EP18 David Crabb, MD of Rovex

Welcome to the STAT AI Podcast. Welcome to the future of emergency medicine. A realm where artificial intelligence isn't just a buzzword, but a beacon of hope and a tool for transformation. I'm DJ Apokma. My co-host Ethan Abin-Night are both emergency positions, and we have witnessed firsthand the burgeoning role of AI in our field. Our everyday experience grappling with health care disparities and social challenges have led us to a pivotal question. In the wave of AI applications, what changes are coming to emergency medicine, and will they be for the better? This podcast was born not just to ask these critical questions, but to seek out answers from the experts to understand and to demystify the potential of AI in emergency medicine. We believe in bringing this conversation to the wider community, to embrace AI as a tool for positive change. This podcast is our exploration of that future, but journey we invite you to join. So let's embark on this path together to learn the question and to envision a new era of emergency medicine enhanced by the power of AI. On this episode of the STAT AI Podcast, definitely realized as I went along, "Oh, this is going to require an army to build something like this, of people." And so I think the biggest surprise is just seeing the people that are willing to come alongside you when they believe in what you're working on and they believe in your team. Hey, Ethan. I have a confession to make. I finally joined a robot revolution. Oh, really? Did the hospital get one of those fancy surgical robots by chance? No, so much better. I got a Roomba. And let me tell you, watching this thing navigate my apartment has given me a whole new perspective on robotics and healthcare. It gets stuck under my couch. It runs out of battery mid clean. And sometimes it just spins and circles for no apparent reason. But you know what? It's still vacuum's better than I ever would. That's actually a perfect metaphor for where we all are with healthcare robotics right now. It's good at what it does, but it still needs you. Autonomy with oversight, not autonomy instead of oversight. Exactly. And I think when most people hear robots in the E.D., they picture a terminator doing tree odds or something. But the reality is way more like my room. Heldful, occasionally confused and definitely not ready to work on supervised. Which brings us to today's topic before we talk with Dr. David Krab, CEO of RoVex, about the future of healthcare robotics. We want to dig into a fascinating paper that shows exactly what it does or what it takes really to bring a robot into the emergency department. So the paper is called bringing a socially assisted robot to the pediatric emergency department, designed development and usability testing. It was published in frontiers in robotics and AI in October 2025, led by Dr. Mary Ellen Foster and a team from the University of Glasgow and sick kids hospital in Toronto. And full disclosure, when I saw the title, I was skeptical in robot in the pediatric E.D., really? What changed your mind on that? I remember this shift when I was a resident. I was trying to start an IVL like a screaming five year old. The mom was filming me on her phone, which was great for my anxiety, by the way. And I ended up spending about 20 minutes just trying to calm the kid down enough so I could even attempt the stare. And you thought to yourself, you know what the situation needs? A robot. Yeah, actually. I was saying that they're doing my best took me on my own, impression, trying to distract this kid while also maintaining stare, sterile technique. And I'm thinking there's got to be a better way. So when I read about this robot that's specifically designed to distract kids during procedures, I was like, okay, maybe there's something here. What I like about this paper is it's not just we build a robot. It worked. Hooray. It's a really honest look at what it takes to make this work in the real world. They spent years collaborating with clinicians, kids and families across two Canadian children's hospitals. Yeah. And what struck me the most wasn't the technology. It was the process. They used something called co-design, which is a fancy way of saying that they asked everyone involved what they actually needed, not just with sound at pool. So what does the robot actually do? Let's call the N-A-O, maybe now robot. That's about two feet tall. It's humanoid looking and it works alongside healthcare providers during IV insertions. It can dance, tell jokes, teach breathing exercises and it adapts its behavior based on how the kid is responding. Can you describe that adaptation? I'm kind of curious. Yeah, that's the interesting part, right? So they built in AI planning and social signal processing. So theoretically the robot can detect if the kid is getting more anxious or distracted and then adjust accordingly. But here's the key. There's always a human in the loop. So a research assistant has a tablet which basically has a big red stop button on it, which I think makes sense as we know from working in the emergency department, what sounds good in theory doesn't always really work in practice. Oh, absolutely. They learned that the hard one. In the first round of testing, there were some, let's say, learning moments. Give us a few examples of that. One case, the robot just froze mid-dance. Can you imagine your scared kid? This robot has been entertaining you and then suddenly it just stops. It's kind of like a toy running out of batteries. I think of the old energizer bunny. That might be worse than no robot at all. That might be super scary. Yeah. And that's why they did two full cycles of usability testing with 25 kits across two EDS, which is something, I guess, small numbers, but something. And they tested it. They got feedback, fixed problems, tested again. And then the robot would show repetitive behaviors, where we have awkward pauses or play the wrong music. All that stuff only becomes obvious when you're actually using it with real patients. I think this is why you can't just buy a robot on Amazon and then just drop it into the ED. Technically, you can buy a room book on Amazon. Fair point, fair point, fair point. Okay. So after all this testing and iteration, the research team came up with six specific design guidelines, but they really just boiled down to three principles that apply to pretty much any technology or brain into the area. And on my read, honestly, these are useful for whether you're thinking about robots or just trying to implement a new, for example, electronic health record feature or a triage protocol. Exactly. So principle number one is involve everyone from day one, not just doctors, include nurses, child life specialists, patients, families, everyone who's going to touch this system needs a voice in how it's designed. Because I think otherwise you would end up with technology that works really well in the lab, but falls apart in the real world setting. So number two was keep humans in control. So the robot can't override clinical