Melina Davis 

Hello, I'm Melina Davis. I'm the CEO and EVP for the Medical Society of Virginia. And I'd like to welcome you to the next episode of the SafeHaven JoyMed Podcast. I'm so excited.  

Today we have somebody really special, Dr. Mark Townsend, who is Chief Clinical Digital Ventures Officer for Bon Secours Mercy Health. He is a physician leader, an innovator who's dedicated his career to advancing patient-centered care and embracing technology to transform healthcare delivery. 

Dr. Townsend currently serves as a physician leader at Bon Secours and has been a driving force behind initiatives that integrate cutting-edge medical technology into everyday practice. With a background in internal medicine and pediatrics and years of experience shaping clinical education and leadership, he brings a unique perspective on how artificial intelligence can empower physicians, improve outcomes, and enhance the patient experience. In this episode, we're going to explore Mark's vision for AI in healthcare and what it means for the future of medicine. 

Welcome, Mark. 

Dr. Mark Townsend 

Thank you. Thanks for having me. 

Melina Davis 

I'm so glad. I'm spoiled because I get the chance to talk to you about AI on a regular basis, and I'm thrilled to share the opportunity with others. You're such a subtle person that I don't think people understand how much expertise about AI and digital innovation is sort of packed inside of your brain. 

Dr. Mark Townsend 

Thank you. You're kind. And once this type of technology turns into your day job, it does start to change your conversations and your outlook perhaps a little bit. So no, that's kind though. Thank you. 

Melina Davis 

Why did you find your way from direct patient care into exploring this space? How'd you get there? 

Dr. Mark Townsend 

Great question. I got there through a series of small steps that in retrospect makes sense. But as I began my clinical career, I'm a med peds cardiologist, so I got into essentially an emerging specialty on the ground up as it was formalized as a specialty. And so at the time, out of necessity, we didn't have access to the tools that we needed for patient care, particularly for patient imaging. And so I started tinkering and I knew smart people in in medicine, who also had some background in technology development. So out of necessity, just started developing some tools. And that would have been 25 years ago. And always had, as a creative outlook, a project that I was working on. In parallel with my clinical career, I also had an administrative career, beginning as a medical director in an academic medical center, and then moved on to then a COO as my first administrative job in a heart and vascular institute. So the operational St. smarts of essentially taking a problem at its inception and then trying to create a solution, or in some cases, many cases, partnering with someone who had a solution to your given problem statement was truly how I ended up in this space. 

Melina Davis 

That's so interesting. I know you well enough to know how much you care about humans and people. So I think it's such a beautiful blend when you talk to you about digital advancements in patient care, because you're really looking for how to help people, but you're also looking how to help the humans in healthcare who have to deliver that kind of care, right? 

Dr. Mark Townsend 

Absolutely. In our line of work in healthcare, If we're not making people's lives better, we're not doing our jobs. And being in the unique intersection of technology and patient care delivery is a really fun place to be in right about now. 

Melina Davis 

Yeah, it's an exciting time, especially with so much going on with AI, right? Scary to a lot of us who don't know the unknown, but what excites you most about AI and healthcare in the moment? 

Dr. Mark Townsend 

It's a really fun time to be alive and in this line of work because AI truly is at an inflection point. And we've talked about AI since Dartmouth and there has been this logic if then logic for years. AI is not new. But in the experience of Bon Secours Mercy Health, we had gotten into co-developing a digital assistant pre-ChatGPT, and we did that in a partnership with a company called Brado, a technology that really saw the future coming fast. Post-ChatGPT, let's just say that the whole conversation has been transformed, evolves, it feels like, by the moment. But we truly are at an inflection point with AI, not only in terms of adoption, the number of Americans that are using it as healthcare consumers is really a fascinating construct. But the inflection point of the technology is just in the past week, for example, you have OpenAI releasing their healthcare version of ChatGPT, and then you have Anthropic responding with their version, and then you've got Google responding all within three to four calendar days, which is really amazing. And so it is a fun time to get up and quite frankly, read your favorite news source in the space and then bring it home to a health system. 

Melina Davis 

Wow. And do you think clinicians can rely on those resources as terra firma, sources of information at this point? I know the debate a few years ago was, are they looking at good source material? 

Dr. Mark Townsend 

Yes, garbage in, garbage out applies here, right? So a resource library, an LLM is only as good as, you know, the data that feeds it. In the instance of ChatGPT, where the resource library is the whole internet, right? These models that we're talking about, you've got to obviously be smart. 

