Subscribe to The Podcast by KevinMD. Watch on YouTube. Catch up on old episodes!

Your patients are asking ChatGPT what’s wrong with them before they call your office, and getting an answer in seconds. In this episode, Tod Stillson, a board-certified family physician, medical device inventor, and health care entrepreneur, explains how AI chatbots became the new waiting room, and what that means for the doctor who used to be the first call. You will hear why patients turn to ChatGPT, Claude, and Gemini instead of waiting days for a reply, how AI companies are racing to own the front door of a system that is 20 percent of the economy, and why staying passive hands that ground away. Stillson makes the case that physicians can shape how these tools are governed, through medical advisory roles and digital health committees, so the guidance patients get is accurate. This episode is based on his article “AI chatbots and patient safety need physician design,” published on KevinMD. Press play to hear why you are becoming the second opinion, and what you can actually do to stay in the patient journey.

Partner with me on the KevinMD platform. With over three million monthly readers and half a million social media followers, I give you direct access to the doctors and patients who matter most. Whether you need a sponsored article, email campaign, video interview, or a spot right here on the podcast, I offer the trusted space your brand deserves to be heard. Let’s work together to tell your story.

PARTNER WITH KEVINMD → https://kevinmd.com/influencer

SUBSCRIBE TO THE PODCAST → https://www.kevinmd.com/podcast

Transcript

Kevin Pho: Hi, and welcome to the show. Subscribe at KevinMD.com/podcast. Today, we welcome back Tod Stillson, family physician and health care entrepreneur. Today’s KevinMD article is “AI chatbots and patient safety need physician design.” Tod, welcome back to the show.

Tod Stillson: Kevin, it’s great to be with you, and I’m so thankful to be able to chat about so many dynamic things going on in the world of health care when it comes to how patients are accessing care, how we’re using AI, and how physicians are responding to that. So it’ll be a fun chat today.

Kevin Pho: All right. We were talking offline about how AI is almost becoming that waiting room, and patients are engaging with AI even before they see their physician. You wrote about that in your article, of course. So tell us more.

Tod Stillson: To be honest, I think we all recognize this, because even as consumers and users of AI, all of us physicians will now often turn to AI to ask the common questions and to look for answers. And the things patients are asking AI nowadays are, “Why does my throat hurt? Should I go to the ER? Is this medication safe? What’s causing this rash? Do I need an antibiotic?” That’s really what a consumer is looking for, and this is what the studies show over and over again: immediate answers, some reassurance, some education, triage, or some next steps.

What AI is filling the void of is what we would oftentimes describe in medicine as, “I have a friend who’s a doctor, and I can just send them a quick text message and ask them this question.” And they just answer it for you, right? And you know it’s trusted, and you know it’s good.

But patients are now turning to AI in an increasing fashion to answer those questions before they even get to us as physicians. And so that waiting room where people are thinking about these things has now moved from the physician’s office to the patient’s smartphone. Honestly, that’s where it’s all happening now. And I think it’s very fascinating when you think about that.

Kevin Pho: And whenever I think about this, what you say is exactly true. Patients, of course, consult GPT or Claude or Gemini or whatever their AI of choice is before they come and see me, and they already come in with questions and tests they have in mind.

And I think this issue is partly of medicine’s own making. Because we made it so hard to access health care, because there’s a primary care shortage, and because it’s just so hard to get medical information from a physician’s office, patients are going to go to where that information is. And oftentimes that information from these large language models is good enough. They’re not perfect, of course, but they’re good enough. So I think what you’re describing is partly an issue of our own making.

Tod Stillson: Yeah, no, it’s absolutely true. Sort of our own making, and sort of the system’s own making too, though, right? Let me just point out that you’re in the trenches doing medicine still, just like I am. And so in your EHR, there’s usually some sort of secure messaging system where patients can asynchronously ask you questions. Now, you and I both know that’s a win for patients, but for doctors, holy cow. When you start getting inundated with those basic question messages that come from patients, and there’s a recent study that showed that has increased by almost 20-fold for physicians, you need almost an exclusive administrative day just to take care of those asynchronous messages. Patients sort of know that, and depending on how quickly you respond, you might respond to them in 24 or 48 hours.

So patients, even though they might be trying to access us through those new asynchronous methods that do drive us crazy, they’re now turning to those instant answers. If you put that same question into GPT, you’re going to get an answer in five seconds. Well, what’s the patient going to do? And then, as they grab that information, they’re going to compare it with what you say and so forth.

So it’s just a very interesting space that we’re in right now. And I don’t know about you, but I don’t mind patients going to Claude or GPT or Gemini or whatever to ask some questions. I mean, it sort of sets the table for our communication when they come into the office. And that sort of redefines the front door.

