How Can AI Help Solve the Misdiagnosis Epidemic?

[00:00:00] Jared: Hello, and welcome back to another episode of the med+Design podcast, where we journey into the heart of healthcare's most innovative and transformative ideas. Today, we have the privilege of hosting Dr. Steven Charlap, a name that resonates deeply within the realms of medical innovation and artificial intelligence.

Steven's multifaceted expertise with an MD and an MBA to his name positions him uniquely at the crossroads of medicine, technology, and business. From his mission to save millions of lives to his pioneering work with SOAP Health. Steven's journey is a testament to his dedication to revolutionizing patient care.

His early forays into AI, long before it became a buzzword, showcase his visionary approach and his ability to anticipate the future needs of patients. In this episode, we'll unravel Steven's insights into conversational AI, the transformative power of the perfect medical interviewer, and his perspective on the ever changing landscape of healthcare.

So let's get started on this fascinating journey. Welcome to the show, Dr. Charlap. Happy to have you.

[00:00:57] Steven: Thank you. Appreciate it.

[00:00:59] Jared: Yeah. So just getting started talking about your background a little bit. You started off as an MD later on, got your MBA and got into business as an entrepreneur.

And it's interesting because for people like the regular folks, like me, you think of physicians as being a very stable job profession that pays very well, more than the average Joe. And for a lot of us, you'd think, Hey, being a physician would be absolutely fantastic in and of itself. And then going into business is something that's a little bit more of a risk on endeavor.

And you've successfully navigated that time and time again throughout your career. What motivated you in the beginning to make that transition?

[00:01:39] Steven: I think if you go back to college, actually, I was a speech and drama major and a film minor. And so what I most enjoyed

when I was younger was things that involves creativity. I ended up going to an undergraduate college called Yeshiva University, which is a religious institution. And many of the students didn't want to become doctors. And I fell into the wrong crowd of close friends who wanted to be physicians.

And the next thing I knew I was the only one with a non science major that was applying to medical school. So I've always been a little bit more eclectic than the next guy. Yeah, I love that.

[00:02:16] Jared: And so you have a really interesting mission statement, which is saving the millions of lives of people that you've never met.

And so where do you feel like you are in that journey so far?

[00:02:25] Steven: Look, I describe myself now as the singularly most obsessed person on planet earth to prevent misdiagnosis after having lost my brother, a practicing cardiologist to a misdiagnosis. And then subsequently I lost a very dear friend, my wife's best friend, another physician to a misdiagnosis.

And finally, I almost lost my own life to a misdiagnosis. So for me, this is an obsession right now. And losing my brother was so painful that I don't think anybody should have to experience that when it's clearly misdiagnoses are preventable. So we are already in market. We're working with some of the biggest names in health care, such as the National Cancer Institute, University of Miami, One Health, Mayo, etcetera. We are determined to build software and applications that no one has seen before. We have a broad patent on the use of a digital human as a medical interviewer. That means that it talks to the patient and the patient can talk back and it can recognize virtually anything that the patient might say.

It's been trained over seven years, even though the company is only four years old, with over 100, 000 unique terms. That includes slang, idioms, synonyms of all kind of medical terminology.

[00:03:43] Jared: We do have a lot of follow up questions there. And I can see why this has been such a personal mission for you.

And I really do hope and I'm excited to see that eventually this technology really changing people's lives and, leading to longer sustained lives. I want to dig into SOAP Health's vision first. Before that, I do have a question that also is just around the value of an MBA as well.

You're a physician that got an MBA, the debate around the value of an MBA, I feel like changes based on who you talk to. Some people are like, Oh, it's all about, networking and other people are like, Oh, I really wanted the skillset that you get from it. And particularly as a physician what, where do you feel like the most value is out of an mBA?

[00:04:24] Steven: So first of all, my reunion from HBS was last week, my 35th reunion. So I got an MBA a long time ago when it wasn't so common for MDs to pursue a joint degree. I was in a surgical residency and during that residency, I would often be exposed to senior attending physicians and medicine had changed in 1982.

They instituted some radical new regulations that required patients to leave the hospital quickly. And everybody was complaining. And then one day, the chief surgeon in my program got up at what's called the grand rounds, where all the faculty and students are in attendance. And he said, it's the Harvard MBAs who are ruining medicine.

And he completely demoralized the staff. And of course, at that point, I had been accepted already. And I always say that what I heard during my residency was that the doctors were complaining that business people were making decisions that were impacting medicine and the business people were complaining that doctors didn't understand that medicine was a business.

And so I said to myself, if I'm going to have a greater impact, I'm never going to have as an individual doctor that if I can get both degrees, and I'm going to understand both perspectives, and I'm going to bring a unique perspective to changing healthcare. So the MBA, I found invaluable. I'll tell you on a funny note that when I told people that I was leaving medicine to go into business school, people said, I don't see it.

You were such the doctor type. And by the time I finished business school, people said, I can't believe you were a doctor.

