文字起こし 🔗
Doug Rogers: Hello and welcome to The Counterweight. My name is Doug Rogers and I'm joined by my co-host Russell Stover and we are very pleased to welcome Dr. Michael Turkin, founder of My Doctor Friend Oh, by the way, we're going to explain that. It means just what it sounds like, but there's a whole lot more to it. Michael, we are extremely pleased to have you. Welcome to the pod. Why don't you take a moment, introduce yourself, describe what My Doctor Friend is, and then we'll dive right in.
Michael Turken: Sounds great. First of all, so delighted to be here. Thank you both for having me on. So my doctor friend is a co-pilot design for patients. I, great, yeah. ⁓ glad.
Doug Rogers: I love that. I need that. But actually, I have that I downloaded the app. It is on my phone. I just want to put a plug. I did download it. I created an account. So yes, love it.
Michael Turken: ⁓ thank you so much. Yes. We launched the app just a few weeks ago. Really? You know, I out to create something I hoped all of my patients would have access to. So I work in an urgent care clinic and I just see every week people get sort of lost before they even engage with the healthcare system. And so that was, that was the original we tried to solve. And since we launched the app, we've learned a lot from, from seeing what people need beyond just sort of navigating that first mile of care to. better understanding their health plans, understanding when to reengage with the healthcare system, managing the health of other people and their family.
Russell Stover: Yeah, so and Michael, you practice right now, you work in an urgent care center. Tell the audience about your background, because I know that you've worked in technology before. So this is not something that you've been dreaming of for some period of time. Well, maybe you have been, but it's not that you've been dreaming of jumping into the world of high technology. You've had some experience in this arena.
Michael Turken: Yeah. Yeah, thanks for that. Yes. my first ⁓ in the tech space. I, you know, I've been in health really since I was a medical student. think what got me really fired up about the opportunities in health technology was just seeing how really for me, was kind of the end for health care technology. I just saw how unsafe this process was for patients and how even the best medical centers really struggled to get timely diagnoses people. and so, you know, short, shortly after finishing, my training worked as a hospitalist at Stanford, and then I joined a place called forward. So they were a tech enabled direct primary care practice. Sadly, they, shut their doors, you know, several months ago, but there had, you got the incredible experience of essentially building a healthcare system from scratch. built personal health records for patients. built electronic medical records for our. clinical staff and really thought we're trying to think ahead to the time we're not living in when AI would be more powerful
Load full transcript
than it was at the time. then after Forward went to Omada Health helped launch their programs for people with diabetes and hypertension and really sort of got familiar with how does remote monitoring work? Where does it work? Where does it not work so well? And then my next stop after that was working at Google, worked on wearable devices, data platforms. The version of the personal health record, we can talk about that. And then place I worked was on the Google search team, which really gave me sort of insight into what has Dr. Google the team behind Dr. Google been thinking for a long time, and have they sort of seen that were not yet addressed. So yes, I. I will say I came into my doctor friend already having a lot of sort of strong ideas about what was needed and where our technologies were falling short, but I've still learned a tremendous amount since starting this journey.
Doug Rogers: With my doctor friend, just take a moment, when like myself downloads the app, what is the experience and that you're hoping that I have as a consumer of healthcare and in the engagement that you're wanting me to have with the app from a healthcare perspective?
Michael Turken: Yeah, it's a great question. you know, without doing sort of a live demo, know, the idea behind my Dr. Friend is, know, I think at its surface, it initially looks similar to any sort of AI chat interface, although the system behind is quite a bit deeper. The basic problem we're trying to solve, think, across, not knowing what your specific health concerns might be, I will say that the primary problem we're trying to solve is
Doug Rogers: Yeah.
Michael Turken: people worried about their health. I think in general, at various times in our lives, depending upon what health concern we might have, we just have a lot of worry. Am I doing the right thing? Am I talking to the right people? Am I taking sufficient action? our health care system has been designed around scarcity. And so the cost that is that people have to with a lot of worry between visits. And so think the primary problem that we try to help people with is how can we help people
Doug Rogers: Yeah.
Michael Turken: be less worried. sure that a new symptom crops up today. They have at least the beginning of a plan, whether it's just to monitor that new symptom, say a cough that you've had for two days. Can we hey, watch this for few more days, see if it gets better, if it gets worse. And then if it doesn't come back, then we can talk about next steps. I think that's at fundamental level what we're trying to help people with. And then there are all the things that sort of ladder on top of that.
Russell Stover: so that's been my experience when I've used it and I kind of I as a plug I would recommend anyone listening this to go give it a shot because it really is very easy to use and I you know whenever I think about your history I I I've tried to you know put myself in the perspective ⁓ of you the patient and the provider as much as I can I've never been a doctor but I think that you've spent a lot of time with pretty large organizations that have been, or smaller organizations that are growing, or a very large organization like Google that are approaching consumer health technology with the ideas of kind of trying to transform the nature of the relationship between the health system and the patient around technology. tools and technology. And I think that's it's an interesting space to be obviously right now because there's so many things that are going on. But ⁓ curious, what some the lessons that you think have been learned over the years from approaches that large organizations have to try and engage with consumers? What are some of the things you think they've gotten right and some of the things they've gotten wrong whenever they to actually engage with consumers? maybe, we talk patients, consumers, but just people, whenever they try to engage with people to ⁓ take ownership their.
