Latest / AI Meets MedTech / When Machines Are Better, Should Human Surgeons Step Aside?
Transcript
- Dr. Amir Nasir v2: Welcome to AI Meets MedTech Podcast. My pair is Jeroba. Welcome to AI MitsMetTech. My guest today spent his early career as an optical engineer. Then he became a board certified plastic surgeon and a business owner. And then he wrote a dystopian novel about the last human surgeon on Earth. Welcome, Dr. Amir Nassim. So happy to have you on AI Mitsmetic. Well, thank you, Anna. It's a real pleasure to be here and I look forward to our discussion. That's awesome. I have so many questions and maybe We should start with a little bit of your background, because obviously just from my intro, it appears that you have a very interesting career path. And I definitely want to talk more about your novel, because I think it's right on the topic of AI and the future of healthcare and what may or may not happen to medicine to surgeons and I guess the world we live in. But before that I'd like to ask you, so you've gone from the engineering background to plastic surgeon. Back to to your work as a plastic surgeon. How did that path shape the way you think about AI right now and healthcare using AI and novel technologies at the moment? Of how machines work and how to measure things. And when you do optical engineering, there's a lot of mathematics involved. We did, you know, calculus, differential equations, quantum mechanics, we did a lot of mathematics and all of AI right now is based on mathematics because it's all algorithms. So I can see kind of how AI works from a really fundamental point of view. Also, when you're a surgeon, you're doing things that are very similar to what engineers do. You're solving problems. That's all you do all day. That's what engineers do. You know, they look at loads, they look at lots of very concrete things, and they try and figure out how to improve those things. As a surgeon, you do the same thing. Somebody comes to you with a problem. They have a tumor, they have breast cancer, they have an issue. And so the way that you are trained as an engineer easily translates into surgery because it's all problem solving. And that's what AI does as well, is it's solving problems. And I think having the benefit of an engineering background and a coding background, because I did a lot of coding when I was an engineer, I think I've got a pretty good understanding of how LLMs work and how AI the limitations to AI, basically is what I'm trying to say. I'm so glad that you mentioned that word because I'm so honored to have you on this podcast because you're the person I really want to talk about AI and medicine and how we can see it from the perspective of, let's say, a surgeon, from a clinician, and also how we can see it from a perspective of, let's say, software engineer. Because there those are very two different perspectives. And you have both. And that's why I think your conversation with you, so valuable, I would like to, you know, gather more from you. Where is that line where we what what I hear, let's say, on the news is true or not true? And where do we s need to be skeptical and where we don't need to be skeptical? So I would love to dig deeper with you on the conversation about AI and medicine. But before that, can I ask you one more question? So what pushed you to write a novel? Because your novel is fantastic in terms of the idea because I don't want to spoil it too much, but maybe we can talk a little bit about the premise about the novel. because it's called The Last Surgeon. And the premise is the premise is that it's the last surgeon on planet Earth after the era of AI and technology going forward. Is that correct? Can you tell me more that's yeah, that's pretty close. So it's called The Last Surgeon, a dystopian novel. Yep. And I you know, I was I I meditate every day and so I was meditating one day and this idea just came to me and so I started writing and I haven't really written a anything that wasn't clinical since high school. So it was a really interesting process to to go through the writing experience. But what it's about is it's about how AI very slowly encroaches upon all of the different institutions within a hospital, within a hospital system. More importantly, you have a lot of skills as a surgeon. You have to be compassionate towards your patients, you have to have manual dexterity, technical skills, broad knowledge. And so I I thought that if a machine can replace a surgeon, then I think everybody's job is at risk. And I wanted to see how or to demonstrate how that can slowly happen within a hospital system and make it as realistic as possible. I think we would all agree that medicine as we know it is a it's broken. Healthcare in the United States is broken. It doesn't work. And I think it's getting worse and I think it's going to continue to get worse. And I feel that hospital administrations are going to look for ways to salvage the system and automation may be that maybe the way they do it. So I I paint a very what I think is a realistic scenario of how hospital or or how AI and robotics are delivered into hospitals and how they slowly i take over, not in an evil way, but because they're very good at what they do. so that's kind of the premise of the book is, you know, is it possible for