[00:05:59] Jared: Oh, man, love that. Just so humble and easy to talk to, that's why. So talking about SOAP Health and the vision for it. Nowadays everyone and everybody now they're all talking about A. I. Chat GPT changed the world in regards to the everyday average Joe actually getting to use this technology and see firsthand the potential.

And for yourself what inspired you to actually go into this, I think it was 2019 or so to understand that AI was the solution that you were, trying to get after for early disease detection and, the prevention of misdiagnoses.

And also has the rise of AI made the adoption of your product a little bit easier as a grease, the wheels a little bit for conversations now?

[00:06:42] Steven: I tell you, it's actually the opposite because everybody has attached the word AI to their efforts. And so all AI, regardless of it's true AI or fake AI or advanced AI or beginner AI, or someone who played with generative AI for 10 hours, everybody's now an expert in AI.

Those of us that have been working on it for almost 8 years now are thrown in the same lump as those who have been working on it for 10 minutes.

[00:07:12] Ty: You have heuristic now that you mentioned that between to separate out the 10 minute experts from the 8 year experts. Is there like a quiz question you have for that?

Or how do you.

[00:07:23] Steven: There isn't. It comes out in the course of a conversation. For example, I consider myself an expert in diagnosis. I recently attended a podcast by somebody presenting diagnosis information from the society to improve diagnosis and medicine. And he finished the podcast earlier and he allowed questions and I bombarded him with questions.

He couldn't answer. And he said, boy, I got to talk to you because you seem to know this topic better than I do. And I've actually now become friends with the guy. But the point of the matter is, if you talk to an expert long enough the difference between an expert and a would be expert. I've been working on this for 8 years.

And in fact, after my brother passed away, I went to Stanford for 3 and a half years and then 2 years at Harvard to basically gather the knowledge and understanding of solving the problem that killed my brother, which is why did doctors not identify his cancers earlier, despite an extensive family history of early cancers and deaths.

And the answer is doctors don't spend enough time with patients anymore. And there's all kinds of consequences of that, but it's made worse by physician burnout, administrative burdens, financial incentives, et cetera. So the problem is medicine isn't what it used to be. I think it's beyond repair, and it needs some really significant overhauling in order to be improved.

In fact, I just wrote something on LinkedIn yesterday that insurance premiums have gone way up 7 percent just since last year, $9, 000 for the average patient, $24, 000 for the average family, huge deductibles. It's not even one of the top topics of conversation for the upcoming presidential election. So we got a broken system that costs more than any other healthcare system in the world, but we're not ranked anywhere close to number one in quality of care, and it fundamentally needs change.

And AI is the first true promise of change to come along since I went to medical school in the 1980s. I believe that.

[00:09:27] Jared: Oh, something I also saw on LinkedIn. Was it your birthday, like yesterday or a couple of days ago, two days ago? Yeah. Yeah. Happy birthday to you. Forgot to mention that. Something that you were alluding to within that was also talking about part of SOAP's mission is to bring the joy back to medical practice.

And I think that was a really interesting point as well as. Just in what ways has the joy of medical practice, been diminished over time for a lot of physicians like yourself and many others?

[00:09:56] Steven: Yeah, I went to a cardiologist the other day and he was sitting in the corner staring at the screen during the entire appointment.

Why? Because if he didn't finish the note by the time I walked out the door, he said he wasn't getting home for dinner. So where's the joy in that? Where's the joy in knowing that you don't have time to provide the absolute best quality care to your patients because you won't make a decent living. If you don't basically move the patient assembly line along. It's unfortunate, but it's realistic and it sucks the joy out of practice.

[00:10:34] Jared: I'm just intrigued how you're drawing a line between, the research you've done in AI, and then the solution you've gotten to, to try to address, because we've heard this time and time again about how physician burnout, how there's 26 hours of work required each day in order to complete the tasks put on physicians these days.

[00:10:52] Ty: Yeah. Paying more for health care than and getting less value back than we ever have before. And I'm just curious how you're setting that up.

[00:11:01] Steven: Look, those are very well known problems about physician burnout. But here's what people don't know. People don't know

there's a study to show that more than half of Americans lie to their doctors about things that might embarrass them. So doctors are not getting the full story. They don't know that five weeks ago, Johns Hopkins published a study that shows that 375, 000 people die every year from misdiagnosis and 400, 000 develop permanent disabilities.

They don't know that calculates to every 40 seconds somebody dies or develops a permanent disability. They don't know that the National Academy of Medicine now predicts that everyone will experience a diagnostic error. They don't know that for every 100 doctors, there are 97 lawsuits. They don't know that according to the American Medical Association, 65 percent of doctors are sued by age 55.

And I could go on ad nauseum, they don't know that adjusted for inflation, primary care doctors make less money than they made in the 1970s. They don't know that there's no financial incentive for a doctor to prevent disease or earlier detect disease. They don't know that diagnostic errors cost over a trillion dollars.