But the interesting thing, I think, that we don't really talk so much about, and certainly the common nomenclature, you know, what you read on your favorite media outlet, you're not going to hear about curating your own, essentially, your own resource library, if you will. which is an intentional design construct that is focused on making sure there's not garbage coming in and you really can control the environment and quite frankly, then the quality of the material that is produced by these models. 

Melina Davis 

So, the individual provider can control that or the source that you learn to trust, you can help guide or help evaluate that. 

Dr. Mark Townsend 

If you're a builder as an individual, you could, but more so the latter, yes, you start to learn essentially what tools and what versions of what tools are going to be more credible than others. 

Melina Davis 

It's very helpful. And so it helps to be part of a community that has its line of sight on trusted materials so that you as a clinician can also take part in that community of trust and good information. 

Dr. Mark Townsend 

That is a great way of phrasing it, Melina, because that community arguably is healthcare. Because it is incumbent upon us to get these tools right. Because if we don't, then people can get hurt, right? So yeah, I mean, we are that community. 

Melina Davis 

What do you think the biggest opportunities are to improve patient care? What do you see? 

Dr. Mark Townsend 

I love that we're having this conversation under the rubric of JoyMed, right? 

Melina Davis 

Yeah. 

Dr. Mark Townsend 

Because what AI is doing is arguably restoring the joy in medicine on behalf of physicians and behalf of providers. And if you can do that, you can improve the experience of not only our providers who are languishing under moral injury, what's burnout, you know, I mean, it's arguably more than just being overworked and overburdened. But if you can intervene there and you can restore joy in medicine, you can measure the impact in terms of quality and safety of patient care. You can measure the impact in terms of care delivery as a health system. And quite frankly, we get this right in terms of profitability of health care as well. 

Melina Davis 

Wow. I've always believed that if you build it well enough to make a reasonable profit, but you build it well for the humans, then you're really at sustainability, right? But I love that you brought up well-being. How can AI add joy to medicine? Tell us what you're thinking. 

Dr. Mark Townsend 

Yeah, so here's a specific example. Chatting, I was on site at St. Francis Medical Center last week, talking with some physicians. And so we have rolled out ambient dictation or the notion of an AI scribe, right, which is very commonplace at this point. But it doesn't stop me from asking physicians about their experience. And I heard the story of one of our docs who left her device at home. You need your device to turn it on, essentially, before you go into a patient room. And she got to clinic and realized she didn't have her device. So she made the decision that she was going to start her clinic late, turned around, went home, got back in 30 or 40 minutes with her device, but she was not going to go through the day. without her device and without the ability to use AI as part of her documentation is so much part of the workflow. 

And you ask why. It's not just a time savings. It is absolutely improving the quality of documentation. So, Joy in Medicine is defined by sending physicians home so they can eat dinner with their families. It is measured by their willingness to see another patient, when it's the right thing to do for that patient. It's their willingness to adapt, quite frankly, to the technology and adopt it as their own. So, it's all of that. 

Melina Davis 

Yeah, and feel the empowerment that the tool could potentially give you versus the frustration, I think, that some have experience with more rudimentary technologies in the past. 

Dr. Mark Townsend 

That's right. And I love how you call that out because adoption is the measure of success in this space, right? An example I just used was a physician who has adopted a technology and can't live without it. 

Melina Davis 

Right, enjoys it. 

Dr. Mark Townsend 

And enjoys it. And right. And so that can also conversely be used as a measure of failure. So when we roll out a technology that isn't adopted, we're going to know that very quickly, right? And so Imagine a day, again, using a smartphone as an example. Imagine a day without your device in your pocket, right? I mean, you don't feel like a whole human being if you walk out the door without your device in your hip pocket or wherever it is, right? 

Melina Davis 

Yeah, you feel it. So funny, we talked about this recently at a dinner party I was at. We used to all walk around without being somehow connected to other people. Was that better or not better? But at this stage, if a network, like happened in the US recently, a network goes out of phone outage, what are we all, we're all a little bit panicked. Like, how am I gonna call the pizza guy? I guess. 

Dr. Mark Townsend 

Right, I mean, how is life gonna work, right? If I can't make an outbound call. 

Melina Davis 

Yeah, that's true in practice, right? You want those devices to be useful enough to miss them. 

Dr. Mark Townsend 

Yes. 

Melina Davis 

So I'm going to ask how you spend your day looking for technology? How do you know what would work and what are you seeing out there that you find compelling? 