Michael Turken: Hmm. Yeah, that's a fantastic question. And the last, I would say, five years have been just tremendously exciting. I think some of the things we learned in the past, have to maybe think again. We're living through a major paradigm shift in technology. I think large companies have always seen healthcare as an enormous opportunity. How can you not? mean, it's an enormous industry. I think the question has always been for the bigger companies is how much, what is their appetite for risk? Right? And the risk comes from all different kinds of places, whether it's regulatory risk, legal risk, reputational risk. you know, Google has taught, I think the whole industry, you ways that can go wrong. And I think it, tremendously affected Google's position in how it navigates the health space. Because this is no secret, but health-related search, and search was the main revenue driver for Google, still is, was anywhere from 5 to 7 % of their traffic. That's enormous. I mean, it seem small, but that's a huge, that's a lot. ⁓
Doug Rogers: 5 % of the biggest number in the world is a really large number.
Michael Turken: That's right. Yeah. Yeah. It was, was, it was enormous. And then the question is always, well, what do we, what do we do with that attention that we have from people, you know, Google to its credit, I think cautious about monetizing that attention. And some that was sort of hard learned lesson. You know, they, they realized that, you know, the bigger you are, the more, the more people start to worry about what other motives you might have. And so ⁓
Russell Stover: I didn't know it was that high. That's amazing. Yeah.
Michael Turken: I think traditionally most of the big companies have been pretty risk averse in the consumer health space. And a lot of what they've been trying to do is sort of avoid negative attention and hopefully get some positive attention. And Google famously did that. the COVID pandemic, they correctly navigated people to places where they could get
vaccines and sure really reliable information was front and center. And I think that's generally been the level at which a lot of big tech players, at least in the consumer health space have tried to be useful, right? Which is, let's make sure there's high quality reputable information that people can sort of see upfront. Amazon is sort of famously different. They've sort of charged headfirst into the health space. And that's why I say the last five years have been sort of interesting because they've really tried to change possible in the consumer health space. then more recently, think, know, ChachiPT and Anthropix sort of entrance into the consumer health space. also is going to be teaching us a lot of lessons. I'll try not to be long, it's super long, yeah. Yeah.
Doug Rogers: Well, I think lessons is the right word because lessons implies we're going to learn good things, we're going learn bad things, and then the balance is going to be does the bad influence the good or vice versa.
Michael Turken: That's right. That's right. I will say from like a strategic approach, I think one of the challenges for all tech companies, it's actually rarely the tech, it's rarely the talent or the resources, invariably it's the culture. ⁓ And that's the part that they find hardest to get right. I think that like the culture of healthcare and the culture of technology are just, they're really opposed and they...
Doug Rogers: Yeah. Yeah.
Michael Turken: I've seen it over and over again and I've heard it from friends in this space. It's just like butt heads. And where this happens, you can have really incredible productive things happen, but you can also have less favorable outcomes.
Doug Rogers: Yeah, mean, I think that's a really great point. And this not to say that we should slow down on the technology side. And I do think that there is an intrinsic ⁓ in some the mentality that tech first, if you will, organizations that are going into health care have a disruption mindset. Why does it have to be this way?
Russell Stover: Yeah.
Doug Rogers: Why are we doing it this way versus that way? Why do we have this level of risk or these many people involved, et cetera? I think actually there's real value in kind of breaking some of that thought process down, but you can't miss the point that at end of the day, we're impacting health and wellness for a person. that you... You can only break so many things while you're going fast before you do real harm. And that has to be at the core mindset of everybody who is in this space.
Michael Turken: Well said, very well said.
Russell Stover: Yeah, yeah, as somebody and I, I sometimes forget this, but so I obviously I worked for a high tech company too. I worked for Cerner and they were acquired by Oracle and I worked for Oracle for three years. So, in that ⁓ you really get to see the differences in culture. I, ⁓ Doug I both worked for Cerner. they'll probably, you may not know this Michael, but there's probably several people
that will listen to this podcast that did work for.
Michael Turken: Mm. Huh.
Russell Stover: for Cerner. Some probably may be working for Oracle. Yeah, yeah, yeah. And many that are probably working for Oracle right now. So they'll kind of understand firsthand the differences. think whenever you there's term health tech and I think that that is there are health tech companies out there. You operate one. I work one. Doug works in one. I think that
Doug Rogers: Probably more than several, but yeah, that's right.
Michael Turken: haha
Russell Stover: in the Venn diagram, like collision course between healthcare and technology, I think that the successful companies have to figure how much they overlap between each of those. And inevitably, I don't want to necessarily say that, you know, ⁓ you have to more of a healthcare company whenever you're in healthcare. But I do believe that if you are in healthcare, you have to be a healthcare company.
Michael Turken: Mm.
Russell Stover: You can't be a tech company that does healthcare. If want to be successful, you have to really think about it as a company that's actually being a healthcare company because the values and the relationships are so much different.
