Robert Pearl MD – Fmr. CEO Kaiser Permanente
[Music] Network bricks on the screen. Healthcare in between. Physics art and math got us down this path. Historical tales come alive. Where science and stories dive. Blended facts with heart. Knowledge from the stars. Meet bridge blues. We sing to the future memory. [Music] Health and science in the base. M’s got the fix. Complex every way that we actually I want to lead the way and ways that make me more efficient in the world of digital streams of old stories. Thank you so much for joining with me. It is truly an honor and a privilege. MedBick’s webcast is a new kind of healthcare media channel to break the silo mentality and foster interdisiplinary thinking. Our mission is to fill that gap and to create a curiosity through different lenses, tapping the passion that guided the historical figures in medicine and bringing present-day groundbreaking research to the front lines. We interview physicians and leaders in the industry about present day healthcare challenges on our live programming. Hey, welcome welcome to the Medrics webcast. This is Raman Anamar Raju. I’m the host. I have a new addition to the medrics. Her name is you know I tell you she’s a better looking better talking and she’s a doctor of nursing practice DNP. She is a Dr. Susan Davis is joining as a co-host is a very welcome. She’s a you know very interesting background she has she has a you know she while she’s not saving lives she saves lives you know rescue she’s a founder of rescue RN rescue RN is a she’s a founder while she’s not saving lives she raises cattle so she who else who doesn’t want her to be co-host welcome to the medics subcast Susan welcome welcome to the oh yeah welcome thank Thank you, Romana. What a what a fun pleasure. I I’m thrilled to be here and and what a what a great night. Thank you for having me. You survived three hurricanes and lost home and build back it, you know. So, you have all the all the best things in the world you have behind you, you know. Really, really, you did everything. You’re a great survivor. The season is young here in Florida. So, we’ll we’ll we’ll hold our breath on that one. Okay. Yeah. So let me introduce our guest and just let me talk about a little bit about the background a little bit. Healthcare is in crisis you know everybody knows about that you know I don’t have to you know every I don’t know why I’m not getting the oh that’s not she’s not there. Okay. So, healthcare is a is is a kind of in crisis because we are in a we are in a big doctors are not happy, patients are not happy, rural hospitals are closing. So, when there’s a crisis, what we do here as a humans, we look for we look at the past, we look at the sages, teachers, visa guru. We have such a teacher we have with a full of wisdom. He can he’s an Nostradamus I call him. He he can see the clearly the future of the healthcare and I learned so much from him. He is a crisp writing. He is one of the one of the the pioneers in healthcare. Dr. Robert Pearl is a he is a he is a plastic surgeon and he is um he is studied in Yale Stanford. So qualifications you know out of this war and then he became the the former C CEO of Kaiser Permanente the largest healthcare physician healthcare group. After that he you know that he retired now but now he wrote a he authored several books. I met him actually one of the book ceremony in Commonwealth Club you know 10 years seven years ago I think. Uh so I met him before the pandemic. Uh so now there is so much changes. uh we have a new player AI in the in the game. So Dr. Robert P wrote authored a book called Chad GPT. So let me bring him to the to the table and let me let me MD MD yes Dr. Pearl welcome welcome to the Oh just a minute one second. Welcome sir. Welcome to the Medwix webcast. You’re thank you for coming back. you know you are you’re coming couple of times here on my podcast. Honored to have you. Thanks Roman for having me and welcome Susan to the show. Thank you sir. I have a question for you. I introduced everything but I have a question for you. We have a after the last time I spoke with you there is a big changes came in artificial intelligence progressing enormous speeds is accelerating in a in a in a positive direction. But there is something crisis going on. The crisis is Geng generation and millennials are leaving the doctors in droves and and they’re going to rescue to the to the to the LLMs logic language models or chat GPD or Google. They’re risk going to these kind of things. What what is going to happen to the doctors? That is a crisis. What is going to happen to the doctors that if you can if you can help us? Well, first of all, let me just tell listeners and viewers that all the profits from my book Chat PTMD: How AI empowered patients and doctors can take back control of American medicine goes to all the profits go to doctors without borders, a great charity that we all should be supporting right now, particularly around the globe. So, the question you’re really posing is that as has happened for millennia, you have evolutions and generations. You have your baby boomers, you have your Gen X, but then you have the millennials and the Gen Z. And these individuals, by the way, it’s 40% of the population. They’ve grown up expecting to get information from other sources besides from doctors or nurses, uh, from getting it from standard textbooks. They want to be able to get the information um, quickly, social media, on their phone. uh and they trust a lot of people who are there particularly now when you have a tool called chat GPT and I by the way I use chat GPT just a shorthand it’s really any large language model could be Google it could be uh Gemini I have uh no relationship sorry it could be claude it could be Gemini I have no relationship with any of these companies I just liked all the letters flowing after each other so I think that we’re going to have to evolve to provide care to them in a different type of way than we have before. But I also believe that what will happen across times is the other generations will come along because the generative AI tool not only provides information provides expertise. You know, you go to Google, Dr. Google, and you put some information in place. You you get a lot of links. Doesn’t necessarily mean they apply to you. You can’t be sure what’s accurate or not. When you finish, you still need the clinician to be able to interpret it for you. But with a generative AI tool, it’s very specific. And because you put all your information in place, it’s about you. It’s personalized. It’s about one person. And that is remarkable. It’s never happened before in terms of the patient in the history of medicine. We have a different another crisis. I want to talk with you about that is the health the iron triangle you talk about frequently about the iron triangle health access and cost is unbreakable either one of them you can solve the other one always going up so how do you health care system how do you solve that problem of this you know American