This transcription is provided by artificial intelligence. We believe in technology but understand that even the smartest robots can sometimes get speech recognition wrong.

Today on This Week Health.

Since AI is math, not magic, you can't just say it's this panacea that's gonna fix everything. Right. But what I am seeing over and over and over is when the right data, to your point, curated with the depth and the breath and the spread is applied to the right measures or outcomes, the predictive capacity of this stuff is extraordinary .

Thanks for joining us. This is Keynote a This Week Health Conference show. My name is Bill Russell. I'm a former CIO for a 16 hospital system and creator of This Week Health, a set of channels dedicated to keeping health IT staff current and engaged. For five years now, we have been making podcasts to amplify great thinking to propel healthcare forward. Special thanks to our keynote show partners for choosing to invest in our mission to develop the next generation of health leaders. Now onto 📍 our show.

All right. Today we are joined by John Halamka, president of the Mayo Clinic Platform. John, welcome back to the show. Well, hey,

always an exciting place to be because we can speculate about the future and then create it .

we talked about bias in AI last year, and I know you've only done a handful of other interviews since our interview in January of, of this year. So I'm just reminding you a little bit of what we talked about. We talked about bias in AI and how we're going to get through that. You talked about labeling.

On ai models and those kind of things. So that's some of the stuff we talked about. We're gonna talk a little bit more about that the state of data, ai, innovation and healthcare. But I need to start with this question cuz my listeners have asked me to start with this question. Tell us about your health system.

My health system? You mean about Maya? Yeah. I know you, you assume everybody knows about it, but it's interesting cuz you'll say something and people go, I didn't know that about Mayo.

Yeah. So I have worked as a physician for 40 years and most organizations look at the needs of the provider or the needs of the professor or the needs of the payer.

I would argue Mayo is unique in the country and always putting the patient first in every decision to be. It. If Mayo were a government, Mayo would be a socialist society because it is everyone, 73,000 people asking how do we make things better for every patient we encounter every day? So I mean, I know that may sound a little unusual, but it's very different from many organizations.

Mayo doesn't think so much about how can Mayo be number one. It's how can healthcare be better for everyone. So it's very much global in scope, so Sure. It has a destination medical center in Rochester, Minnesota. In Jacksonville, Florida, Phoenix, Scottsdale, Arizona, in Abu Dhabi and in London. And we have a Mayo Clinic Health System, a set of affiliated organizations, mostly in the Midwest, Wisconsin, Iowa, as well as organizations that are working with us as collaborators called the Mayo Clinic Care Network, 73 different hospitals of interest.

The 73 Collaborative Hospitals report. And why is that? Because the assumption is if you are gonna create a learning laboratory for innovation, working directly in the community with emerging digital health is a great place to start. So think of Mayo as global in scope, 73,000 employees and driven by the needs of the patient everywhere.

It's interesting that you start with that and, and some people will probably take exception to what you said. Having worked in healthcare and whatnot, I, I'm not sure I take exception to it. I, I remember a number of times as a cio they would say how is that going to serve the clinician and how is that going to serve?

And there was billing things and all that. We were constantly trying to, to balance all those things. Are you saying that Mayo will will put those things on the back burner and just say, no, no. First, first and foremost has to be driven by the patient and then these other things just support that, that.

And that's exactly true. And so lemme give you just a personal example. So you know, I have glaucoma. I have glaucoma because my father had glaucoma and you can't choose your parents . So I went to Mayo and said I'm losing vision in my left eye a little more rapidly than we'd think despite maximal medical and surgical therapy.

Yeah, you just, you just described a four month process somewhere.

Exactly, so literally I show up at, and this isn't just because I'm a Mayo Clinic employee. I mean, this would be true of any Mayo patient. You show up at 8:00 AM you're discharged at noon and four months of care done in one morning.

Yeah. That's amazing. I want to talk about the Mayo Clinic platform. Your ceo the CEO of Mayo recently wrote a LinkedIn post about the Mayo Clinic platform, and I was thoroughly impressed. I just love it when CEOs. Are beyond Conversant and they're actually taking a leadership role with the use of technology.

