The Future of Medicine: How AI is Reshaping the Healthcare Landscape

[00:00:00] Welcome to the Chalk Talk Jim podcast where we explore insights into healthcare that help

[00:00:08] uncover new opportunities for growth and success. I'm your host, Jim Jordan.

[00:00:19] Welcome back to the show. Today our guest is Dr. Harvey Castro and he's a visionary physician

[00:00:24] and entrepreneur at the forefront of the artificial intelligence revolution in medicine. With 20

[00:00:30] years of ER experience, Dr. Castro brings unique perspectives. He's not only managed hospitals

[00:00:35] and emergency rooms, but he's also actually founded his own healthcare system. His passion

[00:00:40] for leveraging technology to improve patient care led him to become really a pioneer in

[00:00:44] the field of artificial intelligence and healthcare. He's authored over 15 books

[00:00:49] from diagnostic tools and predictive analytics to virtual assistance and robotics. Dr. Castro is

[00:00:55] paving the way for future where AI and human expertise work hand at hand to develop exceptional

[00:01:01] care and optimize healthcare outcomes, but importantly bring the human touch back into

[00:01:05] the relationship. Join us as we delve into the transformational potential of artificial

[00:01:10] intelligence and its impact on the evolving landscape of healthcare business models,

[00:01:14] products and innovations. Dr. Castro, tell me in the audience a little bit more about yourself.

[00:01:20] I have a nutty professor in that. I have a lot of the different portions of healthcare,

[00:01:25] obviously the front line work, literally location and I literally was managing doctors,

[00:01:31] managing a brand and then as well as having my own marketing and then billing and then

[00:01:37] staffing for it. And so that gave me a lot of information on how to look at it from a

[00:01:42] CEO's point of view or that angle. And then the last it was and this all describes who I am.

[00:01:48] It was November of 2022. I'm literally playing with chat GPT and I think, oh my God,

[00:01:53] this is going to change medicine. This is going to help so many people. It's going to help

[00:01:58] physicians and healthcare professional burnout. And I thought to myself, literally I've been

[00:02:02] typing on the first time within minutes, I said, I'm going to write a book about this.

[00:02:06] I literally went running to my wife and like, hey, I'm writing this book on chat GPT

[00:02:10] and she's looking at me like, what are you talking about? And then I was explaining to her

[00:02:14] and she started chuckling. She's just no one's going to listen to you. No one's going to understand

[00:02:18] this. This is like way too advanced. It's not going to catch. And I was like, hmm.

[00:02:21] So I said, you know what, I stall for a week and then after that, I'm going to write the book

[00:02:27] and fast forward. I literally wrote the first book on chat GPT and healthcare.

[00:02:31] So tell us when did it go out and where could people find it?

[00:02:33] Yeah. So that book went out January 2023 and you can find it on Amazon.

[00:02:39] Since then I have written probably over 15 different books on AI and healthcare, AI and crime, solving

[00:02:46] different cold cases using AI. And I'm having a lot of fun. Actually, this was not prompted,

[00:02:51] but I have these books next to me. And this one's the latest one. It's how to use the Apple Vision

[00:02:57] Pro and AI and healthcare because I'm just really passionate about these different types

[00:03:02] of technology and all of them are also on Amazon. So that's a lot. Let's pause and

[00:03:06] take a moment. Where do you think AI is going to fit in healthcare in general at this phase

[00:03:13] in time and maybe in the future? Because obviously you've got the goggles there too at the same time.

[00:03:17] Yeah, I see it two flavors. We have an AI basically, I see it as two worlds. We have

[00:03:23] the structured data and unstructured. And just to break it down to something simple,

[00:03:27] think of it as AI everybody's heard of it in radiology. It reads x-rays,

[00:03:31] it's seen CAT scans, it's helping. But what we really haven't seen much of is how is it working

[00:03:38] inside the electronic medical record? What is it doing? How is it helping patients?

[00:03:43] We're really accustomed to seeing those all these algorithms for x-rays. And so to answer

[00:03:48] your question, where do I see? I see that it's going to continue to grow. But now the rest

[00:03:54] of medicine is going to start to adapt. And so when I look at the rest of medicine,

[00:03:58] I look at both the patient and then the physicians. And so within the different verticals of medicine,

[00:04:03] AI is going to be used a little bit easier versus others. For example, the big one is family practice

[00:04:10] has really not adapted AI, whereas other parts of medicine have. For example,

[00:04:16] the administrations of all the hospitals are really now looking at these tools of

[00:04:20] let me help with patients. Can we create a bot that's going to help me create a schedule?

[00:04:25] Can we create different billing codes and send it quicker to the insurance companies?

[00:04:31] And so what I'm seeing now is a shift towards let's use this type of technology and let's use it for

[00:04:37] that repetitive work. And so that's the first slow hanging through. The simplest example

[00:04:42] is I'm looking at this keyboard right here. And can you imagine if I never had to use

[00:04:48] this again in the hospital? How many times have you seen your doctor literally look away

[00:04:52] from you typing because that's where the computer is at, but really it's the doctor should be facing you.

[00:04:57] And so automatically this is old technology and sadly it's not being used enough,

[00:05:02] but the simple transcription of AI going through this conversation and then putting

[00:05:07] it into the EMR. And so that's the next phase that I'm embarrassed to say that hospitals are

[00:05:12] still using fax machines. If they're still using fax machines, then this is sometimes

[00:05:16] revolutionary for some of the hospital systems. So I think when I break down AI and where its

[00:05:23] risks and rewards are, I like you think this is administrative side where really if we get

[00:05:28] something wrong it's not life or death, right? Whereas obviously in the clinical setting

[00:05:33] it is can be life and death. So I think that's going to accelerate. And then I think when I

[00:05:38] look inside of the hospital, you really see it in what I would call closed loop systems

[00:05:44] like radiology packs. You don't see it going horizontal yet. And I think that the AI that's

[00:05:50] there, even though it's been there, it's been perceived as a tool and not as a decision maker.

