Okay, everybody. Thank you so much. While we got ourselves organized up here, welcome to Day 4 and to today's session, 10.30, for the Anesthesia Patient Safety Foundation session, The Evolution of Artificial Intelligence and the Intersection of Patient Safety, Continued Quality Improvement Program Grounded in Implementation Science. A couple housekeeping notes before we begin. My name is Jeff Darm, and I am a member of the ANA Professional Development Committee. And for everybody here in the room, a couple of housekeeping reminders to please download the ANA app and select the ANA 2024 Annual Congress and to log into that app to access the meeting notes and presentation. It is also where you will complete your evaluations. Be certain you get your evaluations completed at the end of each session to claim your CEUs. You will have until Monday, September 9, 2024 at 12 noon Pacific time. Please mark that date and time down and tell your friend to submit all individual session evaluations and the overall conference evaluation. Following that date, you will no longer be able to claim CE credits for Annual Congress. So that's really important. Make sure you tell a friend and you get those CEUs for credits for attending this session. We have three incredible presenters today. We have Patty Riley, who is a CRNA at Chester County Hospital, Pennsylvania Medicine in Westchester, Pennsylvania. We have Dr. Steven Greenberg, who is Secretary of the Anesthesia Patient Safety Foundation. And Desiree Chappelle, who is a CRNA from Louisville, Kentucky and is the Vice President of Clinical Quality for NorthStar Anesthesia. And I will then turn the programming over to Desiree. Thank you so much. Thank you. Thank you. Hello. And welcome to the ANA 2024 and our annual presentation for the APSF or the Anesthesia Patient Safety Foundation. I'm Desiree Chappelle and thank you for the lovely introduction, but I am a CRNA practicing in Louisville, Kentucky, and I'm Vice President of Clinical Quality for NorthStar Anesthesia. I also have a perioperative medicine podcast called Top Med Talk. And so those are all kind of my relevant disclosures, I will have to say. So we are so excited to be here. This is an annual presentation and conversation that we have every year here at the AANA. And because it is a conversation, we really want that piece. I would really urge you, if you're sitting in the back of the room, come on up. Don't be shy. We love, we're like, we love everybody and we want to have a little conversation. So please, if you can, move up and that would be great. If not, it's okay. No worries. There's going to be lots of different ways to interact with us today. First of all, please feel free to come up to a mic. I know we're in this cavernous space and that makes it a little bit difficult, especially for you guys sitting way in the back back there. But thank you for joining us anyway. But use a mic. We're also using the app that we know and love for the AANA. So we'll be taking our questions on the app. And we'll also be utilizing a platform called Slido. And so this is slido.com, so you pull out your phones. It's the one time during a presentation, actually, I think there might be a couple times, that it's good to pull out your phones. Pull out your phones, shoot this QR code. And this is what we're going to be using for polling of our presentation today. So throughout our presentation, you'll be seeing this come up, but it's slido.com. So all of that good housekeeping stuff out of the way, and I'll show our disclosures here in a second. I wanted to take a moment and first of all introduce the panel that we have here and then talk about one of our very dear friends who hasn't been able to make it today. So if you're watching at home, Lynn Reed, we love you, we miss you. But we'll go ahead and hand it over to Patty Riley. If you can go ahead and give a little bit more introduction and tell our audience today a little bit more about yourself. Hi, everyone. Thanks for being here. I'm Patty Riley. I am a nurse anesthetist for a very long time. I also have been in the industry beginning with Nalcor through five different companies, now Medtronic, but I work closely with to bring the clinical voice into industry. And I've had the pleasure to be on the APSF, be with the APSF for probably 25 years, board of directors for the last six years, committee on technology, as well as other committees. So I'm very honored to be here. I'm glad we can bring our message from APSF to all of you and hopefully recruit all of you to join. And I see Dr. Van Pelt, Maria, thanks for being here, past board member of APSF. Thank you for being here. And I'd like to say one more thing. Lynn Reed did so much work on this presentation. Unfortunately, she couldn't be here. So I have to say, and I hope she's listening, thank you, Lynn. She needs to be recognized for all that she does for AANA and APSF. Yeah, absolutely. And this is for Lynn. All right, Steve, why don't you introduce yourself? Thank you, Desiree. I just want to thank the AANA and my esteemed colleagues on this panel for inviting me here today to talk about emerging technologies and artificial intelligence. I'm