Dr. Mark Townsend 

Yeah, so as Digital Ventures Officer, my job is to take the problems of Bon Secours Mercy Health and then to source solutions. And then we take those solutions, create partnerships, sometimes co-develop technology when we can find a sort of fit, but not quite the exact fit relative to the problem that we're addressing. But then we take those partnerships and, in some cases, do strategic investments as well. So it's a blend of VC technology, but it is all sourced by the problem statement of the health system. 

And so, my background as an operator, as a COO, essentially, and having served in that capacity within the organization, within Bon Secours and Mercy Health, is such that I have the connectivity to all of our regional markets. I've had responsibility in those markets in some form or fashion. And so, what we did was we recognized early on that as we were getting into this space, we could approach the problem statement, which is we need to help people work smarter, not harder. 

We could approach that in one of two ways. We could get behind it in terms of a directive or a mandate of governance. or we could, or we could bring people into this conversation. And so the analogy we used was you could herd cats in one of two ways. You can get behind them and crack the whip, in which case they'll run a million different directions and hide in the bushes, or you can turn on the can opener. So we have created a can opener, which is discretionary funding, and the operators from the health system then come to us with their problems. because they need some funding in many cases. 

We're a budget-driven culture. And then sometimes they have sourced the technology. If they've done that, they just hand it off to us and we do an introduction and we get to know the partnership. Sometimes we source the technology. And so that's how these relationships evolve. But I mentioned VC as well because our partners in VC are a tremendous source of information as well. And so we are constantly meeting in series of conversations with fellow VCs to say, what are you seeing? What problems are you working to solve? And so, yeah, that's a great source of information and connectivity as well. 

Melina Davis 

I don't think people know that health systems are doing this kind of innovation work. It's really exciting to think that you guys are out there looking for these solutions and that your leaders have an Ave. to find solutions. 

Dr. Mark Townsend 

It is exciting. I love that you use that word. Not all health systems are in the space, given the constraints of their individual cultures. I will just say that our CEO is tech friendly, is an innovator at heart. and very much sets the tenor of the organization. And you see that, you see CEOs like John Starcher featured as leaders in this space because it is well recognized in terms of emerging technologies that health systems like ours that are willing to partner in terms of co-development of technology or even early adoption of technology, we're not all there, right, as health systems. 

Melina Davis 

Yeah. If I'm a clinician or provider, what's the downside to… the thought of these new technologies or integrating them or AI by itself? 

Dr. Mark Townsend 

So here's a fun one. I was at Northwestern and meeting my counterpart, having a fun conversation about the projects we were working on. And she used this example when I asked that question about their culture. But they had co-developed essentially a technology that empowers the workflow of radiologists and their training hospital. And when they had gotten far enough into the project, the attending physicians, the teaching physicians, turned to the physicians in training, the residents and fellows, said, you're not allowed to use it anymore. And they weren't allowed to use it anymore because it was too good, right? And so it was kind of a fascinating story. 

And so that begs the question, will there be atrophy of skills, right? Or if you start to rely perhaps a little bit too much on these tools, could it create some cognitive atrophy? Or, said differently, could it then change the way that the up-and-coming generations learn medicine? It obviously is going to, and it is changing. It is changing the way they learn medicine. And then we're going to have to figure that out, right? So when is it too early? to use an AI that is very good in diagnostic radiology. And back to garbage in, garbage out, you have to validate the product, right? And 100% of the time in healthcare, there has to be a human in the loop that validates the output of the AI. And so that's the fascinating push and pull the tension in this space. 

Melina Davis 

Yeah. it's interesting. This is the conversation I hear all the time is that this humanness with the technology is really important to patients. And frankly, it's really important to the clinicians to feel good about because ultimately they are responsible. The doctor is responsible for these decisions. Is there another reason to keep humans in there besides those? 

Dr. Mark Townsend 

Other reasons to keep humans. And while I approached it from the perspective of the physicians, another very human perspective is that of the 200 million patients per week who use ChatGPT to ask a medical question, right? And so from a consumer or a patient perspective, that's quite a space to start thinking about as well. relative to accessibility of information and sometimes misinformation. And, in medicine first, do no harm. It's easy. It's not easy. It's actually very difficult to build in the safeguards, the checks and balances within governance of a health system. It's deliberate, but very, very important for, you know, the ChatGPTs of the world. And, you know, having had this conversation with Kevin Wheeler, the Chief Development Officer, It's fascinating because they recognize the responsibility of what they're doing. And it comes with tremendous risk as well, right? I mean, harm can happen. 

Melina Davis 

Unintentionally. 