system healthare system is a paradox cost going up and access going down it doesn’t make any sense so what do you think of this iron triangle how do you break it And you think the AI can play the role is some role in this area. You think it’s no longer an iron triangle. Uh generative AI can break it and if we use it well it will break it. Let’s just take a one aspect of what it can do which is the the management of chronic diseases. We’re talking about diabetes and hypertension, asthma and heart failure. And what we know is that if every patient with chronic disease had it as well controlled as we’re able to do through the best organizations today, we’d have 30 to 50% less heart attacks, strokes, cancers, and kidney failures. Now, think about that. What would medicine be like with 30 to 50% fewer heart attacks, strokes, cancers, and kidney failures? The estimates are it would reduce course cost by 1.5 trillion dollar. Now all of a sudden, I call that higher quality. The access can now improve because you’ve now created 30 to 50% more capacity and costs have come down. We have to get this notion of the iron triangle triangle out of our minds in this era of generative AI. Susan, do you have a you want to follow up some questions on that one? Well, you know, I wasn’t privy to the original triangle conversation. So, you know, it’s I’m from I’m summarizing and and skipping ahead like Dr. Pearl just described. And first of all, I’m in complete agreeance with, you know, smashing that triangle and moving forward because what are our options anyway? So, um I I love the prediction if and it’s not even a prediction. We’re here, right, Dr. Pearl? I mean we’re here well let’s let’s take an example you know you’re you’re a nurse doctor too but a nurse as well probably before you became a doctor of nursing and you know you take care of a patient with hypertension so what do we do in the medical profession today patient comes in sees a clinician we diagnose hypertension we start on medication and what do we say come back at three to four months another it’s elevated what do we call that white co so now we’re already six months nine months in the hypertension which by the way accounts for 40% of strokes which you know but the listeners and viewers may not um it’s not controlled now let’s take a generative AI tool what you what we know is that you can have mobile blood pressure cuffs the patient can have transmits information using Bluetooth it now would go to a generative AI tool and over the course of a month three times a day would take the blood pressure it could graph how whether it’s improving or going down and after 30 days we know is the medication If it’s not working, you need to either change the medication, change the dose, do something different, and now all of a sudden hypertension becomes controlled and now 40% of strokes disappear. That’s a pretty good win-win. higher quality, better access because the patient’s not sitting in the hospital or the ICU with a stroke and dramatically lower costs because we’re talking about literally pennies for some of these medications maybe load single dose single numbers and dollars and we know generative AI tools essentially are free if you want them where you can pay $20 a month. Three lattes a month at Starbucks will get you access to the totality of medical information and the expertise that’s required. Now, you got to work with a clinician. This is not doing it totally by yourself and whether that’s going to be a nurse practitioner, whether that’s going to be a physician, someone is going to have to get you the different medicines and make the changes. But now you don’t have to see the person. You can do it on a text message. You can do it on a email. You can do it on a tele medicine phone call. We we’ve we’ve taken Dr. Pearl. Dr. Pearl, but but what you’re But what you’re saying is is happening now and it’s miraculous and I’m on board and I’m on your team and and let’s go. However, I’m going to go ahead and speak to like my own family. I mean, they have me to live with and they’re not adapters, right? I’m an early adopter, but they’re not. And so, I mean, I’m wondering about access, not just rural access, but people who are are not they’re, you know, they’re butting the system here right now. They’re thinking this AI thing is scary. It’s it’s horrible. Much less their health care. I mean, they barely trust their provider, much less this AI machine. But to to your point, I’m it’s it’s miraculous to me of what we can do to fill these gaps. world right now you the first visit, second visit, third visit with our different hypertension medications, but like I can’t get my mom to even barely listen to me about you know how well wonderful this AI can help her. So to your point, how are we going to get this to the people who are not listening to how amazing this is? So, your mom couldn’t get your grandma to try ATM. Uh, and you can go down the entire line. You know, I don’t know if, uh, you know, your kids could get you to get an iPhone or whatever. I don’t know how old they are, but this is the nature of technology. There’s always the early adopters. What I’m telling people right now, both clinicians and patients, is experiment. Don’t trust it. Just play with it. It’s fun. asking questions, pretend it’s an expert, you can use the voice uh feature that’s simple to use. You don’t even not need to know how to type and just put the information in place. You know, if someone has a your mom has a pain in her hip and is going to go see the clinician, put all the information in place and see what it provides. You know, I u was on a podcast a while ago and the host at the end of the visit who never used it for medicine said to me, “My husband fell skiing. He slipped 100 feet. His arm was over his head. His shoulder still hurts. It’s three months later. You’re a skier, Dr. Pearl. You’re a doctor. What’s going on?” And I said, “I’ll tell you what you said. You never used it. So why don’t you do the following? Put all the information in place at the generative AI. see what it says and then give me a call back. Five days later, she calls and she says, “Thank you so much.” It said he probably had a rotator cuff tear. It said he needed an MRI to establish a diagnosis and he should see an orthopedic surgeon because he almost definitely needed surgery. Now, compare that to Dr. Google or whatever else your mom likes to read to figure it out. And of course, went to the surgeon, says probably a rotator cuff tear, but now they knew exactly what the rotator cuff was. as you know, a very complex mechanism in the shoulder. Number two, when he said he needed to get an MRI, he said, “Yeah, bring it on because I we you need that.” And after surgery, he said, “If you had waited three more months, I probably could not have reattached the tendon because the muscle would have shrunk.” That is expertise. Once you have one of those experiences, you never technology again. Bring her over and get her to do it and then let me know what happened. How’s