Subscribe to The Podcast by KevinMD. Watch on YouTube. Catch up on old episodes!

In this thought-provoking panel discussion, orthopedic surgeon Yoshihiro Katsuura and premedical students Kie Shidara, Maria Llose, and James Schmidt come together to examine the growing influence of artificial intelligence in the health care industry. From the potential benefits of AI-driven efficiency to the concerns surrounding job displacement among physicians, they navigate the delicate balance between technological advancement and patient-centered care. Drawing parallels between AI integration in health care and recent debates in the entertainment industry, the panel offers valuable insights on how health care professionals can advocate for ethical AI usage while prioritizing the well-being of their patients. As the health care landscape continues to evolve, it is crucial for stakeholders to engage in meaningful discussions about the future of medicine and the role of AI in delivering quality care. Join us as we explore the complex intersections of technology, efficiency, and compassion in health care, and discover how professionals can leverage AI to enhance, rather than replace, the human touch in patient care.

Yoshihiro Katsuura is an orthopedic surgeon and author of The Spine Encyclopedia: Everything You’ve Wanted to Know about Back and Neck Pain but Were Too Afraid to Ask. Kie Shidara, Maria Llose, and James Schmidt are premedical students and research coordinators.

He discusses the KevinMD article, “What doctors can learn from actors about artificial intelligence.”

Our presenting sponsor is Nuance, a Microsoft company.

Do you spend more time on administrative tasks like clinical documentation than you do with patients? You’re not alone. Clinicians report spending up to two hours on administrative tasks for each hour of care provided. Nuance, a Microsoft company, is committed to helping clinicians restore the balance with Dragon Ambient eXperience – or DAX for short. DAX is an AI-powered, voice-enabled solution that helps physicians cut documentation time in half. DAX Copilot combines proven conversational and ambient AI with the most advanced generative AI in a mobile application that integrates directly with your existing workflows. DAX Copilot can be easily enabled within the workflow of the Dragon Medical application to bring the power of ambient technology to more clinicians faster while leveraging the proven and powerful capabilities used by over 550,000 physicians.

Explore DAX Copilot today. Visit https://nuance.com/daxinaction to see a 12-minute DAX Copilot demo. Discover clinical documentation that writes itself and reclaim your work-life balance.

VISIT SPONSOR → https://nuance.com/daxinaction

SUBSCRIBE TO THE PODCAST → https://kevinmd.com/podcast

RECOMMENDED BY KEVINMD → https://kevinmd.com/recommended

GET CME FOR THIS EPISODE → https://earnc.me/ibtiU3

Powered by CMEfy.

Transcript

Kevin Pho: Hi, and welcome to the show. Subscribe at KevinMD.com/podcast, and get CME for this episode by clicking on the CME link in the show notes. Today we welcome Yoshihiro Katsuura, Kie Shidara, Maria Llose, and James Schmidt. Yoshi has been on the show before, he’s an orthopedic surgeon. Kie, Maria, and James are premedical students and research coordinators. And we’re going to talk about the KevinMD article “What doctors can learn from actors about artificial intelligence.” Everybody, welcome to the show.

Yoshihiro Katsuura: Hi Kevin, thanks so much for having me back.

Kevin Pho: All right, so Yoshi, talk about this article and introduce the group that we have here.

Yoshihiro Katsuura: Yeah, so my name is Yoshi Katsuura. I’m a board-certified orthopedic spine surgeon, I practice in Northern California. I’m also the author of “The Spine Encyclopedia,” which is a book that I handwrote and illustrated myself, which is somewhat relevant to the topic today.

But more relevant, I think, is I’m also the founder of the human performance and spine surgery research pod, and some of the co-members are here with us today. Kie, James, and Maria are all premedical students and help me performing various projects within the group.

And our mission is basically to maximize human performance in surgery, in a world that has ever increasing technological sophistication in medicine, where many of the skills and tasks are being automated and basically taken away from physician hands and minds. And we’re kind of running countercurrent to that and trying to develop new ways to improve human performance, basically with increasing enthusiasm for artificial intelligence and robotics.

Kevin Pho: All right, Yoshi, in your field of orthopedic surgery, tell us how that intersection of AI, orthopedic surgery, how’s that evolved in the last year or so?

Yoshihiro Katsuura: So I think it impacts what I do in a lot of different ways. I mean, I can just tell you a little bit about the week that I’ve had this week, and you can maybe see some examples about how AI may help in the future and different areas of technology are impacting what I do.

So I generally spend about two to three days in clinic per week, and then another two to two and a half days in the operating room. I’m on call for trauma services as well during that time, so I can be seeing clinics all day long and then get called into the operating room in the evening. And some days can be quite challenging with a high patient volume.

