Designing for Health: Rethinking Clinical Workflows in the Age of AI with David Kirk, MD

Intro:  

Hello and welcome to designing for health. I’m Nordics chief medical officer Doctor Craig Joseph. AI can find information, clinicians mis write better documentation and potentially give physicians back time with patients. Great. Now comes the harder part figuring out what it does to clinical judgment. My guest today is David Kirk, a critical care physician and chief medical officer at Regard. 

We talk about using AI to uncover important information hidden deep in the air, reduce documentation burden, and improve quality and financial performance. But we also explore the emerging risks, including confabulation and cognitive spoofing. When AI gives someone the language of an expert without the underlying expertise. For health system leaders wrestling with AI strategy and governance, there’s plenty here to think about. 

Let’s plug in. 

Craig Joseph, MD, FAAP, FAMIA: 

Doctor David Kirk. Welcome to the podcast. Where do we find you today, sir? 

David Kirk, MD: 

So I am sitting in one of our glorious phone booths in regard. I’m in Manhattan and have been here for a year now. 

Craig Joseph, MD, FAAP, FAMIA: 

All right. So we’ve established where you are. I’ve established, I think, that you’re a physician, I believe. You’re an intensive and that you work for a regard. I have that all down. 

David Kirk, MD: 

Yeah. ICU doc for 25 years. Regard chief medical officer for a year and enjoying every minute of it. 

Craig Joseph, MD, FAAP, FAMIA: 

So maybe give us a little kind of some background of. You went to college. Were you thinking you were going to be coming ICU doctor, like, how’s that story? When you were five, you thought you wanted to be a dentist. And then that changed for some reason. For me, it was an architect. I wanted to be an architect. And now I ended up in medicine. I don’t know how, but why don’t you give us your story? 

David Kirk, MD: 

Yeah. So, my parents owned a dental practice in small-town Mississippi. 

Craig Joseph, MD, FAAP, FAMIA: 

I was joking about the dentist thing. Oh, wow. Look at that. All right. 

David Kirk, MD: 

Yeah. So, you know, because it was small town Mississippi, there weren’t a lot of funds were often paid in livestock and vegetables. And everybody came in the door, got a big hug, and it was a great family. So, my mom and dad was always like, David, you seem smart. We need to get you out of the small town. You should be a doctor. At the same time, I was playing with every technology I could get Ahold of, my dad had this computer about the size of a refrigerator that I started playing with as a young kid in high school. It was long distance to get on the internet, so I had to find unique ways to get on the internet despite being long distance. 

I installed the security system for my high school. As a as a high school kid, I helped, you know, design the network and security. And then in medical school, in gross anatomy over the cadaver, I met my neuroanatomy professor. He and I started talking about tech stuff, and ultimately he started working on the first web application that we would have for my eventual tech company. So, I’ve always had this journey of the love of medicine, loving, take care of people like my I saw my parents growing up doing. At the same time. I’m just a huge, massive nerd and I’ve always tried to find a way to marry all these things together. 

Craig Joseph, MD, FAAP, FAMIA: 

I’m interested. What was this neuroanatomy connection with software like? What was some of that? Did you create something that you sold? Like, I’m just I’m interested. 

David Kirk, MD: 

Yeah. So, Quinn, who is a neuro anatomy professor, also had a very conflicted life growing up where he wanted. He was a PhD. He really had this passion, passion for medicine and neuroanatomy, but he also was just tech, tech, tech. And we were in there talking about various aspects, and he was teaching me the very detailed aspects of neuroanatomy, which were always very difficult for me and were also saying, hey, is there a way to educate people using. At that point, the internet was just starting to put ads out there. It was just starting to expand out to the masses. And we said, hey, how could we do this? And that ultimately expanded out to us doing tutorials for people around various tech things. And then we have some of the first iPhone and iPad applications for medicine that were out on the market and just continued over multiple years to write. 