eventually robots and machines to do what surgeons do? I I think so. a lot of my colleagues think that that's that will never happen. And they could be right. But I think we all want to believe that whatever job or task or role we have in life is irreplaceable. And I think we're all replaceable to some extent. And I think also as a surgeon or as a physician, if there is a machine that does your job better than you do it, then, you know, it it kind of goes against the Hippocratic oath to continue to practice if something else does it better. And and let me give you an example of this. So I love radiologists. I've I've worked with many radiologists. I have a lot of respect for them. They're incredible physicians. They're brilliant people. But, you know, we've seen recently that, you know, some so AI software can detect breast cancer earlier than experienced radiologists. So if we ever get to the point where machines can do that better than humans, would we be doing a disservice to patients to not allow those machines to do To be able to read films and CTs and MRIs. Maybe you could make a strong argument that that that we should allow machines to take over. So it's it's kind of those ideas are are what's in the book. This is incredible because that's exactly what I'm thinking too. When I actually not long ago I had a lawyer on my show and we discussed that. You must use it. You can't say no because it is superior for the safety reasons, et cetera. So we are right on top of what I am thinking about. So let's let's talk more then. So in your clinical practice, where do you see that line for trust in AI tools right now? Because obviously, like you said, radiology is radiology. We understand it's more of a software systems that can read images and analyze images, and AI is doing a great job there. But when it comes to surgery, it's a very different line of work. And how Would you describe right now the line of trust? Do you find that surgeons say like we only have maybe ten percent automated systems and ninety percent is still hands-on manual? Or are we kind of pushing that line? What's what's happening? I think right now within within my field anyway, I think everything is manual. People do like surgeons do all the surgery. I mean, there is there are surgeons who do laparoscopic surgery where you're working in conjunction with a machine, but the surgeon is in one hundred percent control of the machine. when when they're doing that procedure. So where is the line of trust? Well I would say that the bar has to be set very high. When it comes to patient safety, you have to have you know you can't there's no compromise. You can't risk a patient just to try out a a new technology. So I think if there's ever a point where machines get to be good enough to be able to do surgery, what you'd have to do is you'd have to do hundreds or billions, hundreds of millions of cases virtually first, and then you would then train the machine on animals like pigs, and then and only then would you allow machines to operate on people with their consent. And I and I all of this is kind of written in my book and I because I thought about all this stuff, like how would I how would I do it? So I think it's a while away. It's not it's not something that we'll see probably in the next five years, but it's not impossible that we'll be seeing, you know, machines operating autonomously on people. And what makes me say that is, you know, there are there's a machine called the Da Vinci machine. It's a very well known robot that general surgeons use for laparoscopic surgery. Well, I'm assuming that the the manufacturers of the the machine are taking all the data that the surgeons are using when they're manipulating the the devices in the machine. And so the surgeons are basically training machines to do what they do. Now there's a lot of variability in human anatomy. There are things that come up in surgery that are unpredictable. But I think with enough training and with enough time, that's where it's going to start. It's probably going to start with laparoscopic instruments and lap a a machine that does laparoscopic surgery. Very simple procedures initially. But then I think more complex procedures will will will be done thereafter. Yeah. The question that I got from this is what do you think do you find it's a little bit troubling that surgeons are trading machines and they're willingly giving away their skill set, their knowledge to robots, knowing that in five, ten, twenty years those robots and machines can replace human surgeons. So I have mixed feelings about that. I mean I I think, you know, the the the job of a physician, a surgeon is to help their patient recover, get better, improve, heal. And so if you're training a machine that will ultimately do that better than you, then, you know, you're kind of doing the right thing. It means your career is over. You know, like if if tomorrow, you know, there was a like if Tesla came out with a machine like robots Optimus, if Optimus came out and could do surgery tomorrow and it could do it better than humans with better outcomes, then you know, that would be terrible for me, but it would be great for patients because the cost would go down, it would be safer, it'd be quicker, it could have you could do surgery twenty four seven, you