Tell us a little bit about the Mayo Clinic platform and how you engage your CEO in that conversation and, and how he's moving the vision forward.

To be transformative and to be global in scope. And so he outlined this idea of, well, if we were to look at platform concepts various kinds of things that might have been done with Uber or other industries where they didn't. Just be a means of production. They were a facilitator of bringing producers and consumers together where the network effect and the secondary effects were so large.

These things became transformative and he then wrote a job description. Which is wanted a physician engineer with 40 years of digital transformation experience, public policy and economics degrees, international travel writing and speaking to change healthcare globally. Now you can understand it's a slightly rarefied job description.

and so it was. Such an amazing opportunity to work at such grand scale that I said I think that's what I should be doing with the next 20 years of my career. So his vision, but then handed to me and a team without micromanagement to execute in the way we think would be most impactful.

Wow. So the Mayo Clinic platform is what?

Well, sure. So when you, I'm gonna ask your question a couple of ways. So the word platform, if you look at Sloan or Harvard Business School, says, oh, it's a business model where every participant benefits from the existence of other participants. And so John Rico said, well, why is it in healthcare?

We de-identified all the Mayo data, moved it to Google Cloud, built secure containers and sub tendencies, and built legal agreements, privacy compliance agreements, so we could bring any partner from anywhere in the world. Together with Mayo Clinic in two weeks, , so we have now 65 partner. And we have all of Mayo Clinic data, but we've also signed an agreement with Mercy to bring its 15 million patient experiences birth to death.

It's a way of saying, oh, we can represent the data in a standard schema and then build the privacy and IP protecting mechanisms so that organizations can work together collaboratively to generate new algorithms and validate algorithms. So think of platform with its 60. Collaborators as a virtuous group of data generators, solution providers, and healthcare system delivery personnel having a very significant impact to bring machine learning to a global stage.

📍 📍 Let me throw you a curve ball and ask this question. You described your experience, glaucoma, optimal, that whole morning set of sessions. That's the platform I want. I want everyone in the country, all of my employees, let's just start there. All of my employees to have that experience.

If they have some sort of situation, I want them to have that healthcare experience. They, they go from specialist to specialist to primary care physician. They get it explained and. They're getting treated. They're getting the right diagnosis because it is, it filters through many lenses and gets to that point.

How do we get to that point where the platform delivers a care experience across the board? I understand the data and the AI aspects of it, and we're gonna move cures and those kinds of things forward, but how do we change the experience?

Well, sure. So one challenge, and let me give you a couple of examplers, is that the, what I'll call pipeline models we have today, our one clinician sees one patient don't, they don't scale so well.

So if we say, oh, we have 300 million people, 330 million people in the US and everybody with an eye disease get to see Sally . Yeah, that's not gonna work. Right? So what you have to say is, well, What are the aspects of what Sally does? What are the knowledge elements? Oh, well, she is able to interpret retinal images with a facility no one else has, or she is able to understand from a visual field test subtleties of changes.

Oh. Can you instantiate some of that specialty knowledge into algorithms that Democrat. And make a whole lot of the capabilities really scalable. And then there are those that will need on-prem care and that's okay, but you're figuring out the right patients to go to the right person in the right place.

So much of what we've done in creating now more than 60 algorithms is be able to deliver the specialty knowledge that is what you describe, but deliver it. Deliver it in a rural community, deliver it in the home. And so there is in effect some aspects of the kind of model you want that will be dependent upon technology deployment. And that's where we've started.

We're at a, I don't know if it's a unique time. It feels like it's a unique time in the history of healthcare where it's not only clinician shortage, it's shortage of staff, period. Across the board we have financial pressures, we have other pressures that are going on within healthcare. There is this need for what you just described, the ability for technology, for algorithms, for AI to step in and say, Hey, I'll take some of the cognitive load. I'll, I'll take some of the workload that is normally handled by humans. Tell us about the progress of AI in the clinical setting at Mayo.