And if you take that trillion dollars away, you know where that comes from? Out of the pockets of people. They don't know how bad it is. They don't know that one out of 100 doctors attempted suicide. 10 out of 100 thought about suicide. 27 percent report depression, 55 percent report some symptom of burnout and 63 percent report at least one symptom of burnout.

They don't understand how bad it is. And how hard it is to be a doctor, and how few doctors recommend their children become doctors themselves. We're in a state of crisis and it's not getting the attention that it deserves.

[00:12:50] Jared: No doubt. And it goes back to even how, for example, if we were just to go our family medicine doc, and you get to see them for maybe five to 10 minutes max in a lot of cases.

And that's the amount of time that this physician has to really get to understand your issue that's going on here. Within that 10 minutes of time, you're talking about the margin of error, there's razor thin.

And so it's no wonder that we're having this sort of like misdiagnosis epidemic, I guess you could say across healthcare. And for talking about early disease detection, I also am someone that could have been a a statistic in this as well. I almost had sepsis from misdiagnosed appendicitis that my appendix burst. I had been sick for some days leading up to it. The physician that we went to told me that, oh, you're what was it? That I had a stomach bug or something like that. But I can imagine that would be a lot harder to diagnose and get it right but where SOAP technology I think comes into play as cancer detection, heart disease, type two diabetes.

Is that really the focus of SOAP's technology and your innovation so far?

[00:13:59] Steven: Our focus is on all of medicine, you mentioned something very interesting. You talked about this 10 minutes. That's the crazy thing. Our competitors think that they need to copy those 10 minutes. In other words, they create an application where the intake is 5 to 10 minutes long.

You can't do a good job. A human being can't do a good job, nor can an A. I. application do a good job in 5 to 10 minutes. That's why our application on average starts at about a 25 minute intake because there's a lot of things you need to know about a patient. And I'll give you a perfect example. A patient comes in with a stomachache.

Your example. Okay. Pain in the abdomen. Now, knowing whether or not you had an appendectomy is important to rule out possible appendicitis. Knowing that you have an extensive family history of colon cancer should increase suspicion even though you're a young guy, there happens to be an epidemic now of early colon cancer and you suddenly are presenting with abdominal pain, the doctor now should think about colon cancer as one of the possible diagnoses because of your extensive family history of colon cancer.

Most doctors don't collect detailed family histories, so they miss early cancers, heart disease and infections. What you're talking about is very common. It's one of the common misdiagnosis. The answer is not trying to duplicate what a doctor does under pressure. The answer should be building something that's comprehensive, accurate, truthful, and here's the most important part, effective.

Cost effective, time effective. That's what we need. That's what I'm obsessed about. That's what SOAP represents.

[00:15:43] Ty: You mentioned that the majority of patients will lie to their doctors. If it's something that they find is embarrassing, I'm curious, you mentioned truthful is a key part of how you're getting at that.

And so one of my favorite books is called the mom test, which is where if you're testing an idea that you have for a new product and you go and you ask your mom whether or not that idea is a good one, she will lie to you. And the whole premise of the book is how to basically get information without setting yourself up for being lied to.

And so I'm curious how you've gone about and addressed that problem in particular from like gathering truthful information from patients.

[00:16:20] Steven: That's a great question. But what got me excited about SOAP and why it took me several years before I formed the company was waiting for the technology to catch up with the vision.

It turns out that when human beings talk to a digital human, the brain lights up differently. This has been shown on scan studies. And what it does is you're not afraid of being judged. Why do patients lie to doctors? Because they don't want to see that look in the doctor's eye or the words coming out of the doctor's mouth that says, you need to eat better, you need to exercise more, you need to stop sleeping with multiple people, you need to stop the drugs, you need to do this, you need to do that, don't do this, don't do that.

People don't want to hear that, not even from their doctors. So they don't tell the doctor the truth. But when they talk to a digital human, studies show that for some reason, they're compelled to tell the truth. And not only that, but they're actually compelled to speak at greater length, the normal etiquette of waiting for them to respond and not respond.

You just open up and you're more forthcoming. But that's not the only way we get accuracy. We also get accuracy by reflecting answers back to users. We also get accuracy by using intuitive imagery that helps the user, even of low literacy level understand what's fundamentally being asked of them. For example, if we ask you, do you have a productive cough?

The word productive cough is a medical term. Many patients may not know what that means. So we show them a picture of somebody holding a tissue with a little glob on it. And we say, are you bringing up anything? And there's a picture. They understand. Oh, yeah. When I bring up something, I cough it into a napkin.

That's an intuitive image. That's why we're obsessed with every detail of the application and people know they show me an image. We do all our own custom art and you show me an image. I might iterate on that image 37 times to make sure that it conveys exactly the sentiment that we want the user to understand about what we're asking them.

It is that obsession with detail that differentiates us from other people.

[00:18:26] Ty: Yeah, that's very much a human centered design approach to things and just making sure that people are really going to understand what's being asked of them like that's so few forms that I've encountered take that into consideration.

That's brilliant.