Dr. Mark Townsend 

Unintentionally. Right. Yeah. And so, from a human perspective, we're seeing the safeguards evolve very quickly. We are seeing the industry adapt very quickly. We're also seeing the regulatory environment adapt. 

Melina Davis 

Yeah, that was really my next question is sort of this ethical regulatory environment. Do you see it evolving at the same speed that the technology is? And what do we need to do to make sure that it does appropriately evolve? 

Dr. Mark Townsend 

The pace of evolution of the regulation is lagging the evolution or the growth of these technologies? 

Melina Davis 

And that's normal. 

Dr. Mark Townsend 

And that's normal, right? It's the way it happens, right? You don't know what you need to regulate until you've rolled out a technology. And then it's unfortunately sometimes adverse outcomes that then inform the regulatory process. And I don't think this is different. 

From the FDA perspective, just starting at a federal level, the FDA has gotten into regulation. But there was a little bit of tension, well, what's the role of the ONC? And, you know, the ONC traditionally, you know, governs interoperability, electronic health records. But should they have a regulatory role to play here as well? And then what's the role of CMS? Because, you know, what CMS reimburses all of industry will follow 

 And then what's the role of the states? And you see the states starting to react, for example, New York reacting to, an adverse outcome related to a consumer who should have been a patient, and there was harm. And, so written up in the lay media, New York responded with state-specific legislation. And then what's the tension relative to states, should they be regulating? And the most recent executive order in this space is, it is not the, it's not the purview of the states to regulate AI. 

Melina Davis 

Yeah. 

Dr. Mark Townsend 

And I think we're going to see that challenge, right? I mean, it will evolve. 

Melina Davis 

Because the states, that's where medicine is overseen. That's right. There's a direct conflict there that we're going to have to debate and resolve. 

Dr. Mark Townsend 

Yeah. 

Melina Davis 

So interesting. What steps, if clinicians are new to all of this, what can they take? Steps can they take to get used to and start integrating AI in their practice or learn more about it so they can do it comfortably, thoughtfully. 

Dr. Mark Townsend 

Yes. You know, I will advocate for going back to how you learned to ride a bicycle. You learned by getting on the bicycle and maybe fell off a few times and skinned your knees. but you learned how to ride a bike and his muscle memory. Most of us can jump on a bike and still ride a bike, right? It might not be as pretty as it was in the past, but I would advocate for clinicians and physicians to jump on in and start using these tools. 

Open evidence, for example, you know, has exploded. And back to adoption, once I started using it, I couldn't stop using it, right? Because I wasn't going to take some of the questions that I was asking open evidence and ask ChatGPT. You don't put patient-protected information into these models, right? That's important. Enterprise GPT models are absolutely coming along. 

And so health systems like ours, we have not yet adopted one. Some health systems have. And then you can outsource those as well for patient-protected information. You can create specific relationships with guardrails that protect your information so that it does not go into, for example, the open internet or, you know, a open evidence, you know, type environment, right? But within all of those parameters, jump on in and start using them, right? So if you don't use ChatGPT to plan your next family vacation, you're missing out because those of us who do are not going to go back, right? We're not. 

Melina Davis 

We just did this. It was impressive. 

Dr. Mark Townsend 

It is. It's hard. 

Melina Davis 

It'll tell you restaurants. It'll tell you where to stay. It's amazing. 

Dr. Mark Townsend 

Yeah. And price points. And you can tell it what you want and it will absolutely customize and cater to your needs. 

Melina Davis 

Yeah. It's interesting because when you need to go back to the conversation about having a human involved, do you think there's ever going to be a moment where clinicians are not part of this relationship with the patient and that they've been pushed out of the way by technology? I think there is a fear of that and that might be holding people back from trying to learn and see how they can be part of the process with technology. 

Dr. Mark Townsend 

The Harvard Business Review does an annual report of how, and it started with generative AI, it's now agentic AI, but how AI is used internationally. The #1 use case is actually in the behavioral health space. 

And if you look at all of South Africa, It's tremendously underserved by behavioral health and mental health specialists. So in the absence of expertise, what are people doing? They're using these tools, right? And is more good than harm being done? We would hope so. It's not been studied that I've read. I'm sure there's some, you know, experience. But it's interesting, back to adoption, is there ever going to be, you know, the use of autonomous uniforms of AI. 

It's possibly, I will tell you that as a health system, we're looking at robotics and we're looking at humanoids. And, we're working with companies and technologies that have developed MVPs. And, we can see them in action. They have been used, by the way, as couriers in healthcare for now 10, 15 years, right? So starting with that model and what we know about using robotics for couriers, carrying up to 30 kilos, figuring out how to use elevators autonomously, I mean, that's here and now. 