an anesthesiologist and intensivist at Endeavor Health. I'm vice chair of research education there. I've been there for 18 years. And the APSF is near and dear to my heart. I've been involved with the Anesthesia Patient Safety Foundation for my entire career. I'm vice president of the organization and editor of the APSF newsletter, which goes out to readership over 234 countries and really is a multidisciplinary, multi-professional modality in which to educate the world on patient safety. So I'm delighted to be here, and I can't wait to have this discussion. As Desiree said, please come up. We want to hear from you about this emerging trend topic that we're going to talk about today. Absolutely. Well, today, like I said, is going to be what we call a presentation, but it really is a discussion between the three of us talking about AI, the use of it in our world of anesthesia and a greater health care system. And so we're going to go ahead and kick it off with our first question, which is not about AI, but hopefully it'll get you used to using Slido. So as I run through our disclosures here and our conflicts of interest, I'll pass through those very quickly. All right. Actually, I'll go back so we can see our questions. I didn't put that in the right place today. All right. Favorite Olympic sport. So shoot the camera at it. Anybody voting there? There we go. 27 participants. We'd love to see those numbers get up. All right. And we've tried a new system here, and it should be coming up. Oh, shoot. Oh, there we go. Well, I can tell you what the poll answers are, because it's not showing up here. All right. Well, I'll go ahead, and we'll get going in this, and I will fix Slido while we're chatting here, because we have a lot to talk about today. So what we wanted to cover today is really, like, the intersection of artificial intelligence, patient safety, and how we implement that into practice, and really thinking about how we utilize that to also improve patient outcomes. So kicking it off, Patti, why don't you tell us a little bit more about the Anesthesia Patient Safety Foundation and our role as CRNAs within that organization? So great. Thanks, Desiree. First, I want to comment, we attended Dr. Gu's lecture this morning on AI. Thank you for that lecture. We learned quite a lot, and we walked away with the future of data, data, data. I don't know if you're in the audience, but thank you. So moving on to the APSF, I shared with you that the APSF is going into its 40th year. It's the oldest patient safety organization right now. They're known because we actually transferred the anesthesia care back when they started, and a group of individuals got together and decided they needed to make anesthesia safer. I've been a member, as I shared, probably about 25 years, participated on many groups. I shared with you there were other CRNAs on these groups that impact what we do, how we work collaboratively, and it truly is collaborative. So as a CRNA, please, if you're not involved, try to get involved, at least read the newsletter. I know that Lynn is not here. I don't know if Josh is here. Josh contributes quite a lot to the APSF. Maria's there, Drew's there, and, of course, myself. So anything that you can do or participate in, there's lots of opportunities. Please join us, APSF.org. And Patty, I will say, I did. I took Patty's advice, because she urged me to participate, and I recently joined the Infectious Disease Committee. We published there, and then a colleague of mine, we just published a paper in the APSF. So it's very easy to do. It's a wonderful organization to be involved in. So we'll give another plug for it later, I'm sure. But... Yeah. And we'll go through some of the work that's being done. So the newsletter, I'm going to ask Dr. Greenberg to talk about a little bit. As I said, he is the editor of the newsletter. If you haven't read it, it's a great article. I used to say it's U.S.A. when people read newspapers. It was U.S.A. today as a newspaper. And I think it's probably the most widely read, and you can probably speak to that anesthesia paper. Yeah. Yeah. So certainly the most widely read safety piece of information in the world. About eight years ago, we went global, and now we are in, as I said, 234 countries. We translate the newsletter in eight different languages, which includes English. And our international constituency is even as much as our national one. So our foundation has changed into a global foundation. I think the most important thing about our foundation is that we're inclusive, inclusive of all folks that deal with anesthesia and perioperative care. And I think that's a real focal point on how we get things done, and get things done in a very positive way. We sponsor a lot of research. We've given over $13 million of grants to scientists to improve safety in the perioperative space. We build and enhance patient safety partnerships with other national organizations to try to improve the care at the bedside. And again, we promote national and international exchange of ideas, and we're continually enhancing those relationships around the world. I just wanted to comment. It's also made