Alter Everything Pod: Welcome to Alter Everything, a podcast about AI analytics and the future of work. I'm Joshua Burkow. know, AI shows up in every industry, many times as a sort of scale play, whether it's efficiencies, economies of scale, smoother systems, but healthcare. Healthcare is different. The blaring obvious example is that you

Joshua Burkhow: Dr. But I'm really happy to have you here on the podcast. ⁓ so many questions for you that I think I'll start with that if you could ⁓ briefly describe how you think about role today. It's sort of, I've seen clinician, I've seen operator, I've seen innovator. How do all these sort of perspectives collide?

Alter Everything Pod: can't tolerate failure now can you? A patient isn't some bot that you can throw away and start over if you mess up. But at the same time, waiting for perfection often means either a missed opportunity to benefit the patient worse, no care at all. Today's conversation is about that tension where AI helps.

Ami Bhatt, MD: Yeah, so I think all of those things are true. And whenever people ask, they say, wow, well, you really kind of had a course that got you here. And the truth is when you're looking forward, it is anything but a direct line, right? It is a very windy way that all of us get to where we are today. ⁓ But I will always be a clinician at heart. was and raised in the kind of ⁓ slash pre-med slash med school ⁓

Alter Everything Pod: where AI maybe shouldn't and how leaders decide does good enough really mean. I'm joined by Dr. Amy Bott, the Chief Innovation Officer at the American College of Cardiology. Someone who sits at one of the most consequential intersections in healthcare today. Clinical practice, innovation leadership, and federal digital health governance.

Ami Bhatt, MD: And so that's always gonna be who I am. But at the same point, I think all of us have been patients in our life. we've been caregivers for who take care ⁓ of ⁓ And think experiencing that healthcare system from the outside, addition to the inside, has kind of led me to where I am now. And so if I have to describe what I do currently, it's really I spend most of my time either talking with companies, big and small,

Joshua Burkhow: That's right.

Ami Bhatt, MD: who are bringing promising technologies forward, especially in terms of how we deliver care, but are struggling with scale, which is essentially everybody. It kind of doesn't matter how big or small you are, scaling into our healthcare system is challenging. And so how do we think about workflows? How do we get that done? At the same time, our clinicians and health systems, nobody has taught anybody about AI.

Alter Everything Pod: If you're interested in the healthcare space at all with AI, then which I think you should all be, this episode is for you. Let's get into it.

Ami Bhatt, MD: And so we are learning and teaching our team, you know, as we speak, is how do you actually take what's happening and make it relevant to your practice? And why should you change your practice sometimes, right? Like what is the advantage? And so I spent a lot of time creating what we call the AI enabled clinician.

Joshua Burkhow: Yeah.

Ami Bhatt, MD: And then the third part is, you know, not really a part of any job anyone's given me and a little bit of a job part of like all my jobs, which is I spend a good amount of time with the government formally with the FDA as the Digital Health Advisory Committee's chair, informally with the Center for Medicare and Medicaid Innovation, with the GAO and others, just starting to think together about are there things at the national level? that we should think about with AI, ranging from regulation to what I'm really focused on is infrastructure that can help us do population health better. Because I was kind of trained in a, don't want to call it ivory tower, but tertiary center, complicated care, and we wouldn't have all that complicated care if we could just do prevention right.

Joshua Burkhow: Yeah.

Ami Bhatt, MD: And so I guess that's how I end up where I am today. And what I do is my North Star and I tell my team, anybody who considers themselves part of my team any part of my life all knows like our North Star is high level cardiovascular care in the communities where people live. Like that's my mission. We got to get there.

Joshua Burkhow: I was gonna bring this up later, but it seems pointed Because your comments with you know working with the government is I think you were in just in Davos just recently right and so it's on

Ami Bhatt, MD: I was, I was. I wasn't like in the in-in room. I just want to be clear on that. Like there were world leaders. I was like outside where I happened to get a couple meeting with different world leaders only because they're all thinking about AI and infrastructure. But what an environment to be in and what a shared experience for like the entire human race to figure out how can we use AI to make things better? How can we use the compute power to actually reach more people? And the key thing that everybody talked about was access.

Joshua Burkhow: Yeah? Yeah, yeah, Sure. I bet.

Ami Bhatt, MD: Access to care is limited whether you're in a central African country or you are in rural America. Same problem. And so how do we solve for it?

Joshua Burkhow: interesting. Yeah. Yeah. Yeah. Love it. That's actually sort of where I wanted to start off first. think there's a lot of people that are deep into AI, into the AI world, they're building stuff or doing stuff. there's automation, there's generating new products and but it feels like the healthcare space is maybe different. And I often talk to folks about sort of impact AI is making, there's a sort of expected improvement in most industries. But when you hear about it in healthcare, for me, resonates different. Like it's the feel good stories, for example, of detecting cancer sooner and the impact in areas that you're obviously passionate cardiology. Can we sort of there as a a framework or as a guiding here is ⁓ state of healthcare and AI as you see it. Like what are the some of the challenges that you love and any stories that like to hear about that as well.

