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

This episode is brought to you by Dr. First. Their AI powered solutions improve the accuracy of medication history, enhancing patient safety and reducing clinician administrative burdens.

From e prescribing to medication adherence, Doctor First Solutions integrates seamlessly with your EHR, informing clinical decisions and improving care. Visit thisweekhealth. com slash doctorfirst today to transform your medication management with Doctor First's innovative tools.

Bill Russell: Today on Keynote

Lacey Knight: (Intro) I don't even have to remember my wallet. I don't have to remember a credit card. As long as I have my phone, they can sell me something. Think we could do it with healthcare with a bit more of a concerted effort on what the value piece was.

The harder part, I think, is us, maybe not the technology.

Now, let's jump right into the episode.

Drex DeFord: (Main) This is the Dr. FIRST Fireside Chat. Today we're having some cutting edge conversation with some of the nation's top CMIOs.

Thanks for being here. I'm Drex DeFord, a long time recovering healthcare CIO, and I'm now one of the principals at This Week Health, and I'll be your moderator today. I'm in the virtual room here with three doctors, and usually when I said something like that during the course of my career, that usually meant I was in some kind of trouble.

But not today, I hope. About a week ago I was with a small group of CMIOs gathered in Washington, D. C. at the this week Health 229 Executive Summit. They talked about successes, they talked about challenges, they had some really great insights. for each other and now you get to hear a little bit about what really happened in that room.

Colin Banas: Thanks Drex. Glad to be here.

94 percent hit rate for patients 65 11 percent of patient queries have been enhanced by AI. And by the way, Piedmont Health is a 25 hospital system that's been live on Epic. since 2012. And you also hear from Dr. Jake Lancaster, CMIO at Baptist Memorial Health Care. They originally went live with Dr. First in 2018, and since then they've provided over 1.

of course, our other panelist today is dr. First CMO Colin Banas, the famous Colin, and we're really glad you're here too, Colin. You want to take a few seconds and introduce yourself and Dr. First?

Colin Banas: thank you. Thanks for the opportunity and a special thanks to Lacey and Jake for joining.

I'm happy to call them not only colleagues, but friends, and really appreciate them taking the time. As you mentioned, I am the chief medical officer for Dr. FIRST. I am an internal medicine physician by training. Just like you, I'm also a recovering C suite from a large academic health system. I served as the CMIO for Virginia Commonwealth University Health System.

So the things that you just recited are related to medication safety, medication reconciliation. We also still do a fair amount of e prescribing. And then we'll, we're branching out into other places like patient adherence, prior authorization, price transparency. I really appreciate you guys taking the time and I appreciate you moderating and can't wait to get into it.

Drex DeFord: thanks to everyone who's attending and again, thanks for submitting questions and you can continue to submit questions through the webinar side panel. Okay, let's get the party started. We'll talk a little bit about AI, although we don't want AI to like completely dominate this conversation.

It is the thing that lots of folks were talking about. It was a lot of stuff that we talked about when we were all in the room together. I'll start with Jake. A lot of talk focused on ambient listening. Can you Replay some of that conversation for us and answer the question. Is it table stakes now?

So yes to answer your question, I do think it's becoming table stakes in that it is going to be very hard to recruit physicians in the future if you are not offering them an ambient solution, it's just we talk about ROI and what it means to the company, like how to measure that.

But at a certain point, it's one of those things that if you're not offering it your competitor will be. And yeah, I've just been very stunned to see how fast it's really been adopted by healthcare systems. It's, gone from nothing two years ago to ubiquitous.

Lacey what, did you get from the conversation specifically around ambient listening? But we can certainly expand from there.

Lacey Knight: I think how quickly it's changing, I think is one of those things. Jake mentioned a couple years ago, it was in one space. I think in most cases it required a, literally a human in the middle that would do the transcription to do that.

and now certainly with generative AI, month over month, the capabilities of these solutions are just getting faster and faster. And I'm, I think we're trying to spend some time thinking about what that paradigm shift looks like. I think those of us on the call are used to learning how to take care of patients by reciting the SOAP note and entering this debate over the APSO note where you have the assessment and plan first so that it's more readable and things like that where you're Ambient listening is really about how you have the conversation with the patient that enables language capture.

