← Embracing Digital Transformation
Loading summary
Transcript70 lines
- [00:01]
A
I've come full circle. When I first started, I was in an office that felt like a home. I had families coming in. And now I'm actually enabling care that takes healthcare back to the home where people live. And it's not the last mile, it's the first mile of healthcare starts in the home and I get to do that.
- [00:25]
B
Welcome to Embracing Digital Transformation, where we explore how people process policy and technology drive effective change. This is Dr. Darren, Chief Enterprise architect, educator, author, and most importantly, your host on this episode. From Telemedicine to AI A New Era in Medicine with chief Innovation officer from the American College of Cardiology, Dr. Ami Bhatt. Ami, welcome to the show.
- [00:59]
A
Thank you so much for having me.
- [01:01]
B
Hey, when we talked, it's been a while since we talked. It was before the holidays and I went back and looked at my notes and it said on there, great interview. That's what it said. Plus the subject that we're going to talk about today, which is AI in the medical field, specifically around clinicians. But before we get started, everyone that listens to my show knows that I only have superheroes on the show and every superhero has a background story. So Ami, what's your secret identity and your background story?
- [01:32]
A
Oh, my goodness, I love this question. So let's. We'll go back a little bit then, I guess. I was trained as a kind of boots on the ground cardiologist in the office every day seeing patients of all ages. I trained in pediatrics and adult to be able to just kind of see anybody that needed to be seen. And I actually started in literally an apartment. So it was a two bedroom apartment and one of the offices belonged to the senior physician that I had joined fresh out of my training in practice. And I had the other one. The kitchen was turned into a little EKG room and Sharon mans the EKG room. Sharon was 82 when we started practicing. When I started practicing and. And Mary was at the front desk and. And everybody knew everybody. People would come in, they would sit in what was literally the living room of the apartment and even patients kind of knew each other because they oftentimes picked the same day the same time. There were families, we had younger people. So. So that was my first experience. It was really like hanging a shingle. And I went from that to, you know, the growth of medicine in general to administrative positions while I was still practicing where I would take care of adults with congenital heart disease, people who were young but they were kind of growing up and they didn't have a home in the adult world. I did women's cardiac health before the field even existed. So that's always interesting to say. It definitely dates me and how old I am. And started to run outpatient cardiology at the Massachusetts General Hospital in Boston. About 60,000 visits a year. 134 clinicians, 43 rooms. And of course, two years into running it, Covid happened, and the whole operation shut down. And I was like, oh, God, what will we do now? But fortunately, those young patients I was telling you about before I took care of, they had really insisted that I learned to do telemedicine because they wanted to be able to FaceTime with me. And I was like, well, I can't use actual, like, FaceTime, right? So we've got to do something else. And so my patients had gotten me to learn how to do telemedicine. They were average age 24. They just wanted to connect from work or from school or from home. And not everybody lives in downtown Boston. And so when Covid happened, they said, well, what are we going to do now? And I said, I think I know. And so in 48 hours, we turned all of cardiology at Mass General back on. Virtually that year, we were above 70,000 visits. So we provided more access to more patients than we had in the previous years. And it was just. It was kind of a revolution. And coming out of that, I realized that I loved seeing patients one at a time, but I could touch hundreds to millions of patients at a time if given the opportunity. And that's when the American College of Cardiology, which is where I've been working for the past three years, had a position called the Chief Innovation Officer, which was, how do we get exquisite healthcare into the communities where patients live? And I said, you know what? This is? This is the right time, the right place. And so I followed my gut instinct and moved over to this position. And it's been phenomenal ever since. The field has grown. We'll talk about it, how AI has come in. But the most important thing I think to me is I've come full circle. When I first started, I was in an office that felt like a home. I had families coming in. And now I'm actually enabling care that takes healthcare back to the home where people live. And it's not the last mile. It's the first mile of healthcare starts.
- [05:15]
B
In my mind, and I get to do that. I love that. I love that. And I don't know if I told you, about 14 months ago, I had double cabbage.
- [05:24]
A
Oh, my goodness.