judgment. Staff needs to be able to pause or stop at any time, especially in emotionally sensitive situations or if there's a medical emergency. I think you're thinking back to when your room book got stuck under the couch. That's exactly why you need a human in the loop. Autony is great, but oversight, as we know, with these with this technology is essential. Perfect callback and totally true. I definitely had to fish it out. But principle three is iterate constantly. Test, get feedback, make changes, test again, because nothing works perfectly the first time or the second time, apparently. But seriously, there here's a question for our listeners who might be thinking that's this sounds super cool, but was ever actually matter to my practice. Let me pose a scenario. You're working a shift or research team wants to test a robot in your department during peak evening hours. What's your first concern? For me, and I think probably for everyone else, is this going to slow me down and is it going to confuse my nurses? And lastly, what's going to happen when a resuscitation is called mid process? And that's exactly what the research team was worried about as well. That's why they spent so much time on workflow integration, training protocols, and making sure it fit into the actual chaos of the ED, not some sort of idealized fictional fairy tale version. Okay, so did it work? Actually, yeah, overall kids, families, healthcare providers respond very positively. It helped with distraction, reduced the burden on caregivers and didn't cause major workflow disruptions. But and there's always a but it wasn't perfect. Some kids thought the robot looked creepy. Some wanted different songs. Maybe I have no idea who kids listened to. Bluey, some were completely uninterested, which I think brings up a really important point about personalization. I can tell you from firsthand experience, one kid's fear of distraction is potentially another kid's nightmare. Exactly. And that's true for all of our patients, not just our kids. I had a patient last week who found my attempts at small talk actively annoying. And maybe they all do. I don't know. But what I'm trying to say is that some people want to chat somewhat silence and someone a robot that can dance the Macarena. I did this robot actually do the Macarena. I assume so. The paper didn't specify that, but if it did, they missed an opportunity. Agreed. But the thing we haven't really talked to yet about yet, I think is the elephant in the room, which is cost. Your rainbow was say $300 when you bought it on Amazon. These hospital robots were talking tens of thousands of dollars plus training, maintenance, plus ongoing technical support, which does bring me to the biggest concern. Always is equity. Are these robots going to be available at large academic centers like ours? Or is there a way to make this technology accessible to smaller under-resource hospitals? Because I think if we're not careful, we could end up widening disparities in healthcare across health systems in the US. And I don't want a future where pediatric IVs involve a mockery in a and dancing robot while the community hospital in Roam, Montana is still using, here's a sticker, please dot cry approach. - So let's zoom out for a second. The paper focused on one very specific use case, pediatric IV insertion, but the principles apply, I think much more broadly than that. - Yeah, and I would think that's what makes this relevant, even if you're not doing pediatric EM, which I do not often do. The lessons about co-design, iteration, workflow integration, but plays whether you're talking about robots or really any AI tool. And I think there are so many other potential applications such as medication delivery, patient transport, which is relevant to today, a vitals monitoring and assisting with triage. Some hospitals already have TUG robots, delivering supplies and medications. - And those are basically Roombas with car-gate space, think back to helps as a recreation, DJ Roomba, driving around the room play and music. They're not doing tests, they're doing tasks rather, not interactions. And what's interesting about this robot is that it's trying to have a social relationship patient. - Which I think is both the most promising part of this research and the most challenging aspect. It's easy to program a robot to form a simple task, like go down a hallway. I think it's obviously much harder to program it to read a room and know when a kid or a child a pediatric patient needs a joke versus when they need some space and a loan time. - And what I kept coming back to, all this research and development happened on the academic side, but what about companies actually building and selling these robots? Are they thinking about the same challenges? - That's exactly what we're about to find out today on this episode. We have Dr. David Krabb, who is the CEO of RoVex, which is working on robotic solutions for healthcare settings. And I'm really curious to hear his perspective on how companies can balance this innovation and practical implementation. - Yeah, like how do you build trust with clinicians who are skeptical? How do you price these symptoms? So they're not just a luxury item for wealthy hospitals and how do you handle the inevitable technical failure? - And what does the next five to 10 years actually look like because I think there is this tendency to assume robots will take over everything or assume they will never work at all. The truth is obviously somewhere in between that. - Yeah, we're probably not gonna have robot attendings anytime soon, but we might have robot assistance that makes certain tasks easier, just like my robot doesn't replace me, but it does handle the tedious stuff so I can focus on other things. - Exactly, and maybe there's a key, that's the key real insight to this whole paper that we're talking about. These tools really should augment what we do and not replace human judgment, which I think has been a common theme on the podcast over the last 17 episodes. The robot in this particular study wasn't trying to replace anyone such as a nurse or a doctor. It was just there to be another tool in the toolbox. - Yeah, and I would be remiss if I didn't point out that there's this whole other operational side that the paper touched on, but we haven't really dug into. I think that's where the industry perspective becomes really interesting. - Which brings us to our guest today again, Dr. David Krab is an emergency physician and CEO of RoVex, which is working on these robotic solutions for healthcare settings and patient transport. And I think he's gonna really give us a different lens on all this. - Yeah, I'm really curious to hear how someone who's both an clinician and running a robotics company thinks about these problems because he's lived on both sides of this equation. - Thank you for joining us on the STATDI podcast. Today we're here with Dr. David Krab. He is a board, certified emergency physician and clinical informaticist. He is also the CEO and founder of RoVex Technologies, former assistant