that, Susan? Yeah, but you called it because that’s the one thing you have to get them off of Dr. Google and on to Dr. Chat GBT at least once and let them use their microphone and then they’re oh because what comes back is truly miraculous. It truly is. So to to your point I agree. Yeah. Yeah. along those along those lines. Actually, I wanted to do 30,000 ft. Question for you in in Texas, you know, I used to live in Texas part-time at least that that in used to go there anybody’s house in the basement, they have a they have every gun in the in the planet with the guns and bazookas and everything. They think that government some agency is going to come and take over them. In the same time, in the same place now they’re building a10 billion dollar data center by Oracle and the charge GPD open AI they’re building the 10 billion center huge this creates a lot of unsettling things for a lot of people including the educated people where is this intelligence centralized intelligence control that te tell us a little bit about that Dr. Uh Pearl, what do you think about that? Well, what I think what you’re describing is the technology which is, by the way, it’s it’s two years old. In fact, it’s not even two years old. It’ll be two years old until the until November the 30th. So, we’re talking about a technology is less than two years. It’s ve it is um energy and water intensive and the sites to be able to run it uh are being built because people are anticipating the growth that’s going to be sitting in place. What you see with every technology is that each year it becomes more efficient. Each year the tools are able to be redesigned in ways that require fewer chips, less energy, less water sitting in place. But I think the I think it’s really um a red herring if you want to use a very old term because this is going to happen whether medicine likes it or not. This is a tool that is so powerful in terms of being able to propel businesses. We shouldn’t even be asking should it happen or not should it. It is going to happen. Now the question is knowing it’s going to happen. Are we going to jump on board and lead the way or we as clinicians going to seed that space to someone else? because I’ll guarantee you they’re not going to use it on behalf of either the patient or the clinician be they a physician or a nurse. So Dr. Dr. parallel. What what can you can can you tell me what are the two examples you can see you know specific examples where the doctors or the offices like even including nurses like Susan can use at this time you mentioned about the putting all the data and ask questions but are there any in the facility side the doctor’s offices they can use uh anything like that any do you think any any any use at this time immediate use Well, there’s no question that doctors can use it. You know, I was just talking to a surgeon two weeks ago. She had done what’s called a microchia repair, a child born without an ear. And we do that repair, we take out three ribs in order to carve and rebuild the ear. At the end of the surgery, we get a chest X-ray because the ribs sit right over the lungs. And if you tear the lining around the lung called the plura, you get what’s called a numoththorax. And the X-ray is taken. Now remember surgeons are not radiologists so they do the best they can to interpret it but she said anyway she said let me just get him chatt’s opinion and it created a report that was identical to what the radiologist reported 12 hours later. So the answer is absolutely it can be done but I want to go you know I love having Susan on the show bring you some nursing things in place. So, I want to talk about a hospital that Susan knows well. Not a specific hospital, hospital in general. So, when you admit a patient to the to the hospital, you can put them in the ICU because they’re really sick, or you can put them on the floor because they’re not very sick, or you can put them in this intermediate space called teley medicine, teleimmetry. This basically takes all those monitors that are at the bedside of the patient, puts it on a large screen, and a nurse sits there for eight hours in a row looking for a problem. Well, number one, it’s expensive, but more importantly, it’s a lot of patients who don’t get monitored. Now, we shift it around. We take a generative AI tool and we put it attach it to those same monitors and we can now monitor the patient. nurse comes by at 8:00 a.m., you’re doing okay. Comes by at noon, you’re in big trouble. Probably at 9 or 10, there was something there. And now all of a sudden, instead of the nurse doing the usual rounds everywhere, he or she knows exactly who’s in the biggest trouble. That is the kind of opportunity. What what this technology does is it fills in the spaces between the nurse coming at 8:00 a.m. and noon. the space between the doctor seeing the office and seeing you back in four months. The therapist, you know, a therapist sees you and they’re a little worried. You’re a little depressed. You don’t they don’t think you’re suicidal, but that’s the greatest fear that every therapist has and they see you again in a week or two. What happens in that space? We have no idea. This technology allows all of that to happen. And as I say, it does it at a reasonable cost and we can implement it quickly. Remember you before we had generative AI, we had what’s called narrow AI. These were tools trained on large databases. They were they’re called narrow because they had a single problem very expensive by the way to create. The tools that I’m talking about we could create very fast because we actually don’t need the tool. We need the training tool to teach the patient how to use it. I believe this is going to be a major part of nursing in the future. That part in the same way that nurses today teach you how to, you know, take care of your health and lower your blood lipids and etc. They’re going to teach you not necessarily that information, but how to use a generative tool to accomplish it. That is how I see a transformation and revolution in medicine. and I’m on that team 100%. So I I’m just going to comment briefly on your your your commentary on recognition of course. So now you’re talking my talk, right? Because recognition is recognize a problem, call for help and act. These are three things what failure to rescue. It’s a huge problem in the United States and globally uh you know failure to rescue especially in a hospital uh setting because they can’t be everywhere. We all know what space staffing is and ratios of nurse to patient is and we no need to get into that, you know, can of worms. But to your point, in that little room, I always call the little guy in the room down the hallway. It’s not necessarily a nurse. It’s a tech. It’s it’s an EMT. It’s a paramedic. They’re watching 15 screens of cardiac rhythms. And often we get a cold blue in the hospital based on, you know, the patients leads are off or they’re or they’re tra they’re they’re transferring the patient and we’re calling a code blue because of what’s being