- [05:26]
B
Yeah. Full, full Full thing. I did not have a heart attack. I had a very good functional medicine doctor that diagnosed it and said, get to the hospital right away before anything happened. And when they ran the scans, they saw and they said, you're going into surgery right away.
- [05:42]
A
Wow.
- [05:43]
B
So I appreciate.
- [05:45]
A
Fabulous.
- [05:46]
B
I don't know.
- [05:46]
A
You know, the audience can't actually see you right now, but. But you look far. Fabulous. Fabulous.
- [05:50]
B
Thank you. Thank you. I. I credit a good cardiologist, good functional medicine doctor, and a great heart surgeon. That I appreciate, obviously. Excuse me. I owe my life, too, cardiologists. I appreciate you, and I appreciate that you're bringing that care to the whole world in the way that you are. And I want to. I just. Thank you just from the bottom of my heart. I appreciate it. I want to touch on. To stop Darren from crying. Let's. Let's touch on the catalyst. The catalyst, it appears to me, because I saw this not just with what you described, but with a lot of different things. The catalyst was Covid. Covid forced us to adopt technology in a very rapid way that before, because of cultural change that we needed to overcome, because of regulatory issues and the fear of that, all those things disappeared. And you've turned this into something old. Awesome. Which is I can now. I can now move into people's homes. You called it the first mile. I love that. What a transition. That must. That it must have not just happened, though, right? I mean, there was a big catalyst, but what took. What made that happen?
- [07:24]
A
Yeah. So, you know, it's interesting. Plato, I believe, is the one who said, necessity is the mother of invention. Right. I coined a phrase actually, before COVID that reflected my practice. So I'll tell you the truth about the practice in telemedicine, my patients stopped coming to see me. They would graduate from college, and their parents wouldn't, like, insist that they come when they come home anymore. They. And I was losing patience as they were going out into the workforce because they couldn't afford to live in Boston. They were young adults, and they couldn't necessarily afford to drive in. They didn't want to. I kind of had to do telemedicine. They suggested it, but. But I had to do it. And so the phrase I used at the time, which applied again in Covid, is desperation is the mother of adoption.
- [08:09]
B
Ooh, I love it.
- [08:10]
A
I adopted telemedicine because I had to get my patients back. We adopted telemedicine, digital health remote monitoring, after Covid because we had to get to our patients. And this was oftentimes the only way. And I'll take it forward to AI for a second. Back when I was in that office, right, with the two bedrooms and the patients in the living room, the amount of information for any one topic was like in a textbook or two. And there were a couple seminal papers. Today, you take any one subject in cardiology, there are more papers and research than you know what to do with. And there's no way the human brain can parse through that. Darren, I can't see you in 20 minutes and give you the best scientific care possible because my brain can't absorb that. And so we are now desperate again to adopt AI to help me get the information I need to help you.
- [09:09]
B
So what do you. What do you attribute to this explosion in data and research in cardiology specifically? I mean, we're seeing it across the whole medical field, but in cardiology, is there that there's more cardiologists, there's more people doing research, or what caused this explosion of data?
- [09:29]
A
I think there was a really great phase of fiscal support for research, both on the industry and academic sides. And largely the US really did contribute to a lot of this research. Of course, there's global efforts, but. But I think that that honeymoon phase, or whatever you want to call it, of really people being allowed to experiment and think that was a phenomenal phase of research for us. And that has, you know, pulled back as the world has pulled back a little bit. But I know now as I look at AI and next week I'll be at the World Economic Forum and in a side meeting, and we're talking about AI for drug design, right? And so just think about the ways in which, again, I think we're going to have that boom where we get even more investment in and efficiency of figuring out how to treat disease. Um, and so I think there was a. There was a high for that. And we saw all of the information increasing and the scientific rigor increasing. And now I think there is a second wave of increasing scientific information that's going to come because of the efficiencies of how we do research.
- [10:43]
B
So isn't there a fear? Because we see. We've seen it out there. There's a fear of my doctor. Now, I don't believe this because I want my doctor to use AI because I feel the same way you do. But isn't there a fear out there that my doctor is not going to. They're just going to relegate everything to an AI and they're going to lose some of that Intuition, some of that. The skills that they learned in school of exploring and discovering. Because the AI is just going to say, well, this is what they got, and I'm just going to sign off on it and go.