I think That's the part I think we're going to find out in the next year or so, how we have to change the input when we know that the output's going to, be dramatically different.

Colin Banas: going to chime in, Rex, because I actually think that's a really interesting observation, Lacey, because I find that, and I think like you guys, I was fortunate or unfortunate enough to practice on paper, practice on hybrid, and then practice fully electronic.

There are things that you think of in the creation of that document that leads to changes in your orders and changes in your differential, etc. And I worry, or at least I think your comment is spot on. We're going to have to change the way we educate and we're going to have to do it pretty quickly.

Jake Lancaster: Yeah there's a lot of concern about de skilling with a lot of artificial intelligence, losing various skills. And I also think back to when I was in residency, we got electronic health records there was worry about residents using the order sets that were pre built that walked you through how to treat a certain illness and worry that would lead to de skilling.

Drex DeFord: It's interesting, too, to think about that problem in the context of Every medical school and every residency and every fellowship, like they're all different, no matter where you go there's consistency, but there's also where that organization is when it comes to AI and will there be a haves and have nots over time about this?

Lacey Knight: I, think that as the costs come down, it'll shift more towards something that maybe is a little bit closer to a commodity, or some of these types of things. I think one of the biggest challenges in medicine has been that we, all learn through apprenticeship, But yet, we think that clinical variation is the, challenge that impacts outcomes.

And so we imagine that a lot of the technologies that we use are a means to encoding, clinical practice standards. So Yeah

Jake Lancaster: I

They're not providing these things. They're not going to get them there. So they almost have to do it more than a large healthcare system that has monopoly on physician practices.

Drex DeFord: That really drives back to that table stakes kind of conversation. Maybe we're getting there. Colin, what's next for AI and healthcare? As you listen to the folks in that room, as you go out, you're exposed to a ton of stuff. All of you are. I'll start with Colin now. What's next for AI? What do you look forward to?

Colin Banas: Yeah, I I think we're at that, peak of enlightenment, or inflated expectations right now, I should say.

I still think we're way behind any other industry. What I haven't seen yet, but what is indefinitely coming is the diagnostic stuff, the clinical care stuff, right? So we've seen the pilots of. Radiologists being assisted in detection. I don't think that's wide scale yet, but I think it's coming, and I think it's going to be more widely adopted.

Similarly complex algorithms around things like sepsis and oncology. Sepsis algorithms mostly until recently have been like a coin flip. And now we're getting better because of the maturation of AI. So I think we're going to shift from the administrative and the automated to the, clinical.

And I think you're probably going to start to get bigger bang for your buck in the clinical, but it's a tread carefully proposition. Lacey,

And I remember Distinctly looking forward to my first email address and I would get email once every couple of days. I remember that in the middle of the night there were some groups of people that would talk to people they didn't know in online chat rooms and thinking that was never going to take off.

So I think really the opportunity is anything that is somewhat challenging. to process either through human effort or challenging to understand because of the density of data, I think are where the opportunities are going to be. Certainly a variety of different complex clinical operations.

I'm sure scheduling, I'm sure perioperative space I think the reverse of what ambient listening is doing, which is collecting information and summarizing it where you actually have the AI. Do some outreach on behalf or simplified follow up type of thing. I think a lot of that we'll probably start to see.

I We're already seeing that with the ambient scribes. It's coming to nursing next. It's coming to inpatient. That's going to happen and fairly easily since the EMRs are actually integrating AI inside of them. That's one of the big issues with AI programs in the past is getting them integrated with the EHR.

And so now that Epic and others are doing that, it's going to be far easier to deploy than what you had to do in the past. I think it was something like 1 percent of all AI algorithms that have been developed or actually deployed in operation. I think now that the EHRs are helping with some of that development, it'll go a little further.

You couldn't also look at prior imaging studies. You couldn't look at the patient's history of the labs at the same time. With GPT 4 and these newer transformer models, you can incorporate multiple modalities. So the multimodal AI is where the industry is really going. That being said training multimodal data is going to be a little bit of a roadblock, and then implementing that into an actual healthcare system is an even bigger challenge, as you probably know from your CIO background.

But that is where the future would go. And you can already see this a little bit with GPT 4. I did a demo where I took a picture of my son who has eczema, his rash, and sent it to ChatGPT and asked it for a diagnosis, didn't give it any clinical background, but it came back with atopic dermatitis.

