ROI. So if they can see that you paid a provider $1,200 but they prevented, let's say, two readmission rates for a CHF patient, I think that's when you can say there's a positive outcome, both for the patient and for medical cost reduction, and that's when commercial plans will start to jump on board. But I think that takes some innovation and some innovators going at it and demonstrating patient outcomes and demonstrating reduction in costs. So Dan, we'll go to the original question back to you about the biomarkers. Yeah, no, I'll answer the original question. I'd like to come back to what you said, Liz, because what you said about medical grade technology, both in the consumer market and in the hospital market, I think that's a really, really important point. But to go back to your question, I think the digital biomarkers that we've mastered today are the basic vital signs. Pulse oximetry, heart rate through motion, low perfusion, those types of things. And as you mentioned, CGM, glucose monitoring. But where I think our company has its latest innovation is in continuous hemoglobin monitoring. You mentioned finger sticks to measure glucose. Imagine drawing all the CBCs. I mean, how many times do you CBCs daily? We're exsanguinating the patient with all these blood draws in addition to all the discomfort. So what we have with our optical sensor, with our latest generation optical sensor, we're pushing light at the hemoglobin molecule, as crazy as that sounds, and we're assaying it in the blood non-invasively with extraordinary accuracy. And it's a continuous collection instrument. So you can put it on a dashboard, say, for example, in a patient that undergone cardiac surgery who's coming out of the OR with a baseline amount of anemia. And you don't have to draw CBCs. You can just see that non-invasively and get a sense of, has that patient deviated from their baseline? So this is just an example, and I can't give you the ones that aren't released products for us, but you can imagine the other meaningful laboratory tests and biomarkers with our force sensors and with our optical sensors. We are going to get that level of sophistication to non-invasively collect this data. Abhinav? It's a fantastic question, and I think that I certainly align with what's been said. Vital signs certainly have been done quite well. I would argue that potentially measures such as atrial fibrillation through digital means, there's now a huge body of literature looking at arrhythmic detection. So because of the physiology of detecting arrhythmias through pulse, I think that has certainly advanced a lot. But I think we can even expand our thought process into what is a digital biomarker. Can actigraphy be a digital biomarker, which has been passively collecting physical activity data through your phone. So we're looking at a lot of actigraphy-based biomarkers and correlating that with gold standards as a way of non-invasively tracking people for very low, if almost no cost, because it's on everybody's phone. So I think that can really expand our horizons and what we think is a digital biomarker per se. The other thing is also coming from Canada, we have a socialized healthcare system. So the incentives and the financial drivers are very different. And the ability to drive down costs is so paramount that it's very difficult to get the government to pay for anything unless there is a tremendous amount of evidence showing reductions in hospitalizations or other costs. And even then it may not happen. So I think there's certainly something to be said about competition and about innovation and industry playing a role in stepping in where governments are just either not mandated to or have a difficult time moving at that pace. I should also quickly clarify, the reason why I was mentioning the multi-payer system is because people switch out of their insurance all the time. And that's why the 12-month ROI exists. If people didn't switch insurance, if you were with your insurance for five years, your insurance company would be incentivized to, even if they paid something that would have a return on the investment in three years, they would say that makes sense. But because we're in a multi-payer system where people switch from United to Anthem to Elevance to Cigna, or they lose their job and they switch to Medicaid or they go on to Medicare now because they're over 65, there is an incentive to only book and build and reimburse for technology that is within 12 months. And so that's why I talk about the multi-payer system. Jean-Philippe, I'm going to ask you a slightly different version of the same question. We've heard a number of the panelists talk about AFib as a biomarker. You're obviously very interested in this. Now, AFib, I find a little bit unique because if we use it as a signature of something we're trying to identify, different people are using different approaches for the same endpoint. So there's, of course, you can try to acquire an EKG through some handheld thing. There's PPG. You've introduced VPG. So talk a little bit about this whole notion for the same biomarker, there may be different approaches to capturing that biomarker and how will clinicians and patients and consumers think about what's the optimal way to procure that end goal, which would be detecting of AFib? Yes. Thank you for this question. It's very interesting and very important indeed. So there are many solutions you're right now today for detection of AFib and you have just at your finger, you can record an ECG and define your rhythm, et cetera. But I think that what we are facing very soon is really a problem of compliance with this kind of equipment and devices. Today, we have a problem with compliance of patients with treatment. For example, we know the patient do not take the lifesaving drugs that they