Hey, welcome folks. It's stay off my operating table with Dr. Philip Ovadia and we are joined today by Dr. Ami bot who Phil this one is both different and pretty intriguing to me. So I'm just gonna step back and listen and then I'll butt in when I, you know. When my brain says I gotta do that sounds good. I'm really excited to have this conversation with Dr. Bhatt. Dr. Bhatt a cardiologist but we're probably not gonna talk all that much about the traditional sort of cardiology stuff. We talk about, you know, cholesterol and inflammation and insulin resistance and all that. Dr. Bhatt came across my radar. I can't quite remember how, but she is the chief Innovation Officer. I hope I got that title right. Yep. You got it. She's shaking the head. So I nailed it at the American College of Cardiology and I've been seeing her on social media talking a lot about AI and medicine and I knew that we would have a great conversation. Before we jump into all of that AMI maybe you can just kind of give your background to our audience and what led you from sort of traditional cardiology world to innovation and ai? No, absolutely. Thank you for having me. It is really great to be on here. And as a cardiologist who works so closely with cardiac surgeons, I have to say stay off my operating table is just genius, right? Because that's all we want. We just want people to go upstream and be healthy. I, I was used to say, the less I see you in the way, the better. It means you're doing so well. I you know, it's funny, the word traditional. I used to call myself traditional, and then sometimes you think about it. And I studied medicine and pediatrics, so I upfront said, wanna be able to take care of anyone who walks in the door. And so that put me a little bit in the minority, but I loved it. And then I took care of adults who had heart disease from when they were children. And so that was a field that didn't really exist at the time that I started. But I, with a group of people who largely had studied pediatric cardiology and said, Hey, we gotta be ready for them when they become adults and we need to train adult doctors to understand them too. And so that was maybe a little less traditional as well. And so I realize now I might be a builder. I might like areas where we don't exactly know how to do it yet, but we know it needs to be done. So I did that for many years. Then I started telemedicine. Because I had a lot of young patients and they wanted to FaceTime with me and I was like, oh gosh, no. Like I'll get arrested, we can't FaceTime. So we started doing some telemedicine and then I happened to be the director of outpatient cardiology at the Massachusetts General Hospital in Boston when COVID hit and they said, Hey, does anybody know how to do that telemedicine thing? And I said actually, yeah. And so it was a really great experience to see how many people we could reach using digital technologies that we were a little uncertain about before, right? Do we really wanna do this? Do we know how to do this? And then, you know, my favorite phrase is desperation is the mother of adoption, which is when you have to do it, you figure out how it works. And yeah, so that's kind of how I ended up, you know, towards the end of COVID. And that's when this position of chief innovation officer of the American College of Cardiology opened up and it was a chance to offer high quality healthcare in the communities where people live. By meeting with the companies and saying, Hey, can we help you understand us as doctors? Can you, we help you understand our patients and can we help your really awesome technology actually take off and stick rather than end up sitting on a shelf? And that's my current job. Very. What kind of technology are we talking about? Oh gosh. We're talking about everything. We're talking about AI for your coronary CAT scans that tells you've got blockage, you're not, and you don't have to undergo a catheter procedure. We're talking models where you look at an EKG and it tells you things about your heart. The human eye can't tell, like liver disease. I don't prefer that one because, you know, I like to have a drink now and then, even though I'm all about prevention and so I kind of don't wanna know. But but yeah, things from imaging to predicting, you know, your risk in the future. And now with generative AI to you know, can I help people in rural America get healthcare faster? Because I'm going to the clinician closest to them and I'm giving them the tools to say, Hey, I'm gonna ask a complex question, but can you just tell me enough that I need to know to get this patient to the next best place? Because there's so many people who just have no access. And gosh, we can upskill a lot of people using AI to start figuring out, hey, sick versus not sick. And that's everything.'cause if you end up in an emergency room, you end up on Phil's operating table. That's what happens. We gotta stop letting people end up in the emergency room. You know, being the heart surgeon that I am. I'm gonna ask you the tough questions first, right? Yeah, I love it. We're sitting here, we're talking about AI and, you know, all this stuff, how it moves medicine forward. And yet, you know, in medicine, we are, I think the last industry that I know of that still uses fax machines and pagers. Okay. And pagers, right? There have been some traditional barriers to innovation in medicine. I'd love to kind of get your thoughts on, you know, what those are, how we start to overcome them. You know, because I agree fully, right? We need to move things forward. It can have some amazing impacts on our patients and quite frankly, on. Doctors as well. You know, there's lots of hope for how this could maybe start to turn around the negative trends we've seen around burnout and, you know, all the administrative burden. But, you know, how do we get past these barriers to innovation that we've just, you know, seem to have had in medicine for at least my whole career and probably yours as well. Yeah. No it's a perfect question and I used to give really long answers to this. And as you grow older, you grow a little bit wiser. Not a ton, but a little bit wiser. Two things, workflow and respect. There's not a lot of technology behind getting people to adopt. It's, Hey, look, I've gotta get through my day. I have a lot of patients who need me. And I need to see all of them in the office. And then I need to go up to the hospital and I need to run on those patients and see them. And my day is busy and I have a way that I was taught to do that, and I probably still do that way 23 years later, you know, so can you just take your technology and fit it into what I'm doing? Can you take part of what I'm doing and make it easier? Can you not make