[00:18:39] Jared: And so something also that I think about as well as the sort of like fear of AI across generations and the various levels of acceptance. And in testing the product did you notice people not wanting to, they're like, I want my human back in a way, or did you, have you seen people being more open to it and just really accepting of the technology?

[00:18:59] Steven: Yeah, so I sometimes give a lecture titled The Rise and Fall of American Medicine. What comes next? And at the start of the lecture, I asked the audience, how many of you would prefer to go to a human doctor who needs to be scheduled, that may not always be accessible when you need him or her to be accessible, that costs a lot more money and is human, therefore prone to error?

Versus an AI doctor that's always accessible, infinitely patient, sometimes funny, validated to be accurate, and costs a fraction of what a human doctor. And most of the people in the audience choose the human doctor. Then I tell them everything that's wrong with medicine, and I re ask the question. And it shifts overwhelmingly over that they prefer the AI doctor.

Look, I wrote an essay on LinkedIn called Medicine Needs Skynet. Now that's a reference to the Terminator AI powered monsters, but the point is not the homocidal tendencies of Skynet. But the scientific advancement and diligence of Skynet to accomplish its mission, that's what we need in medicine. And there's a very common phrase that goes around that everybody's talking about that AI won't replace doctors, but doctors who use AI will replace doctors who don't.

And the analogy to that is we all know that superior AI chess computers can destroy a human being right now in a chess match, but what a lot of people don't know is that a human using an AI computer will beat an AI computer alone. So it is the power of the connectivity between the human and the computer to ultimately perfect the process or be at the highest level of the process.

[00:20:44] Jared: Speaking of, perfecting the process. The technology now you've launched it. I remember, before you used to talk about how you were holding on to it to make sure that you got all the bugs out. And where are you now? And where, what kind of got you to that point to where you feel comfortable now, giving this to the patient side of things. I remember you said before you were just very cautious to let patients interact with this before it was ready.

[00:21:07] Steven: Yes, I'll tell you, I'm not comfortable yet because there's so much more that we can do to improve it, but it has reached a level where I think physicians can use it judiciously to improve patient care, and that's why, based on the principle of first do no harm. I believe our application will do no harm at this point.

It is not yet been optimized to do the most good, but we know exactly what we need to do in order to do that. It's just a matter of time to implement all the improvements that we want to implement, but it's now at a level that it will do no harm. And that was the minimum threshold that we needed to reach for me personally as a physician to be comfortable to allow other doctors to use it.

So we've been getting a lot of questions in the Q and A, and I'm not even sure where to start with all of them. I guess we'll go with something that's a little bit more on the simple side and talking about the feedback that you've got from physicians about how SOAP Health, saves them time and prevents burnout.

[00:22:07] Jared: The thing that I think of that revolves around that is also like the SOAP note that I saw where it can integrate into the EHR really quickly. I thought, man, that's incredible because when we talk to a lot of physicians a lot of their pain points are around the EHR in general, and so the fact that you've already automated a way for your SOAP note to accurately input into the EHR, I think that must be one of the key pain points anyways.

[00:22:29] Steven: Yeah, look, there's multiple pain points in practice, starting with no shows patients who don't show up for the appointment. It turns out that patients who complete our application always show up for their appointment because now they're invested into the appointment and that makes a big difference.

Okay. When a doctor knows that a patient is invested in his or her care, because a lot of people basically give up responsibility to doctors. They just want a prescription. They don't want to really do anything for their own personal well being. But this application is so thorough that it really engages them getting them to think about things that they may not otherwise think about and identifying the problems that might be missed.

For example, we always talk about this one case study of a very young ethnic woman. I won't give the details in a area of the country where many poor people live, who presented to a clinic complaining about a headache. And she kept talking about the headache on and off for a few days. She talked about some weight gain.

The doctor saw an obese woman in front of him. She talked about some anxiety. She talked about some family history, whatever he asked her about. And based on that intake, he wrote, refer her to a neurologist. What she didn't tell him or the nurses or anybody else for that matter, but Genie, the digital human asked her, had she ever been sexually abused to which he said yes.

And Jeannie then asked, is it continuing? To which she said yes. And then Jeannie asked, would you like a referral to a specialist? To which she said yes. And Jeannie took a far more detailed description of her headache than even her doctor did. And Jeannie also identified that the reason there was an obese woman sitting in front of this doctor, who didn't even bother to ask her how much weight she gained, was because she had just gained 15 pounds and crossed into obesity. So based on this additional information, the doctor canceled the referral to the neurologist and instead referred it to mental health counselor, but Genie wasn't done. Genie also identified that her relative risk for type two diabetes based on a father and mother who were diabetics in their 40s, and her crossing into a BMI above 30, which is obesity, that her relative risk had increased by 450 percent, and her absolute risk for type 2 diabetes had now increased to 30 to 70 percent absolute risk.

And in addition, Genie still wasn't through. Genie flaggd that she might also have something called polycystic ovarian syndrome, which often goes missed for many years among young women who suddenly gain a lot of weight. Genie completely changed the trajectory of this woman's health care. That's incredible. Yes, it was incredible.