are supposed to take. So is it really fair to ask them to now adopt in their everyday life a device that they need to integrate into their daily routine? When you are a chronic patient, that's a problem. That's cumbersome. You have to charge the wearables, you have to wear it, you have to put it on yourself. So I think the future of this digital health is really into the equipment that will be encompassed to embed it into everyday things, right? Having sensors in your bed, having sensors in your armchair or your chair when you're watching TV, having sensing your pulse from your smartphone or on your computers. Otherwise saying what we really develop is not only a contactless measurement, but it's an effortless measurement. It's running in the background. So all the time you spend looking at your smartphone and think about how many hours you spend every day looking at your smartphone or even working on your laptop. You have this camera that faces you, looks at you and can extract information while you're having this interaction with the device. That is your routine interaction. That is what you do every day and you don't want to change that. You want to leverage the patient's behavior and capture information without putting burden on them. So I believe that compliance is a very important part of what's going to happen in the future. So that really segues to my next issue, which is I've heard a lot of you talk about just what you mentioned and most importantly that at the end of the day we're trying to improve patient outcomes as lower cost, that we have to be able to do that. And many of the things we've talked about is really preventing people from landing in the hospital, which I think implicit in this conversation is a big driver of healthcare expense and one easy target to remove it. But when I think about AFib patients or heart failure patients and other patients who bounce into the back to the hospital, to me one common factor is always some variable related to social determinants of health. That there's issues with getting medications, there may be issues with family support, etc. How confident are we that these digital tools that are by inherent expensive, many of the things you just outlined by nature are going to be expensive. You're not just going to get sensor driven furniture to come on the cheap. So how realistic is this in the near term in 2025 and 2030? Is the smartphone so ubiquitous worldwide that you're not at all worried about this or what else needs to be done to really maximize the value of these technologies? Well we definitely have to leverage what we have around us and the natural solution are those smartphone computers that have the resources, that have the power to collect data, they have sensors already that can do that. And you have a family of solutions that are what I would call software only solutions. Those are very inexpensive. Those are just software that you're going to load on this device that you already own and that are going to do the job. So I think such type of solution have probably a big likelihood to help health disparities, to help reaching people in rural area, to reduce actually constraint around digital health literacy. For example, take example of our software. You install it on your smartphone, you have nothing else to do. The smartphone monitors you and when something happens you have a notification and this is journaled in the phone and you can create a report, send that to your physician and he can review the data. So in a sense I'm not convinced that those solutions are going to be more expensive. Maybe if you have complex sensor, maybe they would be. But let's not forget about the fact that those sensors are going to be everywhere. Think about camera for example. I mean cameras, it's very rare when you don't enter a room with a camera. They are everywhere now and it's not going to change. So just think about this room and the number of cameras that are in this room. This is something that's going to be everywhere. So we can I'm sure leverage this kind of intrusion of the technology into everyday life of patients and use this tool to capture information not only that are unique but that are very long time. Think about a smartphone. People can stop using things but I don't know anybody who said oh I'm not going to use my smartphone anymore. That doesn't exist right. I think where health disparities will come into play. and have somebody monitor the data. So I think, I agree that software solutions only. So, I do think that at the end of the day, systems of care will have to change and become decentralized so they can reach patients remotely, and I think that it has to be a bi-pronged approach. Technology needs to be accessible, but I think the monitoring system and action needs to become maybe independent, where it's a closed-loop system is not of an offset to where you don't need a physician. Yeah, I think this goes back to what you were saying earlier, Liz, about the importance of reimbursement and having stable reimbursement models, because the way I think about this is that a lot of these are going to be foundationally grounded in prescribed use cases, okay? So, we have heart failure patients, we have other types of patients, hospice with COPD, pneumonia, and we're prescribing them wearables to go home, and there has to be a stable reimbursement model, that closed-loop. And then lastly, a medical-grade device that has the same type of accuracy and precision of the instruments that we use in the hospital, because we're really sort of getting beyond the brick and mortar of the hospital and rendering that care more effectively at a distance. And I think that's going to be the key to success in this field, and addressing those concerns about disparities is part of it, because I think when people are sick