[00:05:55] Meaning let me scan this, let me get some insights. You think of robotics, right? How do I do

[00:06:00] minimally invasive surgery? You and I were talking before we started recording this great

[00:06:04] company called Ognetics that does virtual reality versus buying procedure. And when

[00:06:09] you realize what they do, the patient gets on the table, you have a plan and you have to

[00:06:13] move the patient. And all of a sudden that plan is not working, right? And it gets adjusted

[00:06:17] automatically and it makes you wonder how orthopedics ever really was so successful when

[00:06:21] they really are doing everything without this technology. So I think that the fear factor

[00:06:26] to me as I talk to people is this AI making clinical decisions. Do you agree with those

[00:06:32] three boxes I made? Oh yeah, I definitely do. And I'm going to throw a curveball for you

[00:06:37] to also subdivide. I hate generalizing because generalizes like it, but I see it as follows.

[00:06:42] Depending what part of the country you're talking to and who the age group that you're

[00:06:47] talking to will also determine your AI widget per se. If you and I were in Silicon Valley,

[00:06:53] man people would be like, how are you guys high-fiving us? Let's go. But I'm going to pick in

[00:06:58] Texas because I'm in Texas. If I pick East Texas and certain parts that are really remote

[00:07:02] and I go in saying, hey, I want to do this. It's your hospital. The administrators will eat my

[00:07:08] lunch. They will say no. I've had some doctors get so upset with me saying AI is the devil. And if

[00:07:14] this ends up at my hospital, I'm going to retire. And there's a lot of fear in this technology.

[00:07:19] So I'm personally taking it upon me to just go out there and talk about the good, the bad,

[00:07:23] and the unknown of AI and healthcare and how to implement it because of this reason.

[00:07:28] It's funny, the dean of Heinz College's name is Christiane. He does AI stuff at the White House.

[00:07:34] Absolutely brilliant gentleman. And one of the things that he talks about is technology is often

[00:07:39] more advanced than the policies, laws, and regulations that support it. So I think

[00:07:43] part of that fear is there's always been that the physician is a final decision maker.

[00:07:48] And if a physician has a technical issue during a procedure, there's always these review

[00:07:53] committees and the review committees have people on it with judgment too. And they talk about

[00:07:58] did you follow one of the recommended protocols? And if you deviated why and tell us about your

[00:08:03] judgment. And I think that one of the fears is that's just going to be a black and white and

[00:08:08] you make a decision, and now all of a sudden you're going to be fighting some artificial

[00:08:11] intelligence engine that quite frankly, we don't have any rules and regulations yet on

[00:08:17] what's in those engines. So I think that the FDA is sorting through that. I think we have

[00:08:22] intellectual property law that's going to be sorted through that because if I have to

[00:08:25] tell you what's in my engine, that's my proprietary stuff. But the reality is probably

[00:08:30] mostly not protected because it's input. And so I think you had your Apple book there with

[00:08:34] your vision. So what's your next step? Yeah, as we move forward, going back to my analogy

[00:08:40] that different verticals are going to be using AI differently and to your point on the Apple

[00:08:45] Vision Pro, literally the first day that the Apple Vision Pro came out, it was being used

[00:08:50] for neurosurgery. And so we already seen some verticals just quickly be early adapters to say

[00:08:57] we're going to use this tool, let's go. And so that's where I see the Apple Vision Pro. Yes,

[00:09:02] it's expensive. Yes, it's cost prohibitive to do certain things. But depending on the vertical of

[00:09:07] your hospital, how you're using it, it actually may be cost effective. And there might be enough

[00:09:13] money, for example, in a neurosurgery department, there's money there to be spent

[00:09:17] because their profit margins are pretty large. They can afford to have this in their OR.

[00:09:21] So I'm not familiar with that story. Tell us what they're using it for.

[00:09:25] Yeah, so the Apple Vision Pro, I'm starting to see different surgeons using it. I use the

[00:09:30] example of the neurosurgeon because that's the first example that I ever saw. And just to

[00:09:34] keep it simple, imagine I'm doing surgery, my hands are sterile. I got to keep it above my heart

[00:09:39] and I'm working and making sure my hands are clean. And then I'm doing surgery.

[00:09:43] If I need to look at something, say I'm working on your back and I need to know exactly,

[00:09:48] I'm putting an injection or working on you. I need to be able to visualize it.

[00:09:53] Imagine if I could, in virtual screen, a 100 inch screen, being able to see your CAT scan in

[00:10:00] real time as I'm looking at you and I'm able to see it but really big and magnify it. And

[00:10:04] then imagine if I can take your structure and just with my hands make the motion of

[00:10:08] rotating and then I'm able to see it. And then imagine if I could take a picture

[00:10:11] and say take a picture of this wound that I'm closing on, I just want to make sure. And let's

[00:10:16] say I blow it up because you see all these surgeons with those little glasses that they're trying

[00:10:19] to make it see. Imagine being able to put the Apple Vision Pro and just verbally tell it,

[00:10:23] make it bigger, smaller, rotate. And so now-

[00:10:26] I think you can see through that. That's a super sauce.

[00:10:29] Yes. I forgot to mention. Yes, sir. I forgot. Yes. So the Apple Vision Pro has a see-through

[00:10:34] and it's nanoseconds delay but your brain won't notice. But I've been using it. In fact,

[00:10:39] I'm the first person in the world at a HIMS conference to stand up on stage and use the

[00:10:43] Apple Vision Pro while I'm giving a talk because I can see through it, which is funny walking around

[00:10:48] through it. Because some people think I'm going to bump into them or I can't see through it.