wouldn't have to worry about fatigue or you know, errors or or whatever. you know, whatever foibles that we have as people. so yeah, I think it I think it's tragic in in a way, but you know, the whole goal of medicine isn't to bill and code and make insurance companies rich and and make hospital CEOs wealthy. The goal is to help people. And so if if machines can help people better than people can, then I think we should let the machines do it. I and I and I say I don't say that lightly. I say that, you know, as someone like what would I do if if God forbid my wife developed breast cancer or or my son was injured, and the options were to send them to a human who could make a mistake or, you know, could make a misdiagnosis, or a machine who does something much better, I would probably send them to whatever would give them the better. the the best outcome. Yeah. And I'm with you on that, but I'd like to play a bit of a devil's advocate here. Because there also there's another way of thinking about machines versus humans. And I am sometimes leaning towards that theory too. There are certain human skills that we can't really teach a robot or machine to really replicate. And in medicine, as you mentioned, the anatomy can change. Maybe there is one off case where a patient represents A completely different subset of symptoms and it's it's definitely one off. And a human can make a decision quickly under pressure, considering I guess it could be even gut feeling. Let's let's call it that way, even with a gut feeling. Obviously, practice and skills, but also a gut feeling. And we can argue that machines and robots can't really have that. Yes, statistically they may perform better, but on those cases where you really need to think on your feet very quickly. under pressure. They can't really do that. I'm almost trying to say that Elon Musk is wrong. And and Lesla robots cannot replace surge replace surgeons completely. But yeah, what do you think about that? Because I do believe humans still have something that is superior and something we can't really teach a machine. But you can tell me if I'm wrong. No, I don't I don't think you're wrong at all. And I think that's a very, very valid point. but what I think about when you say that is You know, speaking of Elon Musk, it was on a p I think it was on Joe Rogan's podcast years ago. And he was talking about how electric cars, when the wheel is turning, each individual wheel can break you know, whether the decision whether or not to break on a s on an icy road is made thousands of times per second, or hundreds of times per second. In a way that a human could never do. Like We we can't process things that quickly. So in in surgery, if someone is bleeding or if there's you know, some anom anomalous anatomy or if something unexpected happens, I think with the right software machines could potentially react much quicker than a human could. And also, what is gut a gut feeling other than the experience that you've had or kind of intuition based on a lifetime of experience. And I think if a machine has done hundreds of thousands of laparoscopic ra laparoscopic cholosystectomies, where a surgeon may have done a thousand, maybe ultimately the machine will ha make better decisions than the surgeon, who just based on the amount of time that they have in their career, they'll have more experience than the human would, and so they will make better decisions, potentially Yeah, it's a pure numbers game. More cases, and of course, yeah, machine can do so many, and they can share information between machines versus humans. We like as a surgeon, you can't really unless you try it and test it and do it multiple times. Yep. Yep. It is a scary thought in a way. It is it is scary and it's it's unsettling. But I think what I what I do is I always come back to what is the purpose of this? What is the purpose of being s a surgeon or a doctor. And again, I come back to it's to try and help people. And you know, if if a machine can do it better, then you have to step aside and let the machine do it. That's that's my opinion. You know, there there are people who say that machines will never have compassion, that you can't establish a doctor-patient relationship with a machine, that for example, they they won't know people in the same way other people do. Let me let me give you an example of that. Let's say that you're in your clinic and a mother comes in with her son and says, you know, this kid has been driving me crazy. I want to strangle him because, you know, he spilled my coffee or whatever. You know, a a doctor, a person would know that she's just speaking jokingly. But a machine, if it's not trained well, might call child protective services and or the police or something and and get take things out of hand. So there are there are things that people bring to medicine and to healthcare and interpersonal interactions that might be much more difficult for machines to understand. But I think in time they they will even be able to see that that people have a sense of humor, that some people are just weird and crazy, that To they'll know the context of a situation maybe as well or better than human beings do. Yeah. So which scenario is more plausible in your opinion? That machines will take over everything when it comes to medicine