FHC #227: Dr. Rod Rohrich on social media, science and AI-powered care

0:11Dr. Robert Pearl: Hello and welcome to the Fixing Healthcare podcast. I'm one of your hosts, Jeremy Corr, also host of the popular books of medicine podcast and CEO of Executive Podcast Solutions. With me is Dr. Robert Pearl. For 18 years, Robert was the CEO of the Permanente Group, the nation's largest physician group. He is currently a Forbes contributor, a professor at both the Stanford University School of Medicine and Business, and author of the best-selling book, ChatGPT-MD, How AI-Empowered Patients and Doctors Can Take Back Control of American Medicine. If you want more information on the book and a broad range of other healthcare topics, you can visit his website, robertperlmd.com.

0:44Dr. Robert Pearl: Today's guest is Dr. Rod Borg. He's an esteemed plastic surgeon, former chair of plastic surgery at the University of Texas Southwestern Medical School, former editor-in-chief of the Journal for Plastic and Reconstructive surgery, a national leader in the application and safety of social media for medical advancement. Hi, Rod.

1:01Jeremy Corr: This is season 11, and it explores what's bothering most Americans about healthcare today. And we're bringing guests on the show who have a wide social network of connections, who are interacting with people in all phases of life. And when I think about someone in medicine who fulfills that definition, it's you. So let me begin by asking you, what are the three biggest things that you're hearing from patients, from people in general? And then we can dive into each of them and explore them in detail. Okay. Well, great, Robbie. It's always a pleasure to be on your amazing podcast. So I think the mindset today is that they want the speed of AI, They want accessibility of social media and they want the compassion of a trusted physician.

1:54So the three questions that I always get from both inferred and otherwise from my patients and on DMs is that, can I trust my doctor or, you know, will my doctor listen to me? And number two is, is this information true or is it merely popular, which is really a problem on social media? And then perhaps the most important thing is, can AI really help understand my health and truly make it better for me? And that is in a capsule about the mindset today because it's a matter of information overload, Rob.

2:33Jeremy Corr: Why don't you tell listeners one more time the three things that you're hearing that patients want? The speed of AI? Yeah, they want, well, they want the speed of AI, but also the accessibility to make sure that the information is true. And is it true or not? Because they get it out there, but many times it may not be true or verifiable. And then the third thing is, can AI really help me improve my health? Because it's out there, but can it make me better? Can these pills or this way of life truly make me age less? Well, as you know, and many of our listeners know, it warms my heart to hear that at least people are thinking about a generative AI tool.

3:21Jeremy Corr: So let me start with that, which talks about the speed of medicine, which unfortunately is not great from my perspective. We have growing weights for access. You know, we have problems with being able to, in a hospital, get care in a smooth, rapid fashion. All the things that are delays I see as being challenging. What's your view on this? Yeah, no, I totally agree. I think that AI hopefully will help transform that. that. But as you know, medicine has always been a laggard in trying to enhance or incorporate, you know, AI into medicine to make healthcare better. But accessibility and availability of physicians today continues to be a major problem because there are just no longer any true family physicians.

4:24The family physician used to be Dr. Google. Now it's Chachipiti, and that's really not a good substitute.

4:33Jeremy Corr: Let's look at some of this. How do you envision a generative addict tool speeding things up? What are you telling patients when they tell you they're worried about this and they've heard that maybe generative addict tool can be efficacious in helping to address it? You know, I tell them that, you know, it is a good asset, but, you know, will it replace a physician? Not entirely, but it can be used as a great asset to, I think, not only get potentially better care, but to help them ask the better questions. I think that's what I'm seeing more and more of is that they're being able to ask more intelligent questions of their physicians when they do get access to care, which I think is very important.

5:23You know, they're becoming more informed more rapidly.

5:28Jeremy Corr: I often tell people early on, I was on a podcast and the host at the end of the podcast, we already signed off from the taping, said to me, Dr. Pearl, you are a skier, you're a physician. My husband fell three months ago. His arm was over his head. He slid down 100 feet. And he still can't use that arm as well as the other side and still has pain. What's going on? I said to her, you know, I think I know exactly what he has, but do me a favor. You haven't used generative AI. This is what I tell everyone I come in contact with, try it. Put all the information in place and just see what it says. And then call me back and I'll tell you what I think is actually going on.

