Hype, Pilot, ROI: What Health Systems Get Wrong When They Buy AI

Attractive right now.

There you go, he just fixed it. Now we're live.

Good morning, everyone. Uh, welcome back to Signals and Symptoms Podcast. I am extremely grateful for our good friend Dr. Young, um, Alan, uh, and of course my bestie Harvey Castro as our co-host starting this podcast. Alan has had a long career in understanding the digital healthcare system, has served at multiple institutes, uh, moving forward. As a matter of fact, if I start going through his profile, it's going to take a whole hour just to introduce him, and I'm going to misspell and mispronounce some of the things. So I'm going to let Alan introduce himself first, and then we're going to start with a very basic question. Given his profound, amazing, intelligent experience over the decade on what the digital health evolution has been And what does he see in the future as the most important things that's going to help patients, providers, cost-effectiveness, value-based care, etc.? So it's going to be a discussion, and that's why we want to bring people like Alan onto this show, is that they provide the historical perspective. And that is important to see in the age of AI. How fast are we going to move forward, if any, or is it going to be helpful or hurtful. So as usual, uh, I am extremely grateful to bring in, uh, Dr. Alan, uh, Young. Alan, if you don't mind letting our audiences know your history and everything, and then I'm gonna ask Harvey's first question, why, and then we're going to start the conversation. Great.

No, good morning everyone, especially friends on the West Coast. Um, thank you, Dr. Castro. Thank you, Dr. Kalia, for having me. Exciting to share this topic because I think, as you pointed out, I've been in this space for over a decade. And I started off as a clinician in training, like both of you. My area of study was orthopedics, and I moved to Kaiser to practice urgent care as a per diem physician. And then that's where I discovered there was a new evolution taking place with the electronic health record adoption. And so being trained at Epic on— at Kaiser, I realized that the practice of medicine was headed in a direction where it would fundamentally change from what we had seen for the previous century or so of doctors and nurses and everyone capturing things on paper and storing them in paper charts. And so that really intrigued me. And I saw that the future was gonna be more digital and technology-driven. And as a result, because of my background, I went to medical school at USC, but did a dual degree MD/MBA program. I was allured into the business opportunities and people said, hey, you know, you've been practicing physician, you work in technology at Kaiser, and you have an MBA. We could use folks like you in consulting. And so that really, uh, opened the door for me to explore the opportunity of working in healthcare consulting. And so I left full-time clinical medicine and continued to practice part-time for the next decade or so, but joined a consulting company that was focused on helping implement these electronic health records into health systems. And so that was the start of my journey. As you can imagine, we've kind of come full circle. I'll explain kind of where I'm at now, but that experience working with clinicians and implementing technology and showing them what to do with an electronic health record really set the stage moving forward. I moved into medical group leadership at USC and then joined leadership at Accenture to deploy our Epic consulting services on the West Coast. And then shortly after, I was in Los Angeles and learned about artificial intelligence. A good friend and mentor, I think of all of our, Dr. Anthony Chang, gave a talk, and this title talk was Artificial Intelligence in Medicine. And at the time, this was before ChatGPT, before COVID before doctors could even type on a keyboard. So I was very skeptical. I didn't see AI was going to revolutionize medicine as quickly as it might have been, uh, but I was intrigued by the topic. I jumped in and pretty soon found myself pulled out of consulting and joined an early-stage tech startup that was using artificial intelligence, machine learning to create the first-generation chatbot, which led to doing, uh, patient triage through automated calls, call center work, and then the digital front door experience and helping organizations reach and improve access for patients through who didn't have enough doctors and nurses and medical assistants to reach out to. So that kind of set the stage of learning what it's like to be in the startup space and working with artificial intelligence and saw that there was a lot of resistance at that time. Organizations were skeptical of AI, they didn't trust it, they didn't see the need, they didn't definitely understand the ROI if there was any. So that was a very unique time of the early adopters for AI. Fast forward, you know, COVID came and went. I spent some time at McKinsey and Company doing management consulting at the corporate strategic level and kind of understood, you know, what people are thinking of when they're making big decisions to buy either assets or technology or talent. And so that's kind of led me back into the startup world. And now I work back at an AI clinical data intelligence platform called Evidently. And the problem we're trying to solve is now that we have these EHRs embedded across most major health systems in the country, if not the world, we have too much data. We have too much data, not just from the EHR, but from everything else associated with it. We have genomic data, we have claims data, we have social determinants of health data, we have wearable data. There's just an overwhelming tsunami of data, and, and no system depending only on clinicians and nurses and medical assistants, etc., can process that data. And that's kind of where it led to where we are now.

Wow, thank you, Alan. Thank you, buddy. That was really good. Wow, where to start?

Uh, so our, uh, another bestie just joined, our co-host, Isha Dr. Manring. She's a pediatrician. Uh, Dr. Young, she's here in town. Of course, we are biased, as you know, we're Texans. And we are in Dallas. Anyways, I just make that joke, uh, but, uh, I'm, I'm glad that she's here as well. Harvey, I'm gonna let you start because he unpacks so many things. Uh, number one, the, you know, the journey of EHRs from, you know, the last 10 years to now AI. And then, uh, of course talked about data silos and now, uh, you know, data, uh, essentially normalization, standardization in the age of AI is, is one of the core issues that multiple startups and And Evidently, of course, is one of the major ones that are trying to solve and quite extreme importance. And I cannot really overstate what Evidently is doing. Allan, we should come back and, you know, one of these days help you, you know, explore the whole Evidently demo and everything and feel free to talk about it as we go about. Harvey, go ahead. First, your question. And then, of course, I have so many tons of things to learn from Allan here. I apologize. Is it okay to call you Allan? Because I know we meet—

Of course, we're all friends here.