read. All right. So, a ridiculous waste of time, energy, and resources. Just that teeny little example. So, in recognition alone, I’m on board. But let’s even go a step further into the preventative area like you were saying. I mean, you know, getting further before let’s talk to these people before let’s fill that gap before they get into our hospitals where this can kick in and help them with their own recognition in and that gap. But in the hospital, I’m I’m on board with recognize the problem, call for help and act. So, the interpretation and the and in in the support for nursing is huge. And in the nursing conversation, you know, no one wants a robot to be called a nurse and nobody wants AI to be called a nurse. Well, they’re not. They’re not a human. We’re the humans. But we are looking for that support and I’m on that team. Yeah. Let let me go beyond that. You know, again, both of you know very well patients with heart failure. And so now we take care of someone in the hospital. They have heart failure. We drain the fluid off. We give them some medications to get the heart to pump a little harder and they go home. Now, what we know at home is that they’re likely to come back and be readmitted to the or go to the ER then be admitted to the hospital. We know that it’s usually a couple of days of increasing symptoms. They can’t lie flat. They can’t climb stairs as well. Their ankles swell. How about a monitor at home looking at them and at the first side of it going south. We use a tele medicine to a nurse in a in a center who can now evaluate the patient and we can change that medication before they have the crisis. You see this artificial notion of what’s an office and what’s a hospital and what’s a home. This is going to change once you have this assistant. We’ll call it an assistant. Whatever you want to call it to be. This is someone who’s there because you can’t do a 24 by7. But if you have a hundred patients you’re following as an outpatient, there may only be three, five that you’re worried about and only two get into trouble. You don’t have to take care of a hundred patients. You got to take care of two or three or maybe five. And that becomes a task that we can do. Very different than today. We’re going to have to shift by the way how we pay clinicians. We’re gonna not pay on pay for volume basis because we’re going to decrease the volume. We want to pay on a pay for value, a capitated basis. a single payment to a group of clinicians to take care of patients and the healthier they keep them. The more we align those incentives then the better they do to say nothing about how much better the patient does. I’ve never met a patient ever who wants to have a heart attack and have a reversed. If you can find me that person, I want to know who they are. They all want it and we can help it 100%. To your point, I always say, you know, cardiac arrest doesn’t care who you are or where you are. Thank you, Dr. Pearl. I spoke with one of the mathematician last last week actually. Her name is Dr. Dana Thomas. She is a prolific researcher in the AI. She introduced a new metric. It’s called a body round index. In the place of BMI, body mass index, she’s doing this body round index. She told me very interesting thing in the AI thing. She was she has a thyroid problem and then they she asked for some kind of a recipe from the AI. They gave the recipe. AI gave the recipe but she didn’t like it. You know the that doctor gave the recipe first. Actually she didn’t like it. So she went to the AI. Can you give me I like Indian food. I like a Meditaran food and they I like ramen and noodles and all those. Can you give me diversified meal structure for me? And then AI gave a wonderful really perfect for thyroid problem. He gave that is one thing I see strikingly immediately patients can see the benefit right there. Right. What do you think of? I do that every Sunday. Every Sunday every Sunday I go to generative AI. If I that weekend I’ve eaten a little bit extra, I say, “Well, this week make the calories a little bit less. This week I particularly want Italian food or uh you know sushi.” So, I’ll I’ll tell it all the things that I want. I’ll give it the budget that I have. And it not only will give me menus for every day. It also will give me a shopping list that I can basically just follow. If I I you know, I’m very healthy, but if I had diabetes, I could put that in there and they would modify it. If I had heart failure, I could put that in there and it would modify it. It personalizes it. This is generative AI. Google is not personal. Even WebMD is not personal. This is personal. I tell it exactly who I am and the recommendation it gives me is different than it gave anyone else most likely in history. You know that I interviewed Dr. Ken Nipple. He said that what he said that was interesting. There is a the patients are coming with a 60 pages of all the notes and then he gave it to the AI there is a company in San Francisco right here close by. Evidently it’s called evidently he takes and put in one paragraph all the summarized that was so easy for him that really see the benefit do you think that is the first one the summarizing these huge chronic problems people have all those things and then how do we stop in general the chronic diseases impending crisis we have we have in the world the chronic diseases tell us a little bit about AI and then influence in this kind of a the doctor’s office and then you well know both of you So, by the time someone has a long hospitalization, the record is about 2,000 pages. And now you got to write a discharge summary or a transfer summary. And it will write it for you. It’ll write it in perfect medical language. It will have pulled all the information out of place. You still got to read it. You still got to make sure it’s accurate. But what otherwise would have been two hours of your time now becomes 20 minutes of time. You know, I think that a lot of the nursing notes, you know, I don’t know how much time nurses spend writing notes. It’s probably a lot of time on a given shift that they could be taking care of patients, the AI can do that because the nurse is going to be entering data all along. It can pull that out. It can add laboratory data. By the way, it can include conversations when the family came by. It can include the information that that the specialist had at six o’clock on rounds. It can have all that information constantly updated. It’s it’s you know at the time it has to be done at 3:00 because at 3:00 one human being is leaving and one human being is coming in. So that’s an important time but instead of spending a lot of time uh providing a a uh update the person coming can read it can ask questions can dive deeper. Tell me more about Mr. Jones. I’m a little worried about this sitting in whatever room it’s going to be because now you’ve taken care of what right now fills a huge amount of the hour or whatever the um g giving the report’s going to be and now you can ask the more important questions just like if you know what that once you know what that rotator cuff is you know about the MRI now you can ask the really interesting questions about okay post-operatively what’s my pain going to be like what can I do how can I prepare care, what should I buy in advance, those conversations never happen because the doctor only has 17 minutes. Once you can move along that process to a more expert level, the way two clinicians might talk to each other with different levels of expertise in different areas. That’s where I think we can now advance the outcome. And one of the other points to me that you know it’s grabbed my attention. 