me learn brand new ways of caring for patients? And I think we used to always say you know, you gotta be willing to flex. And the answer is, we don't have enough doctors and nurses to care for patients. They will always be overwhelmed for the foreseeable future. Right now, as we kind of work on training, you know, more people and new people, and so you have to fit into the workflow. And I know people say you have to redesign the workflow entirely and we're getting there. But right now, if you wanna get the technology to the patient. You gotta work with the workflow that the clinicians have. You have to take things off their plate that they're okay with you taking off their plate. You have to keep things on the plate where they need it, but get them the information to back them up faster. And so I think workflow is number one. And then number two is respect. You know, we used to be graded on how many weeks did you wait to come and see me, Dr. Bott. But that wasn't on me. That was the system, right? Like when was the next available time? I could be in a clinic room where there was an opening where somebody hadn't put a patient and the system didn't have this like block and tackle model of leave two things open every Friday, only fill them on Thursday. Eventually we learned how to do that. But then, you know, the scoring would say, Dr. But doesn't see patients for six weeks. And you'd say, no, I would see them every day all the time. If you actually called me, I would be late for dinner to see that patient. And that's what most. Doctors and nurses are like and so I think you have to respect that people are trying really hard and you have to start saying, what can I do to lighten your load? What can I do to help you get through this? What can I do to make you feel less burdened when you're caring for a patient? So I'll give you an example. Typing and talking to a patient. We always used to say it's better for the patient, right? It's respect for the patient. So true patient feels respected when I make eye contact with you rather than doing this thing. However, if I'm typing what you said 10 seconds ago, I'm not actually hearing what you're saying now, I'm not. I'm gonna miss things. And so it really does, you know, a affect the relationship. And so one of the greatest AI has been, hey, voice to text technology. I'm gonna listen to conversation, I'm gonna type your note for you. You can edit the draft later. Right now, eye contact, patient happy, doctor, happy, right? And so I think that respect for doctors are trying really hard, is really important. And then the last piece of respect is it's not consumerism, it's patient agency. If my patient comes in wearing a wearable and I get irritated, that's not the future of healthcare. So we need to teach our doctors and our patients. What are you supposed to take away from the wearable? What specifically for you is helpful? Is an oxygen level helpful? No, it has nothing to do with you. Fine. Then you shouldn't really pay attention to that as much as your atrial fibrillation or something else. And so we need to start engaging in those conversations. But we can't treat it like consumerism. We have to respect our patients wishes to wanna know their body. And we have to teach our doctors how to understand what those wearables do. That's not part of what you and I learned. Nobody had a wearable chapter in a book, you know? Yeah. And so it's something new we have to figure out. So workflow and then just respect for patient and clinician. Yeah. So on that last point I don't know if it made it across your feed, last week or so it came across my feed and there was some doctor, you know, sitting in his car complaining about how all these patients are coming to him, and they've run everything through chat, GPT, and they're saying you know, and they're kind of, challenging him, right? That, you know, GPT said this and, you know, why aren't you telling me that? Or why didn't you know that or Right. And and you know, even before GPT we had the Google version of this right. Where you would hear doctors annoyed when their patients went to Dr. Google. How do we, you know, start to change our colleagues, perceptions around that. I mean, me personally I love it when a patient comes to me and they're like, here's my GPT summary of, you know, everything I've had done and this is what it says. And then I'm like, great. Now I don't have to go through, you know, all the records quite the same way. You've already summarized it for me. Beautiful. Let's talk about it. Yep. But many doctors don't take that attitude. Yeah. I encourage we, we like to use the phrase AI enabled clinician. And I don't mean just in medicine. I encourage people to use generative AI in every aspect of their life.'cause the minute you do, you know, the other day I came up with this recipe, I was like, oh this is what mom used to make. I found it online this looks great. I'm gonna make it. And I told my mom, I was like, mom, I found it. It looks like what you're gonna make. And then my mom is yeah, no don't use the garlic in that one. I was like, no, but it says on it. And she's trust me. Don't use garlic. It's gonna totally change the flavor of what you're making. Of course I go ahead and I use not just some garlic, right? I'm like, I use like a bunch ofs of garlic and sure enough, like it was not the way she made it. And so it's how I approached her first of all was like, I found it and I know rather than, Hey, I looked this up and this is what I'm seeing, you're the expert. What do you think? And I think the same is true for patients and clinicians. First is just teach our patients like, Hey, you are expert in something. Probably whatever. It's yourself, your, you know, your job yourself. How would you feel if someone came and told you how to do something that you've trained to do? Not great. Remind our patients. But the other is then we remind like ourselves, right? This person is coming, trying. I'm coming to mom saying, I really wanna make that thing you used to make.'cause it reminds me of childhood. And if you see that part like you do when your patients come in. I'm putting things together and I'm coming you 'cause I care about my health. When they come like that, we should be coming right back to them and saying, this is great. I'm so glad you did this. However, right? Yes and yes. And I'd focus on these things, not those things. I'm so glad you brought it to me. Sometimes this is right, sometimes it isn't. These few lines look totally right. This is awesome. These ones weren't now. Anyway, in addition to that, we need to cover three more things in your appointment today that are not the thing you Googled or searched or chat gtd. Let's get to those. So I think again, it's about