Here’s how it’s looked over our generations. You and I have been practicing long enough, right? It used to be that none of that happened, and they would just show up in the office, and we’d talk about whatever the issue was in the office. It was me and them. Then, a generation or two ago, it became Google. All of a sudden there was a Google search that preceded the visit, and they would come in with their articles and things and questions. But nowadays, and this is the interesting part, AI chatbots really do some triaging, and they do it in a different way than Google. Google kind of provides websites and information you read, and you sort of have to sort it out yourself.

But the AI chatbots, as you and I both know, give recommendations. They provide interpretation. They actually prioritize. They actually explain. They do things the old Google search didn’t do, and for patients, that really scratches an itch they have when it comes to these things. And quite honestly, what they do with that information, Kevin, is they decide, “Am I even going to go to the doctor? Do I want to self-manage this? Do I want to go see a naturopath? Do I actually maybe choose a virtual doctor and not my regular primary care doctor in the bricks-and-mortar clinic?” It sort of empowers them in a different way, and it’s just because AI provides that information to them in a different way.

And Kevin, you and I can sit around and complain about that. We can fuss about it. We can decide, “Well, that just isn’t right. I should be the one that owns all that information.” But the reality is that cow’s out of the barn, and it’s done. And so we kind of have to start to redesign even how we interact with patients because of it.

Kevin Pho: Yeah, no, I completely agree with you. You asked me earlier whether I get upset or angry if patients use AI before they see me, and the answer is of course not, because, like you said, just put yourself in the patient’s shoes. If you have a health concern or an issue, are you going to wait 48 to 72 hours for an answer from your physician’s office? Or are you going to go to these AI tools to get an answer in five to 10 seconds? So I think both of us would also use AI tools to get that answer. We would do exactly what our patients would do. So of course we’re not going to blame patients, and of course I understand patients are using AI tools.

So the question we need to ask is, as physicians, how is our role going to evolve now that patients are using these AI tools before they come see you? What is the new definition of what it means to be a physician today?

Tod Stillson: Yeah, that is a great question. And let me just say right away, I don’t know the answer to that one. That’s a big one. And even medical education is going to change dramatically because of all of these evolving changes. But what I see are some big stakeholders in this process. You’ve got consumers and patients. They’ve always been there and should be there. You’ve got us, as physicians, as important stakeholders. You’ve got our health systems as important stakeholders, and you have the AI companies themselves as important stakeholders.

Let me just tell you, and I think we both recognize this and others see it, the AI stakeholders and AI owners are going for the land grab. They truly want to be the front door for these patients and, for lack of a better term, I would guess they want to become their go-to for everything, not just health care. But of course, with health care representing about 20 percent of our GDP, they recognize the true value and the money is in health care. And so they’ve spent the last many years building into this. Some of your listeners might actually have side jobs that are part of feeding these LLM companies with their health information and data. A lot of medical students and residents do that. And they just keep feeding it and feeding it, because these companies want to grab that space, and we’re seeing it at an increasing level in terms of what’s happening.

I’m also going to give our health systems credit, though. Our health systems are not just passively standing back and going, “You know what? We’re not going to do anything about this.” They are responding and trying to resist the full land grab. Health systems are trying to integrate AI into their processes, into their resourcing of physicians, and all the things that go on with that, and they’re trying to ask the right questions about clinical governance.

And then you have the physicians who are getting involved, both independently and employed. If you’re employed, you’re probably going to be on a committee in the health system that sort of organizes this. And if you’re independent, you’re sort of making your own decisions about how you adapt to those things.

And so it is very dynamic in terms of what’s going on. And my point in the article, the point that you and I have talked about, is don’t be passive in terms of your involvement. We do not have to just accept that the AI companies are so big and so strong that they’re going to grab the land and we have no say. Nor do we have to be, and I don’t want to use a negative term here, but we don’t have to just be passive bond servants who do their crawl and pour all that data into their LLMs either. So let’s join into the fray and actually use our brains and make a difference for the patients, so that the AI being presented to the patient is actually governed in the right way to provide the accurate information that’s needed.

Kevin Pho: Now, what are some ways physicians can do that? And I know that you, as a health care contributor, have a product that does something you’re talking about. But short of creating a whole company or creating an app, a lot of physicians don’t have your entrepreneurial acumen. For the physicians who are listening to you for the first time, what are some easy ways they can do what you say?

Tod Stillson: Yeah, that’s a great question. So number one, educate yourself. Just be aware, read. That’s what physicians are really good at. We’re curious learners, and so don’t passively sit back and let this come at you. Read about it and engage and see what’s happening in this landscape.

Listen to your patients and what they’re actually doing in the front-line marketplace. And then, where are the natural fits for you to get involved? If you’re independent or looking for some side work, there are a lot of health tech companies out there looking for physicians to join their medical advisory councils, to serve in their telemedicine groups, to really be a part of defining what the future can look like in medicine. So you can actually get paid, you can participate, you can be involved in these.