Michael Turken: Yeah. Yeah. Yeah, I think it's very true. be curious, ⁓ you've ⁓ it too in sort of a few places, how these actually approach organizing themselves that way. I'm curious how you've seen it. I imagine Cerner must be getting some of this right.
Russell Stover: think Cerner was and this is it's an interesting question. And I, I felt said, healthcare is very relationship focused. Things take a long time to happen. And I think that you can't scale relationships. And think that's very problematic ⁓ healthcare organization or technology companies, particularly in the, weird how like that, and friend ours named ⁓ Dick Flanagan to call relationships like the, or trust the of healthcare. And I think connected to And see it at almost every level of the points between people that exist is trust is so important. Whether you're a technology provider that is providing EMR to a health system or whether you're a provider that's working with a patient. And the of that relationship and the trust that is developed has deep meaning and it progresses over time. And it's why it's so important like You know, we think about like why it's so important to have a primary care physician and most people don't. it's really problematic. But yeah, I think that it's really hard for tech companies to kind of understand that, at least in my experience.
Michael Turken: Indeed.
Doug Rogers: Yeah, I would certainly echo that. think it's a hard thing to also convey. Why? Why that? I mean, it feels right when you say it, but there's a deeper meaning that until you experience it, it's hard to really intrinsically understand Michael, I to go back a little bit because as you mentioned when you talked about the journey you've been on leading up to my doctor friend that you were at
Florida Health, you were at Omada et cetera. Some of these, you know, Omada Forward Health, Those companies had done really great things, innovated, pushed boundaries. Omada clearly still pushing the boundaries and doing really well. But if I think about it, they merged almost this hybrid experience between technology and the delivery of care. Whether it's in the case of Forward where you had some, I think, some physical care infrastructure, with Omada, it's virtual, but it's still there as a care provider that's coming into the equation, et cetera. As you think about and with what you launched at My Doctor Friend, it is all, won't the word disintermediation, not in a bad way, but you are very much focused on the consumer experience ⁓ helping empower the consumer. How does it then connect back to the care provider? Or how do you want it to connect back? How do you think this should connect, whether it's primary care or specialty, et cetera?
Michael Turken: Yeah. That's right.
Doug Rogers: to then join those two sides of the same coin.
Michael Turken: Yeah, yeah, great question. Well, you know, the first step for us, we're to keep this a focus as long as possible, ⁓ is ⁓ I one of the hardest things to, one of hardest reasons, I would say the most common reason why sort of products end not meeting people's expectations in the healthcare space is that you're trying to design for so many different end users.
Doug Rogers: Yeah. Yeah.
Michael Turken: You know, you're designing for providers, for the insurers, for employers, and like, by the time, you know, your product's ready to ship, it doesn't really serve anyone's needs very well. And so we're starting from the premise that the patients who are really well-informed and really empowered are going to themselves have better healthcare experiences, engage the healthcare system in the best ways possible. And then when it is time to meet the providers, whether it's a primary care provider or a specialist, that interaction between patient and the specialist and the provider is going to be as good as it possibly can, as effective as it possibly can. That's the goal. That's the goal. What we're trying to avoid is something that frankly, like I and I know a lot of other people see in clinic, which is the structure of the interaction is getting sort of interesting where doctors have their open and a lot of patients have their chat GPT and Then there's a question of, like, you know, what happens or is your AI and my AI gonna sort of duke it out? And much of this, how much of the ⁓ time are we gonna spend arguing our respective, you know, points of view? I think that's a ⁓ genuine worry some have, not all, but many do sort of worry that, you know, if patients come in having done a lot of research on chatt GPT, but it's not with all the right context, I've done this.
Doug Rogers: Yeah.
Michael Turken: my colleagues have done this, you can spend the first 10 minutes of a visit trying to convince someone about reset exactly and recontextualize that. would not be a good place for us to end up. How we do this, we can get into sort of the details of it. think we're still honestly, we're trying to figure out what is that ideal paradigm look like. But we ⁓ have ⁓ theories, hypotheses.
Doug Rogers: reset our, yeah, yeah. Well, and so let's pull on that thread and I'll probably start with this thread, which is trust. Russell, did, I love you bringing forward Dick Flanagan's phrase, which is trust is the currency of healthcare. it said another way, healthcare moves at the speed of trust. My doctor friend is a tool that I, as a consumer use, I would say to bridge the asymmetry of information in healthcare. between me and the healthcare system. Help me be better informed, answer questions when I can't get to a provider. You've got capabilities in here that not only help me track, but also point me to providers. As I say point, I think it's doing a search, it's pulling forward, et cetera. So I'm getting visibility, I'm getting an understanding of my healthcare. How do you, and I say you being my doctor friend, create trust with me?
Michael Turken: Hmm.
Doug Rogers: And I'll separate that from what other bigger firms are doing, like an OpenAI with NowChat GPT for Health that answers a lot of questions, or just Googling it, or call it, et cetera. How do you create a different experience of trust? Or what does trust mean for you as this app engages me as the consumer?