It is incredible. And that's the type of things that SOAP Health's application can accomplish. And I'll give you one other case study, a woman who carried a notebook into every appointment. The doctor never got her full history because there was never enough time in a single appointment to talk about everything she had in the notebook.

But she spent two hours. with Genie loaded in her entire medical history for the first time the doctor could see it all. And the system flagged that not only was she an increased risk to type 2 diabetes, but she was also an increased risk for Lynch syndrome, which is associated with colon cancer, particularly among younger people.

And in fact, he used a SOAP note for the prior authorization to get approval for the genetic testing of that Lynch syndrome. This is the power of AI applied appropriately, judiciously and a no bs way.

[00:26:11] Ty: That's phenomenal. You think about in order to gather that level of information, there's no way you can do that in the 10 minutes.

And yet, by allowing that reflective time for the individual, like you say changing the mindset that patients have when encountering their physician to then allow them to take more personal responsibility and then come in and use that to triage and flag and prioritize what the most important information is.

It's just phenomenal what you've come up with. And I was thinking about the human centered artificial intelligence, which is one of the areas of research. And their tagline is humans in the group, AI in the loop. And it seems like you're using that interaction with the doctor as that last step quality check to make sure that all of this information that's getting sorted and prioritized is truthfully what needs to happen.

And then to empower the next step in healthcare change. That's just such a amazing, phenomenal kind of solution you've come up with.

[00:27:06] Steven: Yeah. And you mentioned earlier about does AI give us some type of advantage? We are lumped in the same pile as every AI company that's using generative AI to do anything, despite the fact that we've been at this, as you mentioned at the beginning of this program for far longer than most people ever imagined using AI in everyday use. And generative AI clearly has some advantages, but we've experienced it's hallucinations. And so we've taken steps to control for that

by going to other large language models using the same generative AI search capabilities to search those models to reconcile different patient information that typically is not caught by a doctor. And I'll give you a very good example. So a patient presented with the past medical history of high blood pressure and a chief complaint of heartburn.

Genie asked, what medications are you taking? The patient said, I'm taking a medication called Benicar, which is typically used to treat high blood pressure, and I'm taking a drug over the counter called Omaprazole, which you can also get a prescription at a higher dose. Okay, using our technology, we identified and matched up that the Benicar is an appropriate treatment for the high blood pressure.

But here's where it gets freaky. Our application identified that the omeprazole used to treat heartburn can actually cause acid regurgitation, which causes heartburn. I've told this to numerous doctors, and they all agree that never would they have stopped to imagine or think that the patient self prescribing Omoprazole was the possible cause of their heartburn?

That's the power of AI. And we have a pattern pending on that, by the way.

[00:29:07] Ty: That's awesome. Do you mind talking through a bit more and the expertise around preventing hallucinations and you mentioned referencing to other large language models and having a large enough data set. Do you have your own proprietary data set you're pulling from at this point to power this.

[00:29:23] Steven: We were part of the Mayo Clinic program that gave us access to 3 million patient records, and I don't I can't go into it because we have an NDA with Mayo Clinic. But we looked at large databases, and we've looked at some proprietary large language models. I can't go into specifics around giving away trade secrets, but I will tell you that our advantage is creativity.

Okay? We not only think outside the box, we don't see the box at all. And we have an extraordinary leadership team with 125 years of combined health care and technology experience. We have two PhDs, an MD, two Harvard MBAs, two masters of health administration, just in the leadership team. And so we are very creative in the way we think about these problems and we think about how to solve them.

To paraphrase Henry Ford, who once said, although it's dubious, he ever said it, but it's described to him that if he asked his customers what they wanted, they would have said faster horses. We believe that our competitors are building digital horses while we're building a car. And we're not just building the car.

We're trying to build an ecosystem. Which is why we've announced a few partnerships, and we have a few more partnerships coming because we're building the Battlestar Galactica fleet. And Amanda of brilliant people come together to solve a big problem. That's the difference about SOAP versus other companies.

[00:30:46] Jared: Something I wanted to go back to as well was, you're talking about how somebody can go and sit with Genie and share two hours of, their personal data, the personal life, some of the stuff that they wouldn't dare tell even some of their best friends in some cases, and one of the big critics against ChatGPT in particular is you don't want to share any proprietary information because the data privacy has not been really fully locked in there. And so for SOAP Health, how have you guys been ensuring, patient data privacy in particular?

[00:31:15] Steven: So there are specific requirements when you work with large healthcare institutions. One of them is that you're HIPAA compliant, which everybody needs to be in healthcare, but there's also something called SOC 2 certification, which we have.

But it's also the way that we approach everything. It's with putting the patient first and thinking what's in the best interest of the patient. Honestly, we put the doctor second, a very close second, but we put the patient first. Why? Because that patient is us. That patient is you, and you, and me, our family, our friends, our colleagues, it's everybody.

We're all patients, including doctors. So why not put everybody first, and then put a subgroup second? And that's the way we think about it.