[00:10:51] I think that's having been involved in a couple of robotic startups. When you work with

[00:10:56] physicians and you talk about the specifications, one of the challenges has been that the

[00:11:01] physicians say, I never want my eyes to have to leave the surgery site. I never want to have

[00:11:06] to, in theory, look up or take a second because every moment that I'm not focused on that is risk.

[00:11:12] So obviously they stop what they're doing and they're used to looking up at fluoroscopy screens

[00:11:16] or whatever. But I think the point you're bringing up is, wow, you can actually

[00:11:20] meet that vision. And it's always been a specification that we could never

[00:11:23] 100% solve with robotics and stuff because we are great visual systems but you had to

[00:11:28] look at another screen. You had this emergency room experience. How did you

[00:11:33] get this adaptation? First of all, you're an emergency room person, which is a different

[00:11:38] personality in and of itself. But tell us about a time when you've adapted these strategies to get

[00:11:43] here, when you went from being in an ER to running an organization to doing what you're doing now.

[00:11:49] I see myself as a true scientist and I'm really passionate about stuff that I do. So

[00:11:54] if I see a pain point, something that I think can be improved or even if I don't think

[00:11:59] it can be improved but I'm passionate about the problem and I want to fix it, then I literally jump

[00:12:04] in. And so that type of personality has always been in my DNA. When I was 16, obviously didn't have

[00:12:11] money for a car, didn't have money for college. So I literally started my own company and that

[00:12:15] company that I started was selling vitamins and I would make money on the margin. And

[00:12:19] I created all this passive income to the point where I paid for my first car,

[00:12:22] my insurance company back in New York City, which was expensive for a male that was 16.

[00:12:26] But to the point, I was to give you a quick example. I was in the emergency room coding a patient.

[00:12:31] I'm literally holding the iPhone one in my hand thinking, working with the nurse, I'm like,

[00:12:35] hey, this person's crashing. We need to start this IV drip. Just one second doc when I got a

[00:12:40] textbook, something around looking and okay, yeah, I got it. Let me get a calculator. Okay,

[00:12:44] got it. And I'm looking at her thinking, man, when seconds count, this is not good.

[00:12:49] So I literally took it upon myself to learn how to program. I created the first IV med app

[00:12:55] in the world and they pushed it out and it went viral. Every place around the world was buying it

[00:13:00] left and right. But because I saw a problem, something needed to be done and I did it. Fast

[00:13:06] forward when I created all these ERs, same issue as a doctor, my passion is to serve others, help

[00:13:12] patients. And then I won't say where I was working but the administration there was like,

[00:13:17] hey, we need you to work faster and we need you to push more and we need you to order less

[00:13:22] so the profit margins were higher and I look at them and thinking, I didn't sign up to do that.

[00:13:27] I signed up to help the patient that doesn't equate in my brain to do that. And I would call them out.

[00:13:32] I'm like, look at my average. I'm not like below average. If anything, I feel like I'm above average.

[00:13:38] And so that upset me to the point where I said, you know what, luckily in Texas had these

[00:13:42] things called freestanding ERs and I said, I'm going to start building these.

[00:13:45] And then I built about 20 of them and eight of them I packaged and put them as one brand.

[00:13:49] And I said, now I'm going to teach doctors how I think medicine should be done.

[00:13:53] And it went really well. And again, it went to some, my personality of here's a pain point,

[00:13:57] but something's not right. Let me see what I can do to help.

[00:14:00] Now, do you still have those practices or did you?

[00:14:04] No, I exit out of them and now I'm 100% on AI and healthcare. I'm just so passionate about

[00:14:10] this field. I feel like the more I can personally create products, put them out in

[00:14:14] the market, teach doctors, help patients, the more I feel like I'm doing more for people than

[00:14:19] just one clinic, one OCR visit. So when you work with clients, what would you say is there a general

[00:14:27] beginning status of knowledge of AI? Because one of the things that's striking when I work with

[00:14:31] some of the universities on the medical curriculum for medical school, and you start

[00:14:37] talking about that there's more teamwork, there's more digital health and how do we get

[00:14:41] those courses in, but then you recognize at the same time, you need the physician to know what they

[00:14:46] generally know. So a lot of times these pieces of education are going to probably happen after.

[00:14:52] When you're running after reimbursement codes as a physician, you really don't have a lot of

[00:14:56] time to be able to do these things. So what's the profile of your client that's curious

[00:15:00] about this? Yeah, it's literally the bell card. It's really interesting because using my bad

[00:15:06] analogy again, if I'm speaking to someone out in California, they have a different idea of AI and

[00:15:13] what they want versus other parts of the countries when they call, they literally call me and say,

[00:15:19] hey, we need you to come talk to our staff because this is a real case. Our particular

[00:15:23] state didn't want to have driver's license for the longest time because they don't want

[00:15:27] identifiers. The hospital administrators know that AI is coming but they have no idea of

[00:15:32] anything. They literally want you to come in and just tell them where it is today, where it is

[00:15:36] tomorrow. And if you can't don't use the words AI because it scares them. That's where we're at.

[00:15:41] But then I've had other hospital systems call me and say, hey Harvey, we have a particular

[00:15:46] problem and we want to use AI. Can you come and just tell us what tools are out there that

[00:15:53] could help us solve our problem? And then I've had other hospitals that say, okay,

[00:15:57] hospital across the street just implemented this way of AI. Can you come in and help us

[00:16:02] create something that we can be competitive with that other hospital? Fantastic. So how many

[00:16:09] clients are you serving at this point in time? How many people are? We've done over 10 different

[00:16:14] clients that have called us to do this kind of stuff. I'm really selective too of who I'm

[00:16:18] working with because there's two sides of this. Sometimes some of the companies are really

[00:16:23] small and they're also startup phase outside of the hospital that are wanting to get into

[00:16:27] the hospital and they're wanting for me to talk with them. And some of the startups are in my

[00:16:32] opinion, going into the wrong field when it comes to AI. For example, they are like, oh,

[00:16:36] electronic medical record. I want to create my own. And I'm like, you got Epic, you have other players.