You know, as much as a two person shop could write applications. We did that and really enjoyed kind of growing out the very early kind of health tech side of what the platforms would give us available, what, you know, at that time, the phone was there, like, what could you do? Like we could do a medical dictionary. The iPad was available. Like we could bring in more information. So as new technology came out from the web to the iPhone to the iPad, we just continued to have fun building out small technical and medical applications. 

Craig Joseph, MD, FAAP, FAMIA: 

That is how it starts. The combination of healthcare and tech. That’s amazing. I think that’s a fairly unique story from starting that early, right? From a medical student when there was just the internet was in its infancy. So, you ended up if we fast forward a little bit, you ended up at Wake Met, and that’s where you practiced intensive care for, what, 22, 23 years? Something like that. Yeah. Okay. And what were some of the things that you did while you were at? And I think you left Wake Med about a year ago when you started with Regard is that did that was that like a sharp break? 

David Kirk, MD: 

Well, I mean throughout my career, I had enjoyed building the ICU there at Wakeman, then enjoyed leading the ICU awake, mad. And the more I was in the organization, the more I wanted to help fix problems, fix problems from people and process standpoint coaching and help bring teams together. But technology was certainly always a big component of that and ultimately came through during the pandemic. I was head of Covid operations and learned a lot from there and then moved on to as AI started entering the organization. With my background, I was blessed to have Wake Med asked me how they can start entering the AI age and started vetting applications from there. And that’s how I came to the awareness and started getting a lot of exposure to AI and regard. I enjoyed my thought leadership and some of my ideas and ultimately brought me on board a year ago. 

Craig Joseph, MD, FAAP, FAMIA: 

So, what were some of the I think we all know what some of the problems were from, from Covid, but what you know, what were some of the things that you were seeing from a technology or an AI perspective that were causing problems at your hospital, I presume? Not just in the intensive care units, but elsewhere? I think when we were talking to prepare for this, you were mentioning, hey, the electronic health record gave us opportunities to do better documentation, but better documentation made us spend a lot more time. And then we were spending more time looking up information for the documentation. And it was just an ugly cycle. Can you talk more about some of that from your perspective? What were you feeling? What were you hearing and did you fix it? Did you fix it all? 

David Kirk, MD: 

I’m not sure we ever fix it all, but that is the journey. I remember our chief quality officer asked me to join a group of cardiologists because the cardiology scores at Wake Med were lower than anybody expected, and went into the meeting, and the cardiologists were like, oh, one more quality person telling us that we’re not doing a good enough job. And they said that we really think we’re doing great. Like, these were doctors that had trained in, you know, world leading organizations. They said that we know our stuff. We’re doing amazing work. While the why does the quality score is not ultimately reflect that. Okay. We did this very deep dive, and it turned out that doctors were doing a great work. 

They were just not doing good documentation. And when you say that, of course the docs are like rolling their eyes because they want to be touching the patients, they want to be doing the procedures that they know save lives. They don’t want to do more typing and dictating. So it was that journey to ultimately to teach them and for the organization to understand that man, those doctors were really crushing it. All the work that they were doing was paying off. We were saving lives. But the physicians and the organization was not getting credit for documentation. And there’s no way that we could take cardiologists out of the lab to spend more time in the electronic medical record. So, we had find ways to improve documentation without the manual processes that we’ve asked doctors to do for so long. 

Craig Joseph, MD, FAAP, FAMIA: 

So that makes a lot of sense. So there seemed to be basically a disconnect between what the scores were saying. But what you actually knew was happening. Sounds like you can’t get credit for doing the right thing if you don’t write the right thing down at the right place using the right formula. And so how do you fix some of that? I could imagine some of that’s going to be as opposed to, you know, hey, we’re going to have we’re going to pay people to follow you around, which would be great actually, but apparently those people would like to be paid. So, what were some of the things that that you tried to do to allow your specialists to and procedurals to kind of get credit for the, for the work that they, they were doing, just not didn’t look like they were doing from outside looking in. 