400,000 Americans die every year from misdiagnosis. Now, why do misdiagnosis happen? I’m sure sometimes it happens because the clinicians doesn’t know enough. But I’m going to bet that most of the time it happens because the clinician doesn’t have enough time. We know that doctors uh interrupt the patient after 11 seconds. The history has a lot of information. the doctor’s too rushed to sit there and listen for five minutes. But now if the patient’s coming in with the background, we create that time. I think it’s going to increase fulfillment and reduce burnout. Why do doctors and nurses burn out? Because at the end of the day, they want to go home feeling they they did the best they absolutely could. And if the whole day they had to rush and cut corners, they don’t go home feeling great. This creates the most precious entity in medicine today, which to me is time. Go ahead. Indeed. Indeed. Well, I mean, that’s a lot to digest, but I can tell you I was at a hackathon just this weekend, and I was super impressed. It was it was a a hard tech hard and soft tech hackathon. They were all engineers, types of engineers I’d never even heard of, and they were creating these amazing things. But one of them stands out to just made me think of what you were speaking of, Dr. And that was, you know, traditionally it’s been the doctors who have been sitting in their little booths, you know, dictating after every patient encounter and you sit and you dictate, you sit, dictate. Not nurses, you know, we have all of our patients and and they’re after dictating or or typing, typing, typing. But now there’s a dictation device, an AIdriven dictation device for nursing. And right here in Florida on the east coast of Florida, they’re using it in ERS right now. And the team that I I witnessed in this hackathon was creating a better device for clearer non-garbled nursing dictation. I was like, “Oh, winner. I don’t care what everyone else is talking about. I’m down with the nursing dictation.” The amount of time that we do before, during, and after, and frankly, let’s circle back just a brief second to that recognizing the problem in deterioration. You know, we have a lot of early warning devices that are in in place for us to use, but they’re only as good as they’re used in a timely manner. So, if we’re dictating slashcharting postfacto, well, guess what? It’s not going to help us catch anything. So, so here we go. Just a teeny little, you know, case use and thank you for it. But, so yeah, I mean, very cool. I’m about it. I mean, Roman knows all about this. He’s probably had people on this show before doing ambient AI for doctors taking basically completing electronic health record. doctor used to sit in front of a computer screen, never looking at the patient. Now the ambient AI can listen in and be able to record all that information. What people may not remember is if you go back three to five years ago, I probably got called by a dozen companies. They said, “Oh, we got a great AI tool that can listen to the doctor and save two to three hours a day. The problem is no one’s buying it.” And you know what I said to them? your tool doesn’t save two to three hours a day because if it did, your biggest problem would be managing the line out your door. Your tool is not able to do what it says it can do. Now we have 70 companies that actually can accomplish it because generative AI is perfect for that. It can exactly understand language, predict what’s going to happen, put it into a logical context, completely different than these narrow ai tools of the past. It’s the same thing for nurses. It’s going to be the same thing by the way as I say in the room with the families. I had a p a patient I get a lot of things from individuals by the way use generative AI to make diagnoses that clinicians had missed. They couldn’t figure out but I also got one from a patient who said basically I turned my generative AI on in the morning and when the doctor came in it listened in. When the doctor talked to my dad, it was able to synthesize all that, create a summary for us, help us understand what was going on. And the nurse came by the same thing. Yeah. I mean, this is this is the opportunity. And again, I’ll say the same thing I said before. It doesn’t just give you information. It gives you expertise. You actually understand it all. You don’t come in with 60 pages of Google that is basically unintelligible without having all the articles and links. You come in with a good summary. And if you want, you can ask it to create some references and it’ll do that for you as well. This is a a total transformation. And I’m not worried about the technology. I worry about the humans because I’m just afraid we’re going to be so resistant and hesitant that we’re going to miss that opportunity. As I say, someone else will seize the day for us and we will regret that forever. You snooze and lose in this in in this in this case. I mean, we we listen, we have to just be a little bit on the regulatory, the HIPPA, the one of the things that came up in the hackathon. They’re like, “Wait a minute. Isn’t the speaker better in your phone, right?” They’re like, “Well, we could use the phone. You know, these are the engineers. Like, we could use the phone.” I’m like, “No, you can’t. Not in a healthcare system. There’s lots of safety. There’s lots of HIPPA. There’s lots of rules. And I think AI is moving so fast and it’s so good, but policy, you know, comes way after something happened and we know how policy works to what’s like the reality and so on and so forth. So I think there’s a cautionary measure to the usage especially in a healthc care setting on how the how does our policy rules regs and and all the 911 follow this beautiful thing. Dr. When I met you in the when you were the book mistreated you that that one of the books that that you in Commonwealth Club one of the things that when you spoke about it one thing that stuck with me for lifelong is brain hot spine that that is three things are the essential for the healthcare to survive healthare system to survive we have a brain now we have AI we got tremendous expertise brain we have we acquired that that’s a good and good for good you know No, no problem. your heart. He did a you know the doctors without a borders he went to he went to Panama or some other countries in the Chile South Africa he did a cleft surgeries right so he provided a hard