As that continues to increase, we’re faced with a lot of challenges. I have to see each patient as an individual, each patient is different, but they come with a lot of complexity. And each patient basically has reams of medical records that are available to us on the EMR. They come with 20, 30, 40 studies, sometimes those are all areas of the spine for me, so x-rays, CTs, MRIs that have to be reviewed, sometimes multiple areas have been scanned over time. Then we have to have an interaction with the patient, examination, and then somehow synthesize all this information and make appropriate decisions for patients.

And I see AI basically helping people deal with some of that workload. So going through the medical record, maybe synthesizing it all, reviewing all the imaging, and then basically helping you make a decision. And I think as AI continues to advance, there’ll be a slow attrition away from what I do. Because I see infinite variables in the patients that come and see me. They’re all very complex and they all have unique problems, and sometimes putting all this information together in a way that’s cohesive can be difficult. So I do see AI having a role with that in the future, maybe a dominant role.

And I also see another area, and that’s actually in the motor function that I perform in surgery. So for example, this week we performed some spine surgery and we actually used a robot to help us place some of the orthopedic instruments into the spine. And the robot can do this in a way that is basically very accurate, and in some ways more accurate than humans can do it.

But it also comes with a lot of problems. One is the cost is astronomical for using this type of machine, and basically it will not be something that can be deployed in all medical centers and can be available to everyone. Also it’s very cumbersome to use, it’s very slow as it is currently, and it takes up a lot of time in the OR, which is obviously very valuable. So I do think that there are things that will continue to improve, and definitely there is a role for AI in the future in my field.

Kevin Pho: All right, so before we talk about the KevinMD article, we’re of course joined by three premedical students. My daughter also is a first year premedical undergraduate student as well, so going through the same path as each of you. And you’re going to be doing the majority of your medical training kind of in this AI age, and the pace at which AI is evolving is just so fast, because we’ve come such a long way since GPT, for instance, was introduced just more than a year ago. So I’m just going to ask each of you this, through your lens as early on in your medical journey, what your perspective of AI is as it associates with medicine. So James, why don’t you go first.

James Schmidt: Yeah, absolutely. I first want to say thank you for having us on today, I really appreciate being here.

And with AI, I think it’s really obvious that it can streamline my learning process, it can make me more efficient, it has access to more information that I could feasibly imagine and could get it to me way faster than I could myself. But that really comes at a clear cost to me.

With this research group, I’m working on a systematic review on mental practice in the operating room and in the medical setting, and I did the research myself, I read the papers, and I have a deeper understanding and appreciation for that. And I could have used AI as almost like a coworker or colleague, and ask its perspective and use that to help channel my biases and understand different views, but again, it would just give me a lower level of analysis. It would give me the information without me doing the work and appreciating that, and I think that comes at some cost.

Kevin Pho: So James, as a follow up on that, how difficult was it for you not to use AI as a crutch and to read those papers yourself? Did you feel any temptation to rely on AI?

James Schmidt: Yeah, I actually did. When I first did my PubMed, I was like, oh my gosh, there’s 5,000 articles I’ve got to go through, there’s got to be a machine learning tool that I could use. And I went down a rabbit hole and I looked at different tools, and then I started to realize that this process is important for me to learn, and AI could skew my results and affect my research and learning process.

Kevin Pho: All right, Maria, share your perspective on that intersection between AI and your perspective on medicine.

Maria Llose: Yeah, I just want to thank everyone again for inviting me for the podcast, and thanks James for sharing your opinion as well.

First, I think for me personally, I think AI is great. It’s a great baseline to start, whether it be diagnosis or just to get a head start in terms of your goal. But I also want to think about, it’s efficient, I think overall AI is very efficient and starting off, but I do want to kind of think about the longevity of it. How long can AI last?

For example, I work in San Francisco, there’s a lot of AI self-driving cars. How long could you kind of depend on those cars to take you from point A to point B? Would you take on a six hour road trip from SF up to LA, for example?

And just in terms of the communication as well, as a future physician I value relationship and connection and communication to patients. And with AI, how can we trust AI to communicate those, whether it’s feeling if it’s either wrong, or if it’s feeling tired? How can we communicate that? At least with humans, when we’re feeling kind of blue, kind of fatigued, we can sort of communicate that, and also streamline and also work with other people to kind of get that going and achieve that goal.

Kevin Pho: And Maria, in your premedical education, is there any formal training in terms of how AI can potentially affect those relationships at all, or is this something that you’re just finding out along with the rest of us?

Maria Llose: Definitely finding out along with the rest of you guys. But for me, when I study, especially like hard biology, physiology lessons, I love to work in groups, like study groups. I love to talk to people, brainstorm, collaborate, get ideas, and overall kind of have fun with it in a way.