up of industry expert, as well as, as I said, anesthesiologists, CRNAs make up the congregation. And again, it's very collaborative. Absolutely. So the APSF really focuses on patient safety priorities, and has grouped those into different topics. So tell us a little bit more about those. So back in 2018, our multidisciplinary board of directors used a modified Delphi technique to compare one major patient safety priority to the next. And that's how we actually developed this top 10 list. It was based on a multidisciplinary stakeholder group. And we actually update this every three years to reflect the trends in perioperative care and where safety is really of utmost importance. So you can see them behind me. But again, culture of safety, teamwork, clinical deterioration, you can read them. These are really the focal points for our consensus conferences, which we meet every September. The conferences are named after one of our former presidents, Robert Stolting. And this is, again, a way to convene a multidisciplinary, multi-professional group of folks to say, what are some of the problems based on those safety priorities? Who understands the problem? Getting everybody together, including industry, risk management, dentists. There's all sorts of folks that actually come to the table to figure out what's the problem and how we're going to address it. And so we actually break out in small groups during that conference and develop specific consensus statements that we then publish in our own APSF newsletter and in other journals. So it's a great way to learn. We'll talk a little bit more about the conferences at the end. But anybody can really join in those conferences in September. One thing I love about the conferences that I have learned about over the years is that almost every conference that you guys do, the consensus conference, those are published after the fact, right? So you take all of the information that you gathered before, during, and after the conference and publish those. So those people who are participating in that actually are published as well, right, Patty? Yeah. The other comment I like to make for CRNAs is I've personally led and co-led many of these conferences. And again, we really want CRNA involvement in the APSF. So think about that. We do topics that are of interest and meet the need at the time, but also, again, it's collaborative and we're really a large part of it. Yeah, absolutely. All right. I think I have Slido fixed now. Okay. Just bear with us just a little bit. If it's not, we used a new system this time. All right. So, oh, spoiler alert. I don't know if you guys saw that. So the favorite summer Olympic sport is now gymnastics. No, I don't think that's a big surprise. That was very fun to watch this year. All right. So our next question, how do you feel about the use of AI in everyday life? Smart home devices, personal assistants, social media. All right. I'm going to get it. Hold on. One button. Is that going to do it? Here we go. Is it showing it now? Ah! All right. Well... I will tell you that somewhat positive is what we're seeing here. Somewhat positive. Most people, about 47% actually, half of everyone in the room, say they're somewhat positive about smart home devices, personal assistants, social media, things like that. Are you surprised by that, Steve? I think that's cautiously optimistic is what I would equate that to. And that's, I think, how most people in health care are today, and actually in general society. I think we're hopeful that these technologies can help us, but are somewhat half-eye, one eye open, one eye closed to the trust that is involved with validating the technologies that we use on a daily basis. Why don't you tell us a little bit more about artificial intelligence then? I want everyone to do a little exercise. Take out your handheld devices and just raise them up. We all have them in our hands. OK, so that's 100%. If you have one of these, and we all do, you are using artificial intelligence. Whether it's going on Grubhub and ordering dinner when you haven't made yourself dinner, whether it's asking Suri or Alexa, what's the weather going to be like so I wear the appropriate clothing, whether it's asking Google or Waze to get yourself from A to B in the best way possible, we are using AI every single day. And in health care now, AI is being used based on collecting millions of data points and then providing professionals with optimal treatment modalities. So not only in our everyday lives, but also in health care, we are seeing artificial intelligence exploding. And this is a survey based on Pew just last year, about 12,000 Americans all over the United States, and about half of them really understood at least one of the common examples of where we use artificial intelligence on a daily basis. But interestingly, on the other part of this slide, only about 30% got all six of them. The same survey looked at health care, and they asked these respondents, what do you think about artificial intelligence making clinical decisions for your professionals that would then the professionals would recommend? And 60% of those responded very uncomfortable with that