Be able to pull in not just EHR data, the imaging data, the genetics, the the wearable application data, etc. all into one place and have an AI that is more general purpose that can You know really replace us as positions because we're all tired of working.

Lacey Knight: This reminds me of another conversation in the room.

And then the second part gets to the latter comment that you'd mentioned, which was around this kind of platform of platforms. that is integrated. So there's two approaches I've seen with that. One is certainly like the EHR vendor that kind of gets into all the spaces and brings the AI delivery to you within their platform.

But there's also been this kind of platform layer, maybe more on the imaging side where the idea is that the platform they offer is the connection to the EMR. And so then All of these independent, best of breed niche vendors plug into that platform. So first off, are you just absolutely ruling out anything that's not workflow integrated?

And then the second part which of those two battles do you think is going to win? The, EMR vendor that goes into all these spaces, or the platform vendor that brings all of the niche solutions forward?

We do have physicians anecdotally not, internally, that will use CHAT GPT and other GPT 4 algorithms out there to answer actual clinical cases, but that requires them to go to a separate site and, use it. That's, Few and far between. The majority of physicians are not doing that.

I would not, as a CMIO, bring on maybe a separate product that was not integrated, knowing that our physicians it's very hard for me to get them to go to even a separate window within Epic, let alone a separate application outside of it, unless it's something they've been doing for years, like looking at radiology packs or something like up to date.

You can have several different vendors that all do the same thing. But going forward, niche solutions for AI, it's very hard for a healthcare system to bring them on. And just proving the ROI per niche solution is just way too much work. And most of the time it's very hard to show real dollars there.

But if you had a, general purpose algorithm that I was describing before that could come in and solve multiple different problems. I think that would be something that would be more adoptable.

Colin Banas: can answer this with real world examples. We have the same solution that we offer. One is so deeply integrated that folks don't realize that they're using our solution.

There was a long time ago, do you guys remember Billy tool? For pediatrics, there was this Billy Rubin website where you could input the patient's age in hours and their Billy Rubin levels, and you could figure out if it was safe to discharge them or, you know what, that was a lot easier than looking it up on a little note card or in a textbook.

We've been talking about it for 20 years and it still hasn't happened. And who's winning is the big EHR who is Getting their tentacles into everything and blightlight everything up. The EHR is the

Jake Lancaster: platform, yeah.

Colin Banas: Yeah, I mean they're opening up, but they're also your co opetition sometimes.

If you're an independent solution, just like Apple, every once in a while Apple will say, oh, that's a good idea, I'm going to start doing that myself. It's a really interesting question that you posed, though, Lacey.

Drex DeFord: That's a good bridge into maybe the next topic.

I feel like we could go the full hour, the full 45 minutes and talk just about AI but the next topic, the next thing we had talked about was You know, interoperability, and 16 years out from meaningful use. We're still not there. We've made some progress.

Colin Banas: what I heard from the group was from frustration. There was a couple of examples of, Hey, this health system is right down the street.

Hey, this health system is on the same EHR platform, yet I still can't get the appropriate data. I, also hear the flip of that, which is I have a lot of data, but I don't have a lot of information still, and that data isn't eminently usable. It's you'll often hear me talk about the semantic interoperability problem, where, you know, when the three of us were training, we were getting faxes if we were lucky, and we were combing through this stuff in the middle of the night.

And now you don't have to do the fax so much. You might get this massive dump of a CCD or PDFs. And so you went from a paucity of data to too much data, but you still didn't have information. You still didn't have knowledge and, you still had

Drex DeFord: to sort it. You still had to go through and figure out.

Yeah.

in terms of how easy it is for data to move from A to B and be imminently usable. So I think the key word is frustration, but maybe a tint of hopefulness.

Miroslav Balote, CISO at Valley Health, will share his journey to building an integrated risk management program that automates and simplifies vendor risk. Alongside experts George Pappas and Scott Matilla from Enterprise Health, this session will cover the latest challenges and practical solutions in managing cybersecurity threats.

Drex DeFord: We've gone through a lot of M& A conversations, we talked a bit about M& A. In the room and that is a topic that comes up across most of the places that I go and meet with execs.