medical director at UF Health Gamesville, former director of clinical informatics and data integration for the UFEM department, and has published research on clinical decision support systems for cardiac resuscitation. So without further ado, thank you so much for being here with us. - Thank you, DJ Ethan. I really appreciate it. Love what you guys do. And yeah, just really grateful to be here. - Yeah, thanks so much for being here. - Awesome. You were the director of clinical informatics at UF working on the smart emergency room concept when you founded RoVex. I think that was in April this year. You left UF completely to pursue this full time. So can you walk us through the transition moment when you realize I'm gonna go all in? - Really honestly love working in the ER. I think it's really special. We have a really cool opportunity when we get to take care of patients in a really critical moment. But I've been working on few different solutions and things that we can do to try to help our patients and our workforce. Our workforce is stretched pretty thin as you guys, I'm sure, no firsthand. And so that concept of a smart emergency room, how are we able to have the environment work for us to make our lives a little easier, so we can really focus on the patient. We need that compassionate bedside care. And so that's what it spawned some of the research that went into doing this. Ultimately the decision was made though when really solve it, it was gonna be really difficult to balance being fully committed, working in the department in addition to being fully committed to this mission. And I've seen my journey as a way of scaling up my impact on patients from the patient right in front of me to when I was doing informatics, taking care of all the patients and the come to the emergency department through the medical record. And doing robotics is a way to scale that even beyond the walls of the University of Florida. Tell us a little bit about your informatics background and that seems like a very sort of strategic type of plan. How did that pathway prepare you for this healthcare entrepreneurship in ways that maybe a purely clinical for the person might not have? And how has that sort of contributed to your trajectory now? Yeah, I always had the interest in software development, developing software that can assist our clinicians and even before medical school. Working on a project to create a pipeline for analyzing DNA sequencing data. I just thought that was so cool how you could write something and code and get such a real output. And so I was interested in informatics while I was in medical school. I don't think I realized all the opportunities that it truly opens. You get such a systematic view of how a hospital runs, how the ER runs, how you're taking care of patients and how small decisions that you make within informatics can have massive impacts. When I would measure my impact in a number of clicks saved. That's real time for our clinicians, for the patients that they're taking care of. And I was fascinated by that systematic view. And I think that it gave me a broader view of how innovation can be implemented to the emergency department in the hospital beyond. I love that explanation. I think I had a similar kind of tough myself with informatics realizing that you can really expound your own effect, like you said, instead of one patient you're taking care of all the patients. And I would be remiss if I didn't come in on our share UF history, so go get it or any gators out there listening. While you were on this path, you decided you were gonna take on transport. And I think just a picture for all the listeners, like what does a bad transport day look like in a busy emergency department? And is there a cascade of problems that are coming from that? How does it all affect patient care, the throughput, and I guess most importantly the team? Yeah, the transport problem came up when I was tants with finding the biggest bottlenecks that our ER faces around imaging, around labs, procedures. And our number one bottleneck that was recurring was patient transport. And so we started looking at, are there some other things that we can do to try to assist with that? But that's when the transport problem was highlighted to us. At one large hospital, they might see, you order a CT scan and the patient is ready to go. The CT tech said, hey, go ahead and send the patient over. And on average, they might wait over an hour to be brought 200 yards down a hall to a CT scanner. And that's no fault of the transporters. The transporters are overwhelmed. The number of transports that they have, their injury rate is really high. I know many hospitals, the average length of someone staying in that job is about six months. They'll have, they'll injure their backs, their shoulders. And so I think it's really harming our workforce there that they're having to be deployed on such a manual repetitive task. And they're doing something that they're having to put the patient's needs above the needs of their own body to really transfer patients, push them around, things like that. And I think if we can get them some help, that was what the impetus was. But the way that I look at transport in general is the hospital is like a miniature city that moves around 24/7. So just like in a city, you have the movement of goods, materials, and people that are done by companies like Amazon and FedEx, Lyft and Uber. But in the hospital, we task each hospital with figuring that out for themselves. All the healthcare workers that are stressed with taking care of life-changing moments for patients, they also have to think about operational efficiency. And it's not easy to balance all of those things at once. And my goal was, can we find a way to help with that operational efficiency for the healthcare workers? Because as you said, when the flow of patients, the actual movement of patients, is being halted, not because the workers aren't making the right medical decisions fast enough, because they just physically need help moving down the hall. And that's something that we need to find a way to automate and take care of, so that our clinicians and our staff can stop thinking about those pieces and focus on the empathy part. I love your analogy of the complexity of the hospital. We don't always recognize all of these issues come from just these systems converging together. And especially the emergency department where I think it becomes almost pure chaos sometimes. So our podcast typically covers AI applications like algorithm decision support, imaging analysis. I think you're the first guest we've had to discuss hardware, which I think is super cool and really exciting. And RoVex is building physical robots, which we're really going to dive into here in our next question. Talk about this distinction. Why is AI in hospitals fundamentally different from kind of what you're doing, say, from like large language model reviewing medical notes to your work? Yeah, so it's an interesting transition that's happening. Obviously, we're now in a