Will ChatGPT be the new MD?

Brian Urban:

Yes, this is the Healthcare Rethink podcast. I'm your host, Brian Urban. And today joining our show we have a multi-book author, physician by training, professor by nature and healthcare futurist by name, Dr. Robert Pearl is joining our show to discuss his new book here chat, ChatGPT MD. Dr. Pearl, welcome to our show.

Dr. Robert Pear:

Thank you Brian. And I want listeners and viewers to know that all profits from the book go to Doctors Without Borders a really wonderful global charity.

Brian Urban:

That is wonderful. I was going to ask you, but you took that from me. That's okay. That is fantastic. So with every guest, we love to have our audience become familiar with your background and who you are. So let's go back in the day here. Before your years as an executive at Kaiser Permanente, even before your residency at Stanford and your medical doctorate at Yale School of Medicine, how did you find yourself into this path? Who is Robert Pearl? Give us the story here.

Dr. Robert Pear:

My view about life is that opportunities happen. They happen somewhat in a chance kind of way. I went to college to become a university professor and when my mentor, a great, excellent researcher, failed to get tenure, not because of his academic work, but because of his political views, I decided, I was 17 at the time, pretty naive, that I would go into something that didn't have politics and that would be medicine. Of course, in hindsight, that was uncertainty. There's nothing that has almost more politics than medicine today. And that's when I decided to change my major, go to medical school, as you said, at Yale. And I came to Stanford to become a heart surgeon. And again, serendipity entered into the format. I was doing a rotation in plastic reconstructive surgery, went on a mission trip to Mexico and fell in love with cleft lip repair. The opportunity to be able to change a child's destiny. In a matter of an hour and a half, take a child who would've difficulty in education, in the work environment, finding a relationship, and all of a sudden they can have a more normal life.

And that's where I started as a plastic and reconstructive surgeon. I was in my last year of residency and again, serendipity comes along. The plastic surgeon at Kaiser Santa Clara dies in a plane crash, tragic crash, and they ask me to come and provide medical expertise to the patients at the facility and across Northern California, talking about three to four million people at the time. And I get there, I love the job. My first year I get asked to become the chairman of the OR committee. I think, "Wow, they must have read my bio. They must have known how well I trained in academic facilities. They must know how skilled I am." No, no, I was the lowest guy and the only one foolish enough to accept the job.

But it took me into my, I'll call it my second career. And it evolved over the course of about 20 years of a leader. I was able to go to the Stanford Graduate School of Business for executive training. I was able to take on a higher and higher accountability, became the CEO in Kaiser Permanente, the Permanente half of Kaiser Permanente, the delivery side. And able to help the organization become the national leader, number one in quality according to the National Committee for Quality Assurance, a leader in access and service in physician and patient satisfaction. And after becoming the CEO, I'll say towards the end of that Kaiser Permanente journey, I [inaudible 00:04:36] to my third career, which is where I am right now, professor at Stanford Medical School and Business School. I have my own podcast, Fixing Healthcare. I write for Forbes. As you say, published three books with ChatGPT MD being the most recent one.

Brian Urban:

Wow. That is quite a journey. And thank you for taking us to that level of detail, Dr. Pearl. And I find it fascinating throughout your career, you had these different peaks, you being on your third peak now really as a voice of leadership and influence and change, dare I say, as well into hopefully the academic medical society that is quite large and the backbone of the US in many ways and also into business as well. Now the business side of healthcare, that is really the operating model of how we are today and how we can maybe get better.

So I'm so curious, opening up your recent book here, you take us into modern medical miracles and as you're describing a story between two patients having different surgical procedures, you weave in a physician by the name of Dr. Norman Shumway. And you started to give some great comments, some great love to Norman. And I'm curious, you made a decision, it seemed like, to take your residency to Stanford, to work with Dr. Shumway. Can you tell us a little bit about the relationship that you've had with Norman and what it's meant to your career and maybe taking off in terms of medicine at that time?