Yeah, I apologize.

Sorry about that.

Okay, sounds good. Uh, so Harvey, go ahead.

Yeah, well, I'm gonna start with asking everyone listening to type in your questions in there because I want to start now because I know there's gonna be so many, so many questions. And then for me, uh, my favorite question, and, and Junaid knows this, is, you know, let's just start from the core. Why? Why, Dr. Young, why do you do this? What, what is it about this that's really pushing you? What is your personal why? Because you're doing so much and there's got to be some Something that strong why could be pushing you to do all that.

Yeah, no, great question, Harvey. And I tried to keep my introduction short, but just rolling back a few years, I'm actually Canadian. I'm an immigrant that was born and raised in Canada, primarily near Vancouver, and came down to California to finish an undergrad degree. And that's when I discovered kind of the opportunities in front of me in medicine and business and realized how complex the US healthcare system was at that time. Coming from a country like Canada, we had a single-payer system, obviously much smaller, maybe 1/10 the population of the United States, and saw the kind of, you know, enormity and, and was frankly a little bit intimidated that, hey, one day I might have to work in this healthcare system. And I did not understand it at all. And I think that led me to be curious and, uh, wanted to understand, you know, why are medical decisions made? How do health insurance and government influence these decisions, and how does that limit or improve access for patients? And so that started my journey of understanding both the business side and, and the medical delivery side. And then as I started my journey, I was also on paper. I was writing in paper charts and scribbling my signature that no one could read, and then looking through other people's charts and couldn't decipher what the, you know, cardiologist said or what anyone was writing. And so wanted to try to fix that problem. And as we moved into the electronic health record, I was somewhat savvy with computers. My dad was a computer engineer and moved to Silicon Valley before I finished high school. And so had a, you know, inkling towards technology and thought, yeah, of course we should be using keyboards and capturing data. But then as I started working, I think it struck me that the data that was being captured was kind of growing and mutating and expanding, right? Almost like a malignancy. There's just too much information. And even when I was a physician at Kaiser, I'd be frequently audited and told, oh, Dr. Young, you didn't document enough, or you didn't put enough specificity, or you didn't note that this patient received this thing from an outside physician in a consult note that we had scanned in after they faxed it to us a week before your, you know, the visit. And so you didn't look through all that information. And I think that just started to trigger a light bulb to say, hey, this is going to be a significant problem as we think about the, the healthcare system in this country and, and the requirements for physicians, and then the ongoing and growing demands of patients that are becoming more technology savvy. And then obviously we have to worry about the litigious, you know, nature of Practicing medicine in America, that if you miss something as a physician, you are ultimately responsible for the care of your patients. So that kind of was the why of why I wanted to like explore how AI as one technology could solve some of these problems. And I realized again, there's just not enough human beings, there's not enough hours in the day for people to actually go and review all the information and absorb it and make sure that they're making informed decisions for their patients.

Oh my God, you— I love it that you're here. I'm so grateful. I'm gonna let Isha ask the second question then. Um, I need to learn from you a lot. Anyways, uh, Isha, go ahead.

What do you think, Alan, has been one of your biggest challenges in this process? And I guess, what has it taught you and how has that kind of shaped your direction?

Thank you, Eesha.

That's a very thought-provoking and personal question, but, you know, we're with friends and it's early in the morning, and so I'm happy to be—

I'm sorry to throw that out there. You know, it's just—

I've had my coffee, so I'm okay.

We have so much to learn from one another in this process where there's so much new to explore and for us to all educate each other. So I'm sure, you know, a lot of our listeners will find your responses very helpful in their journey as well.

Yeah. And that's probably one of the reasons I'm willing to share. I think the biggest challenge I've had to overcome was, was making that decision to leave full-time clinical practice to become a consultant. And I maintain my medical license. I still practice medicine at Kaiser, but that, that deviation from the traditional doctor path, which I'm sure many of us on this call can understand, and maybe many of the listeners that, you know, I felt like I had a a bigger calling or a different calling that I wanted to make an impact at a bigger scale, especially coming as an immigrant to this country. I felt like, hey, you know what, I'm practicing medicine, but if I wasn't here, Kaiser Permanente could easily replace me with a PA or another medical student or a resident or a fellow. And so I didn't feel that I was contributing, you know, in a meaningful way on a consistent basis to the practice of medicine. And so when I left and to explore consulting and kind of never came back to full-time clinical medicine, that has been the biggest challenge because I think it allowed me to understand that, you know, within healthcare we do have a lot of limitations in what we're seeing as physicians. And it wasn't until I stepped outside and looked back or looked around that you see that there's so many other areas that we as clinicians can impact and we don't need to be you know, practicing medicine to have that impact in healthcare. So that was the biggest challenge, but it also opened up the biggest opportunities for me and my career.

