Episode #03. Big Data, AI and ML Technologies, The Future of Healthcare Transformation with John Halamka, president of the Mayo Clinic Platform Risk Never Sleeps ยท 8 Feb 2023 Source: https://www.censinet.com/ Transcript: https://podcasts.opennash.com/shows/risk-never-sleeps/episode-03-big-data-ai-and-ml-technologies-the-future-of-healthcare-transformati-3f9d89/ (via OpenNash Podcasts, machine-generated) Welcome to Risk Never Sleeps, where we meet and get to know the people delivering patient care and protecting patient safety. I'm your host, Ed Gaudet. Welcome. I'm Ed Gaudet. I'm the host of today's podcast. Welcome to the Risk Never Sleeps podcast series. And today I am joined by my good friend, Dr. John Halamka. Welcome, John. Well, hey, great to be here. Dr. Halamka is an emergency medicine physician, medical informatics expert, and the president of the Mayo Clinic platform, which is focused on transforming healthcare by leveraging artificial intelligence, connected healthcare devices, and a network of partners. John, that's a mouthful. What are you actually up to these days? So if you ask yourself a question, given what we've learned in the COVID period, and I know it's not exactly a post-COVID new normal, but let's call it that. What have we seen supply chain disruption, staffing disruption, burnout. So if we're going to survive, in fact, if we're going to thrive in the next six quarters and beyond, we're going to have to embrace new machine learning approaches to augment human intelligence. And that means we're going to need robust data sets. We're going to need data that has depth, kinds of data, breadth, numbers of patients spread heterogeneity of that data and those patients. And we're going to create models and we better test those models and validate those models and tune those models. So Mayo Clinic platform was established in 2019 to give us repeatable processes and technologies for the whole MLOps lifecycle while protecting privacy and ensuring what we do is ethical and safe. So I joined Mayo Clinic in January of 2020 to lead that effort. And you joined from where? So having served in Boston for 22 years. There it is. Both, you know, as CIO of Harvard Medical School and Beth Israel Deaconess, and then working in our regional area and nationally on policy matters. Mayo has a unique approach to innovation and risk taking that seemed to be the right place to take on some of these new challenges. Wow. Riskier than Boston. I would argue. Boston has gotten timid. Right? Right. So that is. Oh, no. Listeners from Boston, listen up here. We recognize that we need to disrupt our own business models. We need to try things that may be speculative. Will large language models augment health care or be too unsafe? Does anyone know? You've got to try. You've got to try. And that's where I would just, of course, I work globally. But my colleagues in Boston have stayed close friends. And if nothing else, if I can be an early mover and show what is possible and then bring others who may be slightly more risk averse into the activity, everyone will win. Now, do you see T taking my gallbladder out anytime soon? So, no. I mean, as you probably look at GPT technologies are really at the nascent stage. Now, I think fundamentally, we will recognize that all of health care, I mean, not surgery, but I mean, cognitive health care is about pattern matching. And what does that mean? Well, I went to residency so that I would see thousands of patients. So when I would see the next patient, I would say, oh, well, based on the thousands that came before you, here are the diagnoses to consider and the tests to order. Well, to be honest, pattern recognition is something machines do reasonably well. Machines don't have empathy. Machines don't have sentience. They don't necessarily respect the needs of the individual, but they can augment. So they can say, oh, there's a 64 % chance that Ed has a gallbladder problem. Okay, great. I will, with my human judgment, order some diagnostic tests and not forget to look at the gallbladder along the way. So a tool in the tool belt, but not a replacement. No. Here's what I would tell you, Ed. Doctors, and in fact, financial administrators in healthcare will not be replaced by AI. However, doctors and administrators who use AI will replace those who don't. There it is. That's the phrase for the day, I think. I think there's a lot of information going out around what AI is and what it isn't. And so thank you. I think that clears it up for me anyway. Okay. So you have a very interesting background, and I don't think a lot of people know how you got into healthcare, how you got into IT or cyber in particular. I would love you to share with folks how you did get into healthcare. Okay. This is a slightly longer story. When I was 12 years old, my parents went to law school, and I was a latchkey child in Southern California. Back then, this was the 1960s, Southern California had something I would call free-range children. That is, go out and play, come back for dinner, let us know if you run into trouble, go to somebody's house, because there are no cell phones, of course. That's right. What did I do? I rode my bike to the dumpsters of AT &T, to Raytheon and Hughes Aircraft, because back in the 60s, none of these things were locked. And so what did I do? I pulled integrated circuits, technical manuals. I taught myself in the 1960s and early 70s, analog and digital technologies, then early microprocessor technologies. And in probably about 72, I said, you know what? I bet healthcare is going to benefit from the emergence of these digital capabilities because really any analog signal can be digitized and the human body has multiple bits of telemetry and we should analyze that. So that was in the seventies. And so then all of my education, Stanford, Berkeley, UCSF, Harvard, MIT, all worked at that intersection of engineering and healthcare with the idea that would be my career. And then how did you get into becoming a medical doctor and where did you start off doing that? So again, early on was this notion of being in healthcare and emergency medicine turns out to be a really good specialty. If you want to understand internal