So one of the things you we'll start with is you recognize AI as math, not magic right? Right. And, and so you have to look at each use case and see where math is actually helpful. A as bill, I'm 60, so four weeks ago I had a colonoscopy. You kind of gotta do that when you're 50, when you're 55, 60. Well, do you know that 20% of the lesions. In endoscopy as viewed by humans are missed the missed rate's 20%.

An algorithm developed at Mayo by the GI department has a missed rate of 3%. And so this is not, algorithms are placing humans. It's algorithms augmenting humans to say, oh take a little look at that flecture There is a ditzel there that may be interesting, . And so, okay. Isn't it intriguing to believe that maybe five years from now practice without.

AI Will be considered malpractice because you're accepting a 20% error rate instead of a 3% error rate. And Mayo happens to be the organization that invented that algorithm that will augment endoscopy across the world. And the same thing is true in neurology, cardiology, radiation, oncology. We have produced algorithms that substantially reduce human burden while enhancing safety and.

So the next step, while there's, there's a couple things. I mean, we could talk about transparency and those kind of things cuz we've gotta build this culture of trust. And we did talk about that in the last time you were on the show. But what I wanna talk about is, is integrating the AI into the workflow and it it represents a fair amount of work.

It's very challenging, but also, I've seen so many tools. I'm, I just went to the health conference. I went to the various conferences and you see all these AI tools popping up and I, I just picture my clinicians going, oh my gosh, not another tool stop like this. This is my workflow. Figure out how to get it in here. talk about that process of, of getting AI integrated into how the clinician operates.

Well, it's simple. More alerts and reminders In the ehr, they'll love that. , oh no, no, no. Load a new app on your phone. Oh, go to a new website. Remember five logins, right? So to your point, these are all the things we don't want to do. And so there are a couple of ways of approaching the workflow. So one way is to say the AI is just hidden behind the scenes. You don't even really know it's. It just happens as a side effect. The work you do anyway, and let me give you an example. So I have an svt, a super ventricular tachycardia.

Sometimes my heart rate goes to one 70. I do a deep breathing exercise. It goes away. It's, this is very, very common. Lots of people have it right. Well will I have a heart attack? Will I have a unstable ventricular rhythm? May I have a fib a stroke? Well, Mayo has algorithms for all of those things, and they're predictive.

So what, when I go to Mayo and I get a 12 lead, all the predictors are simply printed on the ECG Oh, no, no, no. Valvular disease, no ejection fraction problem. No cardiomyopathy. They're just there . So do the physicians even really have to do anything? No. It's like I pick up an ecg. It says rate rhythm, interpretation, intervals.

Oh, and look, here are 12 indicators from an AI algorithm that this patient is actually okay in this regard. So it's just sort of like, hidden It's invisible. It's it's passive. One other way, and this, we have to be careful about this. We know that in a healthcare system there are many roles, doctor, nurse, social worker, well, there are also care managers.

and increasingly, especially with value-based purchasing, you can imagine a lot of the care navigation is actually done by a care manager. So you can imagine these algorithms being applied to phenotype, genotype and exposome, and letting your care manager know there's a patient who really should be evaluated for X disease or try Y therapy, and the care manager is facilitating that so it isn't creating more cognitive burden for the.

clinician

one of the challenges I've heard with AI is the quality of data that we have within healthcare. So there's only certain areas we can apply the math to it. How's that work going with regard to making the data, I don't know, cleaner, better able to be used by AI models across the.

A brilliant question. So there are three variables that I describe. One is the depth of data. Okay. I've got structured, non-structured, I've got telemetry, images, genomics, digital pathology. That's depth data types. Another would be breadth, the number of patients I. And the third, something I'll call spread the heterogeneity of those patients.

How many American Samoans do I have? , right? And so what Mayo has done is in its 10 million patients, it's created great depth and 10 million, I mean it's breadth, but it's Arizona, Florida, and Minnesota. So with our alignment with mercy, Where Mercy is bringing 15 million patients, Arkansas, Oklahoma, Missouri, our work in Canada, our work in Israel, suddenly you start looking at, oh, well, I've got pretty good spread.