[00:16:41] Like really if you want to do it, probably Epic, if you had an amazing idea, we want

[00:16:46] to buy you out immediately or just make their own not to call out any particular house,

[00:16:49] just saying it off the cuff. So I'm also working with different startups and saying,

[00:16:53] okay, this idea would work. This would not. But having that doctor and entrepreneur spirit,

[00:16:59] I am creating a lot of things on my own with my business partner. So for HR and AI, we're creating

[00:17:05] different communication tools for hospital systems. At the end of the day, I'm just having fun

[00:17:09] using this technology. How can I help patients? How can I do more?

[00:17:13] And I think when you read up on the state of the physician today and I have a website

[00:17:18] called Healthcare Data Center where people could find this report out there, but

[00:17:22] it talks about the amount of administrative time, the emotional exhaustion that physicians

[00:17:27] are having. And it's generally not patient facing issues. It's more associated with paperwork and

[00:17:34] flow and things like that. And so it seems to me that is the lowest fruit. And it's a ways off

[00:17:40] before I think physicians have to worry about autonomy. I think what they can do right now

[00:17:45] is get their time back and their precious time with the patient. What are the resources that

[00:17:49] you follow to keep current on this that you could share with people if they want to try to catch up

[00:17:54] here? Yeah, I do want to backtrack for your statement about physicians and efficiency. I 100%

[00:18:01] agree these tools will give patients back time, but I'm going to be a play devil's advocate

[00:18:06] and say if our new standard let's just pretend we see two and a half patients per hour,

[00:18:10] but now through AI let's just pretend we can see four. And then you're like,

[00:18:14] you're seeing more, you're not really spending more time. That's one issue. The other is

[00:18:18] I foresee hospital systems and insurance companies going back and saying, hey,

[00:18:21] this is now the new threshold if you want to get the same reimbursement. And so will the doctor

[00:18:26] end up having more time with the patient? I'm worried that maybe not. There's a whole health

[00:18:33] policy debate we could have, right? When you sit down and you look at the cash flow margins of

[00:18:38] a hospital system, a physician practice, and then an insurance company, you'll say why are

[00:18:43] the people that are watching making more profit than the people that are doing? And I think that's

[00:18:48] a policy reconciliation. And I think it's going to end up being you're starting to see physician

[00:18:54] practices leave the groups that they joined when high tech act came, a lot of people said,

[00:18:59] okay, I can't afford this tech. I'm going to be bought out. And they were swayed into

[00:19:03] joining bigger groups. And the promises of joining those bigger groups have not happened.

[00:19:08] And so it's very interesting to me as I interviewed physician practices,

[00:19:13] more in the urban areas right now, they're starting to spin out on their own. And okay,

[00:19:17] now I can get my electronic health record in the cloud. I don't need IT. I don't need a cybersecurity

[00:19:22] expert. So now I can take my power back. And I can see where physician groups are now starting

[00:19:27] to control certain regions of the country and say no to the insurance company and their

[00:19:32] negotiation powers going up. So I think it's going to be interesting to watch that play out.

[00:19:36] But I think the priorities of where the money's going is, I'll say ask backwards. It really is.

[00:19:42] I never swear in this podcast, but I think that's appropriate. So if people wanted to

[00:19:48] follow you, where would they go? I jokingly say I live on LinkedIn. So I would invite you to come

[00:19:53] to LinkedIn just simply type in Harvey Castro MD with that said any social media that you're

[00:20:00] using on all the major platforms. And again, it's the same thing as just Harvey Castro MD.

[00:20:04] I'm literally launching on Monday, my website. And it's again the same as can be Harvey Castro

[00:20:09] MD.com. But here's the kind of geeky fun part that you'll actually like. I have all these videos,

[00:20:15] YouTube's podcast, I clone my voice for my podcast. There's a way that I created this

[00:20:20] algorithm where for the last year, any blog that I see, I put into my own SAS model,

[00:20:25] it creates a picture, a transcription of fake conversation between business partner

[00:20:28] and I, but it sounds very fluid. And then we uploaded. But on the website, I just added a chat

[00:20:34] bot and it'll be live on Monday. And I have literally have uploaded all the interviews that

[00:20:40] I can find that are transcribed all my books, any blogs that I've written. And now you could

[00:20:46] go to my website and then just have a conversation with my books with me. And it's just a bot

[00:20:51] based on all my information. So you can literally, if we have a transcription of this,

[00:20:55] I can upload it and then anyone can go and say, Hey, James and Harvey met the other day.

[00:20:59] And what did they talk about? And it'll go through. Oh, it's kind of cool. So what is the

[00:21:04] help.ai is what's that project? Is that part of this project? Yeah, no, let's talk about it.

[00:21:09] Yeah, I know. I love help AI. And let me tell you what it is first. Help AI is a large

[00:21:15] language model, little device that literally sits over the patient's bed at a hospital.