[00:32:04] Ty: You mentioned HIPAA compliance and SOC 2, and I imagine the FDA has been inundated lately with generative AI devices for approval and all of that. Could you speak to some of the regulatory process you've had to go through?

Because I'm just fascinated by that side of things too.

[00:32:18] Steven: Yeah, so we haven't had to go through a regulatory process because we don't use AI in the risk assessment or differential diagnosis. We use guidelines, and that's very important because for two things. One is AI is a black box, and if you don't understand why something tells you what it tells you, then you can't rely on it.

Now, today, I just got off the call with another consortium that we're working with at Imperial College to submit for a grant using what's called Neural Symbolic Machine Learning, which is a more advanced form of machine learning to solve the problem of early cancer diagnosis. So we're literally on the cutting edge, which again is how we differentiate from others.

But the problem with generative AI, the problem about the regulatory process is that the FDA says if you are solely relying on AI, Okay, then how can we trust that you're going to get the same consistent answer every single time? And so you have to demonstrate consistency, whether it's AI to read an x ray, whether it's AI to read a pathology slide, whether it's AI to make a diagnosis.

But I will tell you, there is no exceptional AI diagnostic tool in the market. For example, K Health, an Israeli company recently published in the Mayo Clinic proceedings that their diagnostic tool was 80 something percent accurate, like the doctors at Mayo Clinic. And I immediately called BS. I hadn't even seen the study.

I just saw the headline. I said, there's no way that they're 80 percent accurate in making diagnosis. And here's why. Because symptoms can only get the right diagnosis 73 percent of the time. So how can they get the right diagnosis agent? So that was the first reason I called B. S. The second reason is I know how complicated it is to get diagnoses.

Plus, we have things in our patent that no one's ever thought about before. And I know that because we did prior art search. But the bottom line is I finally read the study. You know what they were diagnosing at 80 percent and I can't believe it was so low. It was only 80%. Sinusitis. You know how you diagnose sinusitis?

Go like this when you have a bad cold. If it hurts, you got sinusitis. Okay, gastroenteritis, a stomach-ache. They were diagnosing viruses, which lay people could diagnose at with an 80 percent accuracy. How are you feeling? I'm not feeling well. I think I have a virus. Okay. They didn't diagnose in that study one cancer, one heart disease, one diabetes, one autoimmune disease.

They weren't diagnosing. Okay, but it's the headline. The next thing, AI can diagnose as well as doctors baloney. We're not there yet. But by combining risk and symptom assessment, we will get much closer and our goal is to eventually be perfect. But perfection is a process.

[00:35:14] Jared: Yeah, no doubt. Another thing I want to talk to you about as well as.

Just the fact that because physicians are going to be saving so much time is this going to be driving their overall costs down? At like the family medicine level because they just don't have to put as many hours, into working. And the other side of it then is does this sort of more efficient healthcare system then also become cheaper for the patient down the line as well?

[00:35:40] Steven: Look, on my backdrop, it says one free time. So one of the things we give physicians, as you just mentioned, is more free time. But how they use that free time is up to them. Do they go and exercise? Do they see more patients? Do they read more medical journals? How they use that free time is up to them, but it's an important decision for them.

For those doctors who want to make more money, they'll see more patients. For those doctors who want to be more knowledgeable, they'll read more medical journals for those doctors who want greater balance of work life, they'll get home early and they'll actually have dinner with their kids. So everybody has to make a personal choice.

We're not here to make that choice with them. We're here to give them that choice. My mother used to say to me, be in a position to choose whatever you do in life, be in a position to choose. Today, doctors don't have a choice. They got to work like dogs in order to make a decent living.

Okay. And that's their life. They come to the office, they put their head down, they plod through the day, they hope they don't hurt anybody, they hope they help people, they hope they make enough money, and then they go home tired and exhausted. Look, one of my best friends is a surgeon, okay? He is a guy that everybody loves.

Everybody loves this guy, okay? I've known him for 57 years, 58 years. Okay, best friend. He gets a call from the answering service when he was on vacation and I never saw him screaming somebody so much for calling him inappropriately. That's burnout. That's the very essence of burnout when people are so irritable that it takes nothing to trigger them.

I was shocked. This is what's going on in medicine. Doctors have it so bad and people say, oh they make 250, 000, 300, 000. What are they complaining about? They went to medical school for four years. They went through residency. Some went through fellowships and further training. Okay. Something goes wrong. Everybody's ready to sue them. They do their best. The problem is it's so hard to do your best right now with medical knowledge having grown exponentially. Look, 150 years ago, the founder of the Mayo Clinic said, it's impossible for a doctor to know everything there is to need to know about medicine. 150 years ago.

Just think about how much more knowledge has accumulated in the last 150 years.

[00:38:04] Jared: No doubt. Man, we got another really good question coming in through the chat. And It was around, how would you suggest that a physician, with around six years of experience in a field in this case emergency medicine breaks into AI they have a lot of interest in innovation, healthcare, including AI, and they just want to know how they should start out and do you recommend that somebody like that gets into AI?