eventually, we're talking long term. Or it's more plausible that humans will be a supporting power to machines. So yes, machines can do the job better, as you described, but let's say doctors will become like a like a segue or a support person in the i to to support the relationship between a human and the machine i mean i think in the long term or in i think in the sh in the medium term I think human doctors are gonna be a luxury for people who don't want to deal with machines or for people who are scared of machines or people who have nostalgia towards the way things were. I think ultimately machines are gonna do the vast majority of what human doctors do. I mean, if you think about like going to see your primary care doctor because I don't know, you need you have you've got reflux. It is such a horrible experience now. I don't know if you've seen your primary care doctor recently or any doctor recently. It's a terrible experience. You gotta drive someplace, you gotta wait in line, you know they ask you bunch of stupid questions, they ask you another person comes in, they ask you the same questions, they get get labs and it's a really unforgiving process. Now, if you could have if you could do that by telemedicine with a machine from home, that's that's a much more convenient thing. And if insurance companies end up making that cheaper for people, which I think they will, then that's going to, I think, phase out a lot of what primary care doctors do. And I again I talk about this in in my book, excuse me. so I I think what we'll see is, you know, we'll see information based medicine slowly disappear or quickly. and then we'll see more and more and more things get taken over by machines and robots. Yeah. You know what I love about this? on my g on my podcast, my guests, it seems like they fall into two camps. People who say that yes, machines will take over and things will change completely. And people who say that AI and machines will never replace doctors. We will always have that craving for human connection, human interaction when you go into doctor's office. So you want to be heard, you want to be understood by another human. So it seems like you are more falling into the camp. I would probably say Elon Musk is yeah. Mm-hmm. I the the reason I think that is most people don't go to the doctors for fun. They go there because they have a problem. They have an issue, they're afraid, they have a broken but they want the problem to be solved. And they go through all these motions of, you know, finding parking and taking time off and getting babysitters and you know, all of the headaches that you that you have to do. To see a a human clinician, people go through that because the problem is compelling and they need it to be fixed. They don't necessarily go through it to say hello to their doctor. I mean, maybe that was the case a hundred years ago when you knew when when your primary care doc was a doctor who gave help gave birth to you, or didn't give birth to you, but delivered you and then took care of you as you were growing up. But now medicine is very corporate. The doctor-patient relationship is much less important than the doctor hospital relationship. That's what that's what exists now. and that's not very meaningful or very satisfying, I think, to most people. So people now just want the problem to be solved. And I think if again, if AI or machines can do that better, I think they'll they'll opt for that. Yeah. So what do AI and automation then mean for professional identity for surgeons who are, let's say, training right now? Obviously you are kind of far along in your career path, but what about the new surgeons, younger surgeons? Yeah. But I I don't know what kind of career people are going to have within medicine and surgery twenty, thirty years from now. I don't know if there's going to be a lot of lot of options. And I hope to God I'm wrong. I hope that's not the case, but I hope whatever happens is in the best interest of people and patience and humanity. But if what I think is gonna happen then it's gonna be something where you can't let your career become your identity. I think in general that's a that's a good that's a good thing to do. I don't think you should your career should be your identity. But if it is, it's gonna be very hard to move on from from a c a collapse or a reduction in the need for a particular class or or group of people. Yeah, it is definitely something to think about for people getting into medicine right now. And again, it's not gonna change tomorrow, but it is changing, and especially considering how long it takes to train a surgeon. It takes a long time. Yes. Yes. So we might live in a very different world when it comes to medicine and healthcare by the time some surgeons are finishing the training. Yeah, it's a little bit of a scary thought. But can can I ask you a couple of questions on the business side of things? Because outside of your clinical practice. And you being a novelist now, you also run a business. And can you tell me a little bit more about the business so we can kind of move from from there? Because I have a couple of questions. Sure, sure. So we s we started something called Clinical Preflight, clinicalpreflight.com. And what it is, is we look at products, services, and devices that