6:15Jeremy Corr: Well, she called me back five days later. She said, I put the information into a generative AI tool, and it said he probably had a torn rotator cuff. And then it showed me the anatomy. I knew what the rotator cuff was. I never would have imagined it before. Then it said he probably needs an MRI in order to see the status of the tendon, see if it's torn. And then it said you should see an orthopedic surgeon, not just because it tells that to everyone, but because it said he most likely needs the procedure done. And he called me back five days later. She said, I went to the doctor. And when he said to me, it was probably a torn rotator cuff.

6:53Jeremy Corr: I knew what it was. And I could ask more detailed questions. Exactly what you're saying, Rod. Then, you know, then he basically said, you need an MRI. I didn't have to worry. Is he just ordering this to waste my money? No, I knew exactly what he was doing. And then he said, had you waited three more months, I probably could not have attached the tendon because the muscle would have shortened. I agree with you completely, Rod. Are you seeing patients in your practice? You know, you do a lot of aesthetic surgery. Are you seeing patients in your practice who are doing this? And what are some of the positive stories you can tell listeners?

7:28Yes. I mean, that's a great positive story, Robby, because that person was very intelligent and very AI savvy. But I will tell you, many people are not. Because, you know, to get the right answer, you have to ask the right questions. And I think that's the key, I think, with AI, because AI is a learning tool, but it's also a tool that you have to use correctly. So I see that a lot. I see patients coming in much more informed about, you know, for instance, facial aesthetic surgery, which is, you know, facelifts and things like that, because they have done their homework on JAT-GPT or Claude. And they ask all the right questions, you know, that are very, very well informed.

8:12And yet I know where they got them from. And sometimes they'll even bring their chat GPT with them. So it's kind of fun. But it actually, to me, it streamlines the consultation a lot. And also it gets us to focus exactly on what they want and what they need. And many times they'll bring photos of what, you know, Claude will have said. Sometimes it's not correct. But remember, you know, all of these AI products, they are generated from intelligence that comes from humans. And sometimes it's not completely correct. You know, in your friend's case, it was absolutely correct because it was obviously, you know, it was a pretty easy diagnosis for even you and I as plastic surgeons to make that.

9:01But sometimes, you know, you need that artistic input as well, especially in plastic surgery.

9:08Jeremy Corr: There's no question, Rod. I see it the same way you do. These tools do not replace humans. They augment them. They provide expertise when the doctor's not there. I mean, if clinicians were available 24 by 7, I'm not sure we'd be using a generative AI tool. But if you can't reach them once the office closes or you can't reach them between visits and problems and questions come up, what should you do? There's a mindset in medicine, Rod, that I see that we think that doing nothing is better than something that's imperfect. Now, yes, if it's imperfect 50 % of the time, you don't want to be using it.

9:47Jeremy Corr: But if it's 98%, 99 % accurate, that's where I think we should be filling the patient in. So let me ask you, you know, you do a facelift and you send the patient home. You keep them at night, often in your recovery area, and they're doing well. but you have a lot of concerns. Maybe you have concerns about the incision line. Maybe you have some concerns about damage to the ear if they are rubbing it the wrong way. You may have concerns. You may have concerns about a hematoma. Could you envision a generative AI tool looking at your patient and notifying you when they thought there was a problem that existed?

10:28Absolutely. Absolutely. Absolutely. They'll do that. I'll give you an example of a patient last week that came to me that I had done a laser on her and that's a superficial resurfacing of her skin. And she would almost heal. And then I see her like four days later and she's got this incredible, what looked like to me like an infection. And so what I did, and I thought I knew what it was, I put it into Claude. And immediately, it asked the questions, does this patient have fake nails? It looks like a candidate infection, a yeast infection. And it was right on. And then it even listed the things that how to treat it.

11:16And so that was just from a photo that the patient had actually sent me from her home. So I was very impressed by that. And so, yeah, we can use the power of AI. And I totally agree with you that it won't replace us. It will augment our ability to react and actually treat patients, you know, without seeing them or at least treating them until we can see them. So this is the one time I will disagree with Elon Musk, who's a genius, because I don't think AI will replace surgeons in two years. Do you?

11:51Jeremy Corr: No, it will not replace them. In maybe 10 years, I think it will have its first impact in robotics. And again, I think that there will be a clinician involved in overseeing it. Right. But it will be able to learn so much faster. The limiting factor I'm increasingly telling people is that it's not going to be a technological issue. It's going to be a human issue. It's going to be the fact that we're not willing to train the tools to do the things that it could do, because what we know is that there are some surgeons better than other surgeons. Now, when you're looking at an operative robot, it's the perfect learning tool because there's an input and there's a response, and that is what a technology can learn.