specific idea. But then when they were asked the same question, a similar question of whether the technology would actually improve their outcomes, about 40% said, yeah, they would. So there is somewhat of a disconnect there with the trust, as we talked about, with regards to are we really going to rely on this to make clinical decisions at the bedside? So more to follow in our discussion. So AI is not a new idea. And this slide shows that in 1955, John McCarthy, a computer artificial intelligence scientist, coined this term to look at machinery and making it more intelligent. And so that's really where it started in the 50s. And during that time, actually, artificial intelligence was really focused on enhancing marketing. And we went to the 70s and 80s, and there was a creation of neural networks and expert systems to enable marketers to actually hone in on customer behavior to enhance their personalized experience. And then we had the 90s, 2000s, where we found an increase in use, obviously, with explosion of the internet and e-commerce to finally hone in on data analysis to optimize search engines. And then we look at the 2010 and 2020s with a huge increase in big data analysis. Now artificial intelligence can generate its own content. It can potentially predict outcomes and enhance all of our personalized experiences. So that's where we've been in the last 70 years. This is not a new topic, but it's new in that medical field from about 10 to 15 years ago. I would say that in all the different things that I do in the spaces that I run, in the last two to three years, no one really talked about it before then. But now it is exploding, as you said. And so there are a lot of definitions out there about artificial intelligence. I think we have to all kind of be on a similar page on what we're talking about today. And as it relates to health care, it can be broadly defined as the ability of a computer or a device like ours, all of ours today, to really analyze a large volume of complex health care data. And that will help us get knowledge to identify both opportunities and risks, and help us make validated decisions at the bedside. All right, so I know, I've gotten confirmation that you guys can actually see the results on this. So, even though we can't see it up on the big screen, we can all look at our phones, our little AI devices here. What are the clinical applications of AI in the perioperative space? And we have things like management, ultrasound assistance, information searches, diagnostic interpretations, every area of clinical care. Is there anything else you can think of, Patty, that from your experience? There's so much, there's just so much. Yeah. And Steve, what about you? Risk stratification is a big one. Yeah. Figuring out, when we go into that operating room, what are we dealing with in terms of our patient populations and how we can optimize their care by knowing what we're getting into before we get into it. Yeah. We hear a lot of things like wearable devices, prediction, now there's all kinds of technology that can predict when certain hemodynamic instability is upon us, or if the patient is going to be, you know, different things, changes in their status. So, I think there's a lot there to unpack, and I think we're gonna talk a little bit more about that. Second question is, do you believe AI has the potential to improve patient outcomes in the anesthesia space? And it looks like that, oh, it's a little strong, 85% strongly agreed. That's about 60%, about 40% agree. So, I think most of us in the room, there's no one that disagrees or strongly disagrees, so that's good. So, most of us believe that AI does have the potential to improve patient outcomes. What would you say about that, Stephen? Yeah, so, you know, really the focal point of AI in terms of improving perioperative care is really predicated on four or five major ideas. First of all, the idea that we can timely deliver useful information to bedside providers to improve clinical outcomes. We can also assist with personalized care. So, we've all taken care of patients who have liver failure, renal failure, cardiac problems, but we don't know how those patients do in aggregate. And if we use points like machine learning, we can actually figure out how those patients do and then tailor our anesthetic approaches to those specific patients. We can then reduce also clinical variation and hopefully, overall, improve patient safety, outcomes, and their own satisfaction. So, I think I'll be remiss without just focusing a little bit on overall terms as they relate to AI in healthcare, and Dr. Yu this morning went over this, and so I'm gonna kind of give you a global picture of this. So, machine learning is really a statistical approach to fit model with data, to make sure that algorithms work as the way we want them to work. It's a way to put a construct around large amounts of data points to figure out what an outcome or a therapeutic, or optimal therapeutic approach is. And we actually see this in perioperative care today. We see it with evaluating post-induction hypotension before it occurs, using large data sets with electronic medical record. We