Michael Turken: I love that question. So this is sort of the tremendous irony of building with this extremely powerful technology, which I think of it all the time as sort of electricity. And I try to sort of imagine what it was like when electricity first came out and everyone was of wowed by the light bulb. And surely, no doubt, people were electrocuting themselves left and right, trying to better light up their kitchens.
Doug Rogers: Yeah.
Russell Stover: Come on. This is it.
Doug Rogers: Yeah.
Michael Turken: So we have this incredibly powerful technology and its shortcomings have been well publicized. These models, even the best of them, they hallucinate. Dr. Wachter, I like his term better in his book that just came out. He uses the term bullshit. They bullshit a lot. ⁓ And they also are sycophantic. So ⁓ sort of to agree with you.
Doug Rogers: Yeah. ⁓
Russell Stover: with the
Doug Rogers: Yes.
Michael Turken: And they also have amnesia. They forget things. So the three of these things combined are arguably our biggest challenges when it comes to building trust through a product. I think these the problems to solve. And so we tackle this trust minefield in a few different ways. The first is not really necessarily visible to people. And that is that we use sort of multi-agent architecture. So we don't use a single model. We don't just use you know, an open AI model or an anthropic model, we use it a minimum three models often more. and what this does is it decreases the likelihood that, you know,
⁓ model may hallucinate, but the other one is ⁓ likely to catch it. so people using my doctor friend won't see that it's not, we don't advertise it anywhere, ⁓ but, sort of behind the scenes. And I think we hear from, from our members who my doctor friend that they just get better answers.
Doug Rogers: Yeah, it's, yeah.
Russell Stover: Okay.
Michael Turken: My doctor friend asks better questions. It gives better answers. But then I would say the second pillar for us is design. we try throughout the experience to tell people, this carefully. It makes mistakes. And more than just like a tiny disclaimer at the bottom that says, which a of the other providers say, this model makes check its work. We this feature, we it the confidence meter. And it's partly rules-based.
Russell Stover: Yeah, it's really cool.
Michael Turken: ⁓ thank you. Thank you. Yeah. It's partly rules-based, but partly not. And the basic idea is if we know a lot about you, if we have a lot of your, you know, your health history, we know a lot about whatever concern you're bringing. We can have a generally a higher confidence in our answer, right? If you just come in to my doctor friend, first time, first time person using it, you say, you know, I've got, I've got a abdominal pain. you wouldn't want a system like this to do is say, ⁓ well, it could be appendicitis. And you should go to the ER, right? Like that's, that's a bad experience. What might be a better experience is saying, well, you know, it could be appendicitis, but my confidence level in that is about 10 % because I don't know anything about you. don't know if you're, I don't know if you're, you know, 20 years old or 90 years old, right? Like that's, that's a sort of better design. And I think for a while still, we're going to have to lean on design in that way to, to, to teach people how to use these tools responsibly. Cause I think, you know, it is an incredible.
Doug Rogers: Yeah. Mm-hmm.
Michael Turken: capability and power, part of designing trustworthy systems is helping people to understand when do these things work well and when can they fail.
Doug Rogers: Yeah, I'm going to take a step back for a moment and stand on a soapbox, figuratively not, not literally. One, I love the fact that that is a design principle, trust and exposure of trust. mean, if you saw me looking down, I'm in your app, I was putting in information and I see the confidence meter going up and down based on how much more information I'm getting or giving or not giving. And I think that is a really, really important factor. would go so far as to say that there is a real danger in the proliferation of these models and these models being embedded in every single thing that we do, whether it's a software product, whether it's a native experience, et cetera. And no way for me as an person, average can mean a lot of different things, but as an average person, ⁓ to know what can I trust? Is this model good or not? Is it telling me the right things?
And I just candidly, whether it's on the consumer side, on the health system side, I really want to see a accreditation body that goes out and provides some unbiased mechanism that says to a health system and or to a consumer, the model performs this way, it's performing better than it should, you can trust it.
Michael Turken: Yeah.
Doug Rogers: are the model, you may still want to use it, but you shouldn't try it. Kind of like a consumer trends or whatever black book for ⁓ products. But I think we're at a deficit not having an unbiased accreditation agency to help both the consumer and health systems and providers when they purchase or use these things to know, they actually getting what they should out of it? That's my soapbox, by the way. I'll step down. ⁓
Michael Turken: Yeah. Yeah. No, it's a good soapbox. We need more people standing on it. I, yeah, you know, I think, I think the initial reaction to it is just this technology is so, so much better than what we had before. And I still think we're in that honeymoon period. Everyone's like, they remember what it was like to Google search for these things and have to navigate it, you know, pages full of blue links or, and as, as you're right, like these systems can be unreliable. But again, was like quoting, quoting Dr. Wachter again, you know, he sort of invites people to think.
Doug Rogers: Yeah, I agree.
Michael Turken: ⁓ it to perfection because the healthcare system's never been perfect. Compare it to the alternative. ⁓
Doug Rogers: Well, that was Zach Cohane's ⁓ in New England Journal of compared to what. ⁓ fact, we'll put that out when publish this. We'll put that out in the notes. It's a phenomenal article. It's not long. I think it's two pages, three pages. But the premise and the thesis is ⁓ really which is what are we actually comparing these models to? ⁓ And what's goal, which is delivering the best care possible in every scenario?