Ami Bhatt, MD: Yeah, so I would say there's three areas. One is we're actually kind of like any other industry, the structural stuff, right? So the administrative back office work that can be faster. The supply chain work in a hospital. So you have the right number of students in the right room at the right time, So supply chain in the hospital. lot of the just like overall care management platforms and you know, so I think that administrative part is definitely similar to almost any other industry. The part that's most interesting to us is this middle area that I like to call navigating to knowledge, is There is so much we know about, let's take you for example. So you have probably an electronic health record somewhere that has information about you. You have your own drivers in social determinants of health, where you live, how you do things, what you prefer. ⁓ have your own patient reported outcome measures. You might have a wearable, other things about your own symptoms, things you measure. And then there's all the guidelines we write about how to take care of patients and all the additional research that maybe hasn't made to a guideline yet, may be really relevant to you. And so if I'm seeing you in the office in 20 minutes, The human brain is just not able to take all of that and give you the best care possible. But it turns out we can give you the best care possible if we use the compute power and then we use our own clinical acumen, on top of that. And so I like to call that navigating to knowledge, which is getting all the stuff we need, organizing it so that you and I can have a meaningful conversation and make things better for you. People call that clinical decision support. And that's a little bit scary to me because we should not have the computer tell you what to do. It could offer you options of what seems appropriate based on what you see in front of you and based on the data, but then you should still look at that and decide. And so I prefer calling it navigating to a lot of it's not clinical decision support. Fun story, however.

Joshua Burkhow: That's right. Please.

Ami Bhatt, MD: ⁓ If you start using, and many clinicians do, I just gave a lecture at Piedmont Heart Institute down in Georgia, and had the audience ask, you know, they raise their hand, like how many of you use it for your own lives, how many of use it for healthcare, and a lot of people are using ⁓ variations this for asking healthcare related questions. Please don't put your personal information into a non-approved generative AI model. I'm just gonna say that for everybody, okay.

Joshua Burkhow: Mm-hmm. Yeah.

Ami Bhatt, MD: But it turns out that one dog had been asking some questions about something, left their computer, something else, came back, and started asking questions again. And this time, the questions were about vitamins. It was like for their mom who was asking something. The question before was about kidney disease.

Joshua Burkhow: Got it.

Ami Bhatt, MD: these questions because the doc didn't really recognize that they needed to start a new chat or tell it to certain new train of thinking. Start talking about vitamins, but then start to talk about, well, in the setting of your severe kidney disease, however. I would be careful if these things are there." And the doc was like, wait, what is, and they were so confused. And it took them a moment to realize that they had asked this question, that there had been a link, et cetera. And so those little things aren't little when it's somebody's care. Like what if that was a subtle thing in a previous question and now you didn't really notice that it affected your next question about another patient. And so. This is why the AI education, right? We don't want to hear more of these. We want to hear them now so we can create the right models, but we have to teach people how to use the newest tool. You can't assume you put a tool in someone's hand and they use it, right? You wouldn't give me a saw. You probably won't give me a saw at all, actually, but you wouldn't give me the jigsaw and be like, figure this out, right? Like, I think I could try, but things could go wrong. Fingers could fly off. Similarly, you shouldn't just give someone generative AI and be like, go ahead, figure it out when the

Joshua Burkhow: you Yeah.

Ami Bhatt, MD: responsibility that lies with them is so significant. And so that's that second area, navigating to knowledge, not clinical decision support. And then the third area of AI is actually pretty interesting to me, which is this drug development area, learning new things about the body, because there are so many things about both how the body works right now that we still don't know. And there's so many opportunities to actually heal people in a better way. And so that's a really fun, more research oriented, not anywhere near us using it just yet in practice where it's going to touch us, but it will come if we can kind of use it correctly. And so that's a really cool different area that we're in. I'm in Boston. down by the docks, there's a bunch of warehouses. And there's a set of new companies that are doing engineering work. They come out of MIT, a lot of these guys. And they created this little white square that you put on your arm. And it senses your skin and it doses medication through tiny micro needles for you. But like, let's say your skin is dry one day or soft, or let's say I put it on me but I need to put on you, you're and my skin are probably different. They've got the engineering in it. to like sense exactly who you are, get it right. And eventually, this is not so much for cardiology, like imagine sitting at home just having chemotherapy because you put this on and a doc somewhere else remotely turned it on and it senses, it goes at the right pace and you can just watch TV on your couch. and it can do its thing. And so, you know, those are the kind of like futuristic stories that you hear and you're really excited for. In the meantime, like my day job is, hey, can we figure out when to get patients into the clinic so they don't come to the emergency room? Because like that's our problem right now.