It takes organizations working together cuz no one organization is gonna have the depth, breadth and spread necessary for the AI models you described. And it also requires us to ingest novel data types we've never ingested before. So for example, Mayo actually ingests Apple watch data into our AI systems.

And so we have 6,000 patients that are submitting their Apple Watch data into our AI repositories and that you can imagine as we get more instrumentation, we wear and instrumentation in the home and patient report of outcomes and patient generated healthcare data, you're gonna have to think beyond the ehr.

And part of that depth is how to integrate all of the data that is emerging in society. Even craziness. How about your grocery list? I hear you've got a problem with your sodium. Why are you eating Doritos? ,

sir. I mean your grocery list is probably the, best indicator of your health anyway.

Exactly. But yeah, and there's a, there's a ton of sensors out there that are pulling data in, but I mean, you. You bring up one of the challenges for ai. So we, we generate this model at Mayo, and you, you talked about Rochester Arizona Florida and whatnot. Let's assume we now go into New York City with that model.

How, how do we ensure that that AI model is going to work well in that population? Sure. So let me be very

specific without disclosing any intellectual property. So there are a number of world class hospitals in New York, and guess what? I called them up and I said, how would you like to be part of a global virtuous network where you retain all your own data?

But you are able to validate algorithms from outside your organization and locally tune them to your population. And oh there could be a revenue model in that we're not selling data, we're not exfiltrating data. You're just saying, oh, I'm taking this algorithm from this organization, testing it in New York City and it works for this, or it doesn't work for that.

And they said, that's great. So what we're finding is you, you've heard me say this before, sometimes the issues are technology. Sometimes they're policy, and sometimes they're psychiatry, , and, and you overcome the psychiatry problem by telling organizations throughout the world. You retain your own data, you control what it is that you do with that data, but you would want others outside of your organization to test their algorithms with you against your data, and you report. And they are saying, that's great. I would love to do that.

It's interesting there are other models out there where people are bringing data together. I'm not sure it it, and I think people could confuse the Mayo Clinic platform with those kinds of, Data aggregation models, and they are looking at for the good of mankind and we're gonna de-identify this data set and we're gonna do we're gonna look for cures in that data set.

Well, I mean, heck, you have Epic cosmos, which sort of could look like the same kind of thing. You get that, that breadth of data and even the depth of data. How is Mayo Clinic platform different than those kinds of approach?

Sure. And so couple of major tenants, as you heard, the data never leaves the organization, right? So it's not centralized. The data is de-identified. The decision to participate or not participate for any given project is made by the organization, and there are, as you say, there are other models that say things like, oh, we're gonna aggregate things centrally. And then use it as we will . That's a very different approach.

Yeah. And similarly, we're trying to figure out how this becomes a global collaboration and not the hege of any one organization. And, and that's the interesting character of Mayo. It is a facilitator and a convener that can bring people together for the benefit of all without thinking about the benefit of a single organization or a centralization of control.

Let me talk to you about, so some of the work you're doing, we're recording this in early December, so you have some work that's going on right now around AI and the actually just talk about the work that you're doing, you're doing a bunch of interviews right now, even as we speak. So talk about that. Sure.

And so in speaking with academic, government and industry partners, and speaking with the press, everyone says AI has a credibility problem. Right. Oh, I use this algorithm. I ha it has an AUC of three. Three. Flipping a coin is better. It ha, I have no idea who it was developed on or where.

It's useful for purpose. So a year ago, a number of organizations came together to form the Coalition for Health ai. You'll find that coalition for health ai.org tonight at midnight. Yes. Just a few hours from now, we are releasing our blueprint for the. And that blueprint for the country includes how do we define national standards for measuring bias, equity and fairness, testability, usability, safety, transparency, and reliability.

How do we create a distributed data network for organizations to participate in model creation and validation without loss of control or loss of privacy? How do we create a nationwide note, I didn't say national, right? This is not a government entity or something, but a nationwide assurance and validation lab.

There could be many that are able to take any commercial product and go against all this distributed data and say, oh it works for thin people. It works for not thin people, it works for tall people wherever. So the, the metadata around this product now specifies where it's likely to be helpful.