compassion there but the last one that is a spine that is where my problem is where are these leaders spine spine means courageous leaders you know you and and I see Susan but we don’t we have too many that is a big thing missing leadership tell us a little bit about that leadership in the in the AI world on the healthcare world and and overall the policy side policy wise well you know again for Susan’s background you know in my book I was writing about the fact that leaders have to have the intelligence of the brain the empathy and emotional connection of the heart and the strength and courage of the spine so that’s the three parts that leaders have to have I think it’s become dormant I don’t think it I think it exists you know you becoming a physician is 10 years, you know, and it’s not 10 easy years. It’s 10 years of 80 hours a week or something like that during training. You give up so much. You have that ability. You obviously have the intelligence and you go into medicine for your heart because you want to care for people and take care of people. So, I don’t worry that it’s not there. I just think that what’s happened is that doctors they had they were made to run faster because the resources kept dropping actually in a uh inflationdriven kind of way. We’ve watched the average clinicians income drop by 44%. You got to run an office. You start cutting back on your staff. Soon you don’t turn the lights on because you don’t have electricity. I mean it’s really problematic. We never had the tool to make the advance. And what I’m hoping for is that there’ll be people, and all you have to do is get about 30% of people to get over the hump, and then everyone else follows along, who will try this, who will be willing to take a capitated payment, who will be willing to trust their patients. Now, they’re not going to do it because I say so. They’re not going to do it because, you know, Sam Alman says so. Uh they’re but if they try it and say, “Oh my gosh, this is exactly what I would have written.” like I told you about the um the radiology report was essentially identical then at some point you start to trust that as long as this is the point that Susan made earlier as long as you have that relationship with the patient that’s the hard part and you have to then be available to take care see the patient with the understanding it’s not going to happen very often you know I’ll tell you it’s interesting I early in my career I did a thousands of surgeries. And I gave every patient my home phone. How many people do you think called me at home in 10 years? Two. And they both had real problems. Now imagine what it communicated to everyone else. That is what we’re talking about. It’s a confidence that you gain because we, you know, when you’re being made to go faster and faster, on that treadmill more quicker and quicker, you get defensive. And I think some of the mistakes we made of clinicians selling their practices to hospitals, and then all of a sudden having provide that care being told to them by an administrator, that’s really problematic and it’s happening more and private equity is going to be even worse. So now we’re looking at an alternative and I’m just hoping it’s why I come on shows like your own. I will have um keynoted 50 conferences this year. I want people to start trying it to start thinking about it. I want medical schools to start teaching it. Not how do you use it? The incoming medical students know more about it than professors. But how do you improve the health? How do you avoid heart disease? How do you identify heart failure earlier? How do you make obscure diagnosis? We could get on the list of opportunities that are there and they all align with why people choose medicine for their profession. You know, talking about education, Dr. Per, my father is a small small village doctor. You know that, you know, his brain was his MRI, he is a CT scan, he has a diagnos, clinical diagnosis is a village, we don’t have anything. He has stethoscope and a pencil injection. But but now the medical education you know you’re it’s not changing much they education is still not teaching this interdisciplinary fields like a physics and mathematics quantum sensors and all these new fields are coming up do you think there’s a there is a flaw in the education system and the medical education I I think that the medical education has to change it’s you know but it’s always taught of the last generation because your professor professors are the ones who were trained 20 years ago and they’re going to tell you whatever it was that they learned at the time. They’re not going to be looking ahead and I think that change has to happen. But you know what would be fascinating as an experiment for your dad is you know maybe I don’t know 20 years ago I um was trying to explain to people what the internet would do and I came across a great experiment. Remember this is the internet didn’t yet really exist very much and someone took a computer and brought to a small rural village and in this rural village they grew grew soybeans and they always were surprised but they brought them to the marketplace sometimes they were you know 10% more costly and sometimes they were 5% less and they could never quite figure this out and he just gave him the computer he said play with it and they discovered this place called the Chicago Merkantile uh site futures understood what soybean futures look like and they gain 20% more. I am confident that if your dad took a computer and had access to generative AI and let the people in the community do it, he would see that they could become a lot healthier. I think we need to do that in rural areas. I think we need to do that in inner city areas and across time. I think that’s just going to be the way that we obtain medical information, medical expertise, and medical care. I’m sorry, Ramona. I’m just, you know, education has to be fun. Education doesn’t have to be top down. It just has to be fun and has to be accessible and you just have to give a few examples and less is more. And our education, Dr. Pearl’s education, Ramana, your education and mine, you know, we’ve studied to the top of our game, but but the world doesn’t speak that language. So, you know, we we we we have to present things, which is part of the reason why I’m here. It’s like, what can we say meat and potatoes to share this fantastic gift? I mean, to to take soybeans from a from a bucket, you know, to the futures market, it’s a big deal. And it’s fun. That’s fun. Yeah, that’s fun. Back to Roman question around all of this. You know, if you trained five years ago, you had to memorize everything because you’d have to otherwise carry a 50 pound backpack in order to be able to get all the information that you need. It’s now sitting in your pocket. It’s sitting on your phone. You have more expertise there than every library of a medical school had for their entire history. Amazing. We just have to figure out how to use that tool and it goes wherever we are whenever we want to access