I love that, and I think that's such a great message because we tend to box ourselves into certain roles as physicians, as clinicians, and we may not even realize how much more there is going on in medicine, especially today, and how you can impact your patients perhaps on a grander scale if you step outside the box. I think most people, rightfully so, are afraid of the unknown and afraid of change. And I think it takes somebody who is really forward-thinking and brave to step out of the more comfortable and known role and step into something that is going to push us even further. So I have to really admire you taking that step, and thank you for sharing your story.

And of course, take care of your finances, especially in this day and age of inflation. You—I mean, what we call lifestyle bloat—is something that post fellowship we need to or residency we need to help ourselves. And I'm going to be a little crass here, but that's what our good friend from SOP. always says that, you know, have what we call fuck you money available just in case. So if you want to really move into a different career, you need to work hard, save the money, make sure that you have some cushion and then move into. And I'm glad, Alan, you shared that is something that I did personally, worked like my ass off as a director of neurointensive and neurocritical care, director of stroke, and then moved to Texas. more importantly, also doing financial planning in terms of tax savings and everything, and then decided to open up Save Life AI. Harvey is another example as well, and Isha is working on her own platform for patient education as we move forward. Thanks again, Alan, for showing it. I'm going to move a little bit odd, and these are 3 questions bundled into one. Presumably, that you have everything going what we are going at this pace, which is insane pace, by the way. And we can talk about type-safe decision analysis and CDSS and where, you know, honestly, data is becoming so much more important. But let's say 2030, what do you— how do you see the healthcare system, uh, with, you know, a sort of a tiered system in which the data silos still exist, nothing shit happens because of either, you know, medical legal or— data silos, or we are gonna, you know, because it is already happening, transfer some boundaries and how the system would look in 2030 and walk backwards to today in terms of how we should move forward to help patients, providers make money, which is very important to— and I'm very honest directly as well, transparent. And how do you see, Alan, the future And then walk us backwards.

Yeah, great question, Junaid. I think, you know, the title of this discussion is Hype, Pilot, ROI: What Health Systems Get Wrong When They Buy AI. And I think that is what's happening now. So maybe in 2030, organizations will have sharpened their focus and kind of have a clear vision of what AI can do for them. in their unique organizations. Every organization, health system is slightly different in how they function and the decisions they make. So I would hope that within a couple of years in 2030, organizations have like solidified their strategy and are executing on this strategy to deploy technology such as AI in a meaningful way that helps them maintain their margin, right? Stay healthy as an organization. support their labor force, supporting their clinicians, uh, their staff, their executives, and then also giving and delivering hopefully better outcomes for patients and better experiences, right? That's the kind of future state vision 2030 that you asked me about, but I don't think every organization is going to get there, right? I think everyone is in a different adoption phase. Uh, there's kind of the early innovators, the early adopters, the second movers, the, the slow, the laggards, and the, the the negative kind of detractors, right? So some health systems and hospitals will be able to achieve that outcome. Many will probably stagnate and fall into the middle where they may or may not— they're gonna see mixed results of their investments or their focus on technology such as AI. And then some are just gonna be out, frankly, out of business or gobbled up and consolidated by other entities that see that they're distressed asset and need help. And we see corporate medicine has kind of expanded its presence across many parts of America. I'm not sure what it's like in Texas, but definitely I've seen and heard and have felt, you know, the chagrin of my fellow physicians in California where I think this is happening. And so taking a step back now, what, what's going to happen in the years to come to reach that goal? I think you're seeing definitely a strong appetite for health systems to AI. They're either building it internally, they've either had somehow had the the talent and the workforce already. Maybe they knew about this, like like Dr. Castro and Dr. Anthony Chang were preaching this many years ago. So some health systems kind of were preparing for this. Others have recently acquired a lot of talent, brought in data scientists, brought in AI experts to try to help their organization move forward. And then others are just ex you know, outsourcing it to consultancies, to other technology companies, to AI, you know, experts to help them create and build these products. Um, so that's happening at different health systems, hospitals, to a varying degree. Definitely a more openness to try AI as opposed to pre-pandemic or pre-ChatGPT. But then you still struggle. No one really knows how to measure the success kind of accurately, right? You may deploy AI and you see some improved, you know, metrics of some sort. However, the underlying cost is not always clear as well. So you have to then have a period of education and readjustment for your CFOs, your CEOs, and COOs to see, is there going to be a financial impact that they can measure and that they agree with, right? And I think a lot of people have not also baked in the cost of Compute, the cost of cloud, the cost of data centers. These are all things that are beyond, above and beyond just the SaaS subscription or turning on your advanced AI for, for your physicians. This is definitely a kind of larger expense line item that sometimes gets overlooked. And I think we see as organizations are signing contracts with each other, some of this fine print is getting lost kind of in the details. And unfortunately, it sometimes comes back to deter future investment because I think people are caught off guard by the actual expense it takes to actually deploy AI at scale.