So what then about robot-assisted hip replacements, for example, right? So robotics and, you know, talking to a buddy of mine who needed a hip, right? So he sought out the most advanced technology. And it wasn't a traditional lateral approach, let's just say. And his outcome, is arguably better than it would have been without the assistance of technology, which is empowered by AI. If at some point, some of the procedural skills start to be encroached upon by robotics, I mean, that's a fascinating construct. I don't know what to think of that. 

But if you think of it for more of a practical application of humanoids, if they can walk up and down the streets of New York and hand out candy appropriately without getting run over, then they can certainly deliver meals. Should we have them deliver meals in our hospital? So in a provocative sense, I put together a clip, and it's fun to use AI to poke fun at AI, but in this provocative sense, I put together just a short video of a humanoid walking into a patient room with a patient's tray and then starting to feed the patient. And we showed that to our board as a, you know, intentional conversation starter. 

Melina Davis 

Yeah. 

Dr. Mark Townsend 

And, but, it's probably not that far off. 

Melina Davis 

Right. So, but it sounds like what you're describing is, again, technology being part of the process. 

Dr. Mark Townsend 

Yes. 

Melina Davis 

My sense so far of patients and doctors in particular is trust. is between human to human. And that is going to be tough to get rid of, even if we can get farther, faster, get to more people, improve outcomes with technology. But what do you think of that? 

Dr. Mark Townsend 

Absolutely. And that's been well documented that quite frankly, physicians are less likely to be trusted by their patients if their patients know they're using AI, which is, and that was kind of a generic approach. But there is a trust gap. And more specifically, as physicians are introducing AI tools, that just in terms of referencing the past five conversations, for example, I've had with physicians, we expect them all to document that they have used AI, for example. 

Melina Davis 

And how they used it, generally. 

Dr. Mark Townsend 

In the note, right? The patients have consented to the use of AI in their patient consent forms as they come in to receive care. in any health system now, arguably. But then, specific to how physicians then introduce the use of AI, they call it different things, like I'm going to record today's conversation. That doesn't sound scary. Right. It's different than. 

Melina Davis 

Versus I'm going to have the robot do all of this interaction with you. 

Dr. Mark Townsend 

And I'll be right back. 

Melina Davis 

And I'll be back after I've had some coffee. 

Dr. Mark Townsend 

And I have seen that, by the way, the first AI-powered experience at the Shiba Medical Center where a patient walks into a futuristic looking pod, sits down in a comfortable chair and has a conversation with an AI assistant, an AI agent, a physician, essentially. And then the AI agent summons the human, the physician, to come and complete the physical exam, dictate that into the tool. And then the AI agent or physician then puts together an assessment and plan and starts to implement that plan. So next thing someone comes to dry your blood, right? And next thing you're down the hall for a CT scan. But that model's up and running, obviously, under IRB approval. And some patients preferentially are selecting that. They would rather be seen in that experience because it's unrushed. It's more empathetic and it's just a different patient experience. So it's going to be fun to track the trust factor. 

Melina Davis 

It really will be. It really will be. Well, tell me, what advice would you give to clinicians who feel overwhelmed by the pace of this, of AI's movement, but also all digital and technological advancements in medicine? 

Dr. Mark Townsend 

I was asked that question in a room of physicians maybe about 3 weeks ago, and it was meant to be a little bit provocative. It wasn't phrased as nicely as you just phrased it. And so I flipped it, and I said, when's the last time you got on an elevator? And they looked at me blankly, and I said, well, so you recognize when you got on an elevator, you were using AI, right? And so it's an older form of AI, but how did the elevator know which floor to come to? and which floors to ignore, right? That's a form of AI. When's the last time you did a banking transaction online or, you know, on your device, right? So you're using AI. So whether or not you recognize how pervasive AI is, I think a nice way to answer that question is to reference the American Medical Association, which doesn't call it artificial intelligence, but calls it augmented intelligence. 

Melina Davis 

Yes. 

Dr. Mark Townsend 

And so if you use that as a framework and recognize that it is cognitive assistance, right? And it is not replacing me, it is augmenting my work. It will take some of the more mundane tasks and offload them in the future, but it will truly allow me to work at the peak of my license, to work smarter, not harder, and to make healthcare easier. 

Melina Davis 

Isn't that the goal of every provider I've ever spoken to? It's been a real delight to spend time just talking about this and learning from you. Thanks for being here, Mark. 

Dr. Mark Townsend 

Thanks again. Really enjoyed it.