relationship and respect, but if docs started using large language models in all other aspects of their life, gosh, they would learn so much about how awesome it can be, how it can sometimes be wrong, and how it makes you feel like an expert. And then you kind of realize you're not, like you open the hood of the car and you're like. Oh God. Like I am not the expert I thought I was when I was at the end of that c Claude thing about how to fix this. So I think that's a big part of it. I'm, gosh, I'm not even sure where to start. I'm gonna ask the really stupid question that I had earlier, but I decided I'd just keep quiet. The American College of Cardiologists, is that the Professional Association? Yes. Yes. When somebody is, when I see that somebody is a member of such a college, does that mean they've passed some sort of, gotten over some set of hurdles? Yeah. The American College of Cardiology is a global nonprofit, actually, despite the word American in it. We have about 60,000 members, over 2,500 health systems and the largest cardiovascular data registry kind of in the world on procedures and other things. And anybody can actually kind of be a member if you're in administration, if you're nursing cardiology, cardiac surgeon, subspecialists. What about an ad writer? Could a guy who writes ads be a good for, you know, if you decide to like really f on that, only ads that create cardiovascular education, I could probably find a spot for you. Yeah, no, I'm not interested. But but there are designations, there's fellow of the American College of Cardiology associate, and so there are actually levels of Hey, you've really shown that the work you're doing is moving the field of cardiology. People have vouched for you, anyone can be a member and learn from our materials. Come advocate with us on the hill, participate in, you know, registries for your patients. And then if you want to, you know, continue to continue your career as part of, but it's primarily, it is primarily healthcare providers, primarily cardiologists, healthcare providers, and administration. Cardiology adjacent, I guess you got it. Exactly. Okay. So what are the single most common complaints that come in from the healthcare providers? Yeah. About technology. And then on the flip side, what are the what are the most common complaints that are coming from patients? Yeah. Two, the healthcare providers vis-a-vis technology, assuming I love this, you're gonna have the answer to those questions. I do. I we do surveys and I, you know, I get barraged with these every day because people feel that I could fix it, which I try my best to. Three key things for the clinicians or the providers. The first is somebody decided to buy something and implement it without thoroughly vetting it for how it affects the clinical process. Like, why? Can you gimme an example that would make sense to me? Somebody took our electronic health record and they put a hard stop. Anytime you have atrial fibrillation, which is a stroke risk they put a hard stop. You cannot get through your note. You can't go anywhere until you say, why or why not? That person is on a blood thinner. Now that sounds right. If you're an administrator, if you're putting it, you're like, yeah, they have atrial fibrillation, they have a stroke risk. I'm gonna make it a hard stop. You tell me why. Now the problem is if I'm seeing the same patient again and again, and there is a rationale for them being on or not on, I'm having to now explain this again and again as we're doing this and going through it. And so is there a better way to get at the goal of having more atrial fibrillation patients appropriately be on a blood thinner? It turns out, yeah, there are better ways for us to create a mechanism to do that. That isn't every single time I pull up that patient, it's yelling in my face and I can't get past it. So that was an actual complaint that came to us and said, please don't do that. You're making it very hard for me to get past. I think it was something like 17 clicks. And it wasn't, you know, a CC doesn't create this, it was at an institution. It doesn't matter which institution it was, but Right. So that's kind of one thing is administration thinks that this is an awesome tech. Let me give it to the team, and the team doesn't know. The second thing is. You need the team to understand. So when we first started telemedicine, I remember I was walking past the desk and one day there was a woman and she was a front desk admin. And the way she was talking about it was, Hey, this doctor can't see you in the office, but they could do this video thing with you. And I thought, oh God, did nobody give the front desk a script for what telemedicine is, what its benefits are, why we're doing it? And so that got some patient feedback my doctor doesn't wanna see me. They put me on a video. And then we had other patients who came in and said no, I didn't wanna drive four hours to have a conversation. They didn't even need to examine me. I had my test the day before. I just wanted the results. It doesn't make sense. I love telemedicine, right? Because they had a better experience. And so that's the second thing, which is the whole team, everything involved in healthcare, all the people who touch the patient, including the patient. Need to have buy-in the vision's not just for a doctor. The vision is for the system. So then you gotta, you have to share that vision and invest the time. And making sure anybody who's working at a hospital in a clinic understands why are we doing the new technology and what is it? The difference is amazing. The patient experience is so amazing when everybody gets what we're doing. And so I would say maybe so that we don't go on too long a list, 'cause I could do this for the rest of the hour. Like I think those are some of the things that we really see. Yeah. You know, telemedicine is such a interesting example, right? Because, you know, pre COVID you know, if you went to your, you know. Average doctor, right? And said, oh, you know, you're gonna start doing some appointments over the computer instead of having people come in person. I would say the vast majority, you know, maybe it's 99% of doctors would've said, no way, that can't work. It's not good. You know, all of the, you know, excuses. And then COVID happened, right? And everyone sort of got dragged kicking and screaming into this, right? You have no other option. You gotta do this. And they're like, oh, you know, and then we kinda came outta COVID, right? And I think a lot of doctors were like. You know what, that is a lot better. You know, and the patients, like you said, were like, yeah, I didn't have to drive three hours, sit in your waiting room for an hour you know, just to hear you