usually small startups or hospitals are looking to implement within their systems and we look at them from a very practical perspective. Is this a good idea? Is it a terrible idea? Is it something that can be improved, modified, or fixed before it goes out into the messy clinical world? That's what we do. And we do it very quickly. I've seen a number of healthcare systems invest large sums of money in products and services that were not going to go anywhere. That if they looked at if they had an objective person look at them early on they could have saved a lot of time and money. And so that's what that's what we do. Yeah. And my question to you was how do you really assess those companies, especially when it comes to AI, how do you make the assessment and judgment if it's really gonna work where it makes you skeptical? So can you go can can you walk me through the process? Sure. So so we look at the creators of the the service. We look at what their motivation is. We look at the product very carefully from a practical point of view. So I think what a lot of consulting firms do is they are very complimentary towards the product, whatever product that they get, because they want to maintain a contract, they want to continue receiving funds for reports. We don't believe in that at all. We we look at things from the standpoint of a patient, a nurse, a doctor, an administrator, you know, somebody within who who has to clean it, who has to put it in the autoclave. We we look at it from as many directions as we can and and we have access to people that can give us feedback very quickly, to ascertain whether or not it is something that would function well within a system. And if if we find errors or flaws or shortcomings, we're very direct about it. And so I think that comes from experience, that comes from having very good connections in lots of different specialties and lots of different spaces within healthcare. People I've known and trust. So that's that's that's what we do. Yeah. And going back to companies who developing AI tools for medicine. What makes you immediately skeptical about something? Let's say a company comes to you, pitches you an idea, what makes you say, no, there's a red flag there? Well I I think if they have very unrealistic expectations, if they think that this is going to solve all their problems, if they haven't really made an effort to look at its flaws. Because every there's pros and cons to everything. There's pros and cons to AI and robotics. And I think what some founders do and some healthcare systems do is they they only want to hear what they want to hear. And the people that they hire tell them what they want to hear. Yes. So I I think it I think if we if it's we we look at whether or not it's realistic, whether or not the people that are wanting to introduce this product, understand the product themselves, and if they understand the market and they understand the people who are going to be using it. Because a lot of times they understand the market really well, but they don't understand, you know, what it's like for a a nurse to try and learn how to use this product. Or they don't understand that you know no one is going to want to buy a a charge or a a charge for a particular product so they can use it more than once. So we we saw one company that they wanted to sell something for weight loss, but every time you use this machine, you have to buy another gadget to put in it so that you you can use it again. And nobody's gonna want to do that. And you know they try tried lowering the price and like that's that's just not gonna fly. So I think what they're what they're gonna do is they're they've changed their marketing strategy and they've changed their product. I think for the better, that's going to make it much more marketable and much more accessible to more people rather than just to big institutions. Yeah. And what you're sharing resonates with me so much, but I'm looking from the other side of things on the commercial side, and I've seen some companies actually making those mistakes, not understanding the very end user. They think they can go and talk to CEO or COO of a hospital and they agree to some great idea. But then when it comes to doctors and nurses, it it doesn't work. Yeah, with all due respect, I think those are the the people that know the least about what is going on within a hospital is the COO and the CEO. They they don't they really many of them, not all of them, many of them just don't understand what it's like to be a floor nurse or an ICU nurse or a scrub tech or a cleaner or a cook. They just they don't they don't understand the day to day frustrations and headaches that is associated with those different different jobs. And companies very frequently think if they have a buy in from the executive level, that's that's all done. Done deal, we're good. And they even put it in their metrics. I've seen that too. They put in the metrics and projections. But it doesn't mean that yes, you gotta buy in. It doesn't mean that people are gonna use it, or it's convenient to use, or you trained everyone to use it. So yeah, it resonates so much with me with one hundred percent. I agree with you, one hundred percent, Anna. Yeah. So what do you think hospitals and regulators should watch for maybe when the