[00:21:21] And it's got sensors. And the first thing people like, Whoa, it's seeing you. It's not that vision

[00:21:25] of what it's seen. It's not being transmitted. But what it does as an ER doctor, this is dear

[00:21:30] to my heart, it's helping patient and open and clinic because it's called help AI, looking at

[00:21:35] the patient and are the patient trying to get out of bed rolling? Or are they just

[00:21:39] grabbing the remote? And then patients that are fall risk, it'll any patient that

[00:21:45] cameras on will alert the nurse, the doctor dashboard, the alarm that goes outside the bed

[00:21:51] to tell patients, Hey, this patient's about to get out of bed. But what's really cool

[00:21:54] is it literally has predictive analytics in it. So you don't get all these false alarms. It

[00:22:00] literally is able to know if you're getting out and they've put all this things into the

[00:22:04] algorithm to make sure. But the coolest function that I like is that not only does

[00:22:08] it alert people, it actually verbally tells the patient, Hey, Miss Jones, do not get

[00:22:13] out of bed. Someone's coming give us a minute and to hope buy us some more time.

[00:22:18] But what's really cool is we're using AI not only to speak, but to speak to that person's

[00:22:24] language culture. So when they register, if they say, Hey, they're from Europe,

[00:22:29] and they speak, let's say they're from Spain, they speak Spanish, then the algorithm knows

[00:22:33] that it's Spanish. And then when it speaks to them and needs to talk, it'll use examples

[00:22:37] from Spain. And so that they feel more accustomed and more attuned to that. Yeah,

[00:22:41] it's really cool. And so what I'm working with the CEO, we're working on a new addition and

[00:22:48] we're going to create a home addition because I keep telling them as an ER doctor, I have found

[00:22:53] patients down three, five days, literally almost dead because unfortunately that loved one fell,

[00:22:59] cannot get up, could not make a phone call, and they were literally stuck there. And so

[00:23:04] as a result, there's this fancy term called rabdomylicense. But basically the

[00:23:07] muscle broke down, went into their kidneys, put them in kidney failure, and that gets so high

[00:23:12] those levels that it can kill them. And so I've seen literally people dehydrated and close to

[00:23:17] death because of a fall. And what we're creating is a nice tool that will go around the house

[00:23:22] that will basically alert that loved one, say I'm going to make a fictitious example I have,

[00:23:27] say my mom is supposed to be getting closer to a nursing home, but she's no,

[00:23:32] I'm an assistant living, I'm fine, but I'm living on my own. Then I could set up these devices. And

[00:23:37] then if God forbid she falls or she breaks this another part of the AI, she breaks her pattern,

[00:23:43] then I would get an alert saying, hey, mom's down, she hasn't gotten up today,

[00:23:47] or hey, she fell, no one's been alerted, I want to let you know she's been down for more

[00:23:51] than two minutes, you want to call and find out. And then I can go in and

[00:23:55] iterate called 911, all mom, see what's going on. And so I think this is going to save lives.

[00:24:01] I agree. There was a startup company we got involved in that unfortunately was probably way

[00:24:06] ahead of its time. It was probably 15 years ago, but what they did is they put sensors on the toilets,

[00:24:13] the water, the sensor in the bed and I believe the stove. And so it was looking for patterns

[00:24:19] that were breaking and they would, if they were broken, they would alert the family members

[00:24:25] so they can give a call or something like that. But at the time there was who would

[00:24:29] pay for this? There was no medical reimbursement and stuff like that. So it didn't quite make it,

[00:24:34] but it's a beautiful concept. Yeah. Yeah. So I'm really honored to be part of this.

[00:24:38] I'm able to use my AI brain, my ER doctor brain and then just looking how can we get this to the

[00:24:45] masses. So I'm a kind of person that I think behind the sky thing. So I'm already telling

[00:24:49] us, no, we're going to create all these different products. And then I want to go

[00:24:53] international as well as just national because I just think this is going to be huge.

[00:24:58] I think talking to an ER doctor about this is really interesting because when you look

[00:25:02] at both the physician shortage we're about to have and you look at the challenges,

[00:25:07] particularly in rural medicine. To me, the value of AI is bringing the major institutional

[00:25:14] knowledge. So if you go to the Mayo Clinic or Mass General or Cleveland Clinic,

[00:25:19] they've seen every tail of the distribution, right? And all the really unique stuff is on

[00:25:24] the tail. And I think as a, as our physician, you have to work so quickly, right? When you

[00:25:30] assess a patient, I can imagine that AI might be pretty helpful in saying, okay,

[00:25:37] these are the 10 things this could be in addition to your own judgment and maybe

[00:25:41] something in there is a little unique. Yeah. No, I love that you said that because

[00:25:44] that's actually one of the examples that I usually give at the end of the day.

[00:25:48] I'm human and I'm expected to be working at my 100% at 3, 4 in the morning. And

[00:25:53] maybe I'm not a night owl. Maybe I'm a day person, but the part that's really tough on our ER workforce

[00:25:59] is some today I may be working mornings, literally the next week I may be working evenings,

[00:26:04] and then the following I may be working night. And then I may be going on vacation

[00:26:08] after that and then coming back and starting again. So the human factor, I'm not a machine

[00:26:13] and as much as I try my best to be at 110%, it's still, I'm still human.

[00:26:20] And so here's where AI, I think, can hugely help us. I always say it's AI plus human is better

[00:26:27] than just human or just AI. And the case in point with me having 20 years of ER experience,

[00:26:33] my clinical gestalt is strong, but I'm still human. So if I, it's 2 in the morning and let's say

[00:26:38] I'm seeing you and I'm going through muscle memory and I got it, if I have AI to put on

[00:26:43] those symptoms here and say, hey, did I miss anything? Is there anything else I should

[00:26:46] think of? And it comes back and says, Hey, have you checked his eyes because there's this

[00:26:50] particular disease, blah, blah, blah. And then I'm like, Oh my gosh, I forgot to do that. Actually,

[00:26:55] that's a good point. I even think of that as part of the differential. Not that I'm a bad person

[00:27:00] or a stupid person is just that sometimes AI will be able to catch things. And I'm having it

[00:27:05] as a safe place to say, Hey, look at this real quick. And now let me give the best care.