Michael Turken: ⁓ yes, yes, compared to what I love that article. I That's right.
Russell Stover: Yeah.
Michael Turken: That's right. And other part of this, think, is people are ⁓ starting study this in more rigorous ways. And I think that's the positive. I mean, at first, everyone was throwing medical licensing at these models and seeing how well they perform. And everyone was patting themselves on the back saying, wow, look how well these models perform. And then it dawned on everyone, wait a second. People don't interact with these models. Only doctors would interact with models these ways. Most people don't ask questions like a medical licensing exam does. they're starting to think more rigorously about, if is going to have this long-term relationship with AI, how can you measure quality not just over a single question-answer pair, but weeks or months, a year of conversations? ⁓ So it will be fascinating to see. And then maybe Wirecutter will come out with, they'll have like a
Russell Stover: Thank
Doug Rogers: Yeah. Yeah.
Michael Turken: not an accreditation body, at least someone at Consumer Reports will come out with something guide people.
Doug Rogers: I would just shout out to our community. For those who are interested in building lots of things you can build, there is an opportunity to go build a validation engine, maybe open source it, put it out to the community, but put a validation engine because Russell is living this, I live this in my day job. Health systems are rushing to buy systems that provide agent capabilities to interface in the delivery of care. the more that they can validate trust, the better they will feel comfortable doing it and the consumer will feel
Michael Turken: Yeah, couldn't agree more. I couldn't agree more.
Russell Stover: Yeah. Go ahead, Michael.
Michael Turken: I was going to sort of add like I we need. I'll put even more specific plug because I would love this selfishly. We'd like this for my doctor friend. know, we try to, I think am not alone here. think a lot of digital health companies have been starting to do this with AI is they try to validate their models by using AI to a patient. Right. And so they run like an AI patient through their system and they say, okay, well, how well does it do? ⁓
Russell Stover: Yeah.
Michael Turken: We tried this internally and it was just like, it was an absolute embarrassment. was somewhat public about it I try to share this, but AI is terrible at impersonating like real people. It's very good at giving, you you think about the answers you typically get from chat GPT or, know, perfect sentences, no grammar mistakes, really clear. Most people just don't talk that way and getting the AI to of act in the way a typical person would. It's like, it's nearly impossible for AI to do.
Russell Stover: Really?
Michael Turken: So I think someone needs to do is like, rather than you ever was focused on the AI doctor, someone needs to make the AI patient. yeah, I know, please, please.
Doug Rogers: ⁓ hey, well, hey, shout out to the community. You heard it here first. Yeah, that's a really compelling idea. Mike, I want to turn back to, just for a moment, from to where you are now and then obviously where you're going. You have been in some phenomenal organizations. You are still a practicing physician. You are seeing patients. You're delivering care. today and you are building a technology company. What is it like to build? ⁓ Because on the West Coast, right? You're in California. ⁓ You're in people think of as the Mecca for building, ⁓ I perception of people like and I who live in Kansas City that you just walk out the door and there's a VC throwing checks. ⁓ You do you do this, you do this. But I mean, seriously.
Michael Turken: Yeah, yep.
Doug Rogers: What is your experience in trying to build this from the ground up and what are some of the hard decisions that you've had to make to stay true to the mission you've articulated for my doctor friend?
Michael Turken: Yeah, well, first of all, I'll disabuse you of the notion
that the VCs are just throwing, you know, out there throwing money at people like me. I mean, I wish that were the case, but no, it's ⁓ the case, especially, especially consumer health. think it's a different, a different space altogether. Yeah, I, you know, I think the, the day to day is like a lot less glorious. It's, will say it's, it's a tremendous amount of fun. But ⁓ most of it my co-founder and me either, you know, honestly, like working, working in, in, in,
Doug Rogers: Yeah.
Michael Turken: you know, my kitchen, his house, occasionally a co-working space, working late into the evening, you know, on the weekends. And, you know, feel sort I feel fortunate that we can still build for patients. I mean, I this is something that hope... Open AI and Anthropics entrance into the consumer health space will change, which is a sort of perception that consumer health is just like, it's not a place where you should invest. It's like too hard. It's a graveyard. I've always wanted to design a product for patients because I feel like they're just often the, ⁓ the last thought of because they're often not the ones who are sort of paying for the product or the service, right? So.
Doug Rogers: Yeah, they're two steps removed from the buyer, where the person holds the money,
Russell Stover: Yeah.
Michael Turken: That's right. That's right. So yeah, I mean, I would say if there's, know, we're gonna do this for as long as we possibly can. I feel... ⁓
Doug Rogers: Are you and your co-founder bootstrapping this? Did you raise money?
Michael Turken: We've had a couple of angel investors that have helped us get things started, but still bootstrapping quite a bit. And yes, we're looking forward to raising. We're going to raise in the spring.
Doug Rogers: Yeah. Yeah. Yeah. That's awesome.
Russell Stover: Yeah, and it's interesting. talk with people from Kansas City a lot. ⁓ get, there's different perspectives and ⁓ what like to build a company in Kansas City. There's a lot of parallels, but there's also differences just based on the nature the financing and the people arrangements in the different parts of the country.