12:38Jeremy Corr: but there are going to be times when the anatomy is not normal. There are going to be times when there are adhesions in the abdomen. There are going to be situations that are different. And yes, the technology could figure it out if we gave it all the information. But as you say, that's the challenge that's there. And that's why we'll have a human involved, but it'll make it so much, I think more reliable and safe to be able to accomplish that. You know, you talk about family medicine. Primary care is under siege. The amount of things they're being asked to do, it's not possible. I've run a study that said to do everything the primary care is supposed to do would take the average clinician 23 hours a day.

13:26Jeremy Corr: It's just not a realistic way to go. So we're asking the wrong questions, not that it's going to replace the primary care doc. What if it did 20 to 30 % of what the clinician did and they could do it with the same accuracy and success that the physician would have? Now the individual has the time to focus on more complex people, the things that can't be done technologically. I think we're going about it all wrong. wrong. Yeah, let me ask you a question. You know, you see, you see, you do a lot of facelifts, you know, of the patients who come in wanting facelifts, already aesthetic rejuvenation of the face, that's a better way to say it, not the technical procedure, who come to your office right now.

14:09Jeremy Corr: Can you give an example of a time that recently you've been happy that they came in with the information, and maybe a time that you were unhappy that they came in, whether it came off of technology or social media? You know, I mean, most of the time, Robbie, I welcome that because to me, they're well-informed and they really have, that means they've done at least part of their homework and being well-informed. And I think that's been the positive side of both social media and AI is that people at least know at least what to ask for. They may not know exactly what it is or whether it's germane to them, but then you can get them into the ballpark.

14:53And it actually, I think it accelerates the consultation so that they can get right to what they need, which is where we have a role. You know, we are there, as you mentioned, especially in plastic surgery to tell them exactly what they need, whether they need, you know, a deep plane phase of the regular phase or a fat augmentation or whatever, because I think that's the missing link with AI. I think AI really, if used wisely, will help physicians be better and outperform those physicians that do not. I think so it's a time for all physicians and we need to teach it in medical school. And they're not doing that.

15:35My daughter is in a second year medical school. But I think they're starting to incorporate AI into their physical exam, which is, of course, way overdue.

15:46Jeremy Corr: No question about it that when your daughter finishes her training, she has medical school, then residency, maybe fellowship, depending upon what field she pursues, the practice of medicine will be completely different. 100%. There's no question that clinicians will be using these tools all the time, not just in their office. I think they're going to be doing what's called vibe coding and creating applications for individual patients or maybe for groups of patients. Every face-to-face patient, you'll give the same tool that now will help them in the postoperative period, whether it's to be putting the ice into place, whether it's to be figuring out what to do with the dressings.

16:26Jeremy Corr: All the things that we tell patients, we give them a 10-page handout, and they never read, and they often forget. Exactly. You know, like I immediately, for my patient that had the yeast infection, I immediately had AI create a diagram for her for the things to not do, you know, to remove her nail, her fake nails, and to cleanse it and to not ever touch her face. I mean, just gave her like the six points that they should do. And you can do that right in the office and give it to them. It's pretty amazing how it's facilitated, you know, post-op care. And it's made it so much more efficient, at least, you know, in my practice.

17:08And I actually, I love it. I use it, I mean, so many times every day. And, of course, I use it, you know, in my podcast. And I use it in everything I do, you know, on social media.

17:21Jeremy Corr: I mean, just imagine now, and again, we're focusing on plastic surgery because we both know it well, but everything we're saying is applicable in every single specialty out there right now. But imagine if plastic surgeons across the nation all reported the challenges. I'm not sure anyone else has thought about fake nails before you because it just so happened it was your patient. Can we give them a checklist of 100 items? Yeah. But it was so much better to engage in a conversation with a generative AI tool of which one of the questions is just the tool could look at the person and see whether they were likely to have fake nails.

18:00Jeremy Corr: This is the kind of thinking we need to evolve so that we can respond with the first problem. We don't have to wait till there are 100 of them. Someone researches it and three years later, now it becomes part of the routine of our practice. Right. And it actually, it helps, I think it helps young physicians and young plastic surgeons to actually immediately know what to do. Because I will tell you, I presented it to my fellows and really, none of them really knew what that was. But because I said the other key element is that the patient said it was red and it really itched a lot, which is sine qua known for a candida infection.

18:39See, these are the things that are really, you know, AI enhanced that immediately go right to the diagnosis. I mean, it's good for medicine. It's good for healthcare. You can treat them immediately. I mean, I think, you know, technology, you know, should be used to help restore kind of, you know, the immediate cure and give, you know, restore humanity to healthcare to say, hey, you know, my doctor, you know, he or she knows this. and she or he or she uses technology to make me better. I think they embrace that.