like, you know, gimme some test results. And we found like where it fits, and now I think we're on this other side where, you know, we know where it fits and we know where it doesn't fit. And now we can start to sort of, start to use it selectively. And I think now if you ask most doctors, they'd probably put it in you know, yeah. It's a positive thing. That adoption was not necessarily smooth. And like I said, the in my opinion, the only reason we are where we are today is because COVID happened. If COVID didn't happen, most doctors would still be saying, no way I'm ever seeing my patient over a video. It was an accelerator for telemedicine and digital health, like nothing else. Yeah. I mean, it really enabled us to say, Hey, I can provide care. To a patient where they live, you know, it changed the paradigm from you have to come to me in the big white castle to be seen to, hey, we can go back to home visits. Like I can come to you and sometimes in person, a lot of care at home models now, right? Sometimes with remote monitoring, send you a scale, send you a blood pressure cuff, you upload stuff, we chat about it. And then sometimes being able to just discuss things over video rather than driving three hours, fighting the traffic, paying for parking in the garage, finding childcare or adult care at home. All to talk to your doc. Yeah. Now and then, but we're in this you know, we've run into another problem, right? We have regulations that are outdated, right? We have our whole just licensing system, right? You know, so I'm the crazy surgeon who decided actually right before COVID, coincidentally, I'm gonna go start a telemedicine practice. And then I said, okay, I'm gonna see patients everywhere and I now have licenses in all 50 states and Washington, DC and it's a nightmare. And you probably have an idea, right? But, so now we're in this place where great, the technology is awesome, right? Patients love it, doctors love it. But now we have these barriers, you know, on the regulatory side of things that are holding us back. So how do you how do you kind of, start to approach. Regulations move slowly, technology moves quickly. Like how do we start to marry those two things up? Yeah, this is so I gave a TEDx recently, and in it I talked about this idea, right? Technology fails when it moves faster than the humans it's meant to serve. And so in that sense, we don't want technology moving so fast that just the tech is going and we haven't understood is it helping the patients or not. Having said that, right now, our regulatory bodies, at least here in the US, are really recognizing we can't add additional barriers that are slowing things down either because our patients are seeking non-approved mechanisms to get information to think about things, to get better. Because we aren't bringing them the ones that are actually approved. And so actually what you see now is this almost a lovely balance of make sure that your technology is not moving faster than helping the patients. We don't want that. But the same time, instead of regulating technology, create an infrastructure, and that's what our government's doing right now, is trying to create an infrastructure for, if you're gonna use ai, if you're gonna use digital health, let us help you build the way you're gonna use it. And some of those programs, you know, CMMI or you know, the Center for Medicare, Medicaid. Their innovation group has this program called the Access Program that's trying to help people get more telemedicine therapy for things like blood pressure and cardiometabolic. Arpa h is a, is another place that actually funds things and they're trying to fund ways to manage patients with heart failure using somewhat of a more agentic ai, not no, no human at all. I was about to call us adults, which we are, but not no human at all, but like, how much can you automate things to make life easier? And so there are pathways from the FDA, from you know, these places that are trying to say, Hey, if we give you an infrastructure with guardrails now, can you create things for our patients who are trying to use stuff that's not approved? Instead can we move a little faster? The most exciting part is when these things happen, we get to actually see the real world effects and measure them because they're out of a program. We get to see what happens when technology is out there. And so I'm actually really excited right now. I think for the first time we're at a point where we're actually trying to push technology to find an infrastructure and get it out to people. So I think that there's hope as long as we, as clinicians stay involved, right? And don't just let it run. And as patients advocate for themselves, if a tech doesn't work, like we need to have that conversation. It strikes me the thought strikes me, and I could certainly be wrong 'cause I was wrong once, before, a long time ago, that what we're talking is tactics rather than strategy that we've gotta, when in fact we have a strategic issue here. And until the strategy. Shifts, the tactics are going to be although not ineffective, are gonna really face an uphill battle. So let me get specific. Yeah. There's a, there has, and I'm speaking as a healthcare consumer, not a provider. I remember when I talked to when Phil and I were discussing starting this show. Yeah. And I told him I would not I wasn't interested in doing it unless he answered a bunch of questions I had about healthcare. And to his credit he answered every one of those questions. I did not pull any punches. He answered them all on the record. And and it sounds to me like. What you're dealing with is more of the same that I as a consumer, that drives me nuts. And that is in a nutshell the healthcare system as the high priests of health and the lowly consumer as the ignorant peasants that need to be told to come in outta the rain because they're just not smart enough to do that. To be fair, some people really are like that. Some people really don't want to think about their health. Really want to outsource responsibility for their health to the healthcare professionals. My suspicion is there's an awful lot who don't. Yeah. And that hasn't worked. Outsourcing or healthcare doesn't work to somebody else. Doesn't, does it? Does not work. We have to figure out engagement. We have to help patients advocate for themselves. We have to empower them to think about their own health. But as you know, this is why, you know, both of you're so good at this. It's hard. It's hard to motivate me to not eat that munchkin that happens to be sitting next to the coffee, even though I know I should skip it. And we're asking people to do cardiometabolic prevention all the time. And so actually I act, I think that the industry that's kind