Michael Turken: One question I'd curious get both your takes on this is, I'm sure you both followed sort of Doctronics sort of pilot with AI sort of prescribing. this is, it's not in the Bay Area, it's in Utah. And curious what you think, how much of innovation in this space is going be more driven by sort of local regulatory environments or openness to experimentation than
Doug Rogers: Yeah, yeah. Yeah.
Michael Turken: sort of traditional, know, the traditional, like everything happens the coasts.
Doug Rogers: ⁓ What do you think Russell?
Russell Stover: think a lot of it probably will be. I don't that healthcare has ever been lacking for good ideas, but what has at times lacking for are paths to actually capture ⁓ and release the technology in way valuable to those who use it. I think that it's interesting that it just popped up that way. think that there's something to that. can usually find, when working in sales and you're out in marketing, you find specific patches where a particular technology makes a lot of sense. And you go you try to filter through the market way. So I think that there's probably going to be more of that. What's interesting about Doctronic is they found a very specific state regulator that allows them to expand this beyond just that state, which is really interesting.
Michael Turken: Yep.
Doug Rogers: Yeah, I mean, Michael, you certainly you have a more of a lived experience than Russell and I do being a physician, practicing physician, et cetera. I think what what I was surprised by, even though Russell, been in health care, health care technology. I don't want to date ourselves, but many, many moons, if you will. The both the power and the flexibility that resides in state regulatory bodies. as it applies to the practice of medicine, the licensure across different roles and especially around the prescription piece. My default was, hey, this is a federal thing. You had to go to the federal government. So I think that's the unlock and the beauty of what Doctronic did is they found a state, ⁓ by the way, they didn't go through the health department. They went through the Department of Commerce with a sandbox, so an innovation group. know, states have these, but...
Michael Turken: Yeah. That's right.
Doug Rogers: Now, you people shouldn't be surprised about Utah because they did it in Salt Lake where the capital is. Big entrepreneurial community, not that different from the West Coast or other communities, very tightly connected to the innovation arm within the Department of Commerce. They built a relationship. That group then went and broke ground with the ⁓ of Health or whatever department around the licensing requirements. And so created flexibility, helped them navigate, but was seen more as an economic development.
Michael Turken: you Yeah.
Doug Rogers: innovation investment than it was a healthcare thing. By the way, we wrote a sub stack on this. We'll link to it in the notes, kind of the journey of electronic, et cetera. Here's the thing that I looking for. I'm just surprised by how pissed off should Intermountain Health be that ⁓ their are shouting to the roofs, I've got my MyChart message is full of renewals. And yet, you know, the state
Michael Turken: Yeah. ⁓ excellent. You
Doug Rogers: let some company that's not even based there get approval to do AI only renewals. wasn't University Utah, it wasn't an amount of health, wasn't that, I mean, that's the real question I have is why, why not? Why were you not at the forefront and how upset should you argue because you let somebody else take the innovation and quite frankly, the productivity gains that all of your doctors are clamoring.
Michael Turken: is, yeah. What do you think, Russell?
Doug Rogers: Yeah.
Russell Stover: You have to think I do wonder like that's a great question Doug and I don't I wonder
if I wonder if first of all I imagine inner mountains thinking about that. yeah.
Doug Rogers: Well, they have to now. I mean, I imagine part of this is they didn't like they don't have a product, but they have a vendor and they have multiple vendors. They could have brought the vendor to the table that they were It just it surprises.
Russell Stover: Yeah. Yeah. Well, and it is. do what I think if other I wonder if other health systems are thinking about that too. And, know, the opportunities that they have knowing their local market and how can, you know, be specific in the way that they deploy technology to it. I did have a question for Michael. I'm I'm curious. So they've I think so. Dr. Tronic, this is related. They've They've picked off a very specific portion physician workflows, essentially, which is the ⁓ renewals or what they're able to take on, which is, I haven't thought about this, but I'm thinking about it right now. That seems like it's probably the easiest and most logical first step in the process. But there's more complex parts of that.
Michael Turken: Yeah.
Russell Stover: write in new prescription and the diagnostic process are more complex and more upstream. How do you think that unfold? What do you think might be next in terms of the opportunities automating portions of provider workflow?
Michael Turken: Yeah. Yeah, it's a great question. I wonder if a good place to look is the that nurse practitioners and PAs took to increase the scope of their practice. There are few famous studies, but probably the landmark one is this Burlington randomized control trial. think was in mid 70s. ⁓ And if look at in sort of doctronics initial,
Doug Rogers: Okay.
Michael Turken: sort of ⁓ research plan they're really really similar it's like a very sort of narrowly scoped ⁓ in the burlington study they looked i think it was only a couple practices maybe like a handful of nurse practitioners they looked at medication prescription patterns across a really narrow set of medications and it more of a question of sort of is this non-inferior ⁓ to to ⁓ to you physician-led care and they had physicians overseeing
Russell Stover: Interesting.