David Kirk, MD: 

Yeah. I think what many organizations has done for a very long time is to try to do quick things in the electronic medical record, to make the physicians just be able to click instead of type, which does make things a little faster, but not dramatically faster. So that was one of the things that really pushed Wake Med and is now pushing organizations across the United States into AI. Because AI is not just clicking, a AI is actually looking electronic medical record. Pulling in that information and making the work of the doctors so much more efficient, instead of doing a lot of typing, instead of doing a lot of dictating. It’s really like someone that brought you a brief and you have all the information in there. You can say, yes, this is correct. 

Let’s put it in the medical record. It’s much more, more complete. It’s much more accurate, and it’s much faster for the doctors. So it really allows them to make sure the story is correct, because we have to make sure the story is correct. But without all the manual processes that have been there for the last several years. 

Craig Joseph, MD, FAAP, FAMIA: 

One of the things that I know, I think we initially connected on LinkedIn, I had read something of yours. And, you know, one of the things that you seem to focus on, which a lot of people do, is that it’s people process in tech. And you need, you know, tech alone is not going to help. It’s the resources, the people and the workflows or processes all around it that kind of if you don’t have all three, it’s like a three legged stool with one leg missing or two. It’s not going to it’s not going to help very much. Have you run into, you know, hey, the tech seems great, but we’re still not making improvements. Why is that a common thing for you, or is it mostly just. No, no, we needed to put the tech in. We had everything else working. 

David Kirk, MD: 

So, there’s two components and you’re the keyword you actually brought up was why often doctors and organizations are not told why they’re being asked to do something new. For example, those cardiologists understood that a lot of how they’re graded was based on their documentation. Then they were more open to doing something different around that documentation, hearing that we were bringing in artificial intelligence to help them. They were now much more open to that. So, the why, as far as getting people to change, is so key and so often missed. If we could just stress that, then physicians would be much more open to doing things differently. The other thing I will say is AI is allowed hospitals now to adapt the technology to what the doctors are doing. 

So, instead of before, a committee would come together and they’d be like, we need this three, three things done in the electronic medical record. And they would submit a ticket, and they would wait, and they would wait, and they would wait, and it’s in some long queue. And finally, it gets to it nine months later. And by then the committee was like, what the heck are we waiting on? Now with a lot of the artificial intelligence tools, you can sit down and literally just tell the software and prompt out; we want these things. And based on these things, we want these things to happen in a committee can go from, you know, people and process to tech over months. Now its people process tech in minutes. And you can actually see the next day that are you making a difference or not. So it’s really fun that we have sped up that inter cycle so much that teams stay engaged and they really want to push things forward because now we’re adapting to their workflows instead of making them adapt to is workflows. 

Craig Joseph, MD, FAAP, FAMIA: 

So, a lot to unpack there. It’s so correct. Especially I think all humans but especially physicians, you know, we want to know why we’re a little bit difficult sometimes okay. We’re a lot difficult. You know why do you need this information or specifically I think I’m practicing excellent quality, you know, safe medicine. But the evidence that you are looking at doesn’t support it. Where are you getting your evidence from? And the second thing I think that you hit upon that was really great is that finally with AI and I think more, you know, better coding and design, but we’re able to kind of instead of trying to get the clinician to work to, to adapt to the technology, we’re doing the opposite and making we’re now able with a lot with some AI allowed to have the technology kind of work for the for the clinician and the example I remember so many times I had to say, hey, you know, you’re not documenting the patient’s tobacco history and like there’s some, some quality metric that, hey, we’re really focused on their tobacco history. And I would get the same answer every time. Hey, it’s in the note I dictated, I typed it, it’s in the note. I’m like, yeah, we can’t look in the note because we couldn’t, like, we didn’t have we weren’t paying someone to look through notes. It was we were looking at this little box in the electronic health record, and if you didn’t go to that box, it looked like you didn’t ask. 

And now with AI, we’re able to kind of look through. We have many more options to be able to look through that note and, and be smart and be able to interpret without having a human in the loop to say, the doctor did mention it or the patient did mention it. It’s just not in the box where we need it to be. We can put it in that box now automatically. And it’s been a lot of, a lot of happy clinicians around the world. 