fully AI world where that's assisting us in many ways can help just accelerate a lot of the things that we're doing, but only so much because it's still digital, it's still virtual. When we start talking about AI moving into the physical space, now we're talking about some of the most repetitive tasks that we all do every day, becoming automated, and that's a significant, I would say, improvement in people's quality of life and what they can do if they're able to not worry about some of those other tasks. I think physical AI is the next stage. It's really tough. It's not something that just gets done in a day, but its development has accelerated so much when you look at some of the large, for example, humanoid robotics companies coming in. For us in medicine, that's coming for us too. That help is on its way, and I'm really excited about it, and to be a part of it, honestly, is the physical help and movement of people if it's trustable and safe. I think it's something that is really going to revolutionize how we take care of patients. Amazing. You know who also shares that opinion is Dr. Feifei Lee, a builder of world models. And last week they actually just put out the most up-to-date world model. She projects that is actually going to lead to, if there is such a thing, the next generation of, let's say, generalizing intelligence or even super-intelligence them. I think because you get so much more feedback from living in live world, than you would from what's like, protecting the next world's descendants. I think we want to get to it. What? I'm curious, because in my head I'm thinking about basically a slap with wheels on it. What does this robot look like? Can you happen to have shared some images of presentation with us, but Ethan and I have not looked at it yet? If you're okay with it, we'll go ahead and open that up now so I can just satisfy this for myself and try to explain it. Yeah, I know we've been building the anticipation to pause it over and tell just a second. Let me give you a little brief. The RoVex were innovating hospital logistics through AI-powered robotics. Again, focusing on that logistics piece like I had talked about, but our first step along that way is, let's take care of the transport of patients. I think that's really critical. I'm excited about all the movement out there around humanoid robotics and what's happening there. The reality is, it's really capital intensive and really difficult. Instead, I focus first on a task-specific robot, the one that's built just for transport, with the goal later on as we move along, as we move towards FDA-approved robotics that are able to take care of patients, we can move more towards generalized robotics. Arz is tasked specifically for transporting patients. To go ahead and open up the presentation, we're going to go live reaction to what these look like. Here we go. All right, here we go. Okay. And tell me what you think. Whoa, that is cool. That's not right. Hold on, you're so espassy with me. I have it still type in a passage. All right. It's incredible. Roe v is an autonomous mobile robot that's designed to attach to any brand of hospital stretcher, specifically for low-acuity non-behavioral health patients. Grab the stretcher and then toes it along at two miles per hour. I love it. Wait for DJ to get a good look at this. Okay, yeah, I can't. All right. There we go. Click on image 2 for me real quick so I can see it. Go. Oh wow. Roe v. I was completely wrong, guys. Yeah. I guess I was expecting something from like the 80s where it was like on treads or I don't even know. I have no idea. Yeah. So Roe v. This is so much smarter than what I was what I said earlier. So I'm glad. Thank you. Thank you. Yeah, we're excited about Roe v. Roe v is the face is fully customizable. That's a screen, a display. And it's designed to be disarming and friendly. Obviously patients are in a vulnerable state and we want them to feel comfortable, but that is professional and a premium experience for them. So you can tell in the second image it's separate from the stretcher. So it backs up grabs a stretcher toes it along and then we'll release it and then move on to the next stretcher. That way that a hospital doesn't have to buy a bunch of robotic beds. And it's designed to be small enough to fit between in hallways, indoors, elevators. If you go down to the fifth slide, you'll see that for the patient, there's a screen for them. That's cool. Like tracker. Yeah. That's amazing. Cool. Yeah, not quite. We have really talented industrial engineers and Silicon Valley that help us on the overall design of the fairing, which is a really important part. And yeah, it's got like a flight tracker to see where you're headed and how long it takes to get there. And you can even have a staff member call in if they have any questions, video call in and talk to them it route because I still want to provide that human touch and not here to replace that. That's irreplaceable. If you go to slide six, you'll see little demo video. You'll see. Oh, I got you. Go ahead and click that and see if it works. I think it's loading. Let's see if I can get it. I might just describe it visually for those who are listening on one of our audio companion sites, but it looks almost like the cutest flatbed truck you've ever seen with a huge street on bank for the windshield that is smiling at you. And as you said, rather disarming, actually, like I want to get towed around in the rowbacks. Yeah. And then the road, the road is meant to look like it does what it's supposed to do. And you meant you hit it on the head, a tow truck type view. I almost imagine when I first imagined it, I imagine like a golf cart with a robot sitting up there at the wheel of it looking back at you like it's your chauffeur and giving you a thumbs up. That's what I imagined. Rovey as we first started and this is what it became. All right. Here is the video. You want to walk us through it here, David? Yeah, it's just a prototype. The way it grabs the stretcher and lifts it barely up off the floor. It's pretty nimble the way it moves around. Oh, wow. And again, this is just the frame of it without any fairing on it, anything like that. So you can see it can rotate. It can go side to side. It can avoid obstacles. And then it shows in Nvidia's omniverse the way the robot is built in there in a high fidelity simulation, the way it trains within hospital halls. We can actually rebuild a hospital like your ERs halls, the exact halls. You'll be like, there's that scratch on the wall that I saw that it's that high fidelity. And we train our robot in there. You can train it many times, normal time. So we can get thousands of hours of training in your hospital halls before it even steps in. And so that's one of the advantages. Thank goodness for Nvidia and what they're doing for becoming a platform to develop robotics. So cool. Wow. Just and it moves so smoothly. I love the movements are just like very gentle and you could picture a patient being really safe in this on this using this device. Yeah. The way the