Joshua Burkhow: That's right. Right. Yeah. It's an interesting conundrum, which I think is, think if you ask anybody in any industry, it's that sort of, okay, here's what I'm working on now versus I'm starting to see glimmers of hope for just amazing innovation, amazing ⁓ I think it's the, the that still sticks with me is this, the fact that we're just at the sort of beginning of ⁓ the art of the possible going into AI. But even then, just now, we've had amazing breakthroughs in science and medicine and all these different areas with just this short amount of time. It's hard to try to wrap my head around what that could be. Is there any specifically in cardiology that have sort of got you really wound up and excited and hopeful that breakthroughs will happen just in that space?

Ami Bhatt, MD: Yeah, favorite thing right now ⁓ actually the use of AI on an EKG. ⁓ an EKG is probably one of the cheapest tests we have in cardiology, and we can have expensive testing in cardiology. ⁓ can be done anywhere, right? And AI can see based on EKG signals that...

Joshua Burkhow: Mmm. Sure. Wow.

Ami Bhatt, MD: we actually don't really see. And so in TEDx talk I gave recently, we were talking about, you know, ⁓ what of things can an EKG tell you? So it turns out an EKG can actually tell you about liver disease. All the people who had had drinks the night before at that TEDx talk were like, not excited to hear this from me. ⁓ It tell you about, you know, heart failure or like an inability of the heart muscle to pump. Increasingly, it can even tell you things like they studied

Joshua Burkhow: That's right.

Ami Bhatt, MD: students who were going through their dissertation, like, you know, their final PhD thing. And the EKG actually suggested that their age was like higher than their stated age during the hours leading up to and during their thesis defense and then came back down the next day. So like your EKG says, hey, you, Josh, look 32, but when you're stressed, it says, hey, you look 40.

Joshua Burkhow: Can't imagine the stress.

Ami Bhatt, MD: what is going on. And so, right, there's just so many ways EKG is gonna be useful to us. So I'm excited, because that's like a expensive way. And you gotta be able to use these technologies for everybody. I don't wanna just make it for a select few, whether that's like based on ⁓ monetary or education or access. ⁓ so that's why I'm really excited about this kind of EA off of EKG and what we learn and we can do with it.

Joshua Burkhow: That's right. So exciting. I love the idea that we deal with this in AI in general. It's just, okay, we can handful of patients and ⁓ a solution that works for them. ⁓ in the back of our minds, we're thinking like, okay, how does this affect a billion people, a hundred million people, ⁓ is fascinating. That's ⁓ mind blowing. Yeah, right.

Ami Bhatt, MD: That's right. And we can reach a billion people now, you know, which was really hard before.

Joshua Burkhow: just deal with the complexities of those things. So one of the things, sort of shifting gears here to the next theme is one thing AI is really great ⁓ at is us ⁓ to have really uncomfortable discussions, right? And the uncomfortable questions. And ⁓ I think like the sort of Buddhist and me's... thinks way through is the right way to go. We got to address these questions. We got to think through these. Come with our best answers. And the quintessential question around AI is this idea around replacing humans versus maybe partnering or adding to enabling the folks that have. We deal with this question in my work all the time is ⁓ like, Do we just say, okay, here's the last list of things you're doing, let's get AI to replace it and that, don't think that's a great idea. ⁓ think, ⁓ the human is really, really important into this whole equation. ⁓ so, okay, in medicine, they deal with humans the time, right? Like that's sort of the center focus. So is it the same perspective? what's your thoughts in this world?

Ami Bhatt, MD: Thank I think we can. It is the same perspective and I'm looking at the guitar hanging over you, right? And this conversation comes up between me and my daughters and I have two teenage daughters. And we talk a lot about how we think about people who make music and then when people are using AI to make music and we will have kind of vivid conversations about...

Joshua Burkhow: Hmm. Mm.

Ami Bhatt, MD: you know, is that really making music? Is it cheating? Is it fair? What's the best? And my younger one, the 12 year old, she said, well, it's kind of pretty cool actually when you can make your own music, but then sometimes you can then use the AI to understand and make different versions. She goes, I kind of like when it works together. And so I only bring that up because that's kind of how I think about it in medicine. which is the phrase that ⁓ I had coined when I first started this work was collaborative intelligence. a lot of my friends use augmented or assisted intelligence. And I said, that's fine, but that's implying that there's just a tool that's helping you. It's okay. The problem is the really good AI is gonna be the one where we collaborated in its design because we know this field, right? You know your field. When you collaborate in the design of the AI and where it's going to touch and how it's going to do it, and then you iterate on that output. Now it's so much better. And now there's something actually authentic about it again, rather than feel like it's a replacement technology. And so that collaborative intelligence is really important to me. If you think about in terms of healthcare and your patients, we don't want people giving us a technology that we then kind of use in practice and say, this is useless. It's a waste of people's time. It doesn't scale. It's a waste of people's money, right? The venture capitalists who put money into it. So instead, can we help direct people

Joshua Burkhow: That's right. That's right.

Ami Bhatt, MD: teams, engineers money to the right problem and then a reasonable way to solve it and work on that back and forth. So absolutely, I think working together is key and not just assisting or augmenting us. Sometimes that's fine, navigate to knowledge, great. But how exactly do you do that? Then what happens, right? And I think that process, I really want us to be involved in it.