of getting it right is actually the consumer facing industry. It's in fact the, it's the tech companies, it's the platforms, you know, it's the, I don't wanna like use names but it's, you know, it's the whoops and the apples and the noms and the, and I'm just picking names at random. But it's those people that go to a patient say, Hey, can I make this as easy for you as possible? Can I give you trusted information and let you make decisions about yourself? Can I trust you to know your body? It's not capturing everyone. It's not getting to everyone. That's why the governmental infrastructure's important, because that's gonna bring it to everyone. But I think you're exactly right. Like that. We need to start, we almost say that's the last mile of healthcare when you get to the patient. Boy is that wrong. It is the first mile of healthcare. All our focus, our strategy needs to be, start with the patient. Figure out what's happening. Figure out their baseline. When they're off of their baseline, fix 'em. They're you said what I think as the unpaid consultant to the entire healthcare industry. My, my opinion is that the entire practice of medicine Yeah. I'm throwing everybody under the bus, but, you know, assuming sounds good has gotten completely inverted. Yep. The consumer of healthcare, the patient is the point. Yep. And anything that, that takes that away from anything that makes anything other than the patient, the point has failed right out the gate. And it seems like the entire system has been built. To serve the system, not to serve the consumer, the patient. I think the system's gonna be forced to come along. I really think it is because patients are demanding this. I mean, there's a reason that there are billions of dollars every day going into AI and digital health technologies, right? There is a reason 'cause everyone sees that the system's broken and people are starting to fix it in a way from the outside. But the other thing that you see is if, as let's say, you know, the system, the big house, if you start to accept that some of this does not need to be done by you, but it needs to be done by patients, oh my God, in the community where they live and the people who can support them there, guess what? All of a sudden your hospital is not running at 99% vet capacity. Your ER is not overflowing. Your operating room allows your operating team. To take their time, do their checks, get things done. And when you need a procedure, they can only be done in the big house. Hey, guess what? There's room for you. And so I do think that, you know, I'm always a generally positive, optimistic person, but I do think that we're at the point where like desperation is the mother of adoption. Yeah. Like it is happening again now. And AI is enabling people to just move. And it's forcing everyone else who's slower to come along. What I want people to do though is I want clinicians to not just say, oh God, we've gotta go along. I want them to say, Hey, I wanna be part of the solution. Where can my clinical acumen. Come in and make a difference. I wanna work with this, I wanna work on this, I wanna iterate this. And that's really where I want our clinicians to get, is I don't want anyone to say, this is happening to me. If you feel like it's happening to you, I want you to get up and go do something, find a thing you love, find what you're good at, whatever it is. But there is a way for you get involved. And that's what I really, that's the AI enabled clinician to me. I, and it doesn't have to be, the rest of my career is telemedicine. I'm not asking anyone to do that. I'm just saying in your version of the universe, where can you be the expert That helps the technology, that helps the system, that helps the government, that helps the infrastructure. You, there's gotta be somewhere. And so it sounds like you are at the point where you know the system really. Can't be evolved anymore. We're at sort of that disruption revolution point, right? It's gotta be sorta, you know, maybe not completely bro blown up, but, you know, mostly blown up in order for us to really move forward. Yeah. But the problem is this, you know, just to be very transparent about it, there are some cis health systems that are able to test and pilot new things at the same time that they do things the old way to make sure that patients aren't lost. How do you take like a hospital that's running in the red in the community, like serving the area, barely making margins, relying on philanthropy and then say, Hey, you guys are doing it wrong. That's, I feel for those people that's really hard. And who there is gonna do it. So yeah, hospital leadership can think about it, but it's a fee for service system. And if they don't have a lot of value based care. They're a little stuck. And so I, I wish I could say, Ugh, it's the system and people should do it better. But when you get down into it and you put yourself in any one of their shoes, it's gotta be piecemeal. And therefore, Jack, back to the tactics. Yeah. One tactic after another with the eventual goal that we know we're gonna get to. But if you figure out a system where hypertension is no longer seen in the office ever, except in rare cases, but largely taken care of at home, now you've just relieved primary cares and cardiologists of a whole bunch of visits that can open up for other people, tactic one, and then, you know, you just keep adding those things on. Let me follow up on that. One of the things that, this is a pattern that is repeated throughout human culture and that is, systems tend to protect, systems tend to evolve to protect themselves And cease to operate primarily for the reason they got started. We see it in education, we see it in healthcare, we see it the all over the place. Yeah. What specific steps I'm gonna back up one, one step here before I, I go on. Sure. In any industry. You will have a minority of people who are passionate about the purpose, about the reason why they got into that business. Then you'll have, in the fat part of the curve, you'll have the people who just want to go along to get along. And then at the other end of the curve, you've got the people who just really don't care about anything. You gotta the folks are on the front end, on the bleeding edge. You don't have to convince them, you just gotta mostly get outta their way. It's the fat part of the curve that we need to convince them that it's worth the effort to fight this system that has grown up around them like weeds in a lawn where the weeds are starting to take over. Yeah. And, and I think about a personal example. There's a device that can help me. With a medical issue that I'm dealing with. It's not approved for my specific issue by all the people who get to approve it. But