David Kirk, MD: 

Yeah, for sure. 

Craig Joseph, MD, FAAP, FAMIA: 

Well, it’s kind of talk about design. I’ve interviewed a lot of folks, and we’re called this podcast called designer for health. Almost no one that we interview is a designer. Right. But I think they all design. So, if you’ve created a workflow, if you’ve created an instruction manual, you’ve kind of designed some of those things. Let’s talk about, you know, where you’re working now. So, I haven’t really kind of we haven’t talked about regard. What do they do? What do you guys do and what do you do there? 

David Kirk, MD: 

So, for what regard does is we’re AI diagnostic layer that fits on the entire electronic medical record. So, we know about 97% electronic medical record never gets reviewed by physicians. And we know that data is incredibly valuable to patient care and to organization. So regard runs we help doctors build out their documentation. At the same time, we suggest diagnoses that the physicians might otherwise missed those diagnoses. One are amazing to help care. So, for example, when I’m in the ICU, I love to know that regard is running and surfacing diagnoses that I might have missed. Just three days ago. We, one of our clients, talked to me. They installed regard on the very first day regard brought up renal mass in the hospital. This was like this patient does have a renal mass. 

I’ve been taking care of this patient for several days. Regard had found this mention of renal mass and based on our algorithms, it was in the body of a radiographic note, wasn’t an assessment plan that would normally have been pulled over. It was just mentioned in the body. That’s a patient that might have gone home, and that cancer advanced. But because regard pulled that out of a hidden place in the chart, that patient got their biopsy, it was found to be malignant and it was removed in an early time, caught before it, you know, had all the downstream effects of cancer because we were able to surface something from the deeper assesses of the chart. And of course, when we do that, not only better take better care of patients, but those diagnoses help hospitals with their quality scores. 

It helps hospitals with the revenue. It helps them with their safety. So, all those things that are missed in the chart, when those are pulled forward and brought to the eyes of physicians, it provides a lot of downstream effects. And that’s the reason we’re now in 150 hospitals and growing, because we’re that AI diagnostic layer that many hospitals are now depending on to improve care and documentation. 

Craig Joseph, MD, FAAP, FAMIA: 

I can imagine pushback, especially from that story that you just said from the physician who said, wait a second, I’ve been taking care of this patient for several days now. I don’t know about any mass. I let me just assume that this is not true. So, because if a medical student brought up a renal mass, right, you might. Okay. I don’t know where you’re getting that information from. Can you show me the proof? So, do you have the ability to kind of surface that take you right to that report? And so, I believe it. It’s always trust but verify what I think in medicine. 

David Kirk, MD: 

Yeah. So, we validate information two different ways. So, one the doctors can click in the note and we will take it directly to that report where we got that information. But what most doctors do is they just ask the AI with natural language query, what do you mean there’s a renal mass. Where did you get that information in regard would ultimately say in the body of this CT scan from two days ago, this was noted. It was in this location, and pull out all that information. So there’s multiple ways to validate, because our job is always to get the physicians the right information, but is effectively as possible. So they’re not going through the media tab and all the various places that stuff can hide. We’re trying to bring it forward so they can make their decisions and validate it as quickly as possible. 

Craig Joseph, MD, FAAP, FAMIA: 

So clearly that’s improving care, right? We found something that the patient didn’t know about or didn’t mention, and now you were able to address it or at least be aware of it. So from a quality perspective I get that. Do you see AI and the technology you guys are producing and that others are improving physician retention, decreasing burnout? Any good news there or data that you’ve collected or stories that you’ve heard. 

David Kirk, MD: 

Yeah it’s fantastic. So one, a large part of burnout is that physicians go home at night and they’re worried about what was not seen in that 97% of the chart. So if I’m an ICU doc and I’m admitting a patient who’s been in the hospital ten times over the last year, there’s no way I would have time to go through every one of those hospitalizations, every one of those clinic visits. So two things happen. One, you just eventually give up and hope you’ve captured everything or two at night you go in and you’re back digging through the chart, looking through information with regard. Is that back up to surface? What’s important out of the electronic medical record? It really decreases that burden of what did I miss because you’ve got that double check. 