enough to be in the hospital, that's a great equalizer, that's an opportunity for us to step in and to organize their care in a meaningful way. Just to extend off of that question, you know, do these systems of care then need to be compared against standard of care with rigorous clinical trials or randomized trials to prove that they actually not only reduce costs but are more efficient, they reduce hospitalizations, et cetera? I mean, you know, there's lots of companies, you know, like JP and others that are making these technologies and devices. Do they have to then be, you know, evaluated in rigorous randomized trials, not to prove that the device and technology works, which is its own study, but to prove that integration into a clinical care system actually reduces healthcare utilization, reduces costs, and does not actually inadvertently increase notifications, costs, stress, provider-physician burnout. So what is that balance, and how does that get dealt with? Yeah, I think that's a very important point, because sometimes when a patient, you know, comes to me, and I'm sure, you know, Dan's experienced the same in his former life, you know, they come with a notification, you know, that their watch showed something. I sometimes think, you know, when Liz is in her pear hat, that probably the unhappiest person in that whole thing is the pear, because that patient's essentially bought themselves, you know, some ambulatory monitor, an echo, a stress test, and God knows, you know, what else. And so that's a big issue, Dan, a very realistic, you know, problem that, you know, we have to address. Murad, you've thought a lot about this clinical trials thing, you know, what's your answer to that? Yeah, I think you made the very important point. Clearly, you know, classically trained at the DCI with me, so I mean, everything always is trust. So I'm going to give you the real world perspective. So I'm a remote monitoring clinic provider. I have 10 companies every day at the door saying, we have the next thing, you want to have us. Okay, software, hardware solution. Okay, I listen to them, sounds all good on paper, they have validated the technology, it sounds real, it makes sense, it's a no-brainer, but then I have to go to my committee, and I have to tell them that this is the next device I would like to bring on, and what is the question I'm going to get from the committee that includes, you know, general cardiologists, some subspecialized people, and some pure bureaucrats. They only care about the beans at the end of the day, and they will say, show me the data that this will help reduce costs, even though it makes sense. It's a pressure sense. Like, I know pressures, why do I have to show that I can reduce hospitalization pressures? I know how to deal with pressures. Show me the data. And the second you have a randomized controlled trial that shows some benefit for remote monitoring technologies, you're going to be on at our institution, and for me, it's due. Great. But if you don't have that, it's a single-arm study, and you just show that you validated against something, it's tough. It's really tough for the seniors to get that. No, no, absolutely, and I think that for startup companies and small startup companies, this is a big, big problem, right? Because once you have indeed solved the FDA clearance challenge, I would say, which is already very complicated, and become even more complicated now with machine learning, because the FDA is learning about machine learning. So when you come to them, and you have more discussion than anything else, and everybody learns in the room. And once you have, you're right, done that, and the FDA tells you, okay, you're good to go. It doesn't mean you're going to sell, right? It really means now you need to have your business model, and a model that makes sense. And you have to convince people that you act on the three, I would say, pillars of digital health, which is better outcome, lower cost, and scalability, right? And you need to meet these three pillars with your technology. It's very expensive. It takes time, no doubt about it, but we'll get there. We need to go there, because the health system needs this kind of revolution to really become more efficient. As it was said earlier, on the main stage, we really have a problem in the U.S. We need to solve this issue of cost of the medical system or the health care system. So that, I think, lends itself nicely to another issue that people have touched on that I find very interesting, and that's this whole issue of scale, and how important scale is to drive any of this. I'll use what Liz started with her company as a kind of a case example. So let's say that she has a product that tests something, and she can go to the consumer who can buy that product, and that consumer may come to me with some result from that product, and I may have no idea what this company is, right? And I'm saying, well, I don't know what this is, and often I'll just repeat the test or do something based on a convention I know. Conversely, she could come to me, and she could say, her company could come to me and say, listen, I have this product. Jean-Philippe, you could come to me and say, listen, I've got this great product, and I love it, but I'm in a universe of maybe 50 other practices within my health care system, and the other 49 have no idea what this product is, and it means nothing. Or we could go and say, be like the Cleveland Clinic approach, like we're going to go from the mothership. We're going to scale it, try to do it at scale, and all of the people will be informed that it's there, and we'll use it together to build some scale into the system. So Marat and Abid, I know certainly in Heart Failure you've thought a lot about utilizing pressure data and other things at scale. I'd like to