it. Whether we’re making rounds, whether in the office, whether we’re talking on an advice center, it doesn’t matter. We have these tools available to us. Now, I think it’s a question of translating potential into practice. My father would have been ecstatic about it. Actually, he was always vivaceious reader. He was reading all the magazines from uh England and and Russia and then and then United States you know I think that that even now even the villages in India and Africa that still works. We if we give a remote patient monitoring kit for example all the basic kits to these village doctors they can you can transform the healthare system around the globe. Do you think that that is AI with the AI help it can progress into the next level? Do you agree? This is it’s so exciting to me. I mean, I wrote in the past about how Alexa could do this kind of thing. Now, it’s happening. I wrote about taking the smart high school students in Africa. A lot of people don’t go to college just because they can’t afford it. And so, you have someone who’s graduating, they’re really, really smart. And I said, we should give them a six-month training program and give them 20 medications and a certain number of diseases. Yes, they may only be 95% accurate, but today they’re only 15% accurate. And now they have that tool and I had to figure out how I could make available to them tele medicine and a clinician in case they were stumped. Now they have that tool that is there. Yeah, I think I think it should completely change the curriculum of how we provide care. You know, one of my good friends is a surgeon from India named Debbie Shetty. He by the way was um uh the personal physician to Mother Teresa. There’s a pretty famous guy. He runs heart hospitals that uh are sitting there. And what he does is rather than putting people through a nursing program where in the first year or so you’re basically learning all the physiology and all the other parts. He basically puts them in the O immediately. He’s going to train them to do six operations. And by the time they’re done, he says, “I want a 50 suture.” And they said, “No, no, no. You want a 6o.” They know exactly what he wants better than he does. Because they have this level of expertise. We can do the same, I think, with the generative AI tools that exist. We need to completely change how we train people. Not based upon this the superior nature of their memory, but their ability to use the tools and then work with people to be able to help them like Susan’s mom. Uh to be able to have a comfort with it, to be able to know how you get help, how to put information in place, how you create those menus. These are skills that we can help people get. Not using the tool, that’s easy, but understanding how do you prompt it. How do you follow up on the questions? How do you enter into information? And the beauty of all this is that it’s doubling in power every year. Five years from now, it’s going to be 32 times more powerful. You know, a year ago when I wrote the book, you couldn’t enter data with voice. You had to type it into place. Now, you just dictated it into place sitting there. You couldn’t put an image in place. Now you can do exactly that. We’re talking about an entire different way of providing medical care. One that is going to be accessible not just to people who have trained for years in nursing school or medical school, but to all Americans to improve their health. Yes, for sure. and and as well not only the medical care but the medical education the nursing education all professional health professional education can be shortened con concise and and really move us forward on that same mission easily. Another area I want to talk with you about Dr. Susan Davis she is a champion for the you know emergency you know CPR training you know the people need around the we live in communities people need to be trained on the CPR cardiac resistation uh we don’t the schools the families at home and all the we have a na we do we have a national strategy about it and then AI can any participate in this area help people like Susan to grow in globally and then nationally and globally a a kind of a strategy there to to really save the s you know the cardiac disease sudden cardiac deaths. Yeah, I’m not aware of a strategy sitting in place right now, but I could imagine how it could happen. I mean, I think it’s probably better taught with someone like Susan there to help figure it out with a dummy that’s there that you can actually make sure the compressions are working and make sure they go through all the steps sitting in play. The time to learn CPR is not when you have to resuscitate someone. It’s in advance of that. But I also imagine that you could actually prepare people for coming to the course. So before they come to the course, they’ve already use the tool to learn about all the opportunities to figure out how to do it. So the first half hour, you’re not overwhelming them with information. You can actually now just have them start demonstrating the skill. That again is what we’re talking about. How do we accelerate the delivery of care? How do we accelerate being sure that the outcomes are as good as we need? How do we control chronic disease? Not just simply tell people what they should be doing, but actually see it in a better blood glucose or a better blood pressure or fewer visits to the ER for asthma or fewer admissions to the ICU for heart failure. That to me is the opportunity. and it didn’t exist until two years ago and it exists today. So this is the same as an emergency response plan. It’s this is the emergency response plan. It’s recognize call and act and it’s the same whether it’s your education or whether it’s your your living life, you know, in healthcare. So it’s the same thing, Ramana. So you know what we do upstream is is will always serve everybody for our outcomes. And so it’s such a facilitator regardless. But you know all of healthcare begins with education. So CPR doesn’t matter whether it’s uh you know surgery on your your face or I what was your example? It was a really cool example you gave earlier. You were taking was it part of it was an ear surgery. Yeah. Yeah. So pneumthorax. Yeah. Yeah. I mean what so everything’s applicable. The further we get upstream with this the better and it’s more applicable. But if we if we’re going to quicken and shorten and and make better our our decision-m processes in the healthcare setting, it only makes sense that we’ll do the same in the education that leads up to that because the education not only is in the medical need, the nursing need, it’s in the use of these tools. So Romano started this show by talking about the crisis in medicine and that is the equivalent of a disaster sitting in a community and to the extent that we I think this is really one of the key key questions is medicine in crisis the way that Romano has said I think so you think so Susan you obviously think so the question to the listeners and viewers is to answer that because if it is in crisis the