the negative side effects are for all intents and purposes, non-existent, in other words. The worst that's gonna happen to me by trying it is nothing. And the best is it really might help. But I have to go through this, the, this, I have to jump over hoops, jump over, hurdles go through hoops in order to be approved to try this thing for myself. Yeah. Yeah. Why is it like that? I know how it was, I know why it was originally set up that way but it has become a place where the approval body. Their reason for existence is to keep existing. It's not really protection anymore. And that leads me to my question. Yeah. What are the folks on the front lines, what can they do to, I don't know, bypass this self-protective system to ignore it, to get around it? The system's going to fight for its own survival. Yeah. That's what they do. Yeah. I'm gonna go back to the beginning for a second where you talked about, you know, everybody has kind of early adopters, like majority, right? And then the kind of the tail end, the lagging group this came up I I do executive briefings for like large pharma device companies. And we kind of chat about what are we doing with ai? Where is it useful? How are we gonna do things? And one of the thing that things that comes up again and again is, look, you have to take the things that have the most potential benefit to meet some unmet need. Whatever your unmet need is. Yeah. And the lowest risk. Yep. And you gotta start picking those off systematically. Not with a small pilot, but with a series of pilots in different places that grows. It coalesces that becomes part of the fabric of care. It goes back to the beginning of our conversation that gets rid of the frontline guy having to do anything because you are putting into the fabric of care. Because we frontline guys get on the phone and we call, you know, and you're saying the system, I'm gonna just pick on payers for a moment. We call the payers for a prior auth for something that in fact is approved. Nevermind your thing. That's not approved. I can't even get the thing that has an actual f FDA application has, that hasn't done half the time. Yeah. And so I am on the phone. Now, right? Companies who work with the prior auth information, pull it out of your EHR, send the letter for you. Do that. You know, unfortunately, the payers also have that. So that's kind of like then checking what might they give me and how do we give it back? And you're right, robots fighting robots, right? It's robots finding robots. And I also struggle, I don't think I have an answer for you there, but what I will say is, again and again, if I'm doing a keynote or doing these briefings and I'm with these people who run these like large companies, right? Large health systems, all of them come down to the same thing, which is in order to get the majority to feel something, it has to be nearly invisible. And that's what we're aiming for. The early adopters are gonna be the ones who test, pilot, iterate, help you make sure it's safe, help you make sure that the thing out of a box does what it says it's gonna do. A valve out of a box. So like Phil puts a valve in, I see the patient, it should work the way it should work. Like it was intended to work a certain way. You put AI out of a box and I give it to you, Jack, versus giving it, you know, to the doc in the operating room. It's gonna work totally differently in a different place. So there needs to be iteration. You need your early adopters to iterate in this day and age of digital health, AI data. But for everybody else, you just need them to understand that these things exist. You need them to demand that somebody that they respect evaluated it, that there's a safety signal, that there's a privacy signal, and then it gets you the outcomes you want. Not that it makes you the money you want, not that it just makes you efficient, it gets you the outcome. And I think that's where Jack, a lot of our frustration comes from. Sure. Because the system to protect itself. Looks at fiscal bottom line, the system, any system to protect itself looks at efficiency. I mean, this is true for me in the morning, getting out the door with the kids, I'm looking at efficiency. Yep. If you're a little upset with me when I'm dragging you in the car, right? That's actually, my kids are much older now, conceptually I gotta get to round, you gotta get in the car. And I'm not looking at outcome like, Hey, how did she feel when I was, you know, getting her to get into her car seat and buckling her, or whatever it might be. And so it's true across the board, but people really, I think in healthcare, you can't just have a fiscal outcome and efficiency outcome and call that success. You have to push for like patient outcome. As part of that equation. That's kind of the point. But it is, but it's not what, it's not what frustrates us. What frustrates us are the systems where we said, Hey, AI is gonna make us more efficient. Let's go. And we didn't look at outcomes. Yeah. And now we're all struggling and frustrated with it. So I push people, the frontline clinicians, what can they do? They can say, Hey, if you're gonna implement something, I don't want it just for fiscal bottom line or efficiency. I wanna see that there's an outcome coming out of that. Yeah. Phil I guess you, you probably wrestle with this next question that I have but you haven't done it online, or at least, I mean, in this I hadn't done it live on the show. My question is there's gotta be folks like me who are users of, large language models of things like c Claude and Grok and so forth and so on, who are also acutely aware that, at least the commercial ones lie. How is that affecting, how is that fact affecting the adoption of AI by healthcare providers? Yeah. We have to make sure that people understand that the way these large language models work is they do their job. They tell you that dog feels like puppy. Like they're similar words. These go words go together. If they see dog and puppy and then they see puppy and kitten, now they're like 50 50, let's go with like mammal babies, puppy kitten, because you didn't give them enough information to tell them you're looking for dog, puppy. It's not lying to you. Can't lie. It doesn't think. Yeah, it computes. Yeah. It associates words with one another. And whatever the words are out there, the more frequently you see it how you asked your question, all those things affect what words are gonna come out. Interesting thing there, there's a reporter at the Wall Street Journal. She interviewed me two weeks ago and she had put her medical records into a quad and a perplexity and something else, and she asked me to rate it and I was like, oh, this is so deep. Like I can't rate it. I have a hundred