And so even the situation of having the same EMR, but having different instances of that EMR, and those have a hard time sort of comparing anything from lab results to you name it that, that challenge and frustration too. For, Lacey and Jake, what are a couple of things you think would improve if we could get better interoperability?

What would you hope for eventually?

Lacey Knight: feel like we need to start by maybe changing the way we think of the term. I feel like it's looked at as data interoperability. So if I can view the data in one place or the other one, then we've achieved, the goal. Maybe that's a starting point, but to Colin's point, the concern in the room was really about operational interoperability.

So I don't have, I don't really have to worry about whether or not, I have to find a solution to make it work. I think the biggest thing I would see, just a very specific detail, this idea that was referenced in the room about requesting surgical cases and managing surgical patients from the outpatient to the inpatient.

So I think that's a pretty good example. I do wish, I mentioned scheduling earlier, I do feel like the connection between the schedule that happens from a hospital side around which doctor to call. In an emergency, it would be nice if there was a way for that type of information to intersect with the physician's.

Independent practice schedule, because the practices always have to manage call at multiple hospitals. They also assign call different ways in their group. It'd be nice if that was a little bit more, streamlined.

Jake Lancaster: Yeah, I can talk about this a lot. First, I would just want to defend the healthcare industry a little bit more.

I don't think that exists. Probably doesn't. Their databases don't align that way. So it's, not really apples to apples comparison, but at the same time, there's a lot of work that still needs to be done. There's a lot of work that has been done to the point where, yes, we can exchange information from one site to another.

I get information through our, HIE, FXHIE, CARE everywhere. But the issue is it's bad. A lot of it is redundant to stuff I already have in the system and getting the positions. Staff to actually go in and reconcile that and bring it into your system is a chore. And we're finding that it's not worth the squeeze a lot of the time.

If we send a patient over for a colonoscopy and we need to get that scanned into our system to satisfy a health maintenance requirement, that's a lot of manual labor that still goes into that. Everybody's on in EHR and everybody has the same metrics that they care about. Yeah, there's still a lot of work that needs to be done.

A lot has been done. It's a lot better than when we started when you had literally no information on a new patient. Now at least you have a starting point. You may be able to get their problems. Medications, allergies, but that's a pretty good starting point for the majority.

Lacey Knight: Jake, you're a lot more patient than I am, on this.

I'm not willing to give the healthcare industry slack. I can pick a whole host of other reasons, industries outside of banking that

Lacey Knight: It is, but if you wanted to find a product or a flight or a route to someplace, you can put in a Google search a specific item that might be rare.

and have a list of places that might have it within their inventory. But it's, my argument is that it's not that it's necessarily, it might be more difficult than some of these other industries. I don't know that we have been compelled to solve it like the other industries have. The reason why the other industries have solved it is because they have a consumer engagement or a financial incentive to make it easier.

The restaurants I was talking about, you don't, I don't even have to remember my wallet. I don't have to remember a credit card. As long as I have my phone, they can sell me something. that's why they do it. think we could do it with healthcare with a bit more of a concerted effort on what the value piece was.

The harder part, I think, is us, maybe not the technology.

It does make it complicated and I think it does complicate the whole process around interoperability too.

Jake Lancaster: Yeah, there's no shortage of people in the middle.

Even sort of internal operations problems that can take a health system down. Business continuity seems to be a challenge. for everyone. From the conversations that you had in the room and other conversations that you've had, we've all had outages. How do you prepare for those with your clinical leaders?

How are you doing business continuity planning?

Jake Lancaster: I see you can go first. I was like rock,

Lacey Knight: paper, scissors on this one. I think the hard part is

Marker

Lacey Knight: that it requires some degree of vigilance and a fair, a healthy amount of paranoia. To work on this, we have had a number of different tabletop exercises, we are actively discussing how prepared we are with, the business continuity devices and trying to figure out how do we actually make sure that people who are clinical that never imagined that things are not going to be available because their entire career.

I think in this discussion in the room, it came up that some, systems have moved to all the automated tube systems. And sometimes those can be impacted by these downtimes, which means the logistics of getting things from one place to the lab are significantly more difficult because Yeah, where do

Drex DeFord: all these runners come from?

We, I think we, when we do the exercise, we go, and then a runner will appear and all this thing

Lacey Knight: We, had the benefit of a hospital move, and a flooding event that shut the power for the hospital, and, fundamentally what happens is that, when things are not working normally, People's bandwidth do open up.