The second thing is because regard helps you generate the documentation. You’re not having to do as much kind of clerical work. We have doctors time and time again saying, I’m getting when I go home at night, I’m able to close the laptop, I’m able to do the things that I want to do because regards allowed me to finish my documentation much faster. I’m able to round a second time at the end of the day to help get people out of the hospital. I’m able to go back to the bedside and have those difficult conversations with patients or family. I’m able to go back and check on my patients before I leave for the day. That that ability to get back to the bedside and heal and do the things that we went to med school and residency ultimately to do, is such a joy for many folks that we have doctors every day telling us, hey, you are prolonging my career. 

I just didn’t feel like I could do this at this way anymore. By using regard with all the assistance it provides me, it was me to focus on the important stuff, and the stuff that was really driving me crazy is really falling to the wayside. 

Craig Joseph, MD, FAAP, FAMIA: 

You know, sometimes we look at, hey, did you see more patients or did your notes get written so we could close office based encounters faster? But we don’t look at what really matters? Is our quality scores better or improving our burnout scores getting better and going down. So that’s good to see that that’s actually happening. Let’s talk about, you know, some of the concerns with AI. And you know, certainly when I first came about. I know we’ve been talking about AI for decades, but in reality it’s been 3 or 4 years at the most where it’s really been kind of ready for prime time. We heard about hallucinations. And then I think we’ve heard, you know, confabulation, where we’re taking a story and just making it a little bit, a little bit juicy and something when we were preparing for this that you said, a term I had never really heard of before. 

Cognitive spoofing. And so can we talk about some of those things and how do what are they and why do I, you know, why do they keep coming. And doctors of all kinds up at night and what can be done or is being done to kind of minimize some of those things. 

David Kirk, MD: 

Yeah. So one of the wonderful things about all the regard client base is they often come together and figure out how we’re going to handle this new AI frontier together. And these are some of the conditions that they’ve surfaced. So cognitive spoofing is where, for example, a medical student might put something in the chart that sounds like a world renowned cardiothoracic surgeon. There’s no way the medical student has the knowledge to know the experience, but because they used artificial intelligence and artificial intelligence has such a wide body of information, it can really sound inventing. Now, medical students have been doing this for years. I, of course, went and crammed the book before I went to arounds. And, you know, my tendons would ask me a question and they would go to the next level deeper and the next level deeper, until they ultimately got to the bottom of my language. 

And that’s where we all started learning. Okay, David, great job prepping. But you don’t understand the physiology. You don’t understand really the literature cognitive spoofing is very similar. So it can make people look incredibly experience, get very detailed, very compelling answers. But health care organizations, places where learners are you have to continue to dig to figure out where the information, where the truth really lies. But man, it is convincing because AI has so much background now. Confabulation is one of my favorite things. So on large language models people talk about hallucination. But on medical side, hallucinations to us is you’re seeing or hearing. Things are not really there. And that’s not really what medical large language models do. Large language models really confabulation. So a medicine is that it’s that diminished patient. 

That patient with dementia is the patient who has, you know, alcohol abuse and has done some had some brain damage. And they just go on with these very elaborate stories that are filling in the gaps. And if you think about it, we all do that. When I’m telling stories, I probably don’t remember everything. I’m going to fill in the gaps with a little bit of juicy details. That’s the same way the large language models do. And sometimes those details really matter, and that’s where people need to have their sensitivities up that this is a comfortable in the chart and maybe things need to be validated. That’s one of the reasons that regard. We have multiple ways to review things, but also in regard when we make a diagnosis, we use deterministic algorithms that are supported by large language models. 