19:14Jeremy Corr: Let's move on to your second point. And I love this point about finding out the truth. You've been one of the leaders in medicine and understanding both the positive and negative problems of social media. Now we have built upon that, technological tools and information in the person's home. Can you start breaking apart and explaining to listeners and viewers your thoughts on both the positive aspects of these components of modern current care versus the risks that they are posing and how we can have patients sort out what's true and what's not true? Well, that's a great question, Robbie, because it's just information overload with social media.

20:01People are bombarded every day with this is the, you know, the cure for longevity. This will make you sleep better, be stronger, be better. And, you know, and it's all driven by these influencers of which 90 % of them have no health care background. It's just amazing. That's one of the main reasons I'm in social media, because it's kind of like it's truth versus, you know, fact versus reality versus complete bogus. And I think the goal is for us to sort it out. And sometimes it's really hard, Robbie. And so I don't know how the average person can sort out what's true or what's not. I think that many times when people are bombarded with, you know, ads and, of course, selling things, that's easy.

20:55But sometimes they're disguised as, you know, these are the facts and yet there's no data. So I would always encourage the patients to say, OK, take that and put it into your favorite AI engine and ask what is the real data. And that's been helpful. But it is it's a mess out there, Robbie. I will tell you, I don't have the answer, but I think the more people we have like you and I that are saying, hey, these are the things you need to look for, you know, a truly, you know, a true physician or an expert that can give you science or evidence based data to support what he or she says. And then also, you know, when things are too good to be true, you know, they're not.

21:40Yeah.

21:42Jeremy Corr: Rod, I love this idea. I mean, I had never thought about it before. I've always, you know, people worry about generative AI being wrong. The idea of using it as a screen to decide whether the social media is right or wrong is brilliant. Wow. Have you tried it? Have you done it? Oh, yeah. I've tried it. I've tried it a lot. I mean, like you take some of these. There's a crazy influencer that just always talks about how you can prevent Alzheimer's or you can prevent aging. And you put it into regenerative AI and they go, hmm, there's only two articles on this and none of them are significantly peer reviewed and things like that.

22:26So it's very it's very useful. And also the other thing is it helps you educate AI and saying, well, what about this? I mean, I'll just give you an example about how, you know, the recent data that NAD, niacinamide, is actually going to help enhance your longevity. Well, there's some animal data, but not much human data to support that. And yet, you know, there's a lot of people that talk about it and peddle their different types of medications and supplements. So you have to be very careful. So you have to use AI to support whether the AI you're hearing or seeing is true or not. It's kind of a double-edged sword.

23:14Jeremy Corr: Interesting. In yesterday's Medicine of the Truth episode that Jeremy and I recorded, we talked about the peptides that are being considered and reviewed all the data, how that conclusion was reached. And it's the same exact same experience, which doesn't mean it's bad. It doesn't even mean it doesn't work. Just means we don't yet have the science. And we came up with a pretty radical approach to pairing a combination of approval with detailed research. And I would hope that that would be a broader application that the FDA would lead, where if you want to take something that's not fully proven, we can help provide it safely.

23:59Jeremy Corr: But we need to make sure you're part of a rigorous scientific study, because that's the only way we're going to be able to help all the other people. And moving from a single individual to 300 million people in the United States, that to me is what the future can look like. Oh, I agree. Sorry, Roger. No, no, I totally agree. And the peptides are a classic example. You know, they approved six of the seven. And yet of those six, none of those six had any significant clinical trials. They just approved it out of the committee so that the FDA can look at it to approve it from research to compounding pharmacies.

24:38But of course the influencers went crazy and said they've approved peptides. That's not true. They haven't even approved them, you know, to go from research to compounding, and the FDA can't approve any of those until clinical trials are done, you know, obviously to the general public. So it's a lot of misconception out there because a lot of the peptide companies say, well, now it's going to be approved. It's not even close to being approved, truly, by the FDA. They're going to just approve it so that it can be now used in compound pharmacies, which are probably a little more legitimate than a lot of the research pharmacies.

25:14Jeremy Corr: Let me ask you, Raj, you said something earlier that really caught my attention, which is the complexity of using a generative AI tool, because if people want to use that tool to be able to evaluate social media recommendations, they need to know how to use it, which involves both the prompting and the follow-up questions. How would you recommend that we educate our nation, educate patients about how to use a generative AI tool, as you pointed out. It's not straightforward, but it's very learnable. And how do we maximize people's skill with it as we might with any other new technology? Yeah. I think there's some key phrases that each of them like to use.