hear from all of you this notion of how can we scale it in a way that overcomes this fragmentation thing, because Marat, to your point, not only do my people talk about the costing, but they'll often say, well, I don't know so much about this company, or they get inundated with 50 requests from 50 companies, and they just are crippled by not being able to make any decision because they feel like half the companies go out of business, or they may put their chips in the wrong bucket. So the scale issue, I believe, has really kept us from making progress, certainly where I practice. I'd like to hear from you guys how you think we overcome this. I can jump in and say a lot of times, you actually named how typical startups go. They either go direct consumer because they think that there's some willingness to pay, they maybe go high end, maybe they go low end, but depending on where they go, they may go after Medicaid in a certain state, or they may go after certain areas where they say there's a certain provider, like they'll go after Duke, and they'll say, can you pilot with us? I've actually seen a lot also start to partner with academics. So they'll say, why don't we give you some warrants? Actually, this is like an interesting business model where they'll say, we'll pilot for free, we just want to base everything on outcome. If we end up fundraising and we do well, you get to share on the upside of the company growth. So that's been an interesting model that I'm seeing a ton happening within the digital health space. The second area is I'm also seeing, after the consumer side, they now have demonstrated outcome data of engagement, where you were saying sometimes patients don't want to engage, but now they're saying, actually, if we send this box, which is what we've been trying to do too, we send a box kit. Most people, when they're talking to a payer, they'll say, oh yeah, a few clicks, and that's engagement. But we're actually getting them to do a finger prick, put in their things, get their lab results. I mean, that is a lot more engagement, and so I think once they demonstrate engagement within the consumer side, then they can go back and say, hey payer, we now have all these really interesting results, why don't you now work with us for better cost outcomes? I think we're also seeing the third thing, I think you mentioned the three pillars, but I think the fourth thing also is really important is the provider flow, like the workflow. You may not always want to get some random Apple watch information, but you may want to basically work with the patient to say, when someone gets discharged, how do you integrate it with an end-to-end idea of the closed loop you were saying? Can all these different watches basically be integrated into the discharge planning, so that when someone gets discharged, you already know they're going to go out with a watch, and they'll go out with ... I feel like the provider workflow is something that we don't always talk about, but it really is disruptive when you have to train all the physicians to get on this biomarker to read this, so how do we make sure the reimbursement rates are right for the provider? It's also the workflow isn't too much of a hassle. I don't know if you deal with that a lot, where you're thinking through, you've gone through the committee, they're kind of open to it, but like, okay, how do we get providers to change their behavior? I'm interested in hearing also that perspective, because that's another huge area. I think it's a roadblock, but it's something that if we talked about more, we could overcome. I was just going to say that I think the answer with scalability is what Liz mentioned, partnerships. It's also interoperability, and it's the connectivity that you can have when you aggregate data in the cloud, cloud computing. I think those are the three components, and I think the partnerships are going to be between smaller companies and bigger companies, between healthcare and industry, and then between payers and industry and healthcare. I think that there really has to be a collaboration, and then the second piece on the interoperability, look, I mean, there are going to be really exciting novel companies that come along that have found the human signal in the image analysis with video, and to sort of master that domain, and then to have interoperability with the other places that we can collect human signal from sensors, and then put that into a cloud, and to aggregate it meaningfully, and come back to what I was saying earlier about insight, where we're giving disease detection insight to those clinicians, and also to the patients and the consumers directly. So we're getting some good questions from the audience. I see Andrew Cron has said that he has a question. I see him in the back. I think we may need these to hear his question. Go ahead, Andrew. So it's actually not a question. It's sort of a comment that would have been a nice closing comment, so I'll just practically So I usually try not to multitask, but I'm in the audience, and I'm sitting listening to this, and it's been outstanding, and an email ping notification comes in, and the email is titled Apple Watch Death. So I run a cardiogenetics clinic where we see families with sudden death. So last week, I saw a family whose 14-year-old boy was perfectly well, played video games, went to high school, blah, blah, blah, and mom found him dead in the morning, and he'd been gaming until 4 a.m., and he was wearing his Apple Watch. And I got the download, and the parents contacted Apple, who got as much content as they could out of a non-ECG-based acquisition. And so on the one hand, what you see in the heart rate data, without going into too much detail, what you see is you see sort of steady heart rates, a little fast heart rate, probably at