engineers, such an amazing job. These wheels can move 360 and just keep spinning the way they had created those wheels. And so it's amazing how they can go in all directions. So yeah, pretty excited about roving and what it's got in store for all of us. Wow. So honestly, like just recap, you've got basically this non human factor robot you're now using. I'm going to go ahead and call it digital twinning solutions with Nvidia to train it. The head already is so far out of kind of the norm of what I think we think about in the ED. I imagine that navigating a crowded hospital with vulnerable patients on board is going to present unique challenges compared to say warehouse logistic robots when you mentioned like Amazon and they played that. Can you walk us through again, beautiful design? I think you guys have really thought about it. Maybe just tease that out for us. The difference when you're now carrying a human behind you versus just moving a package. Yeah. A lot more variables that you have to think about because of what you said, having a person on board. So first of all, obviously moving at such a slow speed that an immediate stop is not going to harm, not going to harm somebody. So that's always the safety measure is low bar here for us to make it as far as transport time. As long as we make it a slow, intentional, steady journey, we will meet most metrics that are required. But then second of all, just making sure that that patient feels comfortable that making sure it's the right patient that's being transported and then making sure that we have the correct monitoring in place so that we are non inferior to the monitoring that they're already having in the hospital. And so hence the video screen for somebody to be able to talk to them. A camera back on the patient and routed their monitors by staff. Yeah, making sure that somebody has eyes on the patient. Remote patient monitoring. We wouldn't be able to do it without that as far as vital signs go, et cetera. And then we're able to stay in the loop as to what's going on with the patient real time and then react. Sometimes almost quicker than you would normally react in the ER. We can actually deliver a patient right to a nursing station if required or cancel their journey or call people close by. They'll know exactly where Rovia is. I'm still staring at the pictures of it. In awe of it. Forgive me if my next question is not on target. But hospitals are notoriously resistant to workflow changes. How do you think about adopting and change management with this technology and who becomes your champion inside the hospital? Is it the nursing leadership, the C-suite department directors? Have you interfaced with hospitals to try to talk with them about this amazing technology that you've created? Yeah, and you hit the nail in the head as I experienced clinician myself. And having worked with multiple companies in the past while I try to help them integrate into our workflows, I think the key is let's change the current workflow as little as possible. We need to work into what healthcare workers are already doing. They do not have time to have five extra steps added onto what your normal workflows. All you're thinking about is a patient in front of you. I don't know if you guys have any experience if you've had any tools that someone tried to introduce to you and you're like, "No way, I can't do this." Yeah, I was going to say every day. So I feel that frustration and so I am very conscious about we have to fit exactly into the workflow. And that's our goal. Even coming down to things like, "Oh, for transport normally, a brake lever would normally be switched to drive mode." But then after that, someone would go behind the bed and push it. They flip the brake lever for us and then it starts rolling. Then we're good to go. So, no. extra steps required for them. That's our goal. Obviously, that's a perfect world. I know there's going to be a few extra things along the way, but if we're saving them a lot of time, it should help. So that's the first thing is making sure we don't cause any other issues or any extra steps for them along the way. I think you have technology that's disruptive, but then it's momentary, and then it's adapt to it, and it becomes helpful. But then there's also technology. I think there's totally disruptive to workflows, and it's hard to know how to integrate it. And so I think that's a really good point you've made about how to figure out a way to kind of work with the existing systems that are already in place in the emergency department, because I know physicians are very like recalcitrant about changing their workflows. So yeah. Specifically emergency position. We're like, no, I've always done it this way. Yeah. You mentioned you were asking about who the champions are. Yeah. And I would say very much physician hand in hand with nursing staff in providing input on what they feel is correct for patients, what's comfortable. So it requires an innovative institution to do it. Let's be real. Some institutions are completely underwater when it comes to the amount of patients that they're taking care of, and there's a lot of kind of base level problems that they're still working on. They're really tough. And so you have to find the right innovative institution. And we have quite a few that we've been talking to in a few that really stand out to me as institutions that get it. Hey, I'm not trying to step on any shows, create any extra steps for anybody. But if we're going to get to where we need to get to taking care of patients, then we need some input. We need to work together on this sort of innovation. So it's been really cool to see some of that community rise up. And as you mentioned specifically the EM population, I think there's a there is a pretty good segment of really innovative physicians in that group. Informatics specifically shout out to the informatics community. It's been really cool to see they, I think they understand innovation and what it takes to get there as well. And then also really amazing to see the nursing staff coming alongside this. I think they're unfortunately some of the ones that the responsibility will fall on when transport breaks down. The nursing staff then ends up having to take the patient, even when they have many other patients they need to take care of. They end up having to be the ones to transport. And so they're the ones that feel the pain sometimes the most. And physically feel the pain as then can become injured while transporting. So that's a good point. And maybe a segue to something else I want to ask, which is there is maybe that some people would feel that there's a risk out here of losing jobs. So autonomous robots in general could be coming for our jobs. There's a South Park line in there. They're heard. But what has been the reaction from transport staff? I'm sure nursing staff and physicians know who've heard about the Roe. Roe. You've seen it. Yeah, I think the the nursing staff is usually the first ones to raise their hands and say, yep, we need help. As I said, they're