So much to say, so much to unpack. Let me still clarify one more thing. One of the users also commented recently that these data silos and EHRs are still a big mess to connect to. And, um, there is protectionism, but of course then, then there is, you know, HIT governance laws. And every year, uh, you know, uh, I think Obama signed it and now Trump increased the availability of the data and all of that. And such a lot of miscommunication and confusion because, I mean, HIPAA is bullshit, right? It was a port of portability sort of law, which we, I don't know what it matters in the age of AI, especially when the document itself is no longer the source of truth. It's just a summary which is created by AI. And then of course, who owns that AI? Somebody from an intellectual property perspective. We are in like a fricking wild, wild space. But in your opinion, Alan, do you see these data silos coming down or do you think that there will be this oligarchs and— not oligarchs, but gatekeepers still continue to influence future development because we now know clearly. And, you know, you can talk, Harvey, you wanna give 2 seconds of what you're doing with your own health and everything and data access? And then, you know, Alan can answer because we are seeing what we call AI-enabled people, which are, you know, the first curve on the adoption cycle. But then, if you need mass adoption, we need mass data access. Harvey, you want to go first a little bit? You're on mute, I think. I can't hear you.

Can you hear me now? Yeah. Okay, good. Sorry about that, guys. Uh, Alan, I don't know if you know this, but, uh, I created my own digital twin. All my wearable All my genetics, all my information, my electronic medical records, my down to my health, my schedule, my emails. And what I've created is that digital health. And then I created an AI agent that goes out there on the web and finds best practices, clinical trials, things that are about to that have proven medicine, and then it layers it on top. And then I get daily reports on hey Harvey, you need to work out like this. Don't work out like that because you didn't sleep. XYZ. And the reason I share that is I really feel like that is the future for everyone. Everyone will have that digital twin set up for them. So I just love to see what you guys are doing and what your thoughts are.

Yeah, Harvey, a great example of, again, being a pioneer in the field and embracing AI and, you know, learning about it and deploying it. I think that is, to me, going to be the single greatest barrier for widespread adoption, is that probably 80 to 90% of the population will not be as proactive as you. We've seen it in behaviors, why this country is primarily overweight or obese, and why, you know, there's a lack of exercise and more screen time and social media, right? So I would hope that we could move and have more, you know, people inspired by what you do. But that, as you mentioned, The ability to then create a digital twin, you've kind of gone through this, but there's so many different sources of information. Like, you do not go to one place to get information that you're going to use to ingest to feed your twin. Like, even the wearable device you're having is one source. Then you have your, you know, electronic or medical record information. Then you have your changing activity that you're doing daily and decisions that you're making. And I think healthcare systems and enterprise are dealing with that same problem. You know, Junaid mentioned data silos, and I think that's going to persist no matter what. A data silo happens when you have data created in an environment where it's not openly available to everyone. And so every technology company, every vendor, every system will have their own proprietary data. The challenge is though, now you have this huge growing volume of data. Right? And so there's just no way to kind of contain that in any one single container. We even talk about the cloud, but, you know, there's now challenges around cloud cybersecurity and people are being fearful. They don't want to have access to that. So things are moving into on-prem again and data centers and everything else. Then you have to question the veracity of the data, right? The truthfulness and the quality of it. Like you can have 1 trillion teraflops of data, but if you say 20% is inaccurate or, you know, incomplete or partially— maybe a small percentage are hallucinations from the AI world— what is the, what is the value of that data set, right? And so you're going to have this system where you have immense data that needs to be somehow aggregated, analyzed, to pull out the correct insights, recommendations, predictions to help decision-making in the future. And some of the things that are most challenging are looking at discrete versus non-discrete data, structured versus unstructured data. It's fine when it's in digital text and numbers, but what if it's, again, a handwritten note that was scanned in from a medical record a few years ago and is sitting as a PDF, right? No human being is gonna have time to go look through every single file to try to find this handwritten note to realize, oh, the patient had a major, you know, diagnosis that was suspected or had an imaging study many years ago that showed the early onset of something that should have been handled, right? So all these hidden gems are there. I think what's really exciting about companies like evidently the one I joined is they are applying sophisticated AI to be able to read all of that data, understand the structured and unstructured data, and then apply it through the lens of, you know, a general clinician to say, is this valuable? Do I— do we— does another clinician need to know this? And how can I synthesize the data and share it in a way that allows like a Dr. Castro to go and make a decision, or a Dr. Mannerheim to help a pediatric patient? Rather than have them or even their human assistant go and read every single file to figure out what's in there. And then they end up creating a summary document that has everything in there, and then it doesn't do the physician any good that they have to then go read a summary that's not contextualized for them. So I think Evidently is moving in the direction of being able to capture large, vast swaths of data, applying context and reasoning through specialized AI and surfacing the necessary information to allow clinicians to make decisions at the point of care. And so that, I think, is, you know, just like Dr. Castro is doing with his digital twin, that's what's happening. All this data is coming in, he's programmed agents to pick out what he thinks is most important, then he makes a decision based on their recommendation and decides if he's going to do the exercise or not, or how long he's going to exercise, or is he going to walk or bike ride or swim today. or eat vegetables or eat protein. It's one of those things where he's a living example of how having the right data at your fingertips can help you make informed decisions on a daily basis.

That's great, Alan. I just wanted to jump on this question since we're talking about future and we're in 2030 now for this discussion, and we've already seen some of the challenges in terms of accessing data. Where do you see the private practices, you know, in 2030? Like, do you feel like private practice is going to have more of a disadvantage, and are they going to exist in this world going forward?

Yeah, that's a great question. And I think, you know, my prediction is they will exist, but the ones that will exist will have found a way to establish a competitive advantage when it comes to, uh, kind of patient selection of why they would go to a private practice versus going to a major health consolidated health system.