companies are approaching them? So let's give them a little bit advice or maybe direction. So obviously people get bombarded right now. by different products, especially with AI. AI tool does this, AI tool does that. What is your advice? What they should look for in in companies and products I I've always said listen to the people. so if you are, for example, if you are running an an operating room and your charge nurse is saying that they're having a he're having a big problem knowing which what to p what instruments to pull for different surgeons. And that is taking up a lot of their time. Then you can either in-house build a system that makes that that can automate that or you can contract with someone to automate that problem. So listen to the people that are working for you and help them fix the problems that they have, make their lives easier using AI or robotics. That would be what I would do if I was a CEO. I would go around to all my people and say, What can I do for you? What do you need to make your life easier? And and I think I think the executives that do that find that their morale goes up, their retention goes up, the burnout goes down, the bottom line improves, people's people are much happier they 'cause they want to be heard and they are heard. And it makes it makes for a a better hospital community. Because the p everyone knows what the problems are in a hospital. You can ask any nurse, any doctor in the ER what are the problems? They will tell you. They're they're not hiding it. It's just that I think sometimes the peop the leadership either don't listen or sometimes the the people are don't feel emboldened enough or empowered enough to make the problems known because they know that the leadership doesn't want to hear their problems. So I think that's that's What I would do is listen to your people while you have them and they can they can help that improve your systems using AI as a tool. Yeah. Once you said that, I'm like, yes, because one day maybe there will be no people to manage really. You just go and update the software and everyone gets the update simultaneously. No problem. my god. Well, thank you so much. It's been lovely. I I know we're kinda running out of time and I have so many more questions, but let's go to the blitz blitz because I want to ask you a few short questions and you can give me short answers. It should be fun. Sure. so optical engineer or surgeon? Surgeon. Surgeon. I I like I like I like people, so the novelist or the surgeon? Novelist. Novelist, yeah. I I've I love writing. Awesome. will surgeons exist in twenty years from now? As the way we know we know them. That's a that's a very good question. not not as they are now. No, I don't think so. Not in twenty years. Would you let a robot operate on you? Yes. If it was tested, proven, had good outcomes, yeah, yeah, I would. Yeah. Not the Tesla robot right now. Not not the first one that walks off the factory floor. No thank you. Wait a few years, do do a lot of testing, and then and then maybe. One thing that AI or robots will never replicate in surgery. You know, I was I was always thinking like compassion or a sense of humor, but then you think about people, you know, falling in love with Chat GPT. There's people like have small robots that they become companions with. People love their cars, they talk to their cars like people. So I think even like a relationship with a machine, it already exists in in many ways. but I think I think that would probably be the closest thing is like a real human loving connection. I don't think robots will ever have that. One thing AI will never replace in medicine. love. 'Cause like nurses love their patients, doctors love their patients, and I don't know that machines will ever have that. Yeah. And if you had one wish for how AI gets used in healthcare, what would that be? I would want AI and machines to make healthcare more accessible, cheaper, and have better outcomes for patients. I think those are the things that I would love to see. Even though I think a lot of that is not self serving and I think it could be very damaging to the, you know, practice of medicine, I think it would be overall a net benefit to people. And that's what I'd like to see. Well, thank you. That's perfect. ending of this conversation. But before I let you go, can you please show the book so we can show it to the audience and tell me how people can find you and your book and your company? Sure. This is the book. It's called The Last Surgeon, a dystopian novel. It's my first novel. So I I hope if anybody reads it they like it and they I hope it you know is interesting to people. you can get it on Amazon. There's also a website called the last surgeon dot org, which I've made for teachers who wanna, you know, have their students read this, whether they're high school students or medical students or college students. There's some free resources there that can promote discussion. It's the lastsurgeon.org. And then our company is called clinical preflight dot com. You can you can look at it and get in touch with me through there. And yeah, love love to work with people, talk to people. Thank you so much, Amir. It was such a pleasure. Real pleasure. Nice to see you and thank you. Thanks. This is AI Meets Med Tech Podcast with Anna Nestorova.