[00:27:09] So particularly in ER where you have these are always very acute situations.

[00:27:15] And you go back to our sort of 10th grade science, you have a hypothesis you test and it's true,

[00:27:19] right? If you have, so you're trying to form a hypothesis to get to the answer for this thing.

[00:27:24] And in some cases where the situation is one in 250,000 people, this could happen.

[00:27:31] It's pretty unrealistic that you're going to start there. Yeah, you'll get there,

[00:27:35] but the time is your challenge, right? Yeah. And I love that you said that because

[00:27:40] time is a challenge and language is a challenge. I used to work literally at the airport in DFW

[00:27:46] that took care of every international patient that came into Dallas, that was sick. So if you got sick

[00:27:51] on any plane and landed on Dallas, you would come to the hospital that I was at always.

[00:27:55] And what made that difficult is sometimes some of the languages that would come through,

[00:28:00] I then speak and imagine giving healthcare in one seconds count that I still need to get

[00:28:06] the right translator and get on the right line, figure out the language, and then hopefully pray

[00:28:10] that person is really good at that language and able to convey those things. And so I'm biased

[00:28:16] towards AI, I think it'd be great. Now I would love to have that kind of tool. Obviously there's

[00:28:21] hip hook constraints and blah, blah, blah, but assuming that you can get that address,

[00:28:25] man, what a cool tool to help my patients in real time. I'm seeing technology that literally

[00:28:30] I just ordered, I have no stock in this company, the humane AI pin. And it's this

[00:28:34] little pin that can literally translate. So let's pretend you were speaking German,

[00:28:38] I literally could have a fluid conversation with you right now if I had it. And then you would speak

[00:28:43] in German, it would tell me what you said, I would say it in English and it would translate

[00:28:46] it to German to you. And it would be fluid. So the latest release of the iPhone last

[00:28:50] week came out with that capability. And I literally put that on my action button

[00:28:55] because it moves so quickly. So I think the Star Trek universal translator is probably going

[00:29:00] to happen in our lifetime, which is very interesting. So what's the biggest lesson that you've learned

[00:29:05] thus far in your diverse journey? To me, it's stay humble and stay hungry, no matter where I am

[00:29:10] in life, even if I'm the CEO of some big company or not, I think it's just to stay humble and

[00:29:15] true to your value and just be consistent. That's the one thing that I've always tried

[00:29:19] to be that person in it. I think that being sincere and what your passion is will always

[00:29:25] come through and people will want to work with you more. And leading by example,

[00:29:29] it's just so big. And so by staying humble and true to your cause, we'll push you to another level

[00:29:34] and people will respect that. I think one of the interesting things in your journey is that

[00:29:39] you've probably learned some of your lessons by a difference in perspective. So as you've moved

[00:29:44] from an EI doc to actually running a multi facility organization, what sort of views

[00:29:51] change for you? If anything, I understand better the business aspect and the profit

[00:29:57] aspect. If your particular organization is a profit and it's not making any profit,

[00:30:02] then it's probably not going to last long enough to keep those doors open. And so I understand the

[00:30:06] necessary evil of having to obviously you need to make some money. So that aspect, I get it more

[00:30:12] now when you're on the other side in the front lines and you're just working and helping and

[00:30:15] doing, you don't see anything but what you're doing. So being able to see the other side

[00:30:21] of the equation helps because then as I move forward, then I'm thinking, okay, what can I create

[00:30:27] that is cost effective that will actually help save money on the doctor side or the hospital side

[00:30:33] and actually make the physicians in healthcare force happier so that they don't burn out anytime

[00:30:37] soon so that they can continue to help our healthcare patients out there. So when we look

[00:30:42] at healthcare, we've got the efficiency, the cost and those aspects of it, but the healthcare

[00:30:48] workers have not had a lot of job satisfaction for probably 20 years and a lot of pressure,

[00:30:54] a lot of constraints do more with less kind of thing. My physician friends working more hours

[00:30:59] and making less than they've ever made and yet the price of medical school is continues to sky

[00:31:04] rocket nurses not having enough time to take care of the patients that way they want. How do

[00:31:08] you see AI helping with that? Oh my gosh, this is huge, huge. Believe it or not by the year

[00:31:15] 2050, the population will invert where there'll be more older people than younger people being born

[00:31:21] and so literally to your point, we will not have enough healthcare professionals out there

[00:31:25] to take care of us and I'm gonna throw myself in there because I'm 50 already. So

[00:31:29] I'm not far from needing the day where I need more and more healthcare. And so I see AI as

[00:31:35] follows. It's going to make my job better in that I'm more efficient. I'll be happier

[00:31:40] because if I don't have to, I literally have had physicians I've worked for me that have stayed

[00:31:44] after their shift for two hours just documenting because they couldn't document in time. They

[00:31:50] weren't well versed on behind a keyboard. Some of them have asked me to, hey, can we hire another

[00:31:55] person just to type for me? And so knowing that AI can help make that physician efficient,

[00:32:01] that it's not going to cost me and I'm in a leg on the administration side to do that

[00:32:04] service for that physician, that's huge. The other part is having algorithms,

[00:32:10] not many people know about algorithms that are out there that help me evaluate my own team

[00:32:15] that is able to say, hey, nurse X is probably going to quit on you in the next 30 days.