That suggests we need a nationwide registry that is the assurance data for every algorithm that is analyzed by every lab. And here's my dream, six quarters from now. We will have every EHR in this country able to go out to a national registry and pull down an algorithm for the person in front of the doctor right now that is likely to be helpful because of their phenotype, genotype, and expose.

And here's where we have to be a little careful. Remember right now a lot of algorithms are kind of one size fits all. You use it or you don't. But for example, there is an algorithm for car cardiovascular disease progression prediction that works really well. For people with a body mass index under 35 and really bad for anyone with a body mass index above 35.

Now ethically, should we just not use it? Well if a person comes in to see the doctor with a body mass index at 24, sure. Right, and so that's the whole point about creating this National Registry of the assurance data and pulling it down to help the right patients at the right. And ultimately monitor the effectiveness of these models that we deploy. So Coalition for health ai.org will be released tonight at

midnight. What's it gonna take for your vision to become a reality in six quarters?

So the organizations involved in doing this so far are Hopkins, duke, Stanford, uc, Mayo, Emory, Google, Microsoft, FDA onc the White House. And you're seeing interesting, every hca more and there's almost a, a fear of missing out.

It's a fomo this is what we gotta do is to. Yes, be part of this virtuous activity so that we can just spread the wealth, as you said, in an open source like fashion to get more and more adoption and what Mayo is going to do. This is not a secret plan, but just to share with you Mayo's notion, it will take some of our affiliated organizations and it will spread what we're calling a starter pack.

Of AI to the affiliated organizations so they can start to see the benefits and where it's impacting patient quality and physician burnout. And then they can tell their friends who tell their friends, who tell their friends. And before it is a standard of care. That's, that's the approach.

Fantastic. Mayo launched an acceler, I'm gonna call it an accelerator. You can correct me if I'm wrong, an accelerator in partnership with Google and Epic two companies that aren't usually mentioned next to each other, so I'm, I'm curious about that one. Tell us about that initiative.

Sure. So Mayo Clinic platform, accelerate. Has a simple notion. So Bill, how many young companies have you introduced to your readers who are just amazing in their vision, but can't spell hipaa? Right. This is a problem, right? You got these amazing energetic people, but they lack domain knowledge and they lack access to the underlying data sets to make their models more.

So Mayo Clinic said, let's actually every six months run a 20 week program where we competitively bring in young companies and co-develop products with them, getting them access to all of our clinical expertise and of all of our de-identified data. So we've now done that for a few dozen companies, and those do, few dozen companies have created RO products that are starting to get great traction in the marketplace.

And for example and again, I have no interest in any of these, there's no financial issues or conflict of interest. A company in Australia called seer, S E E. Said, we believe we can predict seizures before they happen. We, we, we have no idea if it will really work, but if we only had access to 10 million patients and we brought in some of our sensor data and we built the models together, we think it'll work well.

It worked. And because of their 20 week experience, they got a 50 million series A from the venture community and now they are one of the largest providers of AI services in. And so that's the idea, is serving as a ca a catalyst by removing some of the barriers to success we see in young companies. When

I see Mayo, Google, and Epic does it require Google technology, epic technology, or be a part of the Mayo Clinic platform in order to benefit from the acceler.

I'm gonna, so the answer is no. But let me tell you why. No. So I was a graduate student at MIT in the early nineties, and I had a professor named Hal Ableson. And Hal Ableson said, John, you know that over the course of your life, technology will change, but engineering principles will not remember abstraction layers.

If you really love Product A today, but five years from now, it's awful. You should be able to go to product B without difficulty, cuz you didn't have vendor lock in. Everything is standards based and so this is how we've built it. Everything that we've done, Kubernetes, docker containers, cloud neutrality, so you can run anywhere on anything. It just, it is true Google is a partner for storage and compute. But it is

not a requirement. Yeah. Just outta curiosity, this is more of a personal question. You are the physician doctor who wrote his own ehr. Do you still get enough of the technology fix, like the nerdy playing around with the technology, or is that, are those from days gone by for you now?