If you’re in a health system, oftentimes it’s going to be in the medical informatics group, or maybe in some of the digital health committees that exist in your health system, and you can very much volunteer to be involved in those. And oftentimes the EHRs themselves, when we think about the Epics of the world, are getting heavily into these spaces as well. And they often need physician leads in your organization to be a part of those digital integrations. So jump in, volunteer, actively get engaged in the process, understand what’s happening, and don’t view it, Kevin, as a black box.

This is not a black box. Every AI output has an input that governs what the output is. And so you have the opportunity to influence what that system is that’s built around that AI product, and to make sure the output is properly governed, not by a non-medical person, e.g. a tech leader, but by a real physician or a real medical person.

So that’s what I see in the landscape of opportunities to get involved, and I’m excited to say I think a lot of physicians are doing that in a big way.

Kevin Pho: Now, I’ve seen those job listings where these AI companies are soliciting physician help to train and be involved with their AI models. I’m going to play devil’s advocate here. Isn’t there a little bit of a tension here, where physicians are lending their expertise to these large language models to become more, quote-unquote, “physician-like”? Isn’t there a danger that these AI systems would learn from these physicians and eventually replace what these physicians can do? So on one hand, yes, they are getting involved, but on the other, are they training their replacements?

Tod Stillson: Well, that’s a great question, and I think it depends on the leadership of the AI company you’re working with and what their ultimate endgame goals are. And we can sort of have a conspiracy model that exists here, if you will, that does look like that. Like, “Hey, what are you doing giving all this information to them? Because at the end of the day, they’re going to eliminate you.” Well, there’s probably some truth to that, let me just be frank. But there are also plenty of companies that are built in a different way, built in a way that is for the benefit of the community and the benefit of the culture they’re working in, and they’re not intending to take physician jobs or medical jobs. They’re really meant to assist us to function and adapt in a different way. So those are the companies you would want to identify with, the ones you’d want to support and align with.

And this is how it goes in a free market, right, Kevin? Whoever is going to offer to pay you the most money is sometimes going to be the one who wants to eventually take your job. So be careful about taking the best price for that KB, as we call it, the knowledge base, the information you know. But make sure you dig into it a little bit deeper and make sure you’re aligned, from a missional standpoint, with what’s going on with those companies. And there are a lot of companies out there doing it, so this is a great check that you just said, to make sure you are aligned properly with the direction the company is going when you do it.

But at the end of the day, we’re all challenged by the fact that we are having to learn how to doctor differently, and that is coming at us. We are all learning. And as you and I both know, 80 percent of physicians are using AI every day in some fashion in their professional lives. That’s not even counting their personal lives. Gosh, I’d venture to say 90 to 95 percent are using it every day in their personal lives even. And so we’re all learning how to adapt, to doctor differently, to support our patients. And it just doesn’t have to be a passive, feed-it-to-us thing. It can be an active sort of thing.

I mean, just think about the younger generations, Kevin. I just want to point out that there’s a behavioral shift happening in the younger generation, not just of doctors, but of people, of consumers. They’re digital-native generations. They actually don’t distinguish between the internet and AI. It’s all digital to them, and it’s just natural. And where we would have asked Google 15 years ago, we now see that 50 percent of Generation Z and Millennials ask AI health questions every day. 50 percent, that’s what the data shows. And so they’re leaning into it. Even the young physicians and nurses and pharmacists and other health care professionals are leaning into it with OpenEvidence and others that are supporting our knowledge base in this process.

And so we all have to learn how to doctor differently. We have to be engaged in this. And that’s the message for our audience, and for you and me even, in our own professional life. Don’t be passive. Be active. Go out and be involved in this. Because honestly, in your exam room now, sometimes you’re the second opinion. You’re not the first opinion that patients receive before they get there. They’re bringing the information in to you and saying, “What do you think about this, Doc?” And again, you can either sort of act shocked, or you can already know this is coming when you get there and sort of leverage it to your advantage.

Kevin Pho: We’re talking to Tod Stillson, a family physician and health care entrepreneur. Today’s KevinMD article is “AI chatbots and patient safety need physician design.” Tod, as always, we’ll end with your take-home message to the KevinMD audience.

Tod Stillson: Yeah. You know, health care is being built, or it’s always been built, around access. And whoever comes first usually has an opportunity to define what shapes the patient journey through the health care system. Today, AI is going for the land grab. The AI companies are going for the land grab. And we, as physicians, need to get out in front of that and do what we’ve done all of our lives, which is to think of patients first, and second, to make sure we’re in that sequence of the patient journey and have an ability to govern, organize, and support it. And to be honest, to learn how to doctor differently along the way, but also to define it with the governance that we know is critical to the best outcomes for the patient. That’s what we do all of our careers and all of our lives. So don’t be passive about the opportunity. Jump in and do it. Consumers are already there. Health care companies are there. AI companies are there. Physicians, let’s get on board.

Kevin Pho: Tod, as always, thank you so much for your insights and perspective, and thanks again for coming back on the show.

Tod Stillson: Yeah, Kevin. It’s great to be with you. Take care.