And of course, deterministic algorithms really can’t or hallucinate. We use large language models to do a lot of other work, of course, but those deterministic algorithms are the reason we use that is to keep those risks of confabulation or hallucination down. So these are this in the world of AI, our clients and everybody who’s using AI as a whole are saying these are new problems. We’ve been working around people problems for a very long time, and now people are coming together and saying, it’s going to be a while before there’s literature on how to handle these things. Do we give a learner artificial intelligence on day one? Do we give an intern it after six months? Do we make them like we did at the, you know, fourth year medical student? Make them do the 810 page on paper happy with all the immunization schedule. Are we in this new frontier? We give them AI from day one. These are things that we don’t know yet how to handle cognitive spoofing, how to handle where we do with learners. So at regard we’ve been bringing folks together to have those conversations. So at least hospitals are not going and feel like they’re all by themselves. They can have these conversations. And to make sure that they’re putting thoughts and consideration on how they’re moving forward. Until we have the kind of the literature and experience to support us. 

Craig Joseph, MD, FAAP, FAMIA: 

We one of the things that you haven’t mentioned, but kind of right in line is skilling, right? And the classic example is a gastroenterologist who’s now doing colonoscopies with a machine that has an AI and as because it’s I think a lot of people who are who are not in the business don’t understand that. Boy, there is some skill and experience in moving that camera around and looking for things that are pretty small and not important now, but might become important in five or 10 or 15 years. And an AI can kind of discover some of those things. And, and there’s been a few studies where doctors have experienced physicians who are using these tools have become worse at spotting some of the early tumors before they really become tumors. The question is not if it’s happening, we know it’s happening. The question is, should we care? I think that’s what the question is because we’ve always had some you know, it sounds horrible. 

Okay, well, now we should make sure that the fellows who are training don’t get to use some of these tools so that they get really sharp at finding everything. But once they start using the tools, maybe they’ll lose some of those skills. And the question I think is, which I’ve been talking for a long time. The question is, is that okay? Is it okay to become semi skilled? Haven’t we become less skilled at lots of things because we have the technology to support us and reasonably, we’re not going to do a colonoscopy without these new AI tools. I don’t know if you have any thoughts along those lines. Is it okay if it’s not okay? What are we going to do about it? 

David Kirk, MD: 

No. You and I are thinking about this the same way. When I started Critical Care, we put in central lines in the neck completely based on landmarks. And then 4 or 5 years into my career, we had this small ultrasound about the size of a postage stamp, where you could see the blood vessel, and it dramatically changed how often you got the catheter in the blood vessel the first time. Now it would almost be considered against standard practice not to use ultrasound guidance at all. Did fellows coming up today, did they describe because they can’t do it based on anatomy alone? Yes, maybe. But how often are you actually going to do that in hospital setting? You’re going to have an ultrasound. You’re going to have those two twos available. 

And it’s made the placement of the center lines tremendously safer. These tools, they need to be judged against outcomes. Do they make things better? And if they do need to make things better, even if it takes a little bit of the human component out of it, as long as patients are getting better care from this. Fantastic. I mean, we don’t look at we don’t look at urine under a microscope a ton anymore, and we certainly don’t put our finger in and taste it any more for glucose. Like there’s things that we should move away from. Right. Well, it’s okay. We should not be doing things the way we did just because it’s historical. But we have to measure it. We have to make sure that the best outcomes from the patient are coming from this. And certainly the scaling is an issue we need to watch for. But how much how important it is versus just a maturation of healthcare? I think you’re to be. 

Craig Joseph, MD, FAAP, FAMIA: 

Seen as you were talking about the difference between hallucination and confabulation, cognitive spoofing. I wanted to argue with you a little bit that you said cognitive spoofing is really a new problem. As I think back on my training, I think to a time when I was a surgery medical student rotating through obstetrics and gynecology, and this surgeon kept asking the same question, which was an impossible question to ask to answer of, of medical students. 