Okay, see, I'd only tell Bill this which is I run Unity Farm Sanctuary and it's the largest animal sanctuary in New England, 143 in of things, devices managed in Google Cloud with Google Analytics, voice activated and scripted, and all the pie torch scripts are written by me.

That's, that's great. Now, you did think about a platform for that, right? So this is gonna be able to be used in animal sanctuaries around the world and those kind of things, right?

Everything is standards based and so you are correct. You approach this by saying, for me to incrementally add a new device from any vendor near zero effort, and so one of the things that your readers can't quite see, but you know, it's a good exemplar for you.

So behind me, I have two clocks and one clock is standards based. The other clock is proprietary and closed, and so one clock, which every single part was hand fabricated, is from the Paul Revere era. And for me to fix it, I literally have to go to a metal shop and figure out how to fabricate something. It is not scalable, it is not agile.

And he called it the International Time Recording Company, but then he expanded to other devices. He said, I bet we could automate business machines. So let me change the name to ibm. And so the clock that you see over my shoulder is Thomas j Watson's prototype for his new emerging company called IBM

Wow. I, I did not know that. Yeah, and they're classics. If people aren't watching this on video, I mean, these are, I mean, I could see them moving. I could see the one on the rights actually moving. So these are working clocks, and they are they're very classic. Let me, let me ask you, let bring you back to healthcare a little bit.

There's a lot of talk about automat. In healthcare these days, and I think the automation push is partly due to the staffing challenges and partly due to the financial challenges. Where do you see automation really taking root within the healthcare continuum? Sure.

So earlier today when I was doing a keynote for Stanford's AI conference, I posed the course organizers a question. How many times are you a care navigator or a care traffic controller for some random person in the community? They said, oh my God a hundred times a year somebody says I've got this red dot on my left little finger. Who do I see? He said, well, Don't you imagine that the data of the past, call it sort of ways for healthcare, could inform the journey of a patient in the future and help 'em avoid the potholes along the way and ensure they get to where they need to be in a relatively affordable, expeditious way.

And they say, oh yeah. Right. So what you'd hope is that we are building all these AI models to make care navigation. Accessible to mere humans. So you don't have to call the president of Mayo Clinic to decide where you're gonna go. . Yeah,

there's, there's a lot of those a lot of those stories I got stopped at a party and somebody asked me and I'm like, I'm the cio. I'm not even a clinician. What are you asking me for? Right. But it would be nice to tell them, Hey, we have this, I dunno, fill in the blank. We have this, this tool. This AI chat bot that's sitting here that does these things and for for general things, it does this, but if, if it can't figure it out, it does it, it directs you, right?

It helps you in navigating to the right place for the right setting. And that may not be the best case, but. it's inter I'm hearing a lot of talk about, about automation and again, you come back to what you started with, which is at Mayo you think about the patient first, right?

It's not about, it's not about streamlining. It's not, it's about, okay, how can we get to the pa the patient to where they need to get to in, in the best fashion possible? And it's interesting to think about automation. In that vein.

Well, and here's the example where it's gonna go. So I have some colleagues in South Africa. Who had this notion that they could figure out whether somebody had tuberculosis or covid based on the sound of their cough. So they recorded a hundred thousand coughs, and now they have a model with an auc nine or so. And all you have to do is cough into your phone, right? So you could imagine that somebody goes to a navigation system and, oh, cough, well hear, cough into your phone. Oh, this is likely this disease. Go see this special. That's where we need to get to.

Yeah, that's amazing. Where would you place the state of innovation in healthcare across the board? And so where would you place it and what obstacles do you think we can remove next in order to accelerate the pace of innovation?

So William Gibson told us the future is already here to just unevenly distribute it. So I'm gonna give you an optimistic and a pessimistic. . So the optimism is I see some amazing innovation coming out of the startup environment, coming out of selected organizations that have decided they have to differentiate or die.

But at the same time, I'm seeing organizations retrench. I'm seeing organizations saying, oh my God, the great resignation, the great realignment, supply chain issues, inflation. We better stop our innovation programs cuz we can't afford them. And so my worry is that people will say, I'm losing money, so let me do the same thing faster.