25:58And some of them you always have to paraphrase saying like, please provide evidence-based science or factual data with support reference material, then the regenerative AI, like Claude, will really go deep diving into that. And they'll not give you the usual answer of, oh, this looks great, because it'll go into the NIH, it'll go into the peer-reviewed literature and actually look at it. But you have to prompt it to say that. You have to prompt it to talk about evidence-based, peer-reviewed research. If you don't do that, you're not going to get the result that I think is valid. So you're right.

26:43You have to educate your engine to make sure you actually get meaningful data. Because sometimes, just because it's ChatGPT or Claude, you won't get meaningful data unless you provide meaningful input.

27:02Jeremy Corr: You mentioned your daughter who's in medical school now, and I'm going to guess that she knows more about current technology, generative AI technology, than 98 % of her professors. 100%. And they all use it. All the medical students use it. You know, she was saying that their clinical diagnosis and exam now is so much better because they're told what the patient has and they can actually, they're encouraged to use AI. And they said that it's been a while for the professors. They say, wow, that's pretty amazing. So that's good. I mean, you know, it would have taken a lot of, you know, more work, but, you know, today we live in a very fast paced society and AI hopefully will just help us, you know, not only learn better, but faster, but more importantly, give better health care.

28:02Jeremy Corr: Let me ask a question I'm often posed to me, which is how will students ever learn to synthesize the information if it's being done by a tool that is so good at at accomplishing it. You know, you hear a heart murmur and maybe you see binding under the nails and the ability to be able to figure out that what's going on with this patient is infectious. That takes a certain amount of skill and experience to put all the pieces together. And that's why training lasts so long. If you can just put it into place and have the answer given, or using the example of that rash, if the rash, if the tool tells you the answer and you haven't had to think about that.

28:45Jeremy Corr: People are concerned that the intellectual prowess of our future generation of doctors will decline. How do you see it? That is very valid, Robbie, because like I said, when I presented it to my fellows, none of them went, they went, hmm. So that's a very valid question. And honestly, I don't have a good answer for that because, you know, you and I learned all this the hard way because good experience comes from bad experiences or from learning what worked and what didn't work. So I think our future generation is kind of fast tracking into just getting the answer before knowing what caused it.

29:28I don't know. I don't know the answer to that. I think that is one of the, I think the downsides of AI for future generations, because let's just say what happens if you're in a desert island with somebody and you're having to treat somebody and you don't have AI. You and I would probably do fine because we've been through it, but I don't know about people that use AI as their tool to be better. I think it's a challenge.

29:58Jeremy Corr: My view is that we often in the discussions and debates today want to be on an either or. We have to understand it's both. You know, I think back to when the MRI CT started to become available and physical exam dropped in importance. But there's still something about that physical exam that adds value. And if it's not built into the curriculum, you never learn. It's going to even get more so things like listening to the heart. General AI is going to do it. It's already doing it better than clinicians. It's finding problems far earlier in the process, but that doesn't mean that clinicians should not be listening to the heart.

30:38Jeremy Corr: And so I think it's a cultural piece to understand both. You need to get skilled at learning one set of opportunities, even though most of the time you can rely on this other one. But at some point you're going to encounter a patient. It's not that the general AI tool couldn't do it, but you haven't asked the two, and yet it's sitting right in front of you. That's what I think. I think we need a culture. And right now, the culture of medicine is either positive or negative. And we haven't seen people embracing the fact that it's going to be both.

31:16You're right. You're absolutely right. I mean, I think there are no shortcuts. There are absolutely no shortcuts. And I will tell you that, you know, that's a challenge because, you know, my, for example, my wife is taking care of her 99-year-old father. And she takes him to a cardiologist and he doesn't listen to the heart. That, of course, makes her very, you know, upset because they're saying, hey, you know, they've looked at, you know, the off-fascade and everything. And so I think they're teaching it. They're definitely teaching it because, you know, my daughter, Rachel, they're teaching how to do the physical exam and all those things.

32:01But in practice, I think physicians are so busy, they say, hey, you know what? I have this tools. I have the CAT scan. I have the MRI. I don't need it.

32:13Jeremy Corr: So let's move on to your third point, which I think is really important, which is this notion of compassion of the physician, which let me start by offering my view that it's not a question that doctors are not compassionate. It's that they don't have the time to communicate it, the time to demonstrate it. And as a consequence of that, patients don't feel that they're as being as empathetic. What are you hearing from patients and what are your thoughts about what we should do about it? I totally agree. I think that patients are really, at least, you know, pre the use of AI, patients would always tell me.