So continuity in that space, I think is a, bit frightening, but we do as you mentioned, we do a lot of exercises to try to prepare for it in a virtual environment. And I think if I'm remembering right, somebody brought up the idea in the room, does anybody actually do, a true live mock event?

And all admitted to being absolutely frightened about what that would look like. Because, at a certain point you don't have the experience of not using technology to do things. Measure drip rates, document, Find x rays or review x rays without automated report. There's lists of things.

And I think that's what everybody's really concerned about.

Jake Lancaster: And I think that's part of the reason why we, don't, or most people don't do real mock events is because there's been studies published that downtimes do have adverse patient effects. And so if you do a mock event, are opening yourself up to an adverse event just because you're trying to practice for the future.

But I agree with everything Lacey said. You got to prepare as much, on the back, on the front end for prevention and then making sure that your clinicians know that even without the EHR and the technology, you still have to take care of patients and do those critical things that you know matters.

But it's still, it's something that we all just hope and pray we never have to go through, but almost all of us will go through it.

Colin Banas: from the vendor perspective I have to think about it in ways of how can Dr. First help provide business continuity in these events.

In fact, I gave a decent example 10 minutes ago when I said, We've got one solution that's super integrated, you don't even know it. And then we got another one out here that's web based. Guess what? If you go down, and this has happened, so I can say this, as long as you can get to the web, you can still conduct your some of that business that you were reliant on your EHR for the other things that we think of is providing, duality to our solutions.

And that's actually one of the fun things that I get to do is think this stuff through without having to sit in their seats and contemplate real life mock events.

Drex DeFord: I'm glad you're doing that. I see a lot more partners that are putting a lot of thought into that idea of if we're integrated into the EHR, what do we do if the EHR goes down or if the system goes down?

what's our backup plan and our backup plan to our backup plan to be able to continue to provide those services the health systems. Okay. Thanks for working with me on that one. I'm going to go to some of the questions. A lot of questions were submitted in advance.

How do you go through that process of deciding from a governance perspective, which one makes it and which one doesn't?

Jake Lancaster: I can start. We, like most healthcare systems, have a very small margin. So the first thing, one of the things we're going to really look at is, does it have a tangible ROI that is quantitative? A lot of what we talked about with Ambien earlier, IRIS, FHIR. On the outpatient side, you can find use cases for increasing the number of visits, et cetera, by having that solution.

We talked a lot about some of the other vendor solutions that are coming that look amazing, but don't have as easy of an ROI, to, really present. So we have an AI governance group that is recently stood up to evaluate some of the new things that are coming to help us better understand.

What is the organization ready for? What, where do we want to go first? What do we still need more information on? Just because there are literally hundreds of things that are coming out of this space, especially generative AI, over the last two years. And we need all the people, the VPs and C suite around the table to understand.

Does it replace your old IT steering committee? For us, not really. It just augments it. It sits within it. Because it's really a lot of the same players. And almost every new technology and old technology that is out there has an AI component now. So if you were going to have a separate AI governance structure, You're going to be doing a lot of redundant work, but that's where we are right now.

And it's, still evolving I would say we're not where we want to be quite yet, as far as the governance for it.

Lacey Knight: I agree with everything Jake said. I have trouble imagining that we're going to be where we totally feel like things are working smoothly because Things are changing so quickly.

So in terms of looking at things that rise up to the list a little bit faster, those are things with a very clear ROI. Those are things where the technology seems to match our environment where we are and where we have either clinical or operations that can Adapt and adopt the changes with Gusto, I would say.

Drex DeFord: Yeah.

Lacey Knight: And if we have synergy in those three areas, , then it's still, then I would say the next question is it an opportunity, or is it something that is an urgent item? Urgent's always first. But I would say we're looking at the combination of those attributes to decide where to go.

I don't know if you want to talk about that. These were all great things that I saw , other people in the room taking notes. Like that's the beauty of some of these summits is that there are lots of gold nuggets in those and this was one of them. So I hope you can give us a couple of minutes on that.

Lacey Knight: Sure. Just to give a little bit of backdrop behind this part of, this was driven with this idea that we wanted to make sure we were taking advantage of the types of things AI was offering, but, we're always asked to do things faster with fewer resources. I think that's true for everybody on the call.