- [11:19]
A
So I guess a few things. The first is, although AI can look at you, your wearables, your history, maybe if you fill out some forms and write some things. And then the science, of course, it can't do context, it can't do nuance, it can't do edge cases because there's just not enough of it to study. And so those are the elements that make the clinician patient relationship what it is. It's the context, it's the nuance, it's the ability to see the edges. And so that's not going to go away. But if you ask me, am I going to spend a half hour talking to my doc when I get my blood pressure checked once a year, the answer is no. And you're not going to want to do that. You're going to want to just check your blood pressure at home, by the way, more than once a year in a more continuous fashion. We're getting there.
- [12:10]
B
Yes, I do.
- [12:11]
A
And when you really need us for your bypass surgery and the conversation and the new diagnosis, we are not going to relegate that to AI. Right. And so I think there is a place for some of the continuous preventative care to be better if we can use more data and more care at a distance and for the care where you're waiting. Like, you don't want to wait six weeks for your surgeon to see you, but in some places in our country, you're waiting 812 weeks to get your surgery. And ideally, if the patients who don't need to come in can stay home, be productive where they are, take care of their kids, their parents, it'll free up space for, you know, the Darrens of the world to come in faster to get the attention they need.
- [13:00]
B
Yeah. So this is really interesting. Do you feel that the patient load will then increase, that you can now have a lot more patients? Because I. I was kind of hoping AI would give me more time with my doctor where. Where I could say, you know what? I, you know, help me with my preventative care instead of waiting till, you know, I'm on death's door.
- [13:23]
A
Yep. Right. Yep. So I think a few.
- [13:26]
B
There seems like there's this weird balance that might.
- [13:29]
A
There was a. There's a horrible balance. Not even weird. Right. Because we live in what we call a fee for service model in most of this country, which is how many patients you see, how many procedures you do, and how much money the hospital makes. Right. Or the practice makes that pushes us towards the word efficiency. But it's not really patient experience, it's not necessarily outcomes. And so one thing I've been really kind of adamant about is as we implement AI, I don't want the metrics to be efficiency. I want it to be the outcomes of the patients. I want it to be the patient experience. I want it to be the clinician quality of work. Because we are losing doctors who practice left and right. It's a little funny that I'm the one who says that, but we are for a variety of reasons and we can't afford to because we have more and more patients and fewer and fewer clinicians. So, and so, you know, the clinicians willingness to work, the clinicians wanting to be there like I want to be back and hang a shingle. I don't want to be a cog in a machine. And so we have to design AI and design data and design systems that support that patient clinician interaction. If we don't design it, we'll just do efficiency. And you're right, it'll be cognitive. And yeah.
- [15:03]
B
I have a unique situation because I have a concierge doctor, so I pay him monthly, right. So he is motivated to keep me healthy because if I'm healthy, I'm not in his office. Does that make sense? So his motivation is the opposite of the way the health care system runs, where it's fee, fee per, per procedure, you know, whatever it is, we still need. I still need a cardiologist, right. I still need a, a surgeon, right. That, you know, I want to pay as much money as there is to save my life. Right. But wouldn't it be even better if, if my, my doctor got paid more because I was healthier? Yes. Well, that's amazing.
- [15:49]
A
So this, this is the future of value based care, which is, you know, the outcomes of your patients is what the system gets paid for. And it is, it is kind of that holy grail that nobody seems to know how to get to because you have to completely scrap the current fiscal model and to create a new one. However, CMS is working on this right now. So there's a new program called the Access program. It's for early prevent disease like you were talking about, Darren, thinking about the risk factors for cardio, metabolic, kidney disease, obesity, other things. Right. And the idea is, hey, we're going to pay companies who can go out there and help patients take care of themselves. We're going to pay them directly for outcomes. We're going to measure how the patients do and we're going to pay for it. And the government is testing this over a 10 year period. This is a 10 year pilot. And the reason it's not really a pilot is the goal is at the end of 10 years, this should be a way that we practice in the United States if we do it right. So there's a lot of people who can naysay and problem report about what the government is doing. Not the rest of the government, but like in this program. But what I will say is I think we need to work together to design it better because it's a really great opportunity to let people stay at home where they live. And Darren, if we could have found you even earlier than the acute way. Right. So I'm so happy you are here and you survived and you didn't have a heart attack or an event. But what I'm not happy about is we let you get so far as to your first presentation requiring heart surgery. We have got to move upstream cardiac care. I mean, heart attacks are fully preventable. Cardiac disease is by and large, especially atherosclerotic, you know, plaque buildup. That's what I have, preventable. And so we really need a better preventative health plan. And that doesn't happen in a hospital.