Michael Turken: you know, all of this care too. So there was that additional sort of safety guard in place. And ultimately the regulatory argument that was made is, Hey, there's not enough care to go around. Yes, we don't have, you know, randomized control trials with, you know, tens of thousands of patients or hundreds of practices, but you know, probably good enough. Let's, let's proceed. And so I suspect the same thing is going to happen here. you know, the, the medication refills, I think they, they chose the right place to start. you know, low risk meds for refills. I think it'll be fairly easy to show that this can be done safely. And they choose to like license that technology or to your point, know, Intermountain will, you know, get things together and launch their own version of it. I can't imagine healthcare systems, know, standing while a single company sort of does this. It's just, too valuable. You know, it's too valuable. But eventually, ⁓ I think it's going to be, you know, they sort of scoped it by, by medications, right? There, there are these sort of like low risk medications we can start with. Um, but I think ultimately the way it'll probably get structured is, really being able to capture or understand what is a low risk patient. It's like, you know, calculate the risk of a given patient and then the risk of the med and think of the sort of Venn diagram again here. And more and more sort of low risk encounters will be, will be automated. So. they're starting with refills, just to give you a concrete example, you know, if you're someone in your early forties, the only problem you have is cholesterol that's high and you want to start a statin because you have, say a family history, you know, there's like a simple lab you need to get, want ⁓ monitor, you ⁓ do little basic monitoring, but that's a low risk medication. We can probably start it. And so I think that's the way it's probably going to go is I'll say, Hey, if you're low risk, or like, you know, early diabetes, you know, we can start a simple medication like metformin, right? That's gonna happen faster, I think, than people realize. ⁓
Russell Stover: Yeah, that path makes a lot of sense. It's similar circumstances like you mentioned that have happened in the healthcare industry before. So that's kind of what you, it's useful to study that. And then what I think is interesting is how fast is it gonna happen? Nobody knows that, right? That's the...
Michael Turken: Well, yeah, I think you're right. one knows. I will say there's a story I think that's very instructive here. It's also kind of like a funny story about sort of automation in primary care. Kaiser for some time has, has had these where they call them order sets. You're probably familiar with this. They have order sets for diabetes. So if you're new patient in Kaiser system, you get diabetes. They, you know, they enter diabetes order set and they have a whole standalone sort of diabetes center that's run by primarily by RNs that are following really clear protocols about which medications to start, how to adjust the doses of medications, all very algorithmic, no AI. It's been going on for a long, long time. And, we some clinicians when we first started forward that came from the Kaiser system and had couple of them come to see me early on when we launched our clinic because they were seeing patients with diabetes for the first time and they needed my help with writing insulin orders because they hadn't written insulin orders in so long. They forgot how to write insulin for people with diabetes. so the point this is that these
Russell Stover: Wow.
Michael Turken: there's a really good precedent that doesn't even rely on technology for doing this. And I think that's part of what's going to accelerate it is that we won't
necessarily have to wait for these super long studies. can imagine a place like Kaiser saying like, wow, we've been using this protocol for years now and it works really well. Let's give to AI to do it.
Doug Rogers: Yeah, that's right. Just a couple of stats on renewals and then I've a question back my doctor friend for you, Michael. So part of the value for Doctronic and then any others, they're not going to the last. It'll actually be interesting. What's the timeframe from them to the next vendor? That's what I'm watching. It's not so much when they go to the next state. I actually think you're going to see an announcement in the next three months. Mark my words, we'll find out I'm right. but in the next three months then another state has approved them. The real question is the timeframe from Doctronic to the next vendor. But the value, part of the value of chipping away at this is one for the average primary care physician with a panel of 2000. This is roughly about two hours a day. Writing prescriptions or prescription renewals is roughly two hours of the primary care physician's day.
Russell Stover: Thank
Doug Rogers: It makes up 30 to 40 % of the time that they spend after hours after they leave the clinic. counting productivity gains come now from all the ambient scribes, but after they leave the clinic and they're in their pajama time and 70 % of all the time around, whether it's writing or renewing prescriptions is on the renewal side. So ⁓ just took the one, a big chunk of time and the biggest chunk of that chunk of time and automated it. but abstracted and said, this is also the lowest risk. And so then what comes next? So, fascinating. It's a really interesting space to watch. I'm curious, Michael, I want to turn back to my doctor friend, because once just a really, really powerful asset for the consumer as they navigate healthcare. What is the that y'all are leveraging to get people like myself to become aware and engage with? So said another way.
Michael Turken: Yeah.
Doug Rogers: What's your growth strategy?
Michael Turken: Yeah. Yeah. Well, share some of it. I won't share all of it because some of it's top secret. But I that's going to sound sort of obvious. We're trying to make the best product we can. the distribution is hardest to in any kind of product and no ⁓ easier in the health space. We're
Doug Rogers: Fair enough. Yeah.
Michael Turken: I feel like we've, we've gotten, and I feel very fortunate that we've reached this point where of the people using my Docker friend are already starting to tell their friends about it. They're posting publicly about it. Like we are seeing product market fit. we've, we've met need that people have that's, know, many of these, most of these people have used chat GPT extensively before they, and now they just sort of leave it behind. use my Docker friend and they're telling people about it. And I think that's part of how we've been able to get this.
Doug Rogers: Another.
Russell Stover: Thanks.