Joshua Burkhow: Yeah, so to sort of put it in how I'm thinking about it, you as a cardiologist know the right questions to ask, right? So like you've had the training and the understanding and the knowledge over time and experience and the failures ⁓ say, okay, this works, this doesn't work, this is a good question to ask, this really doesn't make a difference, where you get to bring that sort of skill, you will, domain expertise to AI. ⁓

Ami Bhatt, MD: Yeah.

Joshua Burkhow: and sort of say, focus in on this thing versus hate. Tell me how the heart works. And you know what I mean? Like the sort of generic throw it on the wall and see what I get.

Ami Bhatt, MD: to it. No, that's exactly right. One of the things that we really work on is when I'm in executive briefings with companies or talking with people, we create playbooks for them. And we do this kind of work to say, what is it that you're solving for? give me one or two clear use cases. And then let's work with the companies, the technologies we're talking about to understand how can we solve for that, right? Or sometimes we give them the use case. Sometimes it's a tech company that comes and says, we want to build this. And you say, great, let me give you two clear use cases. Let's put a team together and let's really think about where you're to run up against something and where you're not. And I think that process ⁓ really important of taking that human expertise and do doing something with the technology or AI. Now I will say, you you start by saying, you know the questions to ask. There are times where I don't know the questions to ask. There are times where I'm seeing 20 patients in a day and I may forget a question to ask. You know, we say to err is human, but humans err quite a bit in healthcare. So where we can use AI as a backstop to say, hey, Dr. Butt, I know that you would think about these three things for this symptom, right? If you're having chest pain, I have a list of three.

Joshua Burkhow: soon.

Ami Bhatt, MD: There's also four and five. Less likely on the list, but just FYI. And that's really good for me because all of a sudden number five, I may like, wait, I remember there's a trigger, right? And so that backstop is actually kind of helpful because we're human and we're having different types of days and we slept well or we didn't and it's all part of it. So could we raise that level of safety? Could we, you know, by having a little bit of an AI backstop, that's something that I think people are okay with.

Joshua Burkhow: Sure. Yeah. So sort of getting back to this, this replace versus collaborative intelligence, which I might steal that. That's perfect. There is the reality like calling a spade a spade. There's, there's a reality that there are some things that AI is doing better today than we as humans do. And sort of this, so I of understand the augmentation. mindset is like, Hey, if I laid out my entire body of work that I do in a week, a month, a year, and say, I do all of these things, there's going to be a handful of these things that AI will do it better than me. And that's sort of reality check of all things where we're not, I think that this conversation is like, ⁓ do we go after replacing the human versus Get rid of the stuff that isn't helping us to do the bigger, better, more impactful things. Right.

Ami Bhatt, MD: That's right. That's right. Absolutely. Getting rid of the things that... So, simple example. ⁓ generative AI is used in voice to text translation, which means you can talk and then it, you know, comprehends that it can write things for you, et cetera. And so right now there are quite a few practices ⁓ in the U S using technology. There's several companies who do it so that I no longer face away from you and like type at the computer, right? While talking to you. Cause I used to look at you like this and then progressively over time, this started to happen and like now I'm back. Right. ⁓ and. ⁓

Joshua Burkhow: So good. Like I had to capture all this yeah

Ami Bhatt, MD: And it's great because from our conversation, the technologies now can create a note that's appropriate for you in the appropriate patient-oriented language. It can create a note for billing, so I don't have to spend hours correcting my note so that it's billed correctly for insurance, right? And it can create a short note with the key points that I as a cardiologist need to send back to the primary care. Here's the few key things I did. Not like here's 18 pages of the cardiac history.

Joshua Burkhow: That's right. Right, right,

Ami Bhatt, MD: That is helpful, but when that technology was first created, it just wrote a regular note. And it was when the company started saying, well, wait, I can do it for billing. I can do it for a patient. I can do it for a And that came from clinicians. That came from doctors who actually said, hey, there are different ways that different people need to receive the data, and we can do that. And so definitely, I think that that's really an important interaction.

Joshua Burkhow: Hmm. I mean you can't underestimate that mean I've had the both sides of that coin you go to the doctor and they're just like ⁓ and you said what's your problem? ⁓ you know just no ⁓ of awareness maybe And it doesn't feel good right it doesn't like I don't know if If I could have just texted you exactly what you're typing maybe I didn't have to come here But then there's others that are like you and I today were just laser focused and

Ami Bhatt, MD: Thanks.

Joshua Burkhow: It just, there's that feeling, the communication. mean, you probably know about nonverbal communication as much as I do, like the impact of that. It just enables so much. I think there's like, if AI can help with those sorts of things to enable a human, someone like yourself to be in the game when a patient comes in, I mean, to me, that's magical, right?

Ami Bhatt, MD: It's here. Thank You know, it's funny you say this, but I still remember there was one day, Joshua, was with a patient, this was before any of this AI stuff, and I was just fed up. And that day in clinic, it was like my second patient of day, decided.