Dr. Robert Pear:

Dr. Shumway was the first cardiovascular surgeon to do a heart transplant. I followed the first heart transplant in the world done in South Africa. And as I said, I went into medicine because I wanted to avoid politics. And if you think about it, what could be less political than life and death? And the idea of heart surgery seemed to me the epitome of that opportunity. And that's why I went to Stanford and Dr. Shumway was a pioneer, dedicated physician, incredibly skilled clinician. But as I said, I shifted from that into plastic surgery, plastic and reconstructive surgery, focusing on children with cleft lip and cleft palate just came back actually recently from a mission trip to the Philippines experiencing once again the joy that comes from changing people's lives. So that was really more the journey. Dr. Shumway remained a hero, trained, excellent, remarkable people. His daughter actually now is a cardiovascular surgeon in Minnesota. So it's a legacy that goes on and that's often the case.

I think in medicine when you are at the cutting edge, what you do is you leave a legacy, you blaze the trail for other people to follow. And that's my hope in ChatGPT MD that I've been able to see the ways that this can be a miracle. Now why do I say it's going to be a miracle? Well, it has the possibility of saving hundreds of thousands of lives every year. This was made viewers say, wait a second, how could a generative AI tool save that many lives every year? To understand, you've got to go back to the fact that 400,000 Americans die every year from misdiagnosis. A quarter of a million people die prematurely from preventable medical error. Hundreds of thousands of people die from complications of what's called chronic disease.

We're talking about hypertension, the leading cause of stroke. We're talking about diabetes, the leading cause of kidney failure, major contributor cardiovascular disease. We're talking about the types of problems that first of all are preventable and number two can be better managed. When you look at the United States, you say, well, how well do we do hypertension controlling 55 to 60% of the time we should be over 90 diabetes, 30% of the time we should be over 90%. And if we did that, we'd avoid the complications. What are the complications? Heart attacks, strokes, cancer, kidney failure, peripheral amputations of lower limbs. We're talking about very significant problems. So I ask you what would happen if, as the CDC has said, better management of the problems, better prevention would reduce the chances and the incidence by 30% of people having these problems. We had 30% fewer heart attacks, strokes, kidney failure, cancers, peripheral limb amputation.

What we'd see is our lives would be healthier. There's a twelve-year gap actually between people losing their health and going on to die in the United States, we actually have a mortality rate that's higher than pure nations. We can bring that back down, but why not focus almost 12 years, shrink them down, keep healthier lives, and what would happen to the cost of care if all of a sudden we had 30% fewer heart attacks, strokes, cancers, kidney failures? That would be a miracle. And that miracle is here now in generative AI, ChatGPT, and the other tools. We're talking about Gemini from Google, we're talking about Claude from Anthropic. These generative AI tools have changed and revolutionized medicine. And I predict in the long run, it'll be even more significant than the advances that were done in heart transplantation, as you say, I opened the book ChatGPT with.

Brian Urban:

Yeah, and I love where you're going with that and the story that you take us through from the modern miracles of medicine into the miracle of artificial intelligence as OpenAI released their ChatGPT and now GPT-IV as we fast-forward a little bit in the book, and it's amazing because when we think about the economic shift opportunity, right-sizing healthcare and reducing that gap in improving life expectancy for a variety of populations in the US and also socioeconomic statuses, it's where we need to go and embracing it is what we have to do. And I think a lot of people, Dr. Pearl, see generative AI and large language models doing simple things right now. Doing more efficient retrievals of electronic medical records, being able to maybe diagnose pathology results a little bit faster or more accurately. And then doing a lot of administrative burden, I'd say. Say alleviation for physicians.

And you go into that in-depth here as how physicians in modern United States are overburdened with administrative tasks with a rigid system that a lot of large private health plans and their lobbying partners have had inside our federal government to keep this existing model going. And it's clearly hitting a peak here, inflection that we're not going to make it through. So is ChatGPT for you now trying to super speed past the admin burden relief and go into the diagnostics, to go into disease prevention, to reduce mortality rates? Is that where you see the real impact happening and how long is it going to take for us to maybe see that?

Dr. Robert Pear:

Let me remind listeners and viewers that I have no relationship with any of these companies. I use ChatGPT only because it flowed nicely into ChatGPT MD. It's really the generative AI technology that's been remarkable. When I wrote my first book, Mistreated, I talked about the problems in the system of healthcare. I pointed out it was 19th century cottage industry with physicians uncoordinated with each other using technology that was left over from the last century. Although really since the fax machine is the most common way that doctors exchange vital information, it was left over from the previous century, 1834 invention that they were paid on a piecemeal basis called fee-for-service. The more you do, the more you get paid, whether there's any good or not doesn't matter. And there was not a effective leadership structure. And I expected that change would immediately happen and little changed in American medicine, particularly in our quest to pursue value-based care.