the ones that often will suffer from some of the breakdowns of transport. They're the ones that have to shoulder that burden. Transport staff, I speaking specifically with them, my goal is in no way replacing transport staff. I actually think that they need to be almost focused to the patients that require human touch. Not that all patients don't require human touch, but you're 85 year old, delirious from a UTI patient. Roe should not be transporting that patient. I need a person to be their eyes on with them in route and providing that. That compassionate care. Whereas for the stable middle age patient that has abdominal pain that could walk there otherwise, but we'll sit still if they can go to CT 30 minutes earlier. That's the perfect populations. So instead, I'm trying to shoulder some of the burden. I try to help some of the burden with our current transportors in no way are you going to be able to replace all transport staff not at this stage. In even still, there are still jobs that are created when you do something like this. So a lot of people don't realize that, oh, now we actually need somebody who's monitoring all patients in route. I would say that's a less physically demanding job, less likely to become injured. And I hope that someone's able to enjoy being in that job actually, getting to monitor, talk to people in route and provide a service. And during what is hopefully a really premium experience for the patient. And my goal is really to elevate the jobs of some of those caregivers and really help lengthen the careers of a lot of workers in hospital. Which I think is a central tenet of informatics, right? Human plus or computer system plus human better than either alone. And I think that's what you're going here for. So it's great. Yeah, no, I think he and Lube for sure, especially with those critically ill patients that maybe wouldn't benefit from it. We talked a little bit about your background, David, in terms of your medical informatics. But I think a lot of us are interested as innovators. What is it? What are the skills you need to be a founder of a company like this? And how did you fill your sort of own gaps with either mentors, advisors, co-founders? And tell us about how easy or hard this has been. We'd love to just pick your brain about this because I think that's a real, a topic that our listeners are super interested in. Yeah. So I would say, man, I think a lot of people will see founders as it's a really glamorous job, really cool. And don't get me wrong. It is really exciting. It's such a privilege to be able to do this and help lead a team of talented engineers, just a bunch of talented minds. But at the end of the day, there is a lot of hours in the background that is not glamorous at all. Where you're going through your reading contracts, you're trying to file the rights government forms. I founded it from my computer in my bedroom, sitting there at 11 p.m. at night, one night, grinding. And a lot of times for a lot of founders out there, those hours are in addition to your full-time job. Right. And so I still had full-time responsibilities that I was doing. And doing that on top of that is it is really tough. You have to make sacrifices to be able to do that. And being willing to take responsibility for any and every part of the company, when you first start, you are the CTO and you are the CEO, you're every part of the company. And you hope to be able to recruit alongside you really talented people that can help to show you that burden. But when it first starts, I hadn't looked at so many contracts before in my life and you find the right legal help, but once you do, then that helps. So there's a lot of pieces where you're learning. You really end up becoming a learner. And that's why I think physicians are uniquely suited to be able to do something like that because a lot of physicians are really quick learners. They can pick up information really fast. And luckily now, thanks to AI, education is so much faster. Self-education is so much faster. Yeah, you can verify your sources, everything like that. But it requires learning so much, so fast. It was like going back into medical school again, just in the business world. And then on the other side, I think overall having that business mind, realizing that you can't conquer the whole world at once and you have to do things in stages. You have to do things humbly. We started in a garage. We are in a garage for a long time. The number of times that I brought in high net worth individuals for investor pitches in the garage of the house where the engineers were living. It was I was cringing every time they would roll up and it's yeah, take a look at our robot and the garage. Now we're at a higher stage, but history of garage is doing big pains. I was going to say, I think Jeff Bezos has a picture of himself in his garage with a bunch of books. So I think your, I think your humble roots are okay. Yeah, but you have to go through those the stages. So that's yeah, that's it's a really tough part of it. But thankfully, you've gotten through some of those stages. We still have a lot of parts that you're like, Oh, I can't believe we have to show people this is the best we got right now. And then other parts like you guys get to see the renderings and everything. Really exciting and prep. Cool. I just realized that in New York City, we don't have garages. And maybe that's sending us back in the innovation game. We all need to go on and get a ground floor or something. What's the equivalent in New York? What's the equivalent of the garages? Is there like an attic space? Is there like a underground like seller space? I don't know. Yeah. None of those things existed. It would be the lobby. I imagine. But one thing you said that interested me is that you said you got a little bit of help from AI. And I guess as a founder, are there any applications that they were giving any endorsements here? But you can speak about them broadly or generally that you found to be useful aside from kind of you running the mill chatbot. Yeah. Some pretty amazing ones that I found along the way that you're just like, I have a need and then you find it. Yeah. Obviously, no major endorsements, but I'll say first of all, obviously, Chat GPT. Thank God for Chat GPT. So helpful. And obviously you have to vet everything that it says because it will otherwise lead you down the wrong path. But when you when vetted, it can provide you with some great framework. I know I think one of the one of the founders out there had said that we're entering a time now where someone is going to create a billion dollar company by themselves using AI. And I believe it. I could not have done that for sure. Not in the hardware space, not anything like that, not me, but there are there is going to be somebody guaranteed somebody's going to do it. But that to start. There were a few other really interesting ones. I found a payroll app that I don't think ADP even realized existed called role by ADP. They it's by ADP. I must have been a startup or something that they