can use it with forecasting process EEGs by looking at different concentrations of administration of propofol and other anesthetics. And we can even look at prediction of post-operative in hospital mortality based on individual patient populations by analyzing that large amount of data. But this is customized based on the providers developing the algorithms. Deep learning takes it a whole nother step up where the actual computer itself learns those patterns and then provides whoever is looking at it what the outcomes might be or what the therapeutic responses might be. So for instance, this is actually ongoing now in perioperative care. There are systems that can identify cardiac structures when you're doing perioperative echo. There are also other systems that are handheld that can actually show you where to put the needle and how deep the needle needs to go in terms of effective epidural placement. And also, and we actually sponsored a grant about four years ago, looking at facial recognition, the APSF I'm talking about, and figuring out based on facial imaging who might be a difficult airway or a less difficult airway. And that has been recently published. Wow, that's really cool. Lastly, natural language processing, which is really the next frontier and many hospitals are starting to invest in this. It's looking at trying to put your hand around relevant information based on unstructured text data. So actually we are at our hospital are piloting using natural language processing in the emergency room to identify patients at risk for domestic violence. And so when those unstructured text data points come up in the electronic medical record, we deploy social workers to the bedside to help those patients. So we're looking at what the satisfaction of the patients are and what their outcomes are. So that's one of many ways that now natural language processing is being used. Yeah, we were talking about this earlier too, that being able to use that same processing whenever you're looking up for diagnostic things like a post-op alias, that's not a discrete field in the EMR. So it makes it very difficult kind of to find, right? Correct. And that's where natural language processing really helps is it doesn't have to be a discrete field anymore. It actually looks at language that is put in a progress note or a pre-anesthetic evaluation and then develops those patterns of recognition to help improve outcomes or therapeutic modalities. I got excited whenever you said that facial recognition, so I had to scooted ahead to Slido. So next question here, how comfortable are you with AI assisting in anesthesia decision-making during procedures? Let's make sure I get on here. And what we find is about 25%, 20% say they're very comfortable and actually that number's going down, somewhat comfortable about 55%. 10% neutral, 10% somewhat uncomfortable and we have 1% of the audience say they're very uncomfortable with this. So I think that's really interesting information. Before you go on, can we get a show of hands how many students in the audience? Oh yeah, yeah. How many SRNAs or resident CRNAs do we have in the room with us today? It's a little bit harder to see. Yeah, about 25%. Fantastic. All right, thanks for letting us know that. So why don't you talk to us a little bit about this? So artificial intelligence in the future. I think the survey would lend itself to what most people think about artificial intelligence in healthcare. And that is, I think the future is bright. However, I think there are things that we need to work out. I think we will see a time where we incorporate medical early warning systems into wearable technology so we can get our patients home faster, safer, but that there has to be systems in place to monitor those patients at home, get the data and help those patients who are not doing well to eliminate or reduce failure to rescue. I also see a point in time where we're gonna use augmented reality, not only for training our next brightest anesthesia professionals, but also to allow for early detection of disease process that happen in the perioperative space. That being said, as all of you have one eye open and closed like we do, there are some liabilities and limitations. And I think we all need to understand what they are. The black box results, the fact that there's a lot of lack of transparency on how these things are actually being made, what is going on behind the scenes. And that lends themselves to bias, which we need to make sure that we account for. We need to make sure these technologies are generalized to a variety of socioeconomic background and racial and ethnic groups so that they're applicable to a broad base population. We also, and Desiree, and I think you're gonna talk about this later, implementation is so important. You can have a fancy shiny object, but if you don't ask the people at the bedside how it works within their daily processes, it's going to fail on arrival. And that's really what we need to work. And then the liability aspect, who's gonna own the data? Who's gonna own the liability when something goes wrong? Is it going to be the patient, the providers, the technology