I thought about it. I asked myself why, and that's when I wrote my second book on caring. I concluded that must be the physician culture, the things we learned in medical school and residency left over from the past because that's how our teachers learned it themselves and that if we could change the culture that we'd be able to make in advances, and I pointed out ways we could do so. And again, little happened. And when I asked myself why for the third time, I concluded is that we didn't have a tool capable of accomplishing what we needed to do, which is that simultaneously raise quality, make care more convenient and lower cost. And I almost would flip it around, how do we lower cost by increasing quality, by keeping people healthy, and by making care easy to access? And that's when generative AI 18 months ago popped on the scene and all of a sudden we had an answer that was there.

So how would it work? We're talking about empowering patients. The subtitle of the book is how AI Empowered Patients and Doctors Can Take Back Control of American Medicine. Let's take an example, 60% of Americans have chronic disease. It accounts some would say for as much as 90% of healthcare costs. This is the underlying chronic disease and the complications. These are the heart attacks, the strokes, the cancer, the kidney failure that we discussed earlier. And we have monitors, my colleagues at Stanford have developed these. These tools are very accurate. They can measure blood pressure, pulse blood oxygen, blood glucose perfectly. And yet they're not used at all in medicine today. And why would we not use a tool highly reliable, high quality that could conceivably change the course of American medicine could avoid 30% of these very costly things, could reduce the cost by 30% while increasing the quality clinical outcomes.

We're not talking about cutting back, skipping, we're talking about improving access and improved quality as the means to lower the cost because that data today is useless. Let's take blood pressure. A patient has a monitor at home, gets a hundred readings, 80 of them are normal, 13 are abnormal, seven are borderline. How are they doing? 80 normal ones, I'm doing great. No, 13 negative ones. I'm doing terribly. What can they do? And doctors don't want to have a list of a hundred blood pressures. They don't receive all this information on to the electronic health record. So the data is not used. In fact, the whole monitoring device is pretty rarely employed.

Now we have a solution, put this data in through a generative AI tool and update the patient every day, provide information and opportunities to improve their health with lifestyle medicine. Let them know, let's say a month after being started on a high blood pressure medication, how are they doing? And if it is deviating for what the clinician expected, why wait another four months until the return visit that typically is scheduled every four to six months? Why not be able to intervene sooner? Not by creating a doctor's appointment, by doing in person, but doing telemedicine or maybe a text, or maybe an email. This is the 21st century. How do we apply this tool to change? Let's just start right there with chronic disease.

And we can have the same conversation about acute problems, the opportunity for the patient to be educated, a hospital at home where you can get care with monitors being available so clinicians can intervene, should anything go wrong. The opportunities are almost limitless. I talk about when the iPhone was introduced. I see in many ways generative AI being analogous to the iPhone. I remember my dad got an iPhone the first year they were out. I was really impressed because I never thought of him as particularly savvy individualism. And I said, "Dad, what are you going to do with your phone?" And he said, "I'm going to lock it in my trunk. If I get into an accident, I can get help." That was the max [inaudible 00:17:38] that he could see at the time. And by the way, max, if not most people can see at the time a decade for the iPhone to become the smartphone to become ubiquitous.

But same thing right now, as you said, people are looking, I call it into the past and to the present. How can it help you relieve the administrative burden? How can it allow me to do what I'm doing today easier? And instead, what they should be looking to is the future. By the future, I'm not talking about a long-term future. This technology is doubling in power every year. That means five years from now it'll be 32 times more powerful. If that was your car, it would be going five years from now as fast as an airplane. Let's be getting ready for what's about to happen for what's about to come. And right now I think that people, as you say, are looking at the current means of providing care. And I wrote the book ChatGPT MD to help them to be able to envision a future.