started out of it where it's literally a chatbot that goes through on your payroll and you talk to it. And it was so inexpensive compared to all of the other payroll applications out there. Yeah. And so yeah, it was pretty amazing. It works. But it's like not advertised. Whatever it's real. Like I couldn't believe it. And so that's an example of some of the AI tools that you find that you're like, how is this chatbot real? But it actually does what it's supposed to do. That's amazing. So you left a need guide you basically. And then maybe have Chat GPT run down some leads. It's so cool. Just hearing you talk about all the different tasks you're doing from payroll to everything else. I can't imagine this is just incredible what you've done. So much work. What's been the most unexpected part of this journey with the company? Something you didn't anticipate when you started RoVex. And you could describe either a challenge or an opportunity or maybe a collaboration that caught you off guard that you were like, oh man, I didn't realize I didn't realize that I was going to go down this pathway. Yeah. I was going to expose how naive I was when I first started. But when I first started I was self-funding it by myself. Oh wow. And I thought, oh, I'm going to be able to self-fund this thing all the way through. And I can't tell you how naive that is because hardwood, not only is it very expensive to correctly grow a startup, but it's also hardware specifically. It's very expensive. And it's not just because the materials aren't actually that costly in the grand scheme of things. It's all of the other things that you have to do, right? There is all of this compliance testing that you have to do. Functional safety. It's everything around the FDA. What I thought was enough that I had saved was not going to go very far. It's definitely realized as I went along, oh, this is going to require an army to build something like this of people. And so I think the biggest surprises is just seeing the people that are willing to come alongside you when they believe in what you're working on and they believe in your team. And those are the biggest two pieces when people invest, when people join. It's really because they believe in the idea, they believe in the team, they believe in the mission of what you're doing. A lot of them are not sitting there thinking like just dollars. It's if they wanted to go out and make money, they would just get a second job or something like that. They're coming alongside you. And I think that's been one of the coolest parts of the journey. Yeah. Having those people come alongside. And they're more than that. They're encourages, they're checking in on everybody. They know about the team. It's just been really cool to see that part of it. So it really becomes like almost a community bill that's putting this forth as a solution to help people in the hospital. And it just forms my heart to know that we have such a strong community like that with us. So cool. That makes a lot of sense a long, long, long time ago. I worked as a specialist in the Apple store, which is like the people you see right when you walk in, the breeders who are selling you computers. And it never ceases to amaze me how much overlap there is between that job and the one that I do now where there are a lot of people coming in and they have a problem. They want you to fix it. They want you to fix it now, not yesterday. And so overlap is something that interests me and you've now experienced both worlds emergency medicine start a building. What skills from DM training triandly best to entrepreneurship is the ability to triage is it working under pressure managing teams and crisis. It sounds like the very good one. But you tell me help our linux listeners understand that connection. Yeah. They what's it it's surprising. There's so many things that cross over and you hit multiple of them. Where do I start? Let's start with let's start with tasks switching in the ER as you don't want to task which you rather would plan out when the critical patient's going to come in. Can we just put them on a schedule? Can they come in at 2 p.m. We're all prepared ready to go. Yeah. But the amount of task switching that you have to do and go directly to it because it's the highest priority thing going right now and you have to be all in. And that happens a lot, especially in the startup and early stage startup world. You have to be able to do that. You mentioned triage. People talk about some of the best founders in the world. Their secret skill is knowing the most important thing and working specifically on that because there are as a founder there's 100,000 things that I could work. Yeah. But you have to identify across board and you have to talk to a lot of people. Make sure I don't have any blind spots because there's going to be blind spots. What are the things that I'm missing that are actually the tasks that I have to do right now and people will tell you all different opinions and you have to sift through that and then fine. Okay. I'm going to put my efforts towards this right now and that triage piece as you said, you do the same thing in the ER. You'll have six patients that you need to take care of, but you actually know it's the quiet one over there whose vital signs are really borderline and you know about their history and people aren't paying attention, but you're like that's the person we need to focus on right now and not the person that's standing up walking around yelling whatever. And so having to do some of those similar things. It's been really helpful. And then I would say again, there's a million things that cross over. But having that in the ER, you thrust into this real-time leadership position that is not always the case in a lot of other specialties and really respect and appreciate a lot of the other specialties, but you are in real-time leading a team of 20 plus in the ER. Everyone knows your name. Everyone needs some attention from you. Things like that needs your help with something. And so knowing how to direct that team and handle your own needs, your own emotions, things like that. And in no way, let it be known in no way am I perfect in any of those areas. But as you guys know, you hopefully improve in those areas and really focus on trying to improve in those areas so that your team can be as productive as possible. And that really crosses over here in the same way. Everybody's right next to each other here. Everybody sees me every day and I have to consistently lead them through sometimes some really scary times, but also some really exciting times too. That's some amazing advice right there that you've just dispensed. I think there's so many pearls in there. So thank you for thanks for sharing that with us. Pain is a picture of the future using your crystal ball. So if RoVex and I and looking at the preliminary markups that you just showed us, I think it's, I anticipate it's going to be very successful. But if it does succeed beyond your wildest dreams and in your technologies in every hospital in the