medicine, OBGYN, pediatrics, surgery, right? It's a very much cross-cutting kind of profession, lots of interdisciplinary training. But there's another key point. When you work in emergency medicine, it is predictable shift work. Work eight hours a day. Well, that leaves 16 hours a day for innovation, right? And so I chose medicine early, but emergency medicine because it enabled this working at the intersection of disciplines. That's fantastic. And your job today, obviously, has changed over the years. I'm assuming it's bigger than your last role. So give us a sense of the span of influence and control that you have currently with the Mayo Clinic. The question, of course, if we're going to ethically, in a privacy-protecting and risk-mitigated way, take the data of the past to find insights for the future, what do we need to do? Well, we better build a secure enclave of de-identified data that enables us to bring novel tools and innovators to the data without exfiltrating the data or creating risk. So my role as president of Mayo Clinic Platform is to oversee all of this 150 years of multimodal data on 10 million patients and use it for innovation and model development while keeping it private and secure. And then, of course, as you said in the introduction, we have about 160 partners. And these are companies that are developing novel algorithms, testing algorithms, experimenting with large language models. We also have a distributed data network. Mayo's 10 million patients. Well, it has a fair amount of depth and breadth, but does it have enough heterogeneity? Well, it doesn't necessarily include rural medicine. It doesn't include global patients, but it doesn't include a huge concentration of any one country. So we have worked out with multiple organizations in the United States and the world to have what I'll call reciprocity. You have your data in a secure location. I have my data in a secure location. Can we develop models against each other and validate them without having to centralize the data? And so I oversee the distributed data network and all of these partnerships, collaborations, and innovation activities at Mayo Clinic. And given the global footprint, where do you find yourself most often? Which state or country? Well, so this is probably not going to surprise you, but I work very closely with colleagues in Israel. Because Israel is, of course, the innovation nation. And so Israel brings some very interesting cyber technologies and protections, but also very interesting aspects of data analysis, data visualization and model creation. So obviously we work across the United States. We work in the UK, Europe. I have previously in my life worked extensively in Asia, though during this COVID time, doing a collaboration in Asia was just harder because of restrictions on travel. Yeah, that makes sense. And where's your data source coming from all over the country or just where the headquarters is of Mayo? So I understand Mayo has what I'll call a destination medical center in Rochester, Minnesota, but also in Jacksonville, Florida, and in Phoenix, Scottsdale, Arizona. We also have a location in London and a location in Abu Dhabi. We also, throughout the central part of the United States, have something called the Mayo Clinic Health System. And then we have affiliates, not owned by Mayo, but affiliates of Mayo, some 73 hospitals that are in almost every state in the country. So there you go. It's a pretty distributed source of data. So if I was at the 78 Grateful Dead show in Minnesota area, you might actually have my data somewhere? Well, how about this? I would have no way of knowing. Good answer. Just the test, John. So with the role of the job, what keeps you up at night? Mayo always puts the patient first. And so what does that mean? We always want to look at aggregate de-identified data, respecting the needs and wants of the patient. So we have a patient advisory council to say, oh, we have this idea of aggregating de-identified data and creating algorithms from it. And is that OK? What does the patient think about what we're doing? So really important that we use something called data behind glass is never leaving the Mayo firewall, the cloud container. It's fully de-identified, certified as de-identified. So what would keep you up at night is meeting the needs of the patient and using this in this fashion that I describe to ensure that you're always respecting their requirements. Great. So let's switch a little bit and outside of healthcare, outside of IT, I know you have a lot of interesting hobbies, shall we say, when you're not doing this, worrying about data, patient data or protecting patient data. What are you doing? Where are you spending your time? So, Ed, you remember that my wife was diagnosed with stage 3A breast cancer in December of 2011, and she's doing fine today. But as she began chemotherapy, surgery and radiation, she said, I want to leave a legacy. I want to make sure that we create a community benefit that gives something back and lasts beyond our mortal existence. So 2012, you and I were serving together. We purchased land in Sherbourne, Massachusetts, because in 2012, Wellesley was hot and Sherbourne was not. Who would want to live in a socially distanced, biologically isolated community 25 minutes from downtown Boston? It wasn't a thing in 2012. No, sir. And we established Unity Farm Sanctuary. And we started with alpaca and poultry and organic agriculture. And then by 2016, expanded, doubled the footprint. So today we have 300 animals on 60 acres with 500 volunteers. We have activities for the young, the old, the disabled. We have become a destination for the entire Metro West community for volunteerism and for just anxiety reduction. Right. Time of COVID. Come outside. Walk a donkey. It is very Zen like. I've been there to my listeners. I highly recommend it. Is your webpage still up? Do you have the webcams? Oh, yeah, it's the unityfarmsanctuary.org, 18 streaming cameras. We have blogs and newsletters. And please, you know, engage. We'd love to see you. Unityfarmsanctuary.org. Correct. Excellent. So check it out. It's a magical place. Are you still growing mushrooms in the back? We have 500 shiitake mushroom logs. We