Brian Urban:

I love that you took us there. And those examples are so fitting because we have these advanced pieces of technology that can give us in healthcare big leaps to where we need to go to improve access quality and have an economic shift in terms of cost spend overall in the country, let alone by system or population. And your book does take us into these leaps, and I love the iPhone example. You did talk about Apple Vision Pro in terms of analyzing imaging for physicians in surgery. I think that was a beautiful example, not to give too much away in your book.

I love where you're going and you hit on something just a moment ago that toward the middle of your book you go a little bit deeper in and to give our audience here a tease of that. You start to talk about retail health and innovation there and there's been a lot of noise happening post the closure of Walmart's retail primary care clinics and I wanted to get your take on something, but from a technology enablement perspective. So it seems we're in a trend right now with physical primary care locations. There's obviously a very thin operating margin in terms of reimbursement. You can't scale this very fast. It seems like a very retail-like approach to try to scale quickly and make convenience and experience the lead rather than understanding the bare bones of the payment infrastructure relative to pay or mix in the populations that are being served.

So it seems like primary care delivery on site is challenged when we look at Walgreens and when we look at Walmart obviously. But a CVS, you talked about CVS, they're more of an end-to-end. They have a health insurance arm, they have pharmacy PBM, they have more of a throughput to handle a scale to launching primary care. On the other hand, this is where I want to get your perspective, Dr. Pearl, Best Buy Health, their retail, they have technology that's enabling physicians, to your point, to practice more at the top of their licensure, less administrative burden. But from the data perspective, that data throughput with AI could probably advance the way physicians are seeing patients aging in home or having wellness services at home. So I want to get your take on what I just said there as a context. Is retail health more positioned to be technology enablers or do you think they're still going to try and plug away at having a primary care presence in the U.S. landscape?

Dr. Robert Pear:

The retail giants are poised to be able to transform American healthcare. And as you note we're talking about Amazon, CVS and we were talking about Walmart until they seem to have backed out of the process. But the idea of the retail clinic as just simply an acute care process isn't going to work for the future. We really have a sick care system. What we don't have is a healthcare system. What we have because of a transactional fee-for-service methodology is that all the rewards come to the people who do the most complex opportunities to reverse disease. Once it happens, unblock the arteries to the heart, resect the cancer, provide the chemotherapy, transplant the kidney. We don't go back upstream and ask how do we eliminate 30% of these problems? Not only the cost involved, but the impact on individuals and families, the premature loss of life and the consequences for loved ones.

We don't start there. And my hope had been that these companies, retail giants, would be the people to blaze through and move from a fee-for-service to a capitated model. And when you do that, now you change the entire function of primary care. What we know is that adding 10 primary care physicians increases life expectancy two and a half times more than adding 10 specialists. And yet we continue to train an abundance of specialists and insufficient number of primary care doctors.

So to put it into a context, had Walmart, which by the way had a relationship with United Healthcare, which is the largest Medicare Advantage program, a prepaid program for seniors over the age of 65. Had they decided to create a capitated approach, then they could have led the way using generative AI and other tools to be able to preserve the health of patients to be able to do what we said earlier, use the data off of a wearable device to be able to better manage and control a chronic disease and avoid its complications. They could have promoted health, they could have focused on medicine. They could have made sure that when you got sick that you were able to immediately obtain the consultation and get surgery to center of excellence. They had this ability.

Now why they decided to not do it, I can't tell you, they haven't consulted me so I have no inside information. I think they just said that's going to be more difficult than simply increasing retail. And they made a pure business decision. There was some also internal, as I read in the media, changes in the leadership structure that may have accounted for it. But when I put all the pieces together, I think it's just too bad. I think they were well positioned to be the change force for the benefit of their customers. Many of whom, by the way, suffer from a increased incidence of chronic disease, many of whom have socioeconomic challenges and other problems. So we had a tremendous opportunity and I'd call it almost a missed opportunity.

Brian Urban:

Yeah, you're just absolutely right on, especially with your comments on socioeconomic class, not knowing Walmart's population of buyers and consistent foot track. But the brand that they seem to have has a really good trust in terms of currency with folks that might be middle or lower in socioeconomic class that have a lot of social health needs that go uncovered. That directly correlates to healthcare spend or lack thereof going to healthcare at the right time for the service needed and coming back maybe with an exacerbated condition or a worse condition than they maybe could have had upstream in a preventive view. So they had that opportunity in your hands, you're right. And I'm curious, shifting back into AI now and a little bit into the EHR world, you don't dive deeply into the EHR world in your book, but right now the Epics and Cerners, many techs of the world are starting to consume more socioeconomic data on the virtual intake side, on the upfront end of their revenue cycle management and their process there and even on back end with analytics.