Michael Turken: earned media and were mentioned in the New York Times and then twice in NPR. Thank you. Thank you. Yeah, no, it's a boon for a small outfit like ours.
Doug Rogers: Congratulations, Byway. That's super awesome. Super awesome. By the way, what a drop off to go from the New York Times to the counterweight. So like we feel impressed now that we've got. ⁓
Michael Turken: ⁓ my gosh. No, not at all. at all. No, my gosh. is my first podcast as I mentioned. I'm honored be here. Yeah, product-led growth, think, it's the best thing you can accomplish in the consumer ⁓ space, right? other forms of sort of, there's a lot of noise out there and there's a lot of, and that's only to increase over time. I do think that to the point you made earlier, think trust is the big, is a really, really big thing here. And I think that if we can build up just a product that's more reliable, more trustworthy, word gets around.
Doug Rogers: That very compelling message. And owe for our community. They hear that loud and clear because strikes me, Russell, we've heard this from a number of guests now that as you build a company, there is no substitute for the dedication and just maniacal focus on the product right for the person, ⁓ the entity that selling to.
Russell Stover: Right.
Doug Rogers: you may cover a lot of ills around a go-to-market strategy, a sales team, et cetera, but that is a band-aid on our approach if you don't have just a well-executed, well-built product that somebody truly gets value out of. And I think those are words of wisdom for ⁓ our for our community, ⁓ as they think about their own endeavors.
Michael Turken: Yeah. Very wise words.
Doug Rogers: So, we've covered a lot of ground. One, I would encourage, we'll drop this in the notes when we publish the pod, the link to go download, my Dr. Friend, would encourage everyone who listens to this, you should download it, create an account, you should interact with it. Find out what your own experience, mean, what I've heard from you, Michael, is not so much take this my word, but go experience it and validate that the trust is deserved. So please, ⁓ for our do that, go. ⁓
Michael Turken: That's right.
Doug Rogers: Go see if this is valuable to you. think you will find that it is. But as we bring this to a close, I would just want to fast run three questions, two related to what we've been talking about. Russell has no idea what I'm about to do, so this is even better. ⁓ And one completely unrelated. We'll see how well that fares. So first one, do ⁓ you it's a good idea? ⁓ Do endorse or not endorse AI getting licensed as a clinical provider?
Michael Turken: Okay.
Russell Stover: Thank
Michael Turken: think it's inevitable. I think we be moving in that direction. We should do it safely, but ⁓ we are meeting needs. ⁓
Doug Rogers: Yeah. Yeah, fair enough. So I'm going to take that as a general endorsement that it's going to happen, whether it's good or bad. Although I feel like kind of the delta in terms of resources out there almost speaks to the necessity of it. And we just need to do it right. All right, second one. And we really touched on this with Doctronics, but time horizon. If you were a vetting person, you may be. I don't know.
Michael Turken: Yes. That's right. Hmm.
Doug Rogers: When is AI going to be authorized to actually prescribe without a human in the loop? Doctronics is doing renewals. That's not a prescription. It's a step removed. is AI going to be authorized to actually prescribe a medication without it? Five years, all right. ⁓ Fantastic. Yeah, I not bet against that. I would not bet against that. All right, third and final question.
Michael Turken: I think within five years, possibly sooner.
Doug Rogers: We were talking, Michael, before we started the pod. I commented on how much I loved your microphone. And you said part of this was because you got into music during COVID, et cetera. ⁓ you go listen to an artist, who's the artist that you are most excited to go listen to?
Michael Turken: Yeah. Live or just like on, know, or just, okay. ⁓
Doug Rogers: Let's say live, you're gonna go spend money to go see them in person.
Michael Turken: Okay, Jacob Collier, hands down.
Doug Rogers: All right, I don't know Jacob Coller. Now I'm going to, what genre is...
Michael Turken: a genre-bending ⁓ artist, does everything from folk to jazz to ⁓ pop. ⁓
Doug Rogers: Okay. Nice. Russell, do you know Jacob Collard?
Russell Stover: I don't know this, but I'm writing it down. I'm going to go play some.
Michael Turken: So, yeah, yeah.
Doug Rogers: I'll write it down. When we end this, I'm going downstairs and I'm putting this on Spotify. I'm going listen and I'm going to catch up on it.
Michael Turken: Yeah, you're in for a treat. They call them the Mozart, the Mozart of our generation. So.
Doug Rogers: Wow, so Jacob, say the last name again. Collier. All right, I love that. Well, you heard it here first, First of all, should be licensed as a doctor. Five years or less, it's going to prescribe. And ⁓ everything is around you, hit up Jacob Collier on Spotify. It's going to make the world feel better. Michael, I couldn't imagine a better conversation. You've been a wonderful guest. Thank you. ⁓
Michael Turken: Collier, C-O-L-L-I-E-R. Yeah. That's right. ⁓ my gosh. I thank you both.
Doug Rogers: And thank you for what you're building. We need more people like you building these things for consumers in the space that we're in. So we appreciate that. Thank you for joining the counterweight.
Michael Turken: ⁓ gosh. Thank you, Doug. Thank you, Russell. This was a treat.