And so how do you do things fast in a disciplined way? So we put structure around what's the question or the opportunity. We defined a timeline, In our case, sometimes 30 to 60 days really answer the question or define the opportunity. Defining the opportunity for us means What is the short term benefit?

So each of those phases, are things that we've rolled into. And the first, time I think we did this really was with when we were looking at virtual nursing last year, where the initial question was, what does virtual nursing mean for Piedmont? And that question meant we did a lot of deep research with organizations that had been live on it.

And we picked the sites to start. We picked a time frame to observe the impact, and then we are approaching finishing out our full system wide rollout, again with a well defined scope of impact that we're intending to have, and then expanding it, but I think for us that The time bound discipline around evaluating an opportunity that would be transformative is the thing that we have tried to be very good at, the last couple of years.

We can't do everything. So again, for us, we're trying to pick these gigantic rocks that, that move the organization forward.

Drex DeFord: Jake, as you look at this I know that you're, you have a lot more discipline in this as time moves on. What are some of the metrics you're using? What are some of the processes you're using to focus on ROI and success?

So we do that. We did it with Ambient. We've done it with our virtual nursing project as well as some of our other informatics projects that don't involve an external vendor, just doing an internal project. It, whether it's time for the nurses per shift in flow sheets, et cetera, or if it's time for positions and notes, or even just getting a survey of the staff that are using the new product or, been through the new process about how they feel about it and whether or not they're likely to recommend it to peers in the future, all things that we've looked at, certainly cost of the change is something that we, again we, have to pay attention to, cause we don't have an infinite amount of money other [00:45:00] organizations.

So we really have to pick and choose what we want to do and really, know that what we're putting in place is going to have the biggest ROI for us. But we're not as disciplined as Lacey is, I will put it, I'm going to tip my hat to him. That group is something, yeah, that seems very special.

Colin Banas: Yeah. Your thoughts,

Jake Lancaster: yeah.

Colin Banas: You can't manage what you can't measure, right? It's just important to hear from our partners that, This is what I need to be helping you with during the entire process helping you to make the case when you're during selection. One of the first things I did when I joined Dr.

First as CMO was help beef up the clinical research arm so I'm going to take real world metrics from real world partners and form peer reviewed publications or case studies that not only help you guys when you're going back in front of IT steering to say, Okay, it's been six months.

Because that's the sort of thing that justifies the program going forward. And so I think it's important for vendors to hear, for solution partners to hear that this is the sort of thing that, that is important when you partner.

Drex DeFord: Yeah, I, so thanks. I'm starting to get the hook here. So I'm going to ask one more very short question.

I know I can't believe it's gone this fast. I'm going to ask one more question here. All three of you looking for a very short answer. This one is from the audience. Will AI eliminate jobs in health care?

I'll be short there and stop.

Drex DeFord: Thanks, Lacey.

Lacey Knight: I would also say yes, but I would say that it's not going to be an, it's, not going to be a net loss of jobs. Just to give an analogy, I bet most of the people on the call probably spend more on internet and cell phone. now than they did 20 years ago, and that's replaced other utility costs, but our probably net utility expenses have gone up.

So I think with these advanced technologies, some work will go away, but more complex and different types of work will ensue.

Drex DeFord: Yeah, I like your thoughts around that. Colin?

Colin Banas: Yeah, I think that's a perfect answer. It has the great potential to be a force multiplier, but we don't know what we don't know going forward.

Drex DeFord: Okay. Thank you all for for joining us. Thanks to the audience. Thanks for the audience submitting questions.

Dr. Furst, thank you for for being here. Colin, thanks for sponsoring the fireside chat. And of course, Jake and Lacey, there's nothing that happens on here without without Jake and Lacey. So thank you gents for being on too.

Lacey Knight: Yeah. Thank you, Drex. Thank you for having me.

Drex DeFord: And that's it. That's it for now.

We will see you the next time we round up. for a fireside chat or webinar. Thanks for being here.

Bill Russell: Thanks for listening to this week's keynote. If you found value, share it with a peer. It's a great chance to discuss and in some cases start a mentoring relationship. One way you can support the show is to subscribe and leave us a rating. it if you could do that. Thanks for listening. That's all for now..