the most exciting part of the game, and then all of a sudden, there's just essentially loss of signal. And it speaks to the fact of, you know, what a compelling graphic and image and experience. And the flip side is, we were this close to getting things that have a massive impact on understanding what happened, understanding whether there's an inherited component for their two other kids, on interpreting the molecular autopsy, which is now underway with retained DNA, and in theory, having analytic content that could help us be like precogs in minority report and see this coming. So I just, you know, sort of a shout-out to the forum we're in to say, keep pushing the needle forward, because I think we can make a difference in some of these most, either greatest yield or most catastrophic tragedies. Anyway. Andrew, thanks very much. Another question from the audience has been, do we think that there are any digital biomarkers that can reflect patient quality of life indices? Well, Abhinav knows a little bit about that. Yeah, so, and I think a lot of this, you know, anchors around, so heart failure is a prototypic example where a quality of life is certainly becoming much more emphasized, especially around regulatory discussions, clinical trial endpoints. You know, the idea that quality of life can be assessed through digital means that's not just someone clicking a, yes, no, I feel good, I don't feel good, you know, on a survey, but can be actually longitudinally tracking them is exquisitely important. The FDA actually recently released a broad agency announcement competition to look at various measures, such as actigraphy, as surrogate markers of quality of life. So that actually is already underway, and I think reflects the thinking around regulation of these types of surrogate endpoints for quality of life. I think we do it well from a survey point of view, like Kansas City Cardiomyopathy Questionnaire is one example that's used in the research world, and there's other examples for COPD and other disease states, but to translate that, I think, into a digital biomarker of quality of life, I still, I think still needs a bit of work, but it's certainly in the pipeline, and now even regulatory agencies are, have requested proposals to explore this explicitly. So here's another question that I think is a fabulous question, maybe Dan and Liz want to take a crack at this. The question is, digital biomarkers are often not studied in isolation, but in combination, and in trials we study workflows, not the marker itself. Is the goal for systems to then adopt workflows for reimbursement as opposed to a single biomarker? For sure, no, I think, I'm sorry, Liz, do you want to, yeah, no, I think that's a great question and a great insight. I think that it's this opportunity to blend our sensors, to fuse data, to aggregate it meaningfully, and again, project a line of sight on what's coming. I think that's what the name of the game is, and you know, I'm the guy up on, one of the people on the stage in the medical device industry, but I think that 80% of, maybe 20% is the technology, 80% is the clinical workflows. I think that's what we really have to solve for, and that's why the partnerships between industry and healthcare are essential, because we have to figure out, at a pragmatic level, what's the best way to drive those outcomes and the efficiency, the throughput in the system. I completely agree. I feel like, so I still see patients in urgent care once a month, and oftentimes the workflow is the biggest pain point, right, where I'm trying to get an MRI, but the prior auth comes back and I'm missing multiple things, and we're going through faxes, and it's like, what are we still doing in this day and age with faxes? But regardless, I feel like for workflow, a lot of the discharge planning and the things that we're seeing remote patient monitoring reimbursements are coming from, if you can discharge, again, going back to the highest ticket items of any kind of insurance company, it's ER visits, it's specialty pharma, but it's also hospital admissions, and length of stay. Those are things that we measure. Those are the highest costs. If you can reduce any one of those in terms of the workflow, and you build it in, I think it could be really powerful in the long run to try to get things covered and reimbursed for. Maybe along this way, so I think what often happens when I introduce a sensor technology into clinical practice, many partners often think the sensor will do the work for them, and often forget the fact, well, the sensor is just going to give you the data. It's all about workflow. I mean, a very, very good example is Abbott, which has a technology called Karimas. It's a pressure sensor. I'm coming back to pressures again. It's a fantastic technology. It measures something which we would otherwise get on a tremendous work. We would have to calve people on a regular basis, which we don't do, but here you have pressures every day. How could you not, with this information, keep people alive longer and keep them out of the hospital more frequently? Well, the company has shown that you can do this, but the evidence is not overwhelming. The reason, I think, is that we often get the data, and we don't know what to do with this high-resolution data. We haven't worked out the workflows with it. I think, to the scalability, a company will scale a device, and it will go, and they will sell. Every year, a little bit more. The numbers will look better every year, but a catalytic event to where you exponentially grow will always be clinical trials, something where you can show that the outcome dramatically changed. Unless you show that, and you will never show a clinical benefit in large-scale studies, hundreds of thousands of patients, if you don't have workflows thrown