I think with AI, for example, if I were to use like ChatGPT or something to make flash cards, for example, I don’t think it’s going to be, it won’t be as fun, in my opinion. I kind of like the challenge of looking for things, communicating with other people and just sharing their opinions on that too.

Kevin Pho: All right, and finally, Kie, share your perspective through your lens of medical education, AI.

Kie Shidara: Yeah, thank you for having me. I am currently working on a project on task switching as it relates to human performance and efficiency.

As for artificial intelligence, I have mixed feelings. On one hand, I think that AI has great potential to advance technologies that could help with improving human health. On the other hand, it would reduce the humanistic aspect of the patient physician relationship, which I think is a really important aspect of medicine. And also it would make people generally less intelligent and thoughtful over time if they relied on it too much.

Here’s a story. I was studying for an exam last weekend using a flashcard app, it’s called Quizlet, and it has a feature for creating an AI generated outline, which is something that I used to do for myself when I was in school. I would do it manually by hand, maybe sometimes type it up. If I relied on a feature like that, it would reduce the amount of thought that I would put into understanding the different topic areas, and I don’t think that would be good over a period of time.

Kevin Pho: All right, so right now, Kie, do the majority of your premedical colleagues, how much do they rely on AI?

Kie Shidara: It’s hard to say. This isn’t something that I ask in conversations with my premedical colleagues. But I do work with a high school student, and she does use ChatGPT. That’s not something that I used to use in high school, and it does make me wonder, and also a little concerned, of where her learning process is going to go. Not just her, but her classmates as well.

Kevin Pho: So one of the themes I sense from each of you is really how AI is going to interfere with that human aspect that is so integral to medicine. So Yoshi, you talk about that of course in the KevinMD article, “What doctors can learn from actors about artificial intelligence.” So tell us more about this article for those who didn’t get a chance to read it.

Yoshihiro Katsuura: Yeah, so I was inspired to write this article with my co-members based somewhat on our discussions. We have a weekly research meeting where we go through the various topics that each of the members is exploring, which is a facet of human cognition or human performance in some way or fashion. So we talk a little bit about how AI might impact them in the future, and how it’s impacting me, and how I envision it might impact me in the future.

But also, I was inspired to write this article because of the recent Screen Actors Guild and Writers Guild strikes that happened earlier this year, where AI was very prominent in their negotiations. And some of the topics that they were asking for were basically that writers, for example, could not be presented with manuscripts or materials that had been generated by an AI, so that they could basically put the human touches on it and follow it up to make it into a production without ever knowing that an AI generated it. They might have thought another human might have generated it.

Another example would be using their own source material, so their previous manuscripts or writing material, to train AIs to do their own job. So basically using what they’re doing currently, training an AI that can do it better, and basically putting themselves out of a job.

And I see a lot of corollaries with that and what we do in medicine. So as you and the audience is probably very well aware of, we are tied and chained to the EMR, and we have to put in all of our notes and thoughts into the EMR, and everything is tied to that. All the economics of medicine are basically tied to the EMR, and it’s primed as a major data set to train AIs.

And for those people who may not know, AIs are basically large computer programs where the architecture is based on the architecture of the human brain, and you have millions of nodes which are similar to human neurons that are weighted. And the way they get weighted is by having access to large amounts of data, and the data is used to train the program. And it can be used to train them in a game like chess or Go, it can be used to train them to read an x-ray for certain features, it can even be used to produce medical notes, or evaluate human patient, or doctor patient, interactions to generate medical notes.

And so one of the worries that I have about AI is that we as physician authors of the EMR will basically train these systems to do the work that we’re doing, and there will be no recuperation of that.

For the medical students in the group and for the premedical students in my research lab, they’ve already been through four years of undergraduate training, they’re going to go through four more years of medical training, five or six more years to become qualified as a surgeon, and then another one to two years of fellowship training. And then that’s really when their careers will start at that point, and then they will accrue experience as they practice medicine.

And obviously we can only see a certain amount of patients in a day, we can only do so many surgeries in a week, and there’s a basic limit based on time and human energy. And with an AI, I don’t really see that limit. An AI could look at millions, it could look at every single doctor’s experience in the operating room, in the clinics, and synthesize all that almost instantaneously. And I see that as being both extremely powerful but also having this massive displacement force for the people that are undergoing training and the people that are currently practicing medicine.

And so that really is what inspired me to write this article, and just to maybe be a little bit more cautionary about AI and how it’s used in the clinic. And we need to know when it is and isn’t being used. So when you’re producing EMR charts, when you’re dictating, when you’re using these AI gadgets that listen to you in the clinic, you are training AI whether you like it or not. That data is being collected, that data is going to be used, and it’s going to be used to make a product that is not going to be free, basically.

And the physicians are basically helping to train that without getting any remuneration for it. And what we do is altruistic, and we’re here to basically help our human brethren and promote healing and do all that, but there is a big sacrifice that goes into what we do. And that’s one of the themes that we talk about a lot in the research pod.