industry companies? How is that gonna be worked out? So there are a lot of questions that still need to be answered before we readily adopt AI in the perioperative space. So Desiree, at that point about the wearables, Medtronic, I'm involved with some wearable technology and it's amazing how many hospital and hospital systems and globally, but more in the US are very interested in wearables from our prehabilitation phase so that you can monitor prehabilitation through the periop phase and the med surge floors for early warning so that nurse workflow changes, patient satisfaction changes, vital sign monitoring changes, and then to go home earlier. So it's very interesting to see and it moved very quickly, I will tell you over the last probably year and a half that wearables have just like this. So if you haven't seen them, you probably will see them in your system at some point. You're going to. Yeah. And it's, they're not ICU monitoring, you always have to keep that in mind, they're not OR monitoring, they're outside of those spaces right now. Yeah, but we'll definitely be seeing those. Well, Patty, tell us more about some of the technology. I thought I should share this. I'm probably the only person in the room that worked before there was a pulse oximeter as standard of care. Because you wanted to show a hand on that one. Is there anybody else besides me? Probably not. And I wouldn't work without one. So I wanted to bring this in because at the time that I started using it, AI was not even a word that you talked about, even though Steve said 1955. So you can see the initial release, I wasn't there in 83, but the initial release of pulse oximetry was like any other monitor. You put a monitor on, you get a number, you believe the number, it works. And obviously there were algorithms, I don't even think that we talked about that at the time, but there were algorithms. But AI has been a part of the pulse oximetry world for the last 30 plus years. And how it works is, in a true pulse oximeter, it identifies the patient's pulse, but then compares that pulse to thousands and thousands of data points to determine that that's a true pulse before it reads that pulse with the wavelengths of light to give you the saturation. So that's key, that's AI, that's all those pulse points, data points in the monitor, and I'm making it very simplistic. Your patient's pulse comes in and identifies it's a true pulse and not a venous, it's an arterial pulse that gives you a saturation and gives you the number. And all of us know, I don't think anybody in the room would work without a pulse oximeter and you trust it. So I wanted to share that because all those years, again, I never thought about AI being in it, in that technology. I think it's a very fair point that a lot of us are using it, cell phones, but we're using that every single day in the operating room in many different technologies, definitely in our space. So recent literature, and I think, Steve, you brought this up about equity in healthcare, and that's so key in industry with industry partners, and that's why, again, part of the APSF, you all talk about this and how to solve it, but equity is key. And in particular, do you wanna cite the research being done out of UCSF? Yeah, particularly, we might not all know this, but there are a variety of pulse oximeters out there, and accuracy is really important, and particularly shed light during the COVID pandemic, where people were getting handheld pulse oximeters at home, and we just really didn't know if those were really accurate enough, and particularly when you look at color, different color in different people. The people out in UCSF really showed that different pulse oximeters actually are more accurate with regards to different skin color, and you could see the spectral array of different colors that if the pulse oximeters are not designed to take this into effect, you won't have a reading that is accurate enough to make a decision at the bedside. So I think it's really important to understand this, and this is all generated by AI by looking at large data sets of folks with similar colors of skin. All right, so next question, I think it's along those same lines. What concerns do you have about integrating AI into anesthesia practice? I'll go over the results here in just a minute for our panel, but one of those things, as you said, was equity and bias, and is that inherent in the data points that we're using, because that is a different patient population, or a different type of population that we're looking at? I will say I'm a little bit surprised, but right now, as it stands, patient safety, about 35% of the group feel that that's their concern. Data privacy and security is about 50%, and you can pick all the different ones you want, but reliability and accuracy of AI algorithms, by far, is the one concern that many of us have. Job displacement for anesthesiologists and CRNAs, we have talked a lot about this in previous conversations, and lack of sound implementation, which I think is a huge piece that many of us have been in plenty of situations where we've been sitting in that OR, and a new rep comes through, they drop a