[00:32:21] And to have that algorithm put that person in my radar and then have a sit down with

[00:32:26] that nurse or doctor and say, okay, predictive analytics or tell me you're not happy, you

[00:32:31] may want to leave, what can we do now to make you happy? What can we do so that we

[00:32:36] can make this a win? And that alone has been shown now to expand those individuals to start

[00:32:42] working longer. The other part where a lot of physicians again have said at me for saying is

[00:32:47] I truly, and I've been saying this for over a year now, I truly see a day that we will have

[00:32:52] a robot that will be my assistant and that robot using this camera that I'm using now.

[00:32:57] I literally can tell you what your hemoglobin A1C is, which is your average sugar.

[00:33:01] I can tell you how old you are, what your blood pressure is. And then not only that,

[00:33:05] this whole conversation I can do predictive analytics that'll tell me if you're depressed,

[00:33:09] if you're getting depressed. Even your voice can tell me what your sugar is.

[00:33:14] And then on top of that, it can also tell me if you have early Alzheimer's or Parkinson's

[00:33:18] disease all with just this camera feed. So imagine if I had that robot next to me,

[00:33:23] then now is my assistant and then we're having a conversation and then the robot is able to

[00:33:28] say, hey, have you considered this in your differential? Or, hey, I'm going to go ahead

[00:33:32] and make up a list of labs that need to be done. And I can look and be like, okay, got it.

[00:33:36] That is going to be huge. That it's going to help. I know physically as we get older,

[00:33:41] it's not as easy to lift someone up. And so nurses typically have really bad back positions

[00:33:47] because they're listing patients. We had a robot that can help me assist the patient getting out

[00:33:51] of bed, walk in the patient. That can help save my nurses from retiring sooner because now

[00:33:57] they're not worried because their back's okay. Like they're not going to have to do anything.

[00:34:00] And as far as heavy listing, because we have a robot. So I really see this future coming soon.

[00:34:05] So I wonder, we look at the internet and the personalized messages that we get from

[00:34:10] digital marketing and all that. I wonder if AI can bring back that personal aspect of things

[00:34:16] by being able to keep this history going. I'm struck by, I did a series of qualitative

[00:34:20] interviews a few years ago with neurologists that had Parkinson's patients. And what struck

[00:34:25] me is literally brought me to tears. All these people watch these families because in many cases,

[00:34:31] they have these patients for 40 years before they expire and they're worried about the caregivers,

[00:34:36] maybe the spouse that's taken care of. And several of these doctors I talked to kept copious notes

[00:34:42] on the personal details of people's lives that were applicable. And I can imagine that

[00:34:47] if they can find all that stuff out on us for commercial marketing, there has to be a way

[00:34:51] to bring back that sort of personal approach so that you can be educated when you interact with someone.

[00:34:57] I won't mention the name because I'm starting with this company, but that's one of the projects

[00:35:01] we're working on. We're working on a project of taking data and then turn it around and

[00:35:06] do in predictive analytics and then on cue based on their voice, how they're speaking,

[00:35:11] that we can be empathetic back and explaining. So that way we can create that platform.

[00:35:16] So we're literally working on that. I wouldn't be surprised if we have something like that

[00:35:20] next three months. Yeah, I think unfortunately told the story last week so it's a little

[00:35:24] probably my audience is going to be telling it again. But I was in in stage renal disease

[00:35:28] at a period of time and I happened to walk down south into a dialysis center and nephrologist

[00:35:33] is giving a lecture to this man about he's eating these tomatoes. And he said, I know I'm not

[00:35:39] supposed to, but my wife has this secret family recipe that she does when these tomatoes

[00:35:44] are harvested a couple times a year and it's completely disrespectful for me not to eat it.

[00:35:49] And it struck me at the time that how could a physician sort of record that so when the next

[00:35:54] year comes up and tomato season starts, that he's not only maybe wanting the patient but

[00:35:58] recognizing that educating the patient has nothing to do with that. It's educating the spouse that's

[00:36:03] not disrespectful of the family that this tradition is not good for for mom or dad. And I think

[00:36:10] that's another place where we can get a little more personal or more intimate.

[00:36:14] I love that you said that because on the cultural side there's a lot of things

[00:36:18] growing up Hispanic that I don't agree on the diet portion and I've always had issues with my mom.

[00:36:24] How do I be respectful but tell her, hey, I don't want to eat that or I don't need that

[00:36:30] or I'm okay. I'm full because it's disrespectful not to finish your food. And then just,

[00:36:34] hey, mom, can you give me a little bit less this time? But it's just interesting to that

[00:36:38] effect. I want to explain again how AI would play in that role. I think a lot of life is

[00:36:44] education in healthcare. And so the more I educate my patient, the more you'll increase

[00:36:49] the way you behave and behave have better healthcare outcomes. You may have a better quality of life

[00:36:54] and hopefully you live longer. And I really believe that AI will get to a place where it'll be my

[00:37:00] companion and we'll be able to see and hear and talk to me and say, hey, that meal you're

[00:37:05] having has this many calories has this protein because they'll literally just be seeing it

[00:37:09] and then tell me, hey, this is actually has a component here that you're allergic to or

[00:37:12] this will increase inflammation in your body and you may end up gaining weight.

[00:37:16] So that whole interaction, we're not far from I know it's here now, but to make it to the point

[00:37:22] where it's literally my companion that is helping my health, I think we're getting really close.

[00:37:27] And I think there's articles that say that a lot of our cost is health literacy

[00:37:33] and teaching people things and access. So you have stories like that and then you have

[00:37:37] stories that are my daughter's a nurse. She works with the pregnant mother's inner city

[00:37:41] kind of area. And she is shocked that she literally had a patient a few months ago that

[00:37:50] boyfriend told her that if you drank red wine, you couldn't get pregnant. And this was not

[00:37:55] a teenager. This was someone in their mid twenties. And you just say, how is that,

[00:37:59] how is that education letting the system down? And the other aspect of this is these tools

[00:38:05] can allow I think access is always a big issue access to healthcare and understanding

[00:38:10] how to get access. And I think that AI can probably help a lot with that too.