[00:38:26] Steven: Look, it's often said that we need better tools for doctors and I joke that we need better doctors for AI. Now I just read a couple of weeks ago, the first MD AI program in the United States that is gonna teach AI and Medicine at the same time. So the yeah the answer to your second questions is absolutely doctors should learn about AI. Okay. Just like they should learn about how to use a computer. But that doesn't mean they have to be a programmer. Doesn't mean they have to learn programming code. Doesn't mean they have to become an expert in all the different types of AI and how to create AI.

What they need to become an expert in is how to use AI to practice at the top of their licenses. For example, I participate in Stanford's AI program once a year. I do it for continuing medical education credit and I do it for the knowledge. I'm a big believer in knowledge acquisition. My answer to this individual is look for these programs run by some of the top institutions in the country that are geared towards medicine and medical doctors that provide continuing medical education.

Stanford has an excellent program. And become familiar with the vernacular of AI and the possibilities of using AI in everyday practice.

[00:39:50] Jared: Yeah, absolutely. And would you say that then, at a broader level at sort of the executive level across business that for those that are trying to implement AI become more efficient, save more time in their day, would you say that your advice would be similar as well as to essentially seek education and figure out the right tools to implement based on what you've learned.

[00:40:11] Steven: Look, here's the challenge. There's a lot of AI startups by definition, most startup fail. So chasing AI is very difficult. So you have to unfortunately, there's some organizations that try to do this. There's an organization called AVIA. That we were in their marketplaces. 1 of the top conversational AI companies, but they really didn't give a user and extensive enough understanding of our capability or the other companies listed their capability. So you really have to do your homework. We just signed up a customer who said she researched all the AI companies on her own and concluded that for the use case that we apply to, we were the singularly most outstanding AI company. Most people don't have time to do that type of research.

And here's the crazy thing. Everybody across the United States and healthcare, every health system, every primary care practice by and large has the same issues. But the success of a company like ours comes down to luck, and warm introductions.

It's all about luck and warm introductions. As you can't control for luck, but you can be prepared to take advantage of it when it happens, also to recognize it when it's happening. But warm introductions, nothing beats a warm introduction. We live in an age where people ghost even close friends and relatives.

Okay. Everybody's ghosting. Everybody's missing emails. They don't do it on purpose. They're just overwhelmed. Okay. If you can't get to a decision maker directly and have a warm introduction, the sales cycle is so absurd that you might as well give up. And I don't believe in giving up. Whenever something goes wrong.

I say to myself, fall down seven times, get up eight. Okay. I thought last week we had a major setback about something. And my first reaction was fall down seven times, get up eight. That's the nature of entrepreneurship. But if you can't get warm introductions, if you can't network, forget about trying to operate, because it doesn't matter if you have the best product, you're going to be the beta max of video recording.

Betamax was a better product, but VHS outmaneuvered them by having partnerships with the big studios, etc. It's similar. The best products don't win. It's the best products with luck and warm introductions that win.

[00:42:43] Ty: You brought up luck, and Jim Collins talks about luck as a major factor for the success of different companies he's studied.

In particular, he talks about who luck, meaning you're fortunate by the people that you meet along the way. And one of the things that stood out for me as you were describing yourself and your team is that intrinsic creativity. That seems like it's been a theme and a competitive advantage for your team.

And I'm just curious if you could expand on that a little bit more because your early passion was a creative field. And then you went into medicine and you've been able to sustain a creative outlook and a ambitious vision throughout your career. You just speak a little bit about how you've been able to maintain a belief in your own creativity where so often people, they get critical feedback and then the creative side of themselves wilts a little bit.

[00:43:30] Steven: Yes, I'll tell you, really creative minds are different than the average person. We skip steps, okay, which is very often why I piss off venture capitalists. They asked me a question, and I give them a long winded answer to a different question. And the reason is because I'm answering the question they should have asked me, not the question they asked me.

For example, I was talking to a guy who said to me, do people in Florida like Ron DeSantis? So that was a straightforward question to people in Florida like Ron DeSantis, but I said, he's really asked me, will DeSantis be a good competitor to Donald Trump? So that's the question I was answering.

Why could my mind already skipped to what he should have been asking me? Cause that's really what he wanted to know, because who cares about whether Floridians like Ron DeSantis So people ask me questions and then I ramble on about something else and I think I'm rambling, but what I'm not, what I'm actually doing is my mind has already jumped a few steps ahead and I'm answering the question that they should have asked me. So creativity can also come off sometimes like you look like you're on the autism spectrum. Like you can't answer a question when it's directly asked to you. But the way that I find creativity in other people is I ask a very simple question.

Tell me all the ways to get on the other side of a wall. Now, you might think I'm expecting all these creative answers. But what I'm doing with that question is not only expecting to hear creative answers, but I'm also hearing tolerance for them to keep trying after they feel like they've exhausted every choice they can give me.