And I do think that currently medicine is dependent on people who are willing to sacrifice. Basically people who have a natural curiosity, people who are willing to study, spend a long time in the library whether they’re getting paid for it or not. And they just are curious about biology, physiology, they are compassionate, they want to, like Maria was talking about, she wants to brainstorm and be with other people and help people. And basically there’s a sacrifice that comes along with that, and I’m worried that that will be lost in the throes of AI.

Kevin Pho: So you talk about that tension about how AI can be so powerful and help us sift through millions of pages of data and really extract what is most important for physicians, but at the same time you ask, are we going too fast, are we giving too much power to these AI companies? So do you think that there should be some guardrails? Are we progressing too quickly?

Yoshihiro Katsuura: I think we’re definitely progressing too quickly, and I think as a medical community we’ve been a little bit lackadaisical about, or maybe nonchalant about, adopting AI. And there are people who are already calling for it to just not be questioned at all, we should just accept it as a fact and go along with it and use it. And I agree, it’s going to be extremely powerful, and I also am not naive to the fact that it’s here and it’s here to stay.

But I think that as a community we need to enshrine the relationship between a human physician and human patient. And also there need to be guardrails about decision making in the use of AI.

So I can envision a scenario where, to get a surgery approved by an insurance company, it would have to go through some sort of AI decision-making process. And the insurance companies are probably already doing this. And what happens is, in private practice or even in non-private practice, you rely on an office staff to combat that so that the patients can get the procedures that they need. Well, if an insurance company is using a very powerful AI to basically make denials, I can envision an arms race where hospitals, private practices, individual physicians will also need to employ AI to basically have a counterproposal. So that will take away decision making entirely from the human physician, if that happens.

And so I think we need to ask ourselves now, is this something that is important to us? Is human decision making important? Like Kie was saying, I think over time, if we stop making these complicated decisions we will become deskilled, or even lose confidence in our ability to do it. And I think that will lead to part of the culture of medicine dying, if that happens.

Kevin Pho: So I think that the point that you bring up is pretty salient, because already as a primary care physician myself, whenever I get one of those letters that require pre-authorization, I just run it through AI. And sometimes you get a denial, and who’s to say that you’re not just getting back an AI response, and you’re just going to have AI back and forth? So your point taken about physicians becoming deskilled is certainly very relevant.

Yoshi, is your message resonating? Because you’re on the West Coast, right?

Yoshihiro Katsuura: I am.

Kevin Pho: So you’re kind of in the heart of where all the AI is being developed, Silicon Valley, all the health startups. And I interview so many health startups that are so excited and enthused about the direction of AI. So you’re taking a tack where maybe we should take a deep breath, maybe not necessarily take a pause, but more thoughtful in terms of how we develop AI. How is that message resonating on the West Coast?

Yoshihiro Katsuura: Well, I think it’s not resonating very well. But as you can see from my very enthusiastic students, I think they all have an appreciation for maybe some of the dangers and some of the challenges that AI will pose.

And part of the theme that I think is really important is that the research that we’re doing is all looking at ways that humans can improve their performance in surgery or in medicine without any use of technology, just through basic cognitive techniques and or motor techniques to improve their performance. And I really don’t feel like people are interested in that anymore.

I feel that most of the money and most of the research goes into development of very powerful robotics, like the spine robot I was talking to you about, that costs a million dollars to have and then hundreds of thousands of dollars to maintain, and to keep it working requires this large task force of personnel. And ultimately it is accurate, and I think it will continue to become better and improved, but it also slows things down dramatically.

So most of the energy, most of the technology is going into developing these things that basically take things away from physicians. And what we’re trying to do in my research group is find ways to maintain our skills, to improve our skills, to basically try to go the other direction just a little bit. I know it’s probably counterculture and a lot of people will not appreciate that, but I do think it’s important. I think it’s something that we need to be thinking of.

Kevin Pho: We’re talking to Yoshihiro Katsuura. He’s an orthopedic surgeon, and his research group are premedical students Kie Shidara, Maria Llose, and James Schmidt. And we’re talking about their KevinMD article “What doctors can learn from actors about artificial intelligence.” Yoshi, we’re going to end with you, last question, your take-home messages to the KevinMD audience.

Yoshihiro Katsuura: Well, I would say to the audience, just be mindful of what you’re putting into the EMR. Is it being used to train an AI system? I think these are questions that need to be asked.

And also, if there are any medical students out there interested in learning more about human performance, or potentially getting involved in research in ways to basically boost performance in surgery, we would welcome your queries. They’re welcome to reach out to me via email.

Kevin Pho: Well, thank you all for sharing your perspective and insight, and thanks for coming on the show.