32:57And, you know, when I go see my physician once a year, I mean, he's still old school. So believe it or not, he still listens to my heart and all those things. But I will tell you, that's probably not normal because most of the time when people go to see their physician, they're at the computer screen and they're just talking and putting in data because you're right, they're so busy. But maybe with the advance of AI, you know, you and I aren't primary care physicians, but maybe if you use AI tools correctly, you can streamline all that and you can voice record. There are now places where, you know, a lot of my friends voice record.

33:33They'll use many different devices. They'll voice record the whole interaction. And then at the end, they'll walk out and there'll be a note that has everything they said in there, because in the past, you'd have to spend every afternoon and evening doing that. So I think AI can help you become more humane and more, you know, I think help restore humanity to health care if it's used correctly. But I think that's one example, because right now, as you know, physicians come home, they say, I've got to I've got to do 40 charts tonight. oh and that's burdensome but if you can do it instantaneously and just minimal editing that can give you more humanity to say hey you know what else is going on in your life or maybe do

34:22Jeremy Corr: a physical exam the ambient ai tools definitely are revolutionizing health care they're allowing us to look at the patient not at the computer screen and several years ago i predicted they become ubiquitous and we're rapidly getting close to that. You're right. Right now in the inpatient side, but it's clearly the right application of the generative AI tool that's there. I want to maybe get into an area that people don't associate with plastic surgery right now, which is whether empathy is a skill you can learn to demonstrate, whether it's a skill that's dependent solely upon having feelings, and whether it makes a difference, whether you demonstrate it out of just pure emotion or whether you've learned to express it.

35:15Jeremy Corr: How do you see empathy in medicine? You know, empathy, you know, it's a complex topic because some of it comes from you as being an empathetic and compassionate individual, and that comes from your innate, you know, from who and what you are and how you grew up to what you've become today. So can AI help you become more empathetic? I think it can, but you have to have that basic empathy at heart. And I would like to think that most physicians have a lot of empathy. Some perhaps don't. But you can learn to become more empathetic. Certainly, I'll give you an example. In your written, in the way you write and respect, respond to patients.

36:16I will tell you, I'll give you an example. You know, I get texts from my patients all the time because I give them my cell number. And sometimes they continue to ask the same questions. over and over and over again. And of course, your response probably would not be very kind if you just wrote a text. So what I've done recently, if I had used ChatGPT to provide a more compassionate, empathetic response that I've edited, and that's worked. Because sometimes you need to kind of step back and say, hey, maybe I can say this in a nicer, more empathetic way. And that's where AI has helped me, at least, you know, in the written word.

36:58But to learn it, you know, in a day-to-day basis, that still comes from you.

37:04Jeremy Corr: I love what you just said, Rod, because what you're really saying, if I can paraphrase it, is that whether you feel empathetic or not ties back your parents, your upbringing, and a lot of aspects of yourself. But whether you communicate it, and you can communicate because you really feel it, which is the situation with most clinicians, or you feel it because you just want it, you learn how to do it, that is a different circumstance. You know, they did an interesting study at the University of Arizona, where they took just exactly what you said. They took text messages that came to clinicians and they had the responses that the clinicians did.

37:48Jeremy Corr: And they gave the same text messages to a generative AI tool. And then they had individuals who didn't know whether the message or response was created by the clinician or the generative AI tool to evaluate them, both patients and clinicians. and interestingly enough the responses of generative ai tool were more comprehensive and actually more accurate than some of the clinician responses okay not probably because the doctor didn't know but because they didn't quite know how to explain it but they found the responses of the generative ai tool nine times more empathetic than the humans nine times nine times more empathetic Yep.

38:30Wow. Yeah. Well, I believe that because, you know, like I said, I've used it sometimes, but it really does. You know what? Because it helps you kind of take the emotion away. I like that. That's pretty amazing. That's a great. And I think those are all things that we should learn to do and use, don't you think? I mean, but so many people don't do that.

38:55Jeremy Corr: i keep going back to your daughter and she's gonna be a great example as i follow her progress to try to learn about the education people are getting right now i work with residents but medical students are even earlier in the process um yeah i don't know that she gets a lecture on how to be empathetic or how to communicate empathy no you know i'm not even sure she She gets a lecture on you walk into a patient's room. What's the first thing you say? Yeah, no, they don't. In fact, you know, they don't go to lectures anymore. So crazy. They see it all online or in videos. And it's just it's a different world.