What? You wanna go outta business quicker? I don't get it . Well, I am seeing some organizations unfortunately, reduce their spend on innovation at a time when they should be increasing it.

Yeah, it's, it's interesting. Well, you mentioned it, so coolest thing you've seen in healthcare over the past. You come across way more than I do. I mean, something that gets you thinking, wow, the future is really gonna be awesome.

So remember since AI is math, not magic, you can't just say it's this panacea that's gonna fix everything. Right. But what I am seeing over and over and over is when the right data, to your point, curated with the depth and the breath and the spread is applied to the right measures or outcomes, the predictive capacity of this stuff is extraordinary . And we've published a whole variety of papers and science and nature over the course of the last year looking in various fields where we've decided to attack a disease state. And we're finding we can diagnose or predict these disease states with a far greater accuracy than even the best clinician.

So I think this is why I have such optimism. It's again, not every algorithm or every disease. But I'm seeing enough exemplars that say this is real and this is going to, these next six quarters rapidly accelerate, and we're going to see more and more adoption of machine learning technologies. It's the only way we're gonna be able to get through a pretty hard time in healthcare.

All right. Here's your exit question. I try to come up with exit questions that our listeners have given me some areas they want me to talk about. So you talk to a lot of startups. My guess is from time to time they ask you, how can we be successful in healthcare? Or we, we don't know how to get our message out.

We don't know how to sell into healthcare, and a lot of 'em die on the vine. In that process, what, what coaching do you give them as they are taking off going into the healthcare space with something that is really exceptional?

So I get 50 pitches a week. Wow. And of those 50 pitches, maybe one is viable. So ask yourself, what real world problem are you solving and who will pay for it?

seems very basic,

right? And so you would not believe the number of pitches I hear, and I say, well, what's your business model? Oh, the patient will pay. Yeah, sure. ,

no, that's, that's the number one. And whenever I hear that from a startup, it's like, yeah, we think the business model is b2c. And I go, right, you're gonna have to convince me of this one. Cuz I haven't seen it work yet. I haven't seen, I've seen very few in healthcare B2C models because they expect insurance to pay. And so how are you gonna get the insurance carrier to cover? That. And if not, how are you gonna get the person to separate from their checkbook, some money to you, because you're offering that much additional value in the world of health.

Well, exactly. And so I was chatting with an Israeli company yesterday, and their model is B2B to c. And it goes something like this. A payer wants to buy a product to give for free to a clinician in a Medicare Advantage program so that they can serve the consumer with the services they need. But it's starting at the payer who's actually writing the check. Right. Okay. B2B to C. Oh, okay. Well that could.

It could because they, they get the first dollar of care and they're actually paid to keep people out of the hospital. So if it does keep people outta the hospital, there's a value proposition.

Exactly. So compel me when you present that this is a problem that you didn't make up, that it's a doctor, nurse, social worker, psychiatrist problem experienced every day and there's a clear business model and then you're.

Are most of the, of the pitches that you hear about making the life of the clinician or doing something for the clinician, or is it more about the health of the person in the community?

Yeah, and I think the answer is they're all over the map but there is a, I guess maybe a common theme. It is bringing the right patient to the right clinician in the right setting at the right time.

Cuz that is just a problem all of us feel every day. Yep.

Absolutely. John, I wanna thank you for your time. I wanna thank you for kicking off our new year in style and look forward to the next time we get together.

Well, I'm sure there's gonna be people who disagree with you, but I've talked to enough people who did invest this year who are very sad. So we will see, we will see if that continues to play out in that direction. John, thanks again for your time. Appreciate it. 📍 Thank you.

I love the chance to have these conversations. If I were a CIO today, I would have every team member listen to a show like this one. It's conference level value every week. If you wanna support This Week Health, tell someone about our channels. We have three This Week Health conference, This Week Health Newsroom and This Week Health Community. Check them out today. You can find them wherever you listen to podcasts. Apple, Google, Overcast, Spotify, Stitcher, you get the picture. We are everywhere. We wanna thank our keynote partners 📍 who are investing in our mission to develop the next generation of health leaders. Thanks for listening. That's all for now.