I say anything else? And those who are good natured, I find creativity usually goes along with a good nature. Okay, people wanting to keep their mind open, people wanting to keep trying. And I also tell my kids when they go to interviews, okay, if somebody asks you a question that you don't think you answered well, or you didn't know the answer to, be intellectually curious.

And at the end of the interview, say, can you tell me what was the answer to that? And I tell people at work, be intellectually curious. If something isn't working, ask why it's not working. If you hear something you don't understand, ask for somebody to explain it. Don't be embarrassed. Okay, don't ever let something go by you that you don't understand because you were embarrassed or shy.

And in fact, I'll tell you something that will make me sound crazy. I trained myself to not be embarrassed. I was once at a trade show and they were giving away pens. And they said, these are really fancy pens. We're giving everybody one free pen. Please don't ask for another pen. I said, wow, it sounds like I would be really embarrassed to ask for another pen.

I said, I'm going to train myself not to be embarrassed. So after I got my pen, I went up, I said, can I have another pen? And I was like, ready? And they'll just give me and I go okay, here's another pen. I'm thinking. No, that was too easy. I said, I'm going back again. This time I went back and they screamed at me.

And I was like, yes! I trained myself not to be afraid of embarrassment. Now, what is creativity? Creativity is thinking about things that nobody else has thought about. What is entrepreneurship? Having the arrogance to think that you can create something that nobody else has ever created before. There's no limit to creativity.

Now, I'll tell you one last quick story. When I got to Stanford, I was doing an assignment with other people and they gave us two spaghettis and marshmallows and string and they said, build a tower. So right away I started breaking the spaghetti, which, by the way, is not what you're supposed to do because you don't build this tall tower with the broken spaghetti.

But I was more concerned about building a tower that was going to last as opposed to building the highest tower. And one of the people in my program, my fellowship said, who said you can do that? And I said, look, I'm the type of person that'd rather ask for forgiveness than permission. And he like smirked at me.

And then a guy walks in a few minutes later and says, here at Stanford, we're all about asking for forgiveness rather than permission. And I said, I'm going to fit in real well, because that's the way I think about things. And I try to inculcate other people with that same mentality. Except if somebody keeps making a mistake, I say, please ask for permission before you do it again

[00:47:39] Ty: or make it a unique mistake.

[00:47:44] Jared: So we're getting real close on time here. I had so many more questions and I realized this should have been a two hour conversation, and with our last two questions, first one being what are you really excited about for the next major innovations coming to AI and transforming health care with it?

[00:47:59] Steven: Look, I'm excited and very disappointed that human beings are going to live forever. But it's going to happen, unfortunately, probably after I'm done. So I think that AI is going to figure out things that we can't figure out. And I'll give you just two examples. On 60 Minutes a few months ago, the CEO of Google said two things that kind of blew my mind.

One, he said that they fed a few sentences, I think, of Bangladesh into their AI system, and it learned the entire Bangladesh language. They don't even know how that happened, but he now predicted that they would be able to teach their system over a thousand languages. And then he said something that really blew me away because I was captain of my high school chess team.

He said that AI had played millions of computer games and identified a new strategy to win that had never been identified before in the millions and millions of chess games that had been played, a completely different strategy. So we as human beings tend to think linearly. But AI has the potential to completely change and teach us things that we haven't even the smartest among us, the 200 IQs have never even thought about.

That's the promise and potential of AI.

[00:49:22] Jared: I wish we had more time to unpack that a little bit more about how, humans are going to live a lot longer. I've seen other experts talk about how, in the future, we're not going to be talking about, Oh, it's going to be great to live to a hundred.

It's going to be, Oh, it's going to be great to live to 400. And I'm like, in my mind, I'm like, that's hard to even think about or to fathom. And yeah, just as a final question for today, would you say and this is a loaded question here, but is the future of healthcare human?

[00:49:47] Steven: So I'm a big science fiction watcher and virtually in every science fiction movie and TV show that I watch today, medicine is provided by an AI representation, except for one show called the art on sci fi where the doctor was a human and she became hooked on pain medications. But in every other show, the doctor is an AI.

I believe that AI will be so sophisticated and always accessible, infinitely patient, again, empathetic, sometimes funny, maybe even sentient, probably sentient at some point, but I'm still fuzzy on that. But we'll be far superior to tracking down a human doctor in a time of need. So why wouldn't we use AI when it's clinically validated to be exceptional?

And I have ideas about how AI can become a perfect diagnostician. It's in my patent, so I won't go into it right now. But I believe that the future of medicine is AI.

[00:50:49] Jared: I'm there with you. Dr. Charlap, thank you so much for your time. We really appreciate everything that you do. Really excited to see the impact and the footprint that SOAP Health leaves on just so many lives leading into the future and maybe even my own life later on as well as this is starts to get integrated into, more practices across the country and globe.

So yeah, thank you so much for what you do really appreciate it.

[00:51:09] Steven: Yeah, by the way, we're all over the country or the world right now South Korea. India Europe, Taiwan, UK. The Clara call has gone out. The clarion call has gone out and we've responded.

‍