response can’t there’s nothing we can do. The buildings are burning. We got to do something. If it’s not in crisis, well, okay, then maybe we don’t have to do as much. But if it isn’t crisis, what are we going to do? And I have yet to see an alternative. If people say, “Well, maybe the government should run it.” That’s that’s not going to solve the problem if the costs keep going up. Well, maybe um uh just doctors and nurses should do it. that’s not going to solve the problem if they can’t manage the entire totality of the stream and be able to meet the needs of patients. I don’t see an alternative. Now, will it succeed? You never know. I think so, but I can’t prove it. But my challenge to viewers and listeners is give me an alternative that you really think has a better chance because I have yet to find it. And I wrote about it in my first two books. By the way, it wasn’t that I didn’t come up with ideas around integration and capitation and technology and leadership. It’s that there was not this tool to be able to translate that into action because you’d have to have so many human beings constantly on the phone with patients. It just didn’t make any economic sense. That’s why we had this this iron triangle and now we have the tool inexpensive available 24 by7 filled with expertise. Yes, it needs to improve and few hallucinations and privacy and security I’m aware of those pieces but when I start looking for how are we going to solve it that’s the horse that I’m going to bet on. So I’m Dr. I’m a physics student of physics. So I am going to ask you somewhere in the physics one problem is a blackbox problem that is not solved. We don’t know what that is emergent property people can use this AI to the wrong purposes or we don’t even know what the outcomes are some extent. So what do you think that how do we manage that aspect of it? some people who are, you know, including Elon Musk say the doomsday thing they’re they’re proposing. What do you think about that? We need to slow down the AI progress. What do you say to those kind of naysayers? Well, the first thing that I would say is that I’m much more afraid of the people than I am the technology. And I don’t think I don’t think that I don’t think I don’t think you can regulate it for two reasons. One, you have the blackbox thing. We don’t really understand how it works. But more significant than that, when you when you’re doing a narrow AI and you have two data sets, you can look at the data. You can say, is it representative? How biased is it? There’s a lot of ways you could examine the data. The generative AI tool isn’t trained on data. It’s trained on everything on the internet, every medical textbook, every journal article, soon to be trained on every talk given at a conference. And number two, it’s all dependent upon the person entering the information. So if you look how the FDA approves medical devices right now, they expect the device to give the same answer every time you use it. And the answer is, yeah, if you use it exactly the same every time, it would give the same answer. But Susan’s going to put different information than I am. I’m going to put information than you are. Everyone is going to do that. Why? Because we want to personalize it. And now we’re in a situation where we can’t predict it. If I had to give you a way to regulate it, I would do a blind test. And by the way, it’s been done. Straford just did that. They took the generative AI tool. They took clinicians and they took clinicians with the AI tool. The AI tool, the generative AI tool was 10% better than the clinicians alone and 12% better than the clinicians using a computer. Now, why why would it it go down? Because the human brain has confirmation bias. The human brain has a lot of ways that we distort data. And so I would just do these head-to-head comparisons and when the technology was 10% better than humans, I would authorize its you use if I was an FDA commissioner, which I’m not. And you know, that’s so crazy because often there’s decision support built in and the AI will pop up the answer and then the human’s like, “Yeah, mhm.” And then they go back and you’re like, “I know.” But bing and it’s telling you again. Bing bing bing, here’s the answer. And the human’s like, “You know, I had a really bad morning and I’m not in the mood for you.” to to to the AI which is showing you the answer. So to your point then the human weighs in with all of their morning issues. Thank you sir. You know I can talk we can talk with you for hours. You’re full of wisdom. You really one of the greatest thing about your call is that crisp thinking. Your analytical thinking is you know I’m a fan of that analytical thinking. You write with a crisp uh writing. You write you with a pragmatic solutions. You do that. That’s why I’m a great fan of you. You’re truly at asset and treasure for the country, treasure for any healthcare system. Thank you for s joining. Thank you Susan. You added a excellent excellent dimension for the for the bedrics. You know, always you’re you’re the one of the best things happen for the medrics. Thank you. Welcome to the medrics. And and if people want any more information, my website’s robertpearmd.com. they can go there and there’s links to an unlimited number of uh areas and articles and ideas and I always welcome their feedback. I always learn far more and every teacher knows that your students teach you more patients teach me more than I’m able to tell tell them. So I encourage everyone to go there and uh see what they think. Let us know and I want to thank you and Susan for hosting this. uh your viewers and listeners learn a lot I’m sure from every program and this is what we need. Education is going to be the key to the future and the time has come to maximize that. Thank you for your podcast. Thank you Dr. Pearl. You have a new groupy here. Thank you Romana. Yeah. Thank you. Thank you sir. Thank you sir. Thank you for joining. Truly an honor. Thank you. We are retelling their stories through our patented designs that can be worn on scrubs, t-shirts, printed on conference bags, and endless other promotional materials. Daily Wear for Medicine is the first ever of its kind to create curiosity in medicine through visual art. We at MedBicks are celebrating unsung heroes of the past who made enormous contributions to the field of medicine through our unique campaign called Daily Wear for Medicine. You can wear it, carry it, and give it. [Music]
Artificial Intelligence offers unprecedented opportunities to address chronic diseases and improve health, but also creates uncertainties about jobs and the future. At this crossroads, we must draw on the wisdom of past sages and contemporary pioneers to guide us forward.
Dr. Robert Pearl, a true pioneer in healthcare and former CEO of Kaiser Permanente, is a leading advocate for AI’s transformative potential in medicine. Through his works, including “Mistreated” and “ChatGPT MD,” he explores how AI can revolutionize diagnostics, patient care, and operational efficiency.