Joshua Burkhow: Hmm.

Ami Bhatt, MD: I'm gonna write my notes later. I just can't do this anymore. I love these patients. They've been with me. I followed patients long term, right, because I took care of people when they were turned 16, 21 with heart disease. I followed them through graduation, marriages, kids. they were like family to me. And so I decided I'm not gonna write notes. Like, I'll just do it later. I'm gonna look my patients in the eye. That night was terrifying for me.

Joshua Burkhow: Mm-hmm.

Ami Bhatt, MD: I was sitting in bed with my laptop, like trying to figure out how to like, you know, do what I need to do. And I couldn't remember some of the details that I was certain I was going to remember. And I thought, ⁓ God. And so we don't talk about those days, right? Like Louise clinicians are hesitant to talk about the days where we're like, ⁓ that didn't go as planned.

Joshua Burkhow: That's right.

Ami Bhatt, MD: But that's also difficult. And so as a clinician, you want to be able to look your patient in the eye. And some days when you do and then you didn't capture something, you think, know, like now what? ⁓

Joshua Burkhow: Yeah, yeah, feels.

Ami Bhatt, MD: And so I actually ended up calling a couple of them and being like, hey, we talked about this thing. I didn't write my note because I was talking to you because I was mad at the computer. And they all laughed because they knew me. And they were like, yeah, no, I said this and this is the thing and this is the details. Great, thanks. And then I wrote it in. And so we fixed it. But there's always another side to how these things can be helpful.

Joshua Burkhow: Yeah. What are? That's right. What a world to be in where that sort of enabled you. It sort of lends into the next thing I wanted to throw at you is this sort of idea around good enough. And I'll sort of try to put a box around it and see if it helps. But I think one of the thoughts, questions, ideas that's coming out of healthcare that I really like the idea that some industries will get out in front of a topic, thought, a ⁓ thought leadership, a perspective and sort of share for the rest, right? And one of them is this ⁓ idea in many regions, people don't have access to care at all. I know you've talked about this in the past and they're ⁓ for the doctor isn't gonna be an option, right? It's just ⁓ the ⁓ mechanics of the logistics of it is just not there. And I think this is one of those hard, difficult conversations, questions that a lot of medical professionals have to think about when AI is like, is there a perfect or is there a good enough? Is there a good enough that, you know, especially when the outcome is the complete opposite, right? No care at all, which I can't imagine that that would sit well with you, right? Like, how do you see that perspective is like, do you think like, hey, As long as I get to this point is good enough or do you think that, we only implement AI if it's perfectly buttoned up, gives us, you know, exact diagnosis, all this sort of thing.

Ami Bhatt, MD: Yeah, so I think the most important thing here to recognize is way we do healthcare today is perfect. We act if we are meeting the gold standard. But if you look at most of the specialties, I know it's true for cardiology, we have guidelines. We know what meds you're supposed to be on, what procedures you're supposed to get, what imaging you're supposed to get for certain diseases. Guideline directed medical therapy is kind of less than 50 % for many diseases across the world and in our country. We're just, so we're not meeting some things. The other thing is, if you don't access any care, then when you eventually access care, you're gonna access care through an emergency department. And we know that for a majority of diseases, the minute you come in kind of in distress a disease, your risk of doing poorly is in general higher. That is a statement that one can easily make for just about anything. So if you take that as our baseline, anything that can help us get people to some version of care,

Joshua Burkhow: Yes, has to.

Ami Bhatt, MD: faster than they would have before. So let me give you an example. AI is really good for educating, and it's great for educating people who have some baseline knowledge and just need to figure out sick from not sick. Next basic step, I'm not talking about like acute care in an ICU at Mass General Hospital. I'm saying if I'm in the community, and I'm a community health worker, and I am looking at patients all day long, Can you assist me using this navigating to knowledge, getting me some information to help me make the decision that these five patients seem sicker than those five? I'm going to send them ahead first to the larger city or even appropriate to send them from here in town there because how are they going to get there? Who's going to do the job that they're doing? Who's taking care of their kids or their parents? Right? Like how is all that happening? ⁓ I think that triage area is a really important area for us. to say we're not gonna get 100 % perfect care into some areas. We're just not. And in fact, even if we thought we were and we sent the right doctors and the right nurses, some of the time they wouldn't get it right, but could we get you to the right next best step more often than we're currently getting you there? And I think that's really important. Now, that's at a population level. If you took me as a individual, as a cancer patient, And you said, well, this seems like it's good enough, but your life is on the line. It's a very different conversation. That requires randomized trials. That requires AI in practice. That requires knowing what's happening. That requires no drift of the AI, because that's really important. a whole different level of burden. And so the way ⁓ often think about it is, you talking about low risk, intermediate, high risk? Are you talking about populations and making things better? Are you talking about individual? ⁓ And have to break that down. And once you do, you see there is a large group of that can actually benefit from the use of of AI at the beginning stages of healthcare. And ideally, if we do it in population health and we catch people earlier. We have less and fewer of the people who have this dramatic presentation and this illness that requires that advanced care because we got it right. so creating that infrastructure upfront, you mentioned earlier we were talking about kind of working with the government and mentioned that regulation is one thing, but I'm really looking for infrastructure. What's the infrastructure to support people, to get people to care? to the right next step at the right time, especially in those lower and middle risk individuals. And so I think kind of where we're headed. there's so many different right? And we have case studies, and I give on this. And so I think ⁓ sometimes just telling story helps people understand, because you can say, hey, it's OK. We have to be 100%. We can be 50%. Nobody believes you. But if you can share your stories of when you found somebody earlier, when the pregnant woman who had a child that was diseased was identified earlier by a community health worker who used an ultrasound with AI and sent her to the big city and she got to care in time for her and her baby to survive. Those are the stories that make a difference.