39:37Jeremy Corr: You know, well, let me then back this thing up. You know, you've been president of most medical societies and you're engaged, not just in plastic surgery, but in multiple medical areas out there. Clearly, we are at a problem today, whether it's around the affordability, the lack of empathy, the burnout, the overworked clinicians. The list could go on forever. If you had to give a prescription, a plan to sort of help people. If the dean called you and said, you know, you're a leader in the field, give us some ideas about what we could do different to prepare the next generation or the head of the hospital said, give me some ideas about how you might restructure the delivery of healthcare at this time.

40:26Jeremy Corr: Anything pop to the top of your mind? Oh, boy. Restructuring how, I think, maybe getting the patient's perspective rather than telling them what they need is asking them what they want. And what they want out of their healthcare or their hospital or their stay. And that may be very illuminating because we're so mechanized into delivering healthcare of what we think is right, but maybe we should listen to the patient. I know that sounds kind of very elementary, but honestly, I think we've lost that in medicine today. And people feel that because they feel dehumanized. And I really think that AI can actually help bring that back.

41:16I really do, if used correctly. And I think that would change their perspective on U.S. healthcare, because U.S. healthcare today is challenging, like you always talk about. You know, it's access, it's delivery. And, you know, in the end, it's about what the patient feels and wants and really their true outcome. Are we truly helping them.

41:43Jeremy Corr: Once again, Rod, I love what you're saying, because I think it would be eye-opening. You know, I consult to a lot of entrepreneurial companies, and almost always the problem that I see is they have a great technology, and they say, how can I use this technology? Rather than saying, what's the problem that the user is going to experience? In medicine, the user is the patient. And how do I now use the tools that I have to address it. You know, when I talk to people about a hospital experience, it's chaotic. They can't figure anything out. They can't, they don't know what's happening. They, you know, various, most people now have multiple specialists.

42:27Jeremy Corr: They come in, they tell them something, but they can't synthesize it. And I could imagine based on what you just said, that they could come up with a tool, a general AI tool that would take all this information, as you said, listening into the ambient conversations and synthesize all that for the patient. You're absolutely right, because we should use technology to restore humanity in healthcare today. And I think, you know, people choose between technology and humanity. They should use it to restore humanity. I agree. Wouldn't it be great at the end of every day in a patient's healthcare, whether they're an inpatient or not to say, this is what we've done, this is where we're at, and this is where you're going.

43:11People would actually really be astounded by what they, you know, are learning because sometimes, you know, people aren't on the right, you know, path is what their physicians think they should be on. It's really, it's pretty crazy. It's a lot I like that in plastic surgery. I always try and reflect, you know, what does the patient want and what am I doing? I mean, sometimes they're divergent and they shouldn't be.

43:38Jeremy Corr: You know, the other piece that I think they would say is they want personalized care, which is a combination of superior clinical care along with the empathy that's there. And we have just gotten into the habit. And I don't want to say that I'm any better or you're any better than all the criticism I'm making right now. Now, you know, we operate on someone and we see them back a certain number of days, whatever your schedule is. Everyone's got their own schedule. We never say, you know, maybe there's a patient I should see earlier. And maybe there's one who's doing so well. Why am I bringing him into the office?

44:12Jeremy Corr: Now, obviously, if you take sutures out or a particular intervention, then it has to be on a particular day sitting in play. We don't say that. You know, nurses, they rotate from patient one to two to three to four to five and back to one. Maybe there's some people, and this is going to require doctors write different orders, just let them sleep the whole night. And other people, you should see them every hour, not two, four hours, not every four hours, because there's something coming off their monitor, something about how they're lying in bed, something about them that says that this patient has a higher chance, a higher risk of a major complication.

44:47Jeremy Corr: and we aren't asking, medicine's not asking from my perspective, how do we use this tool to solve the problem that they have? They don't want to be woken up if they're doing well, but they want to be getting a lot more care if there are any issues out there. How do we change the culture of medicine? How do we change it amongst your peers, the people you interact with who are so used to doing it the same way that they never start to ask these questions? Yeah, that's a great question, because medicine is always lagged behind technology. And sometimes in the past, that's been good. But today, we need to be on the cusp of technology, because I think it's going to help healthcare.

45:34And I totally agree with your analogy of saying, hey, you know, there are a lot of patients that, you know, don't need to be woken up every hour or whatever. And, you know, there are new technologies now where people can wear a vest that monitor their all their vitals. So if they're normal, leave them alone. But if they have some problems with their heart, the vest will tell the nurses and they'll be there. So it's just it's a totally different world we're living in today. And we need to use it to, again, make the patient better without disrupting them.