US, what do you picture sort of hospital operations looking like in 2030, 2035? And what sort of changes does this mean for patients not only from the RoVex perspective, but for clinicians and for other autonomous robots fitting into healthcare delivery as well. Yeah, I think and you mentioned that our number one priority is still the patient still at the heart of it. And so my hope is that they experience in what is otherwise a chaotic environment that's really hard to help provide some structure and framework to I hope that we're able to provide a little bit more expectation and framework around it so that the hospital is more able to absorb absorb some of those big bullies of patients and and being overwhelmed by the number of patients. And in my dream, I would have FDA approved robots that are helping take hands on care of patients to do some of the even the little things helping them to the bathroom, getting them a cup of coffee and knowing that they're allowed to have a cup of coffee right now. But even little pieces like they can be done and taken care of so that they have a premium experience so that when it comes to the medical side, our clinicians are nursing staff, our physicians are techs, paramedics are able to really focus on that medical side, providing that hands-on care. Honestly, hoping for a lot less injuries in the hospital. I'm sure you know of different physicians, different staff members, nurses staff. I know some of our paramedics that experience injuries and it shortens their career, makes their career a lot tougher. And so my I would love to see that it is a job where they can they don't have to worry about that side and they're able to stay healthy on the on the injury. And I think that would be really amazing. And just in general, my hope is for a more premium expected flow of patients through the ER as much as we can affect it to automate a few more things. Yeah. It's so hard to address those surges of patients that come into the ED where it's like, oh, you're just sitting there, you're drinking your coffee and then I'll send 12 people register. Three of them are really sick. The rest of them are okay. And then you're trying to deal with all that complexity. So I do like the idea of RoVex offloading some of these other tasks and automating transport and getting those patients with who need like a simple plane film to x-ray quickly and expeditiously and safely. So it's just I think it's a really as a system. I think it really complements everything we do in emergency medicine. And also I'm assuming in the inpatient world as well. Yeah. Yeah, couldn't reward. I think we'll probably end it after this question here. It's been a great interview and thank you so much for sharing some proprietary information. Hopefully we haven't said too much in our descriptions. I guess one thing to leave us with is like we said, you're the first founders we've talked to. You're doing autonomous robots, which is different than I think a lot of the AI that people think about when they think about AI for a learner, a resident attending physician who's starting out or maybe even deeper in their career. If you had one piece of advice to impart for them to get involved or maybe something that someone told you along the way that really stuck, what would that be? Yeah, and I'll throw out there as far as people that that really stuck with me. Adrian Tendell was our chair of our department when I was in residency and when I when I was graduating and Adrian has a just a special gift of casting vision and people's lives as to what he thinks that they could do and he really believes in him. And it's amazing even when you're like, I don't know if I can do something like that and he's no, I think I can see this in you in the future. And that's what he did for me. I had the opportunity to go out and work in the community versus stay in academics and work on some innovation, some publications that I ended up getting. And he had said, you could do that or I want to tell you, I think you could do something really innovative that can really change the world. And if you stay with us, you can, I think you can accomplish that. And he really inspired that belief in me that we actually can make an effect beyond just the halls of our own hospital. And so having as much as you can that broader view that that some of the changes and some of the things that you're doing at your own institution, you could potentially see if it can scale up and help other institutions, making connections and making genuine connections with some of the other leaders and some of the other amazing people out there. They're working on some innovative ideas and ways of improving medicine. I think that's really helpful. I can't believe how many of the connections and honestly friendships that I made all through residency and working in some of those physicians and staff members are working all over the country. And those connections then now coming back and being really helpful in advancing this for this. Just been an amazing part of this journey. You don't know all the relationships that you're making even in your training, even in the early time period. You're going to be friends with some of those people for your life, which is really cool. And to see that then come full circle, just an amazing experience. And so I would say really maintain some of those relationships, really value some of those relationships. The 3 a.m. run to get coffee when you're on with the office service resident who does internal medicine or something else. Like those are real friendships that you will continue to have. I've just been really grateful for that experience. That's perfect. Thank you so much. This has been a fantastic conversation. Yeah, this is amazing. Yeah, so happy. Thank you. Thank you, David, for joining us and really appreciate it and hopefully have you on again. I really appreciate you guys hosting me. I also love what you guys do. And I really hope that just the listeners grow in this because I think you guys have a fascinating mission that you're on. And just so many interesting stories to hear out there. And so I look forward to listen to you guys more. Thanks. Awesome. Thanks David. We really appreciate it. Thank you. Thanks so much for listening to episode 18 of the Stat AI podcast. If you get a chance to check us out on X and Instagram and our newly updated website, Stat AI podcast.com. Make sure to tune in next month for some more amazing content. Thanks so much and take care. Please note that the views, opinions and information expressed in this podcast are solely those of the individual speakers and do not necessarily represent those of their employers, our sponsors or any affiliated organizations. The content provided is for informational educational purposes and is not intended as medical advice. We encourage you to consult with qualified healthcare professionals for any medical concern or treatment decisions. This podcast is not a substitute for professional medical advice, diagnosis or treatment. Listener discretion is advised.