And we kept, you know, people were talking and I heard that, oh, he’s always asking this question. So I knew the answer. There was no technology there. Doctor Kirk, I just want to emphasize I did answer his question correctly, and he stopped the surgery. And because he was so shocked that someone answered it correctly. And then he said, how did you how did you know that? And I honestly answered, well, you’ve been asking everyone. So, you know, I heard that you asked. And he was he was very disappointed that I didn’t lie to him because he, he, he was all excited that someone could answer his question. And he told me, next time, never say that to say I read it somewhere. So I’m thinking maybe that’s some cognitive spoofing. That was not technology. It’s certainly active. I certainly acted like I was smarter, but I did fess up upon interrogation. So it turned out all it turned out all right. 

David Kirk, MD: 

Old school spoofing. 

Craig Joseph, MD, FAAP, FAMIA: 

Yeah, I was I think I’m going to look at it as I was ahead of my time. I just want to let you go ask a question about governance. You’ve mentioned meeting with your clients. And when you were at Wake Med, being the kind of, you know, one of the people trying to wrangle the technology kind of recommendations do you have for dealing with the fact that every physician who goes to any sort of conference comes back with a, hey, I saw this thing from this company, we got to have it. How do you wrangle those cats and make sure that you’re actually getting the best of the best at the right time for the right price? 

David Kirk, MD: 

So, when a physician brings something in one, once again, the CME, oh, the CIO, CFO has got to understand the why. So why are physicians needing this? And at the same time, is this a strategic priority in the organization? And what we’re finding is around the AI side, especially around documentation, quality, revenue; those are components that almost every organization has a lot of fruit left on the vine. 

So, to have a tool that both makes physicians happier and dramatically moves numbers for the organization has been very compelling. So, we we’re seeing a lot of movements. We’re seeing a lot of organizations branching out into this space. And there’s a lot of openness to say; we know AI is coming. How can we make sure that we use it respectively, and make sure that we’re getting great movement on both patient care, the quality scores, and the physician satisfaction? Can we do all of that? And if you get the right AI product, you can. 

Craig Joseph, MD, FAAP, FAMIA: 

Got a ton of more questions. But we are out of time. So now I have time for one last question. We always like to ask the same question at the end of these podcast episodes, which is hey! Is there something in your life that’s so well designed? It brings you joy whenever you interact with it? 

David Kirk, MD: 

I think in the beginning I told you that I’m now living in Manhattan for a farm boy that grew up in Mississippi, now mid-career, to move to Manhattan; you would think it’s quite the jolt. And figuring out the subway, figuring out how to get around the city would be complex, but there’s nothing that helps my lack of internal compass more than stepping out of the subway on that grid system and immediately saying, the streets are going this way, the avenues are going this way. I know what’s north, I know what’s south, and I know where I’m going to go. You would think I’d be scared to death. It’s amazing that anytime you can step out, you get two points of reference, and you know exactly where you are. You can walk and find what you need. To me, it’s just magical. I know every part of New York City is not on that grid, but when you’re in that sweet spot, there’s not a lot better for someone who, as locational and directionally challenged as I am. So, I love it. 

Craig Joseph, MD, FAAP, FAMIA: 

I well, that’s amazing. That’s an amazing answer. The street layout of New York City. And you do use it every day whether you want to or not, if you live in the city. All right. Well, thank you so much. Boy, this was great. We learned a lot and talked a lot about AI specifically, but tech in general really appreciate your time. And thanks for sharing all of your insights. 

David Kirk, MD: 

Yeah, this has been a blast. Do it again sometime. 

Craig Joseph, MD, FAAP, FAMIA: 

Thanks. 

Outro: 

Thanks for tuning in. We hope you enjoyed today’s episode. For more on David Kirk, follow him on LinkedIn and make sure to look at the show notes for more information. Check back for more episodes of designing for health wherever you listen to podcasts or on global. We’ll see you again next time on designing for health. If you enjoyed this episode, please leave us a five star rating and a review. This helps others find the podcast as well. 

Links:  

https://www.linkedin.com/in/davak/ 

https://regard.com/ 

City on a Grid: How New York Became New York 

https://www.amazon.com/dp/030682549X 

New AI risks: ‘Cognitive spoofing’ and fake expertise 

https://www.healthcareitnews.com/news/new-ai-risks-cognitive-spoofing-and-fake-expertise