And it seems like it's more information being shoved into a platform that's going to be thrown at a doctor and maybe not presented in an actionable way. Do you feel AI in terms of large language models will be able to discern non-clinical data and show what action can be taken that's helpful for a physician? Because you can't just show, "Hey, this person lost their job from a credit bureau report, lost an asset, had a loved one passed away, here doctor, try to address this through a screening or through a questionnaire." If you don't have programs set up that doesn't really connect. But I'm curious, do you think AI will to solve a lot of the social health needs that connect so directly into healthcare utilization and spend, do you think that's on the horizon maybe in the more near term or is that still more of a long shot?

Dr. Robert Pear:

I don't think that generative AI can create houses or bring healthy foods into communities or be able to create an employment environment in which people can earn a good living and be able to support their families with a confidence and trust in the future. So I think many of the social determinants of health will still be there. But I also believe that in that environment, what often gets ignored are these chronic diseases. Now, it builds a vicious cycle where when you become sicker, it becomes harder to work. When you becomes sicker, it becomes harder to find the right job. When you become sicker you need to miss work in order to see a clinician, but your work, job that you have doesn't give you the time off to do it. And so you get into a vicious cycle where you have deteriorates and ultimately you are not able to work, not able to support your family.

So I'm not sure that we should separate out the two. generative AI is just the tool to be able to, as you say, handle the data. But more importantly, what the generative AI tools of today are capable of doing is empowering that patient. And not having the patient be dependent totally on the healthcare system, but be able to provide the information to help them to improve their health, to understand the diseases that are going on, to know when they get care, possibly to be able to find the places where the care can be provided that is optimal and affordable. I think it has the potential to do great things. But the key ingredient, and I write about that in the fifth part of ChatGPT, is leadership. That unless the leadership within healthcare is able to figure out how it can incorporate this into the practice of medicine,

I don't think to the point you made earlier that reviewing a lot of data and telling physicians what to do is the big challenge. I think physicians know what to do, they just don't have the time to do it. They know that it's important to look at these issues of social determinants of health. They know it's important to be able to focus on quantum disease. They know they have to deal into the psychosocial issues that contribute so powerfully into medicine. They just don't have the time.

The average visit is 17 minutes in the United States today, and that is why I'm also so excited because burnout, as you know, is 60% of physicians that report the symptoms of burnout. And when I look at that in great detail, I conclude that the biggest driver of burnout is that doctors don't have the time to do the job that they know they should do, like to do. And as a consequence of that, they go home at night knowing they could have done better. Not because they didn't want to, not because they weren't smart enough, not because they had a bad motivation, they just didn't have the time.

And the idea of being able to take 20 to 30% of what happens clinician's offices today, and have the empowered patient at home be able to do it, to be able to monitor and follow their own chronic disease, be able to access diagnoses for many of the problems that can be taken care of without a physician, musculoskeletal-type problems or some type of viral infections. These are opportunities that we have once we start to rethink how we provide care. And I'll go back to the same point I made earlier, and change the reimbursement model because you can't take 20 to 30% out of a clinician's office if the price is going to be paid financially. It's just not going to happen. None of us are going to compromise our families in that particular kind of way.

But once you change the model, now we change all of the thinking. Now we elevate primary care, we elevate prevention. It's not that we undermine the specialists, it's that we make certain that the specialist's work is done in the context of those patients who are going to benefit the most as opposed to those who didn't need to have the problem in the first place. So we can have a very different healthcare system. It's just going to require the combination to try it, I call it, of a dedicated clinician, an empowered patient, and a generative AI tool. And I'm absolutely convinced in the very near future, if not today, a combination of the three together will be better than any one of the three alone.