Joshua Burkhow: Huge. Yeah, huge. Sounds like you're pretty big on the prevention aspect. Do you think AI will fix that in the sense that I'm sort of assuming the end is that prevention has always been such a challenging thing as a society, right? I mean, I know myself. If I put myself out there, it's like, oh. I'll get something and I'll push it off and I'll wait and I'll wait and I'll I'm like, ⁓ it's not that big a deal. Because culturally I was raised to just rub some dirt on it and suck it up. And you get to a point where, okay, you're like, I think this is bad because I can't walk. don't feel great. Like I can't do this or can't do that. Like you always wait till that point. ⁓ you think AI is gonna help close that gap to get us to... ⁓

Ami Bhatt, MD: Mm-hmm. Mm-hmm. Yeah, yeah.

Joshua Burkhow: what maybe you would align on true prevention.

Ami Bhatt, MD: ⁓ I don't think AI is going to get us there. I think what we right now call digital health is going to get us there. probably the easiest way to talk about digital health is remote monitoring our wearables, right? Because...

Joshua Burkhow: Okay.

Ami Bhatt, MD: You and I don't have the same blood pressure. We don't have the same baseline heart rate. Like we have our own, right? But if you look at guidelines, we put people into chunks. You're in this range or that range or that one. Actually, the truth is each of us are unique. And so I think the goal yes, AI, but AI on top of measuring your own numbers is if we have remote monitoring, wearables, ambient ways to say, hey, this is your own personal baseline. Here's how you generally sleep, here's how you generally do this, this is your general way, this is your... And now you are out of range. Now we give patients agency to say, hey, how are you feeling? What do you mean, ⁓ my blood pressure's up 10 points and I've got a headache today. Hey, every time my blood pressure is up 10 points or every time I have a headache, right, and I see the correlation, now all of sudden I think, and then I notice maybe I don't even need meds. I the hot dog and fries last night. It might have been the salt, right? And now you may not know it's the salt, but that's where AI comes in. If that becomes a trend, can I push some information to you in a way that makes sense to patients saying, hey, notice the blood pressure's up. Notice you log headache in your symptom index a couple times. a thing called ⁓ hypertension. Let me tell you a little bit about it what you could do in your diet. You don't have to completely change and go vegan for this, ⁓ maybe you pull back on your salt in these couple days. Here's a couple menu recipe ideas.

Joshua Burkhow: ding ding right you

Ami Bhatt, MD: All of that is possible with generative AI, but we've got to feed it into an infrastructure and the infrastructure has to be, can we really get our patients to have agency? And as doctors, we haven't been great at that because in the old days we knew everything and patients didn't because the information wasn't accessible to everyone. And it wasn't accessible in a way you could understand.

Joshua Burkhow: Right.

Ami Bhatt, MD: Generative AI, first Google and search engines, and now generative AI has changed that so people can get information more readily. And so I think we have to embrace that it's patient agency, it's not consumerism. ⁓ And so yes, I think AI is gonna change things, but I think first, remote monitoring, wearables, ⁓ ambient things like the AirPods you're wearing in your ear.

Joshua Burkhow: That's right. Mm-hmm.

Ami Bhatt, MD: There's a company that actually makes algorithms that can be fed into that so that just you wear it regularly during your day and it tells you if you're developing heart disease. It tells you if you're developing heart disease, right? And that, you wouldn't even have to do anything. That gets to the point that you're saying, ⁓ I should do, like I keep putting off some testing to it, I shouldn't, but doctors make the worst patients sometimes.

Joshua Burkhow: That's That's wild, right?

Ami Bhatt, MD: But there's a point at which if you just make it easier, and we're just monitoring, and we're just measuring, or if we get one test but it tells us four things instead of one, because we can do that now, I think making it easier for patients to just end up getting that care is gonna really make a difference, and that's where AI can help.