out, because nobody's going to act on the data if they don't know what to do with it. That's where it all falls apart. Yeah, yeah. That's a very good point, and it's true that with a lot of wearables today, there is a phenomenon of data overload, right? You receive so much data that it's ultimately not really useful, so there's really a responsibility from this organization to make data informational, right, and actionable data, right? To really filter the data so it is useful to the clinician, and it's another level of difficulty, obviously. I totally agree with your comment. We talk a lot about, appropriately so, quality of life for patients, but I think we can't forget quality of life for the providers themselves, because I think time is very limited, and so a lot of these things don't talk enough about how your time is better spent. We can also no longer just append this to our daily workflow, and the idea is not to save time somewhere so you can just keep seeing more and more in an endless loop. I think a lot of people are just very frustrated, and it contributes to the burnout, just the data deluge there is. So look, we only have a couple of minutes left, and maybe we'll start with Jean-Philippe and come towards the end, and just your final thoughts of what 2030 looks like to you. What do you think will be present in this space by then? Well, when I think about what we would have found today 10 years or 5 years ago, I would never really find out, so it's always a very difficult exercise. What I saw happening with digital health markers, and what I like very much, because obviously I'm interested in technology, is not only they have been able to actually measure vital signs and things we know, but also generate new markers that are very unique, and sometimes very natural and simple to understand. One of the very nice examples, I think it's not in cardiology, is Parkinson's disease, where there was no biomarker for it, but just by counting the number of times patients move their fingers, or how many words they can generate in a certain time, you have a good association with the disease in a sense. So we have with these digital markers also a whole aspect of new things coming up to us that are, I believe, very, very exciting. Thank you. Abhinav? I certainly think that the landscape is changing so rapidly, but I think things in the future, and I would put this out there to a lot of the smaller startups, as Murad and others have said, often in academic centers we're overloaded with requests for collaboration and use our technology, but building together and sharing on the upside I think has tremendous potential. I think we're an early adopter of XYZ companies' technologies, and if they end up scaling or selling then for those initial institutions to share in that I think is a really powerful model to incentivize technologies moving forward. The other thing is that regulatory agencies and other agencies are now thinking about these technologies very actively, so I think driving the needle forward is going to be very critical with good evidence and good randomized trials. Thanks. Dan? Yeah, I think the future is in non-invasives. I voted with my feet based on the principle, obviously, and I think that where we're going to be in 2030 is I think we'll have a very mature clinical decision support to make staffing shortages, clinician burnout less painful. I think that's really going to be essential, as much as the individual technical capabilities of these sensors. I would say, similar to that model, the future of care I think with biomarkers would be if it actually has not just actionable, but an intervention, a change in management that actually is useful for a provider to know, because oftentimes, even if it's a little bit actionable, even if it doesn't change management, you don't necessarily need to know it right away at that moment in time. I think for a lot of these new technologies, it's actually changing what you would do as a provider for the next step. For example, Whoop, this watch, they actually published some recent data. It's indeterminate right now, but they could actually in their data show that they would be able to predict when someone's going to go into labor, a few weeks before they go into labor. There's a lot of really interesting data that's starting. People are picking up on all this stuff with AI machine learning, what they're seeing, but we'll, I think, find out in the future what actually is going to be predictive, preventable, and then ultimately prescriptive, which is changing and changing management. That's what I'm looking forward to. Bharat, you get the last word. I'll be quick. I think in 2030, every healthy patient is going to have multiple sensors on board. It's already the case now. An diseased patient or a diseased individual will have even more sensors. They're going to be a little bit more invasive, less non-invasive, but we'll have by then figure out, I think, more how to integrate data. I think the world will be owned by data integrative centers or companies, but this will not be solvable if the patient will not be in control of their own data because there will be data overload. I think 2030, patients will be more independently active on their own sense of technology and that will drive a lot of the care development and decision-making, as opposed to today, which would be physician-driven. Well, fantastic. I want to thank my panelists for a very robust discussion. I want to thank the audience, whether you're here or someone else in the ballroom, for joining in. I hope you have found it as informative as I have, and I hope you enjoy the rest of HRX 2023. Thank you.

digital health

artificial intelligence

cardiometabolic diseases

chronic diseases

patient engagement

healthcare technology

sensors

data integration

2030