Yeah.

I think that, you know, consumers, patients, their needs and desires are evolving. I think there is a genuine appetite to have kind of point-of-care, at-your-fingertips personalized service versus going to an organization and having to kind of, you know, conform to their requirements of how to receive care. You have to have health insurance, you have to pay a premium, you have to go see these limited doctors. I think there's going to be a market for folks that say, hey, I want to have control over my healthcare, and I want a clinician that can help me in that journey, and maybe a clinician that can do more than just see, take care of one of me, right? So I do think there will be a continued need for like private practice physicians. However, um, I think the, the changing demands of the patient population is going to dictate which one of those private practice clinicians will be successful and able to survive as costs continue to go up, inflation continues to hit, people's access and insurance coverages are getting dropped. But there's definitely been movement. If you look at direct primary care, concierge medicine, there's people that are successfully practicing in those fields and taking care of a large patient base that, you know, to some of them, they're, they're limited because they can only take on so many private patients before the quality starts to diminish and their lifestyle starts to get impacted, right? So there's going to probably be a growing demand in those places, but it's going to be probably focused on a select patient profile, if you will, because not everyone's going to want that type of service. But I think not everyone, uh, definitely not everyone is, is going to be available to access the large health systems as well.

Yeah, I, and I agree completely. I think that's something that private practices do need to pay attention to. How are you going to set yourself apart? And they need to pay attention to these conversations too, because a lot of patients may want some of the advantages that come with being part of a large healthcare system in terms of access to some of these great tools. But at the same time, they want that personalized care. So I think it has to be the right blend of both. And I love these comments coming in too. I think it's so great to hear from our listeners as well.

So Harvey and me always have a little bit of a contrarian workup. So just giving you an example from a technology perspective, I agree 100% with Alan. Initially, you're gonna see a significant jump in cost. And then like, for example, right now, you know, Save Life AI got major funding, got their own GPUs. My cost of delivering AI is now essentially going down to zero. I mean, just FYIing you because things have changed and understanding the future of AI, like the way I see it is that we always need humans, period. And one of the questions was about, you know, human in loop. period. There's nobody moving them, but we're going to enhance them with AI employees, which is essentially agentification of the record, agentification of the process, etc., etc., and essentially all agentification of workflow. But I agree with, uh, Dr. Young that you're going to see increasingly a process, and this is what happened when electronic health records became a problem. All the, you know, all of them had to come through this, and then there was a you know, carrot and stick from the government and everything. And then as we move forward, as you can see now, basically technically EHR is free as long as you buy RCM solutions for them. And I think that may take, as, as Dr. Young suggested, some time. But I completely 100% agree that initially, because of the technology's new initial, uh, what I call CapEx demands— when you talk about business terms, the OpEx and CapEx is is going to go slowly downwards. And then interestingly, Isha, I believe that this will be ushering in a new era of independent physicians breaking away from large organizations because the technology cost is going to go down. So I, I don't know. What do you think? I think it's— he's 100% right initially, but hopefully the same thing, right? I mean, if you really look at it, if the iPhone was initially what I don't know, released at $500. With today's money, it's going to be $1,800. What I was able to get iPhone at that price to iPhone at this price is essentially a supercomputer. Did you understand what I'm trying to say in, in this iPhone analogy? Harvey and Alan, what do you think about that?

No, I agree. It's, it's crazy fun. Uh, somebody in the comments put like the best world is being able to combine your medicine, your entrepreneur slash creativity with, say, technology. And that has just changed my life personally. No, like I just told you, the digital twin, I'm doing it for my personal health. I am— the older I get, the more longevity means to me now. So I'm trying to see how I can stay young or look young, and I'm leveraging technology. But literally, I can just vibe code it now. And today it's quote unquote easy, but tomorrow it's going to be so much easier. I think the next world will be literally in the near, very near future. I wake up in the morning and I just talk to my computer and I'm like, hey, how am I doing? And the digital twin is basically telling you, and it's already integrated. The fact that all of us on this call have the same tools, yet we have different output, speaks a lot to our creativity. And so I'm going to encourage people listening, if you haven't already, start playing with these tools. If you haven't already started coding, start coding. If you are coding, code more. The more and more you do, the more and more you improve. Quick example, everyone out there was scared probably to ride their bike. Now you can ride that bike, you're not even holding the handles because you know how to ride your bike.

100% agree. Alan, uh, thoughts on the technology adoption curve, AI adoption curve, and the cost curve over time?

I think technology adoption, from my experience initially with the healthcare industry, looking at electronic health record adoption, I was kind of there on the forefront.

Yeah.

And initially it was met with resistance. You know, clinicians did not want to change their existing workflows. They either were happy doing paper charting or had even developed their own homegrown EHR and wanted to stick with that because they kind created and customized something that worked for them. There was obviously a mandate, you know, with meaningful use and incentives for organizations to move to electronic health record. And I think that spurred the adoption. Obviously without the incentives or the penalties, it would've probably been a lot slower. We are not seeing the same level of resistance with AI. I would say that early onset of certain tools tools that help really improve the clinician provider work-life experience, such as ambient AI technology, has created a, a pull-through effect where clinicians now are expecting the organization to invest in this or are choosing where to work as long as they have these tools.

Okay.