Joshua Burkhow: Love it. Just my brain's going 100 miles an hour. There's so many great things. I actually have two questions for you. First one is around, gonna try to give them both to you and see if we can tie them together. But the first one is, what are thoughts about the average person or someone who now has chat to BT Gemini ⁓ is saying, I've got this issue, uploads a photo, tell me what's wrong. so that they sort of don't have to go to the doctor. Like, where's your thoughts in that? But then also, I sort of want to tie back to this clinician thing, because I think like you sort of turned me on to how big of a opportunity this is. Is there, if you were talking to a doctor that's listening to this, a medical professional, what would your advice be to them in the AI front? Like, what would you give them to say, hey, we talked about the voice to text thing, which is. just seems so simple, but it's so, so outsized impact. Like what would that be? So Chagybiti sort of self-doctor.

Ami Bhatt, MD: Yeah, yeah. Whatever, I'm just gonna talk directly to your listeners, whatever you do for profession, I mean, it's a side of cardiology, right? Do you want me looking up chat GPT and telling you that I can do your job? Right, like, let me tell you right now, my plumber does not want me going anywhere near that, my electrician doesn't, the piano tuner doesn't, like, nobody wants me going there. Now, if I were to go there, How should I talk to you? Well, I could say, hey, piano seems really out of tune. I looked it up and like, these seem like some of the things. I know it's your expertise though, but I just thought I'd help out and tell you. From what I read, it says this could be a thing and I do notice that kind of thing happening. Now you've said, I respect you for your expertise. I also care and I went and looked it up. I'm not telling you what to do.

Joshua Burkhow: Mm-hmm. Mm-hmm.

Ami Bhatt, MD: But when I looked it up, here's a couple of the symptoms that they related together. Those are the symptoms I'm having, and that's why I think this is the thing. Right? Don't come to me and say, think I'm having, well, I mean, don't call me for a heart attack. Call 911. Like, don't come with an answer per se when you see your clinician. And that's true for your electrician or your panotiner, right? Come with, this is the thing, ⁓ And then ⁓ why I think that, here's why that clicked when I tried to look it up. And then ⁓ it to them and maybe don't show up in their office with it. Maybe use the portal and them if you have an, most people have an electronic health record now, or just call the office.

Joshua Burkhow: Very good. Yeah. Yeah. Yeah. Right. Yeah.

Ami Bhatt, MD: and talk to the admin or the nurse and say, hey, but like give them a heads up. I looked something up. This is my symptom. Cause if it's a serious enough symptom that you're looking it up, they probably really want to know, but they're also going to just look at them. They're going to be having a busy day, right? Their kids going to have had a tantrum in the grocery store the day before and they're embarrassed. like life happens to doctors too. so, ⁓ so that's my advice to your.

Joshua Burkhow: Yeah. Yeah. Right.

Ami Bhatt, MD: to everybody who is a patient, because we're all patients, Is at least do it that way. And whether it's medicine or like your auto body shop, like go in respectfully. Like you don't just because you Googled something about how to fix a dent doesn't mean you're better than Gary is like Gary's the best. Okay. Here's my guy. Separately to the clinicians who are out there. Find that pain point that like bothers you more than anything else.

Joshua Burkhow: Yeah. That's That's right. That's good. I'm going to use that. Mm-hmm.

Ami Bhatt, MD: like that, whether it's an unmet need for your patient population, like find the thing that you can't seem to solve for from a system standpoint in your practice. You gotta pick one. And then what we don't wanna do is we don't wanna pilot, pilot, pilot. And so once you figure that out, you have to get other people on board. with like, do you also see this as a pain board? Now there might be some things that are specific to just your own work, in which case find that thing and then start looking around for ways to solve for it, right? But otherwise if it's in your practice or in your system it's bigger than you, find it, get some other people on board and then don't. try and take a technology that's out there and be like, this is gonna be perfect and solve it. Sell it the right way, which is, we have an issue. This could be a technology that works. We are going to interview that company or that technology. We're gonna talk to them about how they deal with it and we're gonna partner with someone. Now some really big companies will be like, we don't partner, we know how to solve it, we're gonna solve your problem. Okay, if you know what your problem is, if they've solved it in enough hospitals and you wanna buy them, by all means do that. But if we're talking about the smaller size kind of attempts that we have to like really make something happen, find the companies where, first of all, like they really care, right? Like the founder cares about the same thing that you're aiming for. And the second is a team that'll iterate for you and iterate quickly. because there is no AI that you're gonna meet that's gonna be perfect out of the box. I'm so sorry to all my friends who are running AI companies. It is true. The AI never works out of the box. A valve needs to work out of the box. If I put a valve in you, like, it better work the way you said it did, right? But if I take an AI, there's a lot of things. There's a lot of stuff about data.

Joshua Burkhow: That's right.

Ami Bhatt, MD: There's how we put it into play. There's where it fits in the workflow. That's who's using it and how do they prompt it if it's a generative AI or when is that data released to you. There are so many factors that you have to work with it to get to work. It brings us back to collaborative intelligence. It's really important to work with it. So that's what my advice to clinicians would be is find the thing that you're so passionate about that like if that thing got fixed, you would be like, hallelujah, like this is it. And then find a couple more people who like it. Find the company say that they can potentially do it. Meet with them and make sure that they care and they have a good iterative time that they're going to work with you. And then go for it because then you'll probably not have a pilot. You'll probably have the beginnings of something that'll scale.