- [17:44]
B
No, it doesn't.
- [17:45]
A
It doesn't.
- [17:46]
B
Yeah. So how do you feel? We've got all this catalyst for change. How do you feel? And I always talk on my program, people, process policy and technology. That's what makes change. We talked about the cultural shift here. Process changes, they're going to be massive. Process changes, policy. We're seeing government stepping in and creating new policies to make this happen. Let's talk about the technology side. On the AI side, where do you see AI fitting into helping this massive transition happen?
- [18:21]
A
Yeah, yeah. I think there's probably three different areas. So the first area for AI is the administrative side. Right. So for example, I can now see you in my office. We can have a conversation and voice to text. AI technology can just capture our voices and write the note. So I'm no longer facing away from you. Right? Exactly. And so it saves time. And even if it's not saving time, because some studies show it doesn't save time, we as doctors are a little bit type A. We go back and we perfect the notes. We have to get over that. But it definitely brings the humanity back to the relationship. My back is not facing you, Darren. While we're seeing each other. And that's noticed that shift.
- [19:01]
B
I've noticed that shift in the last couple of years.
- [19:04]
A
So those kind of things, insurance forms, the amount of middle management that happens in healthcare and why, you know, fully trained MDs who graduated from their fellowship at 35 need to fill out paperwork. Right. Or even admin assistants who could answer the phone with a live person are instead on hold with insurance. Right? So. So I think some of those places, the efficiencies in the hospital supply chain, in the hospital order, just the right number of stents, just the right number of. So I think those things are great. AI is going to help. Then there's this middle area that people are worried about. You alluded to this earlier, which is clinical decision support. Is the AI going to decide Darren's care, or am I, Dr. Vaught, going to decide in that area? I don't like the phrase clinical decision support right now because it implies the computer will decide. I prefer the phrase navigating to knowledge.
- [20:02]
B
Ooh, I like that.
- [20:03]
A
I am still. I am the clinic. The clinician is the kind of. Is the apex intelligence in the room. But we will get the right knowledge to that clinician faster and more thoroughly. Remember how much information there is now.
- [20:16]
B
Oh, yeah.
- [20:17]
A
And then allow me to make the best decision I can for you. And this is similar to the airline pilot, right? Most people expect, when I get on JetBlue, I expect there to be a pilot and I expect there to be a copilot, and then I expect there to be the autonomous system that is actually flying the plane when the pilot comes back and waves hi to me as he's getting a water. So I would be really concerned if it was just a computer and there was not a pilot, because I would want somebody to make sure things are going right. I would be maybe more concerned if, you know, there's a Boeing with only a guy and no computer system. And eventually we're gonna want that from our clinicians. We're gonna want them to know how to use the compute power to make them the best clinician they can be. And that's the kind of clinical decision support we need to get to. Or let's say you're in rural America, right? I was on Rural Health Radio and we were chatting about kind of access. There are so many community health workers out there. What if AI could help upskill them to say, hey, not the exquisite care, not bypass graft on Darren, but this patient is sicker than this one. Why don't you send that one to the big city first instead of this one. Right. Like those kind of triage mechanisms. So those are the places where we can upskill people and increase our workforce, where we can make the doctor or nurse even better at what they're doing by getting them the information when they need it. Those are the places where clinical decision support makes sense, but it's not the decision the way you're thinking of a robot taking care of.
- [22:05]
B
Right? No, I, I, I like that you're. I call it AI augmented. So a subject matter expert, a doctor is AI augmented.