And so there's a little bit more momentum saying, hey, these are actually proven already to beneficial versus the EHR world. There's like, well, you're doing it because you have to, or we're doing it because we're gonna be incentivized to do so. This adoption is different.

Huge difference.

This is very more kind of functional and impact-driven adoption. However, as you said, there is a cost associated with adopting new technology. I think that's where the resistance is because there are no clear incentives, you know, for organizations to invest in AI. They have to make these choices independently. They are being spurred on by competition. So if a major health system across the street is suddenly saying, we've announced this great AI tool for all our doctors, the other, you know, competitive system probably can't wait too much longer to also make the similar investment, right? Or else they're going to lose some of their competitive advantage and maybe lose clinicians and nurses who are going to go across the street and work at a place where they have agentic AI doing reading their documentation or helping with calling patients back, etc. Um, so the adoption curve is different. The cost curve is still going to be, um, something that's unpredictable. I know you said the cost of AI has gone down, but it's interesting. I'll pull the example of, I think, ChatGPT. ChatGPT offers it for free. We all jumped on, downloaded it, probably like Muse and Dot now, the agentic versions. Everyone's downloading it for free and trying it, and that's the hook, right? That's the hook to get you into the system. Then you realize, oh, I've used my limits. I still need to do some scheduling. I need to do my grocery shopping. I need to send emails. So I better pay for that $20 a month now. That's not a big deal. Then folks like us are like, we need more compute. We're doing things like managing databases of information and running digital twin studies, right? So we need the premium. So Dr. Castro is probably, you know, happy to pay the $200 a month. But guess what? They now upped the ability for the AI and are now charging a $500 a month premium level, which opens access to all of their greatest models that are basically coming out every day. So the $200 users are now thinking, oh man, like, I'm happy, but I'm, I'm not performing, the models aren't performing as well, or I'm not getting access to the latest and greatest. So that cost to that user goes up to $500, but to them it's still worthwhile, right? So you're creating this value-cost kind of, you know, paradigm where we're willing to pay for it because we think it's valuable. And I think that is diminishes in our minds the cost versus before we had to pay for something we didn't agree with, right? So I think that's the difference between AI now. But there's going to be a cost. Like, there's no doubt that these companies are out to make money. The reason they're raising so much money from investors, because the investors think that, you know, the, the population in the market will be able to bear the cost of paying for this. If every person in the world paid $100 per per month for AI, that's a lot of money, right? So we gotta think there's gonna be an inherent cost overall, but to the individual user, it might be negligible because with their AI, maybe they can generate money, maybe that supports their business, maybe they're saving cost. But where is that cost displacement coming from? And that's a whole other topic that we could get into. But, you know, there is a price for the AI that we're using.

Well, time is an asset too. And, you know, you mentioning that there's not as much resistance resistance to adopting AI technology, I think also speaks to how much of a need there is and how much this sense of burnout there is too. You know, when we have tools that can help us get home faster and help us just generate what we need to do in a more timely and just, you know, more organized way, I think we all sense that there's a need there.

So another bestie, Edward Marx, of course, is not here, and he always talks about what we call soft ROIs and hard ROIs. And this is where, you know, this shift is moving. I mean, if I'm sitting as a CEO of a hospital, I'm buying a documentation software, do I actually— do I have to go to my CFO and my board and say, you know what, because of this quote-unquote $2 million a year additional, uh, investment, am I gonna get, uh, additional sort of, uh, funding because the physicians are proven to see more patients, which again is a bigger problem. Are they really going to be able to see new patients, or are they going to go out and play golf? Because I'm talking about— as I'm not talking about as a physician anyways, I sit at 2, 3 tables right now, so I have to juggle my hats. Now we come to the secondary, the soft ROI, and that is like, I need to make that $2 million uh, dollar, uh, investment because Junaid Kalia is the only one who has neurocritical care stroke and epilepsy certifications. There are only 21, 22 kind of physicians over here. To replace him, uh, because he's gonna leave, is gonna cost me somewhere between $700,000 to $1 million. So for God's sakes, give me the $2 million so I cannot have to replace him. That's the soft ROI and the hard ROI picture. And, uh, uh, that's where, uh, I think, uh, The situation is coming down in a lot of boardrooms, as Edward tells us. I mean, of course, I'm not in a boardroom, just being honest and transparent here. But Edward tells us because he's sitting in hospitals' boardrooms and these questions are happening and then going from there.

But it has to be a win-win. It has to be a win-win for everyone involved, you know, and if they're going to adopt this software.

So, yeah, and this is interesting. I mean, people don't know hospitals run at a very small margins. I mean, as a matter of fact, a lot of hospitals, if they don't get grants, they're actually in losses. So, uh, but, but that, that's, that's— but Alan, now you talked about— and I'm gonna step back again, and then I hate the word client and a customer, but, but just forgot, just as a whole, uh, let's just call patients. as a customer service sort of process. And I just wanna— I hate that word, I apologize for it, but I just wanna use it because, you know, as a customer, my experience and my needs, what I expect from Amazon, Google, and even a tire shop down the street when I have to check my tires, I don't wanna call physically, you know, pick up a call, make an appointment, and then go. I need an app that I can just say, find me the latest appointment, blah, blah, blah. It already knows that I have a Tesla and a Volvo, already knows the tire specifications, and then orders it, moves on. These are my expectations changing from the world itself, given that we are living in essentially a new world. Alan, Harvey, Isha, it's all for 3 of you to comment on. What do you think in the next 5 years, Is gonna be the expectation change if you slightly, I hate the word client, customer, but using the client expectations in this terms of patient expectations from systems, processes, et cetera, how it'll change and how will it impact, then Alan can finally say, how will it impact the buying decision at the end and the ROI at that point? point, because that's where the— if— because, okay, nobody cares about Dr. Kalia and Janae leaving, but they're gonna care about 1,000 patients leaving.