have, obviously, it's winter, so not a lot growing right now. Right, right. But we start with lettuces and peas and spinach in the late winter and then move on to tomatoes, eggplant, peppers and cucumbers. So we feed a whole lot of animals from the organic agriculture on the property. What's interesting, too, to me is this whole notion of obviously marijuana being legal, but now psychedelic mushrooms are legal in some states. So is your phone ringing off the hook? Well, yes, but for that, the reason you suggest. So as Ed knows, I am the nation's expert on poisonous mushrooms and plants, and I do 9 ,000 patient consultations a year. And what does that mean? It means anytime there is an overdose of a certain kind of mushroom or a toxic mushroom that creates a medical problem, I get the call. So yes, my phone is ringing off the hook just because there are a lot of people eating mushrooms they shouldn't. That's right. So for you kids listening today, stay away from those mushrooms in Santa Cruz of all places and other places. I'm sure you get them today. John, this is excellent. I love this question because it gets people to think, what would you tell your 20 year old self? Interesting, right? I've written a number of books and articles reflecting on my life, right? So I am almost 61 years old. And what does that mean? It means I've got to experience the evolution of technology, right? As you hear from the 60s to the present. And certainly in my earliest stages of my 20s, we all are a little worried about our own identities, establishing who we are. And you feel pretty insecure when you're in your 20s. like, what am I going to do? Oh my God, it's so intimidating. So I just think back, back in my twenties, how concerned I was about the place I'd end as opposed to the path. And so, you know, I'd go back to my 20 year old self and say, it's the journey that's the reward. And don't worry, you're going to be fine. I love that. I love that. That's so true. It's so interesting too, how perspective is really such a great governor on one's life and one's journey. And the more perspective we have, I think the more just enjoyable life really is because there is ups and downs and everything in between. And it certainly is not without challenges some days and other days you have just a beautiful Zen-like experience at Unity Farm Sanctuary. So here's what I would tell you is that as you point out, you don't know what each day will bring, but there's a process for everything. No matter what happens, good or bad, there's a process to get you to the next step, whatever that is. And that's the thing that has come with age. The understanding, oh, well, this looks like a bad thing. Oh, but there's a process and it'll be fine. That's right, yeah. Time is a great healer. We just give things a little bit of time. They kind of work themselves out. I think there's a lyric in there somewhere. All right, John, we're getting ready to come to the end, but I do have a couple more questions. What are you most proud of this past year, personally and or professionally? Sure, on the professional side, The role I play at Mayo Clinic is really focused on elevating health care quality, safety and access globally. And that is it's about building coalitions and communities. And so when you start to look at these last year, last two years, there have been a number of coalitions that I have been involved in. COVID coalitions, AI coalitions, clinical trials coalitions, where ultimately we've brought thousands of organizations together to make a difference for the benefit of society. So I'm hoping that when all is said and done, you could say, oh, wow, you know, in some subtle way, I made a difference, but as a catalyst for many organizations working together. You know, certainly I see that in the last year. Absolutely. The COVID work you did was just amazing, actually. pulling together that group of leaders to tackle a really insurmountable problem at the time, I think really helped a lot of people out. And one of the things that I would say personally feels very good to me is when I am traveling around the country and somebody says, you know, oh yeah, my niece volunteers at your sanctuary. Oh, I watch your cameras every night. Oh, I read everything you write and it brings me peace. It's again, that community benefit It becomes quite tangible. Excellent. Excellent. Okay. Because this is the Risk Never Sleeps podcast, I'd be remiss if I didn't ask you this question. What is the riskiest thing? And this always gets people, right? I'm going to ask them. What's the riskiest thing you've ever done in your life? Oh, well, that's quite simple. So you know that I've been an alpinist for 40 years. That's right. And so I have done a lot of rock and ice climbing. And the thing you need to know about ice climbing, and a nice New York Times article about that a few weeks ago, given climate change, ice climbing is even more difficult and challenging, is that if you fall while ice climbing, you will hurt yourself 50 % of the time. And because ice is a friable medium, it's pretty likely you're going to fall. So I did extensive amounts of ice climbing in New Hampshire in my 40s. And what you learn is when you're 20, you bounce. When you're 40, you break. And when you're 60, you shatter. And so I don't do ice climbing any longer now that I'm in my 60s. Mount Washington or? Crawford Notch. Crawford Notch, yeah. Or North Conway. Yeah, nice. Cannon. I mean, there's a number of ice climbing. Beautiful area. Beautiful area. All right. Well, very good. John, that pretty much wraps us up for the podcast. Yes, I really appreciate you joining us today and spending time with our listeners. Again, thanks, everybody. Thank you for your service, John, obviously, and on the front lines of health care and patient care. And for everybody else, you know, stay vigilant. There are a lot of bad actors out there. And remember, Risk Never Sleeps. Thanks for listening to Risk Never Sleeps. For the show notes, resources, and more information on how to transform the protection of patient safety, visit us at censinet.com. That's C-E -N-S -I-N -E-T .com. I'm your host, Ed Gaudet, and until next time, stay vigilant because risk never sleeps.