[00:38:15] So when you look at the next five years, what do you see as the biggest opportunity for healthcare

[00:38:20] in this AI world with the biggest challenge that if we solve it is going to have an 80%

[00:38:26] the 80 20 rules going to have a big impact? I think if we can do a better job with predictive

[00:38:32] analytics and healthcare, that will save lives and one case in point, a New York Tron and

[00:38:39] Gator Tron are two large language models out there in New York to the ones in Florida. And

[00:38:42] the skinny is they're using all the data that's out there for their particular healthcare systems.

[00:38:47] And they're creating predictive analytics so that the healthcare physicians or providers

[00:38:53] that are their team can look at the patient and say, this person needs to stay another night.

[00:38:59] This person should go home now because it can say, okay, no, this person won't bounce back.

[00:39:04] This person is going to be okay. And that little tool seems like simple, but having that kind of tool

[00:39:10] because I'm combining, like I said earlier, AI plus human is better than just AI or just human alone.

[00:39:16] If I've had 20 years of my clinical gestalt and I'm like, no, this patient can go home.

[00:39:21] But then if I have AI behind it using predictive analytics saying, hey, these are the factors,

[00:39:26] just as this maybe you haven't considered XYZ and then I can trump that or supplement it.

[00:39:31] Now I'm creating such an amazing patient experience or better outcome.

[00:39:36] And I'm going to argue that then I can save dollars for the hospital, save time and dollars

[00:39:40] for the patient as well, and have a better outcome. Yeah, I think we're going to have

[00:39:44] a clashing of privacy and cybersecurity with this AI issue. So this is vision of a real-time

[00:39:50] healthcare system. And you realize that 80% of the preventative information is outside of

[00:39:55] the hospital system. And if you're looking to connect a bunch of Apple watches and all

[00:39:59] these systems, if you're the CIO of these institutions, you recognize every connection

[00:40:03] is an AI risk. And we certainly have seen particularly recently hospitals being held

[00:40:09] hostage. But then you have the privacy. So one of the great things about having been part of

[00:40:14] Carnegie Mellon University and the University of Pittsburgh here is that medical center

[00:40:19] is always working with the AI scientists. And so they ran a project of just

[00:40:23] trying to predict readmissions. And they had permission marketing. So obviously the

[00:40:28] families allowed access to Facebook. And they realized very quickly, they could see the

[00:40:33] dialogues among the patients and their postings and get a sense that maybe something was happening

[00:40:38] and the result was absolutely fantastic. But obviously that's a huge privacy issue. So

[00:40:44] how do you see us overcoming that? Luckily we have HIPAA. So there's a lot of healthcare

[00:40:49] professional companies that are creating products that are making sure that their

[00:40:53] products are within constraints that they're not violating HIPAA. The other portion is

[00:40:57] that these things start getting implemented at the hospital. The hospital also is worried

[00:41:01] that they don't want to violate HIPAA. So they're making sure that these devices hopefully are on

[00:41:06] secure HIPAA compliant servers. Or I'm advocating we need to create models that literally sit

[00:41:14] on the servers at the hospital level. So that way it's never outside of those servers. And it's

[00:41:18] just inside already what's secured. As we move forward, the other part to this is we need

[00:41:23] to be transparent with our patients. From a HIPAA and privacy point of view, I'm worried that many

[00:41:28] patients aren't aware that AI is being used on their health. And we need as a society to be

[00:41:34] more transparent. And then the next part to this is I'm advocating for what I call like a little AI

[00:41:39] food label equivalent where patients can be like, oh, they're using this AI and here's the white

[00:41:44] papers and it doesn't have to be super long. But it's just a page of saying, Hey, here are

[00:41:47] some of the biases, here are some of the good, here are some of the bad so that they're

[00:41:50] informed of what's being used. And then I'm pushing that we need to have the AI bill of

[00:41:54] rights on every patient floor and every room so that patients can be like, okay, this is what's

[00:41:59] going on because if it's being used for their care and they have no idea, then I think that's

[00:42:04] not right. I think they should be aware. And if they don't want widget XY AI, then they should

[00:42:09] be able to voice that and explain to the administration why? Yeah, I think it's going

[00:42:13] to be a really interesting challenge to watch. That's excellent. So what else would you

[00:42:17] like to share with the audience? At the end of the day, I want people to be excited about AI.

[00:42:22] I want them to think of it like riding a bike. The more you ride it, the more you know. And

[00:42:26] I'm not saying that you have to be optimistic. I'm just saying just take it in so that once

[00:42:31] you understand that unknown goes away. And I'm really wanting power people out there listening

[00:42:36] like how can I personally use AI or my organization use AI to help my cause help my

[00:42:41] patients be more efficient. And if those questions are something that you want to know

[00:42:46] about, feel free to call me. I'm literally going around the world giving talks about this.

[00:42:50] And if there's some particular issue that you're like, look, I really don't know how to solve XY

[00:42:55] in my healthcare system, feel free to call me. This is what I'm passionate about.

[00:42:59] And I would love to help other patients. Just for our audience, you have three books?

[00:43:04] I think I have 15 books now. I can have one that's about to be published.

[00:43:09] Are they all on Amazon? Yes, sir. They're all on Amazon.

[00:43:11] Okay, so I will, for the audience, I will put that in the show notes.

[00:43:14] Thank you so much for taking your time with us today. I really appreciate your insights.

[00:43:18] Thank you so much for your time. Thank you.

[00:43:22] Thanks for tuning into the Chalk Talk Gym podcast. For resources, show notes, and ways to get in

[00:43:28] touch, visit us at chalktalkgym.com.