So— Yeah, I'll jump in real quick. So big picture, the way I see this is, given the technology, given the blockchain, given the power that we have, I truly think we're going to start seeing outside of the US, uh, finding solutions that they're doing because of the regulation, the politics, the top-down effect. Quick example, like Singapore, they're like, this is what's going to happen, and the whole country does it. They have their genetics, they have their information, they can create AIs, they can push and do things for healthcare way faster than we can, even though we have the technology. Because of that, patients— uh, when I used to work for the Ministry of Health for Singapore, they said, hey Harvey, patients are moving so fast, we can't even regulate this stuff. What do we do? And my point is, because again, they have the power to create, they're leveraging that information. And so I think that is the future. I think many patients will be empowered and they're going to be creating products like myself, and that will become— that will push the industry to say, you know what, we need to do XYZ because there's obviously a demand. Thoughts?

What do you think, especially your experience is very different Because you are a pediatrician, you're, you're, I don't know, you, you deal with not just the patient, you deal with moms, which is much harder. Oh my God. So, so how do you, how do you see, I'm sorry, I'm going to remove so that the focus is on you guys.

Well, and, and I think too, just as a consumer of healthcare, you know, what, think about what, what we would want. You know, you, you don't want to keep repeating your history to everybody. You don't want to have to track down your own documentation from specialists, you don't want to keep having to rewrite the whole thing every single time. So you want it to be seamless. You want to walk into your physician and you want them to already know— you already want them to know your concerns. You already want them to know your history. And I think that is going to be more of the expectation as we go forward. Um, you know, we're going to expect that All of that is already there, and we're you know able to to get to the point to a solution faster. I think that's going to be part of the expectation, and and hopefully my hope is that it will also increase the level of connection between physician and patient because it's not like you're getting to know them every single time, every visit. You know, it's like you already have everything in front of you, and you start. from a higher place. And I think that that should be, you know, part of our goal with all of this.

Alan, what are your thoughts? Um, you're working with data essentially, and you have like so much experience. I mean, wow, from EHR. Are the people using their My Health application? Because some of the comments was like, nobody's using the EHR from the patient perspective. Why not? And how do you make the experience better? Or what would the experience better look like? What the experience in the future will look like and how we'll go back on the ROI decisions?

Yeah, great question, Junaid. I think in the end we're still going to see a challenge around what I call and what the industry calls misaligned incentives, meaning that executives at health systems are going to make buy, buy and sell decisions based on what the financials may, you know, project for them, right? So thinking from a business side, how can you increase revenue, right? How can you get more money for the care you delivered? Um, one example is to making sure you're capturing accurate documentation that justifies medical claims, that prevents, you know, um, denials and maximizes reimbursement, right? That's one way of doing it. Another way health systems and corporations do this, and you see it in the headline every day, is they reduce expenses, right? So how can you deliver care at a lower cost? And so with the pressure on supply chain, labor, inflation, gas prices, right, that's hard to do. But what have they started doing? And this is the unfortunate thing, you know, that they've started to reduce their headcount Because FTE carries a significant cost to a health system, right? But once you start doing that, you then start to impact the ability to deliver care. You can't probably do as many surgeries or procedures. You probably can't see as many patients in the traditional way. And you also, you know, decrease or lessen the patient experience. And I saw people in the chat mention that, that You know, that is gonna be a big indicator for whether or not patients decide to continue spending their money at an organization. If, you know, health system XYZ says, hey, you know, your wait times are now 6 months to see your PCP, and to have surgery is gonna cost you this much out of pocket, and you also have to find a parking lot where there's no parking and, you know, take a— take an Uber to come to our hospital, probably not gonna be the best experience. However, if there's an alternative AI technology-driven solution that is being offered, as Dr. Castro mentioned, in other countries where you can engage with a clinician that could assess the data that's available in a shared system, can give you recommendations or prescriptions and care, prevent you from having to perhaps travel and be there and be responsive to your questions through an agent or a digital twin. That is gonna be a big decision-making, you know, piece of information that will make people say, well, why am I going to this health system when I can get care from this alternative? And I think that's where you're gonna see this kind of, you know, perfect storm of conditions where, you know, organizations are trying to make money and save money, but at the cost of losing kind of their customers. and the customers are now able to take on, you know, care in alternative channels which traditionally have not been, you know, well accepted but now are becoming much more acceptable. Um, and but what's that going to do? It's going to be this, you know, continued swirl of like, where does the money go? Do you try to retain those customers, or do you as a large health system now enable those channels and give similar You know, comparable patient experiences at a lower cost to your patients to retain them, right? Um, the money is going to flow in healthcare. Unfortunately, the people that are capturing it sometimes are not the best stewards of the money. And so the reinvestment into healthcare, I think, is where we're lacking.