Brian Urban:

I love that you made that last comment there as we're going into what generative AI can start to fix in our healthcare model in the U.S. today, payment methodologies from a contracting perspective, the infrastructure of payments. I understand there's a statistic release last year from a report that our firm actually released. There's about 15 touch points per given claim that's adjudicated in U.S. healthcare today. And that is just so much time that is sucked away and administrative burden. And then the workforce impact the way that we could advance health outcomes, disease prevention going way upstream. There's so many low-hanging fruit opportunities to your point that you walk through in your beautiful book here, ChatGPT MD, Dr. Pearl. And I wanted to pull out one more tease for our audience to encourage them to look for your book on Amazon. One of the best-selling, the best-selling book I believe right now, in your class.

And there's a part that you comment, actually going back to your book on caring, I believe the committee on cost of medical care. It was a committee that you were a part of at a time, and not to give too much away to our readers, but I want to get your take on something as a future outlook here. Do you think future committee on cost of medical discussions will shift completely because of AI? Will it shift into a space of, "Hey, we've now had total medical costs drop completely, quality of care rise X percent and a lot of health outcomes are great. We're seeing less readmissions, we're seeing less complicated conditions being managed for long periods of time. We're seeing life expectancy happen, lower mortality for maternal health and infant health." Do you think those discussions will happen in a future committee on cost of medical care that physicians will sit around a table and discuss one day or what's your prediction in the future relative to that?

Dr. Robert Pear:

So you're talking about the CCMC, the Committee for the Control of Medical Costs. It convened in 1932, so it's not a new problem, almost a century. And it made the same conclusions that I've reached. I didn't come up with these ideas. As you say, we can trace them all the way back to that committee almost a hundred years ago. It said basically that we needed to have groups of clinicians coordinating care, collaborating, cooperating with each other, needed to shift the payment methodology from fee-for-service to a pay-for-value. But what it didn't have was the tool to make it all happen.

And they were right. We needed to do that. We've known that it's called Value-based care. We've known this for a long time. We've wanted to go in that direction. What I want listeners of viewers to understand is that now we have this tool, and if we use it well, I think that it will be looked at as a modern miracle in the same way that we never thought we could transplant our heart from one human to another, we never thought that we could accomplish all three that we're able to be able to lower costs by raising quality and access. We always thought it was pick any two. You could have higher quality but only at a higher cost, easier access, but only at higher cost, lower cost only by compromising access and quality. No, we can do all three. And they tie together because now we have the tool to make it happen.

The future will be written in one of two ways. Either we had the opportunity and we seized it, and as a result of that, we transformed medicine and we have a healthier nation with medical costs consuming less of the total GDP with the additional dollars being invested in education, being invested in families, prevention of chronic disease. We have a lot of uses for the dollars that would be available. Or write about it as another failure, the inability for us to be able to harness the technology, not because the technology wasn't adequate, not because the technology didn't have the potential, but because humans were unable to do it for a variety of reasons. And we will, I think, regret that significantly and particularly those of us inside medicine, that the burnout will only grow. The downward pressures on costs will only increase and we'll find ourselves in the future being less able even than today to provide the excellence in care that we so value and went into medicine to accomplish.

Brian Urban:

Wow. If that doesn't encourage our audience to dive deeper into your book, to make their own change from what they're learning from your readings, I don't know what Well, that is an amazing way to very much give a conclusion to this conversation. I feel a follow-up coming on though Dr. Pearl. And I will say a big thank you to Dr. Robert Pearl, author, multi-time author, newest book here, ChatGPT: How AI-Empowered Patients and Doctors Can Take Back Control of American Medicine, available on Amazon, of course. And so thankful that you spent some time with us today to go through the book and give some really rich insights for our audience to take away. Thanks for joining the show here today, Dr. Pearl.

Dr. Robert Pear:

Well, thank you Brian. And thank you for the work that you do helping to transform American medicine and move it forward. Together we can make American medicine once again the best in the world. Thank you so much for today. And if people want any more information, they can check out my website, RobertPearlMD.com, where they can find information, not just on ChatGPT, but on a variety of medical issues. And I look forward to hearing from any of the listeners and viewers who purchased the book and read it about their thoughts, particularly areas that they disagree with because this is the dawn of a new era, and we have to make sure that we have the conversation discussion necessary to make certain that we make it the time of great success so that the sun will shine upon us into the future. Thank you so much.

Brian Urban:

Thank you, Dr. Pearl. And for more exciting excerpts and insights, please visit us at finthrive.com.