new piece of equipment in, and they give you one day of an in-service. Has anybody had that experience before, where someone just drops something right in your lap and says, oh, here's the day, only half of your providers are here, so we've all been there, haven't we? Yeah, sure. One of the pieces that came out of, and a great lecture that was part of the recent consensus conference last year for the Stolting Conference about AI was that piece of implementation science, and how do we actually get all of this technology, not even the new technology, but how do we get this into the hands of clinicians and anesthesia professionals, and how do we adopt it? So it's one thing to drop it in your lap, but it's another thing to actually utilize the technology. So Dr. Megan Lane-Fawn, and Stephen, if you could tell us a little bit more about her, she is very passionate about implementation in human factors, isn't she? Yep, so Dr. Lane-Fawn is an implementation in human factors scientist, and so really the way of getting technology or an idea, evidence, into clinical practice successfully, and to sustain that change. And so this is really, incorporates a much higher level of science, because it requires a lot of folks coming together and solving a problem. Even though you might have the technology problem solved, there are several other hurdles that you need to overcome in order to really make a successful change. Yeah, she and I are kindred spirits, so in my role in quality, you know, doing quality improvement in patient safety initiatives, it really requires a knowledge of implementation science. So some of the things that get in the way of reaching our goals for care and outcomes is knowledge, access. Some people don't want to hear it, some people don't want to embrace it at all. Some people know the evidence is out there, but we just can't accept it, right? We can't accept that the algorithms are correct, or whatever that may be. So that's really what implementation science is. It's bridging that gap between the evidence that we know is out there, and actually putting it into practice, right? We say translation of evidence. So our next question here is, what words come to mind when you think about implementation science? And AI just failed us on this slide out here today, guys, so sorry about that. I was hoping that would work better. So some of the clinical applications are, I'm sorry, what words come to think of whenever we are thinking of implementation science? We have four votes. Slow, overcoming inertia, QI. Oh, look, it's working. That's great. See, AI read my mind, or heard me at least. Someone was listening, just like Amazon. All right, reference management, application, it's challenging, absolutely. Anxiety, that is huge. I know that this is something I deal with a lot in some of the quality initiatives that I do. Buy-in, red tape, 100% challenging. You know, I was gonna comment today in the world of anesthesia, all the transitions like the traveling, the staff change, the hours, there's so much around the workplace that affects this implementation. Yeah, it really does. And just time, you have to have time. You know, you're so, again, at least where I practice, there's a lot of things on time and turnover, and where do you fit it in? Yeah, I love this word cloud. I mean, I wanna use this in future talks, but the apps, we all have 10 different apps for what we use in the OR. Inconsistencies with the information that's being shared with you. So we know that there are a lot of barriers to overcome, and it is when you can utilize a theoretical approach to care and then actually putting that into how you implement is very important. So again, translating evidence into practice and closing that gap, that's what implementation science is. So, and sales, so Megan Lane Fall, these are her slides, and I will have to give a big shout out to Lynn Reed, too. These were gonna be her slides, so I'm filling in for her. But Megan said that this was someone who was very inspiring to her, Dr. Ann Sales, and she said, the implementation science is a study of human behavior change under organizational constraints. Stephen, what does that mean to you? So changing behavior, which is really one of the hardest things to do, because people, as we all know, we get into our routines, and we like our routines, and someone comes along and says, no, we can do it better, that makes us all feel uncomfortable initially. But we also have to understand, we have constraints. We work under either organization umbrella, government regulations, but resource constraints. So we may have a great idea, but if we don't have the necessary resources from an organizational standpoint, it will never get off the ground. So we have to balance the difficulties and challenges of changing behavior with the resources and the regulations that we all work under in today's world. And that red tape. Red tape. Someone said it there. That's the red tape. So Megan made the point during the conference last year that implementation science is really, call it a mutt, right? It's a little bit of everything coming into play, and you have to look at public health and program