things to tell you about each of these responses, but the first thing that I noticed is if I put that same information into my large language model that is constantly doing cardiology, digital health, et cetera, like chart whatever, like it knows that stuff, I will get a different answer than she will because her last paper was on homeopathy. The one before that was like on reiki or something, and I'm making it up, but it was like, it was not what I have and therefore, and was trained on different information. It led her down the path that she was most likely to follow according to it. Knowing that in the past she's asked about homeopathy, so I'm gonna lead with that instead of saying, Hey, you need a new aortic valve. I mean, and that's not the case. I'm exaggerating, but, so there's so much about these machines, what they remember, what they hold onto, how they relate words, how you ask you to question that. Anybody who thinks that these things are like truthful, that's not the right phrase. What is the likelihood that it gives you an answer that's close enough to what you need to know to make the next step pretty good, in some cases, not great in others. People are starting to publish and what that is are there now large language models that only have health information? Sure. Are there ones that only have trusted health information that doctors have looked at? Yes. Are they gonna give you maybe answers that aren't as satisfying because I've limited what it can tell you. Yeah. There may be times where it says, I'm sorry Jack, I'm not prepared to answer that. And then you get irritated. So as a company, I don't wanna do that. I don't wanna over restrict 'cause I want you to buy my service. Yeah. But as a doctor, I wanna restrict. You know, what I do now is I say, only give me information based on these papers that I found online in my search. They look like they're from newspapers, whatever that I rely on. Or journals that I rely on. Don't look at anything else. And then I ask you to look through them. And so these are the kind of things I think our patients, meaning all of us, that we need to learn that this thing doesn't lie and it doesn't tell the truth. It finds word associations and oftentimes it's so good at compute power and some word associations are so common. That it knows that dog and puppy should probably just go together. Yeah. You know? So I mean, that brings up an interesting challenge, right? Because I'm sure you heard in med school, just like I heard in med school, right? That, you know, half of what we learned was going to be proven wrong by the time we, you know, finish our career and we have these LLMs that are trained on the prevailing wisdom basically. There's this volume of information that says this is the correct thing, invariably something's gonna come along and oh yeah. How did we ever think that was the right thing? Yeah. You know, and it's it's interesting to see how. You know, or when, you know, you can get past that tipping point. Because if the LLM is just trained on this is what, you know, the majority of the information tells us, even though it may ultimately be wrong. How does that LLM start to recognize, and again, how do we as clinicians start to recognize when Oh yeah, what I thought was true, what I was told was true. What everyone else around me said is true. Turns out it's not true. And this is exactly why I tell everybody, we're not ready for clinical decision support. People use that phrase, and I just keep saying, no, we are navigating to knowledge, we are using it to get to knowledge and then we are gonna use our own brains. Our friends' brains or our doctor's, whatever it is, to look at that and say, Hey, does this seem right? Does this click? Does this seem right? Do I need to double check? Is there a link here where this came from? Is that a real paper? Like those are real questions you need to ask. But we are navigating to knowledge. We are using our own brains to actually make the decision because AI cannot make the decision for you. It also doesn't know, it doesn't know context what's going on right now. It doesn't know edge cases. In fact, it hates edge cases, large language models. Any AI loves to go to the center. It loves to ditch the edge cases. Most of us are edge cases. None of us are, I'm sorry, but none of us, at least the three of us, I can speak for three of us are not normal, right? There's some edge thing going on in most people. And so I think that context, I think nuance, I think edge cases, those are kind of things that AI's not gonna be able to do. But mostly AI can just flat out be. Wrong, and it can sound like it's right, and it can sound like it's nice and it can sound like it's smart, it can sound like it cares. That's the part that scares me the most when it sounds like it cares. So you kind of think that company owes you something when it tells you health information that it's supposed to have a Hippocratic oath where it's gonna take care of you. It is not, it is just giving you words. So no decision making based on AI information gathering. So that's actually a pretty good rule, no decision making. Wow that's fairly succinct and certainly cuts through the junk, huh. Dr. Bott, who are, is your real audience market, I'm not sure how to, as the in your role at the college of Cardiologists, what is who is saying we need to go see what she says? Yeah. I hope it's my clinician and administrator, colleagues. I hope it's the people who are delivering care, who say, Hey, I need to get to know what this is. I don't have to have another fellowship in AI to find out, but she and her team and the a CC are gonna provide me with the things I need to know. They're gonna keep me updated despite how fast things are moving, so I understand where the field writ large is going and where I might be interested in paying more attention, getting vol involved, helping at my institution when things happen, evaluating something. Just being. Part of that solution. So I think that's it. Now what I will tell you is who comes to me, startup companies with a new idea? People who see a gap that they can fix with a technology, large pharma and device companies who have a new innovation arm and say, Hey, we wanna be relevant in this area. We think we can help. We have the dollars, we have the teams help us organize what to do. So I love my job because I get to do all of the above, but if you say what should I be here for? I should be here to create the AI enabled clinician like that. That's the day job. In this particular application, the AI enabled clinician it would seem like the clinician and the administrator the two categories of people that are hopefully listening to you. It seems like they might be at odds in this particular