Joshua Burkhow: Right. funny. That's one that I would like the listeners to take away as not just a healthcare focus, but that in AI use in general, ⁓ have a lot of people out there that are trying to boil the ocean because like I do, I see the art of possibilities. figure one little area that it's good and I'm like, I could do this in 20 different areas. And I think that the message that you're providing is so poignant and so good that just take one thing. that if you got it off your plate, if you got it fixed, if you didn't have to think about it again, would make your life that much percent better, right? Maybe it's 10, maybe it's 50, whatever, but hyper-focus on that because the, and I sort of posted this on LinkedIn yesterday, just nothing will beat time on task. Take that thing, figure it out, make it happen, and then move forward after that because the stuff you're gonna learn in that process is...

Ami Bhatt, MD: That's right.

Joshua Burkhow: exceptional. Love it. All right, so now we're going to move to my favorite part of the podcast. I hope you're ready I sort of called this lightning round for lack of a better term, but I just like to I'm going throw a bunch of questions at you and they're really like a yes or no. If you have a

Ami Bhatt, MD: And that's exactly right.

Joshua Burkhow: If you have some that you know to elaborate it, great, but here's the rule. There's no maybes, there's no playing the fence, there's no, ⁓ depends on this or that. It has to be sort of right out of it. You ready? All right, all right. So the first thing, one thing clinicians shouldn't expect from AI. One that they shouldn't expect.

Ami Bhatt, MD: I love it. Nope, ready. for it to replace them.

Joshua Burkhow: That's good one. Yep. One thing they should absolutely expect.

Ami Bhatt, MD: it's gonna be part of their practice whether they want it or not.

Joshua Burkhow: Have you played this before? Where do humans stay essential no matter how good AI gets?

Ami Bhatt, MD: No, this is great. I love this. and the relationship with their patients.

Joshua Burkhow: one misconception that non-clinicians have about AI in healthcare?

Ami Bhatt, MD: that it's going to give you the right answer.

Joshua Burkhow: All right, that's a good one. All right, two more. In three years, where is the most significant impact gonna be made with healthcare? Three years. I tried not to do 10 years, because nobody can figure that out, but like one, two, three years, I think.

Ami Bhatt, MD: being able to tell patients who measure things at home when they're out of range.

Joshua Burkhow: It's amazing. Yeah. All right. Last one. This is a tricky one. What's the last thing you did personally with AI? ⁓

Ami Bhatt, MD: missed on chat GPT earlier, what was I looking

Joshua Burkhow: Yeah, don't divulge any patient information, but.

Ami Bhatt, MD: No, no, no, and I didn't use touch up for patients. I was just on it earlier. What was I looking? ⁓

Joshua Burkhow: We both have teen girls, so like half of my chaijibiti is like...

Ami Bhatt, MD: No, yeah, it was like something related to that. I'm trying to remember. ⁓ gosh, you got me. stumped me on the very last one. I looked something up on Chat GPT earlier today. I cannot remember for the life of me what it was. It was a restaurant. I was looking for a ⁓ new restaurant.

Joshua Burkhow: Yes, the very last one. Mmm.

Ami Bhatt, MD: to take people to where you can still talk. Because the last time I took some friends to a dinner and we wanted to all catch up, the restaurant was so loud. I think I sound a little old when I say that, but it was an awesome vibe and we caught up zero. And so I was looking up which restaurants are really cool vibe, but not super loud.

Joshua Burkhow: No, no, it's I get it. Alright. Sure. No, there's no Boston restaurants there. Like it's, it's impossible. I've been there. ⁓ wow.

Ami Bhatt, MD: It actually came up with one. came up with, can I say it? I mean, I'm not like obsessed with the restaurant thing, but it came up with this restaurant, Gigi. ⁓ But it turns out that for Gigi in this outfit, you need to make a reservation 24 hours before. So now I need to like wait and then hope the day, so, you know, now I'm gonna hold something on like seven rooms or Brazil, that's gonna happen.

Joshua Burkhow: Okay. Uh-huh. All right, good. ⁓ You can connect it to OpenTable and it'll do your reservation for you. That's the AI nerd in me. Dr. Bott, thank you so much. I appreciate this. Time has flown by. really appreciate you just spending your valuable time with us. I hope we covered a lot today.

Ami Bhatt, MD: That's right. I think we covered a ton. If there are other things that people have seen, like you mentioned LinkedIn, post about on LinkedIn, et cetera, you can find me, Dr. Ami Bhatt on LinkedIn. ⁓ But I also have a website coming up, dramibhatt.com, where we'll put this and a whole lot of other conversations if people are interested.

Joshua Burkhow: Yeah. Perfect. Yeah, I'll make sure we get all this in the show notes so people can reach out to you. But thank you again. I appreciate it.

Ami Bhatt, MD: Awesome. Perfect. ⁓ thank you for having me. I love the lightning round. That was great.

Joshua Burkhow: No, that's awesome.