evaluations, the psychology and behavioral science of your patient, of all the clinicians, of everyone that you're working with. It's community partnered. It is management science and organizational theory, systems and human factors, engineering, and then really communication, dissemination and marketing are huge. So all of those coming together really make up this science. You know, Desiree, this makes me think about, we didn't talk about this morning, standards of care. Yeah. Right? It's really implementation science along the whole thing, but until standards of care happen, we don't do them routinely, right, in practice. I just think about all the stuff we do, so. I think it's also about culture. Cultural critical. And teamwork and the embracing ideas that can all help our practices improve. You know, that has to be a key in terms of process change and implementation science. If you don't have the culture, you'll never have the buy-in. Yeah, for sure. Now, who should care about implementation? So I think obviously patients, families, caregivers, this is very important to them. The researchers, developers, and innovation of the technology and the drugs, all the different things that we're using. Clinician and allied health professionals, really everyone who is involved in that care touches that patient throughout the perioperative continuum. Of course, payers and health systems, regulators, legislators, it takes everyone. So it's truly difficult, but it can be done. We have seen it work very well in adoption of different initiatives. It's cost and quality. Yeah, absolutely. It's cost and quality. Yeah, cost and quality, that's right. So we left this in kind of as a reference. So key features of implementation science is that there has to be an evidence-based thing to be implemented. Implementation science is done in partnership and collaboration with all those different people we talked about before. The outcomes are sometimes different than traditional effectiveness outcomes. So we measure this in the success of the implementation in different ways. Contextual nuances is informative, not problematic. So it's very important. I think you mentioned this earlier about the human factors and the context by which you're trying to implement something, right, Stephen? Yeah, and I would go back to the first part, evidence. I think AI is exploding, and I think a lot of folks are adopting it before the evidence really suggests it might be actually improving outcomes. So I think we really need to make sure we have the building blocks of stuff that actually works before we even get to the implementation side. And sometimes, because this is this shiny new object, artificial intelligence, I know people are willing to adopt a new technology maybe even before evidence. So I think we have to go back to the basics and make sure that the evidence is there. Yeah, there are some early adopters out there who really like to get their hands on things and play around. You know, there's also the concern about FDA approvals and things like that, and what has been approved, and those are all things you have to consider. We talked this morning, Dr. Yu brought up sensitive, sense, Sedesis. Sedesis. It was way before its time. I don't know who remembers that, but that was where the patient could hold a ball and propofol would stop based on the way they held the little ball. It didn't make it. It was probably way ahead of its time. Think about if it were introduced now versus, say, 20 years ago, when AI is so talked about, if it would be understood better and maybe implemented into practice. Not sure, but it's a question, you know? And then also, implementation science uses very specific language, methods, theories, frameworks, strategies, and I think this is important to realize as well, and what we're learning, those of us who are really interested in how to implement and change behavior and human behavior, you really do have to use a different language. You have to bring a whole different emotional quotient into the mix there. So, Stephen, she showed these pictures I thought was very interesting whenever I was looking at it, because I was like, this is not AI, but Megan was really making the point that we have to figure out contextually how all of this fits together whenever we want to implement something such as AI. Yeah, I mean, this is a perfect example of some pictures in the ICU where bedside input is so important. So, whether it's daily huddles in the ICU looking at what we should be doing for every patient so that the quality and safety is of utmost importance, and strategically placing these billboards or whiteboards or any type of board in places that people work and see them and adopt them. Otherwise, if they're back in a corner, no one's gonna use them. So, going to the bedside, looking at the infusion pumps, where we're actually strategically locating technology that works within our daily routines makes adoption that much more seamless or at least effective. And so, this picture shows you gotta go there, you gotta go to the operating rooms, to the perioperative spaces, and figure out how things work. Non-operating room anesthesia is a