situation where the clinician is looking for better outcomes for patients and the administrator is looking for efficiency. Yeah. You know, I think both are looking for both. When I, if we go back to the beginning for med peds training, medicine, pediatrics, we used to approach each other and be like, Hey, are you a, you know, big M little P or Big P, little m which was the side you leaned on? I think that's actually true nowadays for clinicians and administrators. I think the gap used to be wider. It is narrowing and sometimes it's actually overlapping. We see diads and triads, including nursing as well. We see this model of. Hey, I might be the big E for efficiency, right? And you might be the big O for outcomes, but I've got an O in my title too, and you've got e like we both care about the same things. But yeah, our day job probably demands that each of us keeps attention on one. And so I think we're seeing a lot more progress now than we ever did, like 20 years ago, because that's really they're doing the big on little peeing and I think that's really important. Okay, so I'm just, oh, go ahead Jack. It's your show. Okay. I'm gonna put you on the spot and feel free to wiggle out of this because it's entirely inappropriate what I'm asking, but I'm gonna ask it anyway. What is it, is there one particular technology that you are really excited about? That, that hadn't quite got there yet, but. Is on the way. There are two the, oh God, there are three. Okay. But let's just go with the one that's kind of there now is the AI on the cat skin of your coronary arteries. We spend a lot of time saying, Hey, you know, make sure you're checking and we should check your sugar, check your cholesterol, et cetera. But really the AI of the coronary now just tells you, Hey look, you got plaque or you don't have plaque. And we have to bring the cost down in that. There's some work that we need to do to be able to get it to more people. But I think even that long-term, knowing if you have plaque in your arteries or not, is going to completely change how early we treat you, how aggressively we treat you. And cancer's been doing it forever. They screen for disease, right? And so that's what I wanna do. So that technology is not like up and coming. I think it's here and now it's the logistics of like. How do we actually get it done, get it paid for, get it, you know, spread out to all the people who need it. So that's one. The ones that are coming in the future, kind of, I lump into a group, which is this idea that when you're walking around and you've got whatever on you like we can tell your heart rate variability, respiratory rate, other things about you maybe even glucose, right? The idea that we can start to use that kind of data every day throughout your life, even if there's nothing active happening, and understand patterns about you that might be predictive for what's happening next. And I don't mean. I'm gonna predict it in two weeks. I mean, as I go through pregnancy and then I'm postpartum and then other things that's gonna tell me about what's gonna happen later. You know, I think there's so many different physiological signals that happen. Stress. How do we measure stress? Actually, maybe we can measure stress, right? So I think that realm of technology, things on your body, capturing you every day as you live, kind of, you know, do your regular work, giving you your own baseline. My baseline, yours for all these numbers are probably different, right? Sure. Giving you your own baseline so when you're off your baseline and you don't feel good, you see that gel and you're like, oh yeah, I'm off my baseline for something. I think that is really promising, but it is a lot of data. Oh yeah. And so it's at the beginning stages of making sure that we figure out how do we capture that data? How do we capture real world data? We're not gonna be doing randomized control trials. Of that kind of data over 20 years, we're not. So now this is real world data. But I think that area to me is really interesting. And two specific examples is people are starting to create what we call digital twins, right? Is there a synthetic version of, not you specifically, but someone like you who needs a valve or something like that, right? So that kind of building. And then the second is could you not need to do some of the testing that's so complicated or hard or uncomfortable anymore? Because maybe some of this predictive modeling will tell me the likelihood that you need that test or tell me you're in range enough. You can skip the test this time, but now you're out of range. You need the test. So again I think the data on you, and I'll end with saying I'm pointing at the wrists, but there's this new algorithm that goes in your earbuds that tells you about your heart. Just because you're walking around with AirPods or earbuds or whatever in, so it's, there's in your t-shirt, right? So it's everywhere. And so I'm really intrigued by that. I think that's gonna change things. It's at the beginning. Yeah. I think that's a awesome place for us to wrap. That was actually exactly the question I was gonna ask you. So that's why Jack's such a good co-host and we get along so well. After almost five years of doing this now where can people follow you, learn more about what you're doing? Yeah, so they can come visit me@dramba.com or they can find me at socials. I'm generally Dr. Amba, so kind of slash Dr. Amba on LinkedIn or at Dr. Amba on Twitter. And they can also, you know, definitely just reach out and come to any one of the many, either virtual or in-person talks I'm giving. I love seeing people who, yeah, they may be clinicians, but you may not be. I gave a talk at the Boston Marathon talking to kind of marathoners about cardiovascular health and one of the young women, and I now chat regularly. Ever since then, she was a marathoner and she did amazing by the way. And so I'm always happy when people also just come up in person.'cause as much as I love technology, et cetera I kind of a in real life girl as well. Awesome. Very good. This is this type of conversation that we're probably gonna wanna have you back on in six months or a year and see how the whole playing field has changed. And you know, what you know, what is the next new exciting thing that you're excited about. But thanks for coming on and loved having you. Thanks so much for having me. It was a great conversation. Our guest has been Dr. Ami Bhatt. That's a MI, her last name is bot, B-H-A-T-T. You can find her website there, Dr. Ami Bhatt And on LinkedIn and X at those names. We'll make sure that information's on the show notes. Thanks for being with us for Dr. Philip Ovadia. This has been stay off my operating table. We'll talk to you guys next time.