- [22:14]
A
That's right.
- [22:15]
B
And I had this discussion with another guest on the show and I asked them the question, if you had two doctors and you knew one was using generative AI and one wasn't, which one would you choose?
- [22:26]
A
Yep.
- [22:27]
B
And I said I'd choose the one with Gen AI and my guest said I would not choose the one with Gen AI because I feel like that would become a crutch.
- [22:36]
A
So here's the thing. This is the most, the third and most interesting part of what's happening with AI is human AI interaction. We need to actually study it. So we're working on a trial right now where we are taking doctors and nurses and they have X number of AI tokens. They can use AI to answer questions, but they only have so many tokens and they have more questions than tokens. And we are seeing are they interrupting, choosing to use the tokens in the area that they're expert? Are they choosing to use the tokens in area where they know nothing? Like I'm a cardiologist, if you ask me about like the knee. Right. I should not even be insert. I should just say go to ortho. Right. And how are they using it when it's something a little bit adjacent to what they do, but they think they know the answer and that's what we're studying because we want to know. And so there's not a clear answer for every given situation. If you have a doctor with Gen AI or without, there's not a clear answer that this doctor will be better or worse. With Gen AI, it is unique to the situation. And that clinician and eventually our computers are going to have to be smart enough to say, for these kind of questions, I'm going to give this guy AI, he's going to be better. For these kind of questions, I'm going to leave her alone because she's brilliant the way she is. And that's the future of human AI interaction medicine.
- [23:57]
B
That's what we're I like that a lot. Where, where the gen AI is letting us be humans, right and, and guiding, almost manipulating us in some respects to use the thing on top of our head, right? To use our brains to, to, to give more critical thought. I, I teach at Vanderbilt University right now and I'm doing the same things with my students. I'm trying to get them to think and use generative AI to help them think more and to critically think about the problems that they're solving. And that's what I would want with our doctors as well is hey, absolutely bring the most out of yourself instead of just relying on.
- [24:41]
A
I think one of the challenges though I will say is it's busy being a doctor or nurse, you know, in a practice. And the augmented intelligence phrase is one that comes up a lot. I, I, I prefer the phrase collaborative intelligence. And the reason I do is this. When the AI is being designed, we need clinicians at the table. When you are using it, AI will never be ready out of the box. It, it will never work perfectly out. There will always be something. Right now we call it hallucinations. It may be drift, it may be something else, but it's not the kind of technology like hey, I made a valve. You're going to put the valve in the patient. The valve isn't going to change. No, in fact AI is going to change and respond differently. And so we need clinicians to say I'm willing to iterate. It's not going to be perfect. But my clinical intuition is so good that if I see that it's not doing something right, I tell someone and then you need the system to be responsive. My doc said that didn't work. Do I have a team of data scientists that can quickly change this algorithm and make it better? And how am I doing it? We have none of that infrastructure. None. It's not there. So we need to.
- [25:52]
B
That brings up an interesting point. Do you see? I call them community gen AIs. Do you see community gen AIs being developed like a cardiology gen AI that is trained with feedback loops from cardiologists, not from stuff on the Internet though the trimness but researchers and cardiologists where they're the ones contributing and they're sharing information continuously back into the model so that the model becomes the collective of everyone's knowledge and insight. Do you see those starting to be built.
- [26:47]
A
100%. So there are some that are being built, you know, literally still like people's garages. There's some cardiologists that are terminal, their own. But, but we also at the Marian College of Cardiology, you know, we've been talking with Open Evidence, talking with Doximity, Helio, WebMD, all of these larger educational kind of question answer, other thing platforms where, you know, how do we do this? And this is kind of a group thing, right? Like talk to meta. Like there's so many people out there where that's where they're getting their information. How do you make information better with human experience? We don't just want words and word association. That's all LLM is.
- [27:24]
B
Yeah, yeah, just words.