46:11Jeremy Corr: Let me add, we don't have to wait till the nurse gets there. Imagine if the generative AI tool analyzes all that and tells the nurse when there's a problem. So that rather than waiting till the nurse does the usual rounds, that individual can access sooner. Rod, we can keep going on with this, but let me turn it over to Jeremy to ask a question to the two of us on behalf of the patient.

46:36Dr. Robert Pearl: In the past, when people had concerning symptoms, they would go to WebMD or Google their symptoms and often come to the conclusion that something is wrong with them that's catastrophic. That can still happen today with patients using AI. Patients might get a notification that they have some of their lab results back on my chart late at night after the doctor's office is closed. Maybe they didn't get all of the results yet, but they put this into AI and ask what it means, and they might either over-describe the severity of a symptom they're having, or they might be painting an impartial or inaccurate picture for the AI.

47:09Dr. Robert Pearl: Someone also might be finding out on my chart late at night that there's a good chance they have cancer or something serious like that when their doctor's office is closed. We're at a point with technology where health information such as labs, MRI results, CT results go directly to the patient instead of going through the lens of the provider who's giving them that information. This can cause a lot of anxiety for patients. And what are your thoughts around how we can humanize or maybe provide some empathy in these periods of heightened anxiety? I see that a lot. I think the answer should be that we really need to teach AI that, you know, partial answers give you, I mean, basically, if partial data gives you partial answers, which may not be totally correct.

48:01And AI, you know, smart AI will usually tell you that if you teach it, saying, okay, my one lab value is off, you know, I've got liver disease as well. It may just be one of the five liver enzymes and the other ones are normal. So, I mean, I think we need to educate AI to not alarm people because you're right. They're going to go and say, what could this be? I must say AI is much more accurate than Google. Google used to just frighten everybody all the time. And now ChatGPT is a lot more accurate if you ask the right questions. Robbie, what do you think?

48:41Jeremy Corr: I agree with you, but I increasingly feel that we're trying to regulate the technology rather than to support the people. So to me, one of the things we want to do is have all the information into the application, because if you have all that lab data and the generator of AI tool can see it, then the information getting back will be very accurate. If all you give it is the one abnormal lab result, it might not be. So how do we have that going into place? How do we make the applications be secure and make them be private so that the companies that are collecting this information are prohibited from being able to provide it to insurers or to employers or to someone else who might abuse it?

49:30Jeremy Corr: That's why I'm so glad that you've been on this show, Rod. I just think we need to start to rethink this technology as being different than everywhere else. Ask ourselves, how can we use it optimally? Not be afraid of it, but look for the opportunities to improve it and then test it against human clinicians. And when we do that and the technology proves to be as good or better, not see it as a threat, but see it as an opportunity to take some of the load off of doctors today and allow them to do the things that only people can do, the very complex problems, the one that require holding someone's hand and talking to them.

50:13Jeremy Corr: I'll tell you, Rod, you know, I use generative AI a lot, but if I had cancer, I want to talk to a human being as well. Not that I don't think generative AI can do it equally well. In fact, probably in some ways better because it can do a lot more analysis of the genomics. But I want that compassion that's in the human being. And if we don't make sure that stays in medicine today, I think that we will be the losers as a consequence of what should have been the golden age of American medicine. Rod, let me thank you for being on the show today. You've opened my eyes, a lot of people's eyes to the issues that exist, to opportunities that we can embrace.

50:50Jeremy Corr: You've provided wonderful, both warnings about difficulties that may come up and you've been very willing to also break through some of the assumptions and stereotypes that the culture of medicine has built over time. You're a leader in the specialty. You're a leader in medicine. I can't wait to see what you accomplish next. Keep up the great work and make sure all your followers understand what's going to be possible and help them to get better health in the future. Thank you so much, Robbie. It's always an honor and a pleasure to be with you. All right. Thank you so much and have a great, great weekend.

51:28Dr. Robert Pearl: We hope you enjoyed this podcast. We'll tell your friends and colleagues about it. If you want more information on physician culture, you can find it at robertperlmd.com. Please subscribe to Fixing Healthcare on Apple Podcasts, Spotify, or your favorite podcast app. If you like the show, please rate it five stars and leave a review. Visit our website at fixinghealthcarepodcast.com. Follow us on LinkedIn, Facebook, and Twitter at Fixing HD Podcast. Thank you for listening and have a great day.