So beautifully said. I love it. I'm sorry, I'm gonna— my bestie, another bestie's not here, he's flying. Ed always says that by 2030, 60% are going to move. He also says that the adoption is going to be tiered in the sense that— and which I'm going to give you my personal example, I called— my dad is calling me and says, well, I need to go see this patient. I said, you don't. I mean, you can just ask ChatGPT and I'm gonna set it up. He said, no, no, no, no, no, I don't like that. And then my brother, who's 18 years younger than I am, he's saying, why the hell am I going? There's also this adoption curve of age perspective, Millennium versus, you know, Baby Boomers, that we're gonna see. And that's also gonna do— and again, and again, uh, and my perspective is same. is basically that we're gonna see new forms of concierge care and that is gonna catch up. And that's exactly what Isha was telling about, that at some point in time, this alternative care paradigm is gonna figure out. Finalize your comment and all 3 of you, last question of the day is that how is it gonna impact insurances? Because when you have this alternative care and then AI is helping, Am I really gonna just be, you know, moving from medical insurance to what I call hazard insurance? And things are gonna adjust because why pay for a colonoscopy, uh, you know, uh, what is that down payment, whatever the initial cost of $8,000 I have to pay, uh, I can get a cash colonoscopy as Dr. Young was saying, that alternative is $1,200 down the street from my friend. And it's available to everyone, not just me because he's my friend. So what I'm— these alternative treatments is going to highly impact this ROI perspective as well. So how the insurances are going to change in this regard. Harvey, go ahead first.

Okay, Alan, you just said something amazing, and I'm going to just expand on it. Just to give you a quick example, the bus rides that we send the vouchers for our patients— we don't know what we don't know. Leveraging AI, this one hospital system looked at their— the patients and was able to convince the city to change the bus route to the hospital system so that they could leverage that for their patients. And I thought, man, that was really smart because they were, they were looking at the data. Second thing I want to say is, because we don't know what we don't know, I created a website called chatgpthealth.com. It's the exact name of ChatGPT Health product. I just bought it a couple of years ago, and it has all the governance, everything we need to know about AI in healthcare, and encourage everybody to look at it. And I'm not advertising because I'm selling any ads on it. I'm just advertising it because it's going to help with AI healthcare to this talk. Now to the question that everybody— Janae just asked, what's going to be the future? At one point, I used to own emergency rooms, hospitals, and we self-insure so that we could help our patients and our employees and teams. And so what's going to happen with AI I truly already am seeing the writing on the wall with insurances leveraging AI, leveraging the risk, being able to look at the risk to another level and being able to create that self-insurance. So I do believe there'll be another pocket of insurances leveraging AI. For example, this isn't FDA approved, but my voice right now, you would be able to know roughly what my blood pressure is. You would know my age, you know my hemoglobin A1c, emotional intelligence. And with those kind of parameters, I really see the future of AI insurances. And I know that opens up a whole can of worms, but that is going to be the future.

Dr. Young, do you think insurance is going to change as well as we grow with the demands of the consumer expectations as well? Or that might be a new startup we should work on together?

There are, there are new kind of health plan startups However, I think the big insurance companies will continue to do what they can to protect their turf. So you're going to see this kind of constant battle. You know, we live in a capitalistic society where we sometimes don't discourage monopolies, right? There's certain industries where there's— it's dominated by one or two big players. In healthcare, health plans, for example, there's some significantly large large health plans that control physician groups, um, they're going to persist. They're going to find a way to continue operating. They're going to change and evolve and, and come up with new tactics. But I agree, there will be, you know, an emergence of new alternatives. Um, you know, it behooves me coming from a country like Canada and having family from a country like Taiwan, you know, there's universal healthcare. They found a way to provide basic healthcare coverage to everyone, right? And I know the US is complicated. We're bigger, our patients are different. I've heard all these arguments, but there is this future, I think, where we can provide at least a basic level of care, which would then go a long way in, in serving a healthcare system, not a sick care system, um, and also be proactive and help prevent a lot of, you know, diseases and conditions that otherwise have continued to plague this country and drain a lot of the resources for these chronic care conditions that could have been prevented at an earlier age, right? If they had proactive, conscientious healthcare and people had access to, you know, if it's not doctors, at least the AI agents that could give them the right advice and help guide them on their, you know, health and nutrition and, and kind of activities of daily living. So In the future, there'll be a change, but I don't think health insurance plans are going away anytime soon.

Again, I'm so grateful to Dr. Young for coming up, especially on the West Coast. It's too early. You have no idea how grateful I am. Thank you so much for taking the time. The things that we were not able to discuss in this particular podcast is the medico-legal aspect of AI. And then we would love to basically bring you back again. again and discuss that in your experience, how medical legal and regulatory changes impacted the overall digital health course, including adoption of EHRs to now essentially creating both the regulatory and of course, more importantly, medical legal framework, if you're willing. And again, extremely grateful to you that you came up. And of course, we would love to invite the Evidently team to come and present their solutions on this podcast, demo it, talk to the CEO, talk about his challenges and his solutions to it with your presence, and would love to do so as well. Everyone, thank you so much. We are out of time today. Again, very grateful to Isha, Harvey, and Dr. Young for coming, and hopefully Dr. Young comes again and shares his insights with us. Thank you so much.

Thank you.

Thanks.