- [27:25]
A
How do words associate with one another? We want, how does the human experience affect those association of those words? And so, and so that's really, we're aggressively kind of building towards understanding that better. I gave a, a TEDx talk in Boston earlier this year and it's interesting, kind of after I got off the stage, we were talking about preventive cardiology. We're actually talking about longevity and how heart disease is really probably the closest thing to a magic bullet we have for longevity right now. Turns out that if you're in good heart health by the time you're 50, you gain years upon years of life over the average person who doesn't. And so there's nothing else that gives you 8 to 17 years of life but healthy heart. But when people came up to me afterwards, they said, hey, do you think that there's going to be an AI model that's going to take all of my stuff, my genetics, and give me a prediction on what's going to happen? And I said, you don't want a prediction and what's going to happen. You want a prediction on what you can modify so it doesn't happen.
- [28:29]
B
Yeah, I like that.
- [28:31]
A
Right. And that's, so that's that next step that you're asking about, like what is coming next? I think that's that next step that's coming is to understand how do we not just predict, how do we predict? What can we modify realistically to make this person. And then there's a question, do you want to live longer? Do you want to live better? Right. And so I think there's a lot.
- [28:54]
B
To be done and that's something my wife and I have discussed. Do I really want to live longer? I don't. If I'm bedridden.
- [29:02]
A
That's right.
- [29:03]
B
Right. I want to live better. Right. And, and, and then if, if my time comes and I'm happy and I'm, you know, healthy as a horse, and then all of A sudden I'm, I'm done. I'm, I'm happy with that, but I don't want it to linger on for 10, 15 years. Oh, we're keeping you alive. Well, my quality life.
- [29:23]
A
I don't know that this. Yeah. Is this life worth it? It's. We in Davos next week, one of the meetings that we're going to have. Actually, it's, it's a longevity kind of conference, Longevity Investors Conference. And one of the things that we're really kind of focused on is, is how do you find the technologies that promote health span, not lifespan.
- [29:45]
B
Yeah.
- [29:46]
A
And that's, and that's what I'll be talking about. But like, I think you, you nailed it there, Darren. You really just described what you and your wife talk about is, I want my health span to be as long as possible.
- [29:56]
B
Yeah, absolutely. Because we're, we're seeing our parents that are in their late 80s, right. And we're seeing that, what they're going through. And I'm like, oh, my goodness, they go to more doctor's appointments than I have meetings at work.
- [30:09]
A
Right.
- [30:09]
B
And I'm like, dad, I don't really want that from. For my.
- [30:12]
A
No. My, my daughters were, My daughters were just saying yesterday, they were like, wait, we need to get four knees done. And I was trying to understand what they're talking about. They had just gotten off with grandma and grandpa and both of them need like two total knees and they're like, we're going to do four knees. And I was like, oh, God, that's going to be quite an adventure, isn't it, for our family? But, yeah, aging gracefully. I mean, my parents are doing it, but I gotta say, I'm not sure I'm going to be able to. So I hope some of this early prevention stuff works for me.
- [30:40]
B
Yeah, yeah, same here. Ami, this has been wonderful. Thank you for coming on my show today.
- [30:45]
A
Thank you for having me.
- [30:46]
B
And hopefully my. I could keep talking to you for hours, but we don't have that much time. So if people want to find out more about what you're doing, how do they find out more? Especially doctors that are listening to the show. Where can they go to find out more about, hey, getting involved in something like this and upping my game, basically?
- [31:08]
A
Yeah. Yeah. So they can find me on LinkedIn. Oh, gosh. AM. I think I'm. I want to say I'm Ami Bot MD on LinkedIn. Is that possible? We should, we should double check that. We'll put a link or something there later. That would be great. And then. And then soon there'll be dramnybot.com so that'll be just another place to be able to find me.
- [31:31]
B
Awesome. Awesome. Ami. Again, thank you for coming on the show. I appreciate your time.
- [31:36]
A
Thank you for having me, and I'm glad you're in good health.
- [31:43]
B
Thanks for listening to Embracing Digital Transformation. If you enjoyed today's conversation, give us five stars on your favorite podcasting app or on YouTube. It really helps others discover the show. If you want to go deeper, join our exclusive community@patreon.com embracingdigital where we share bonus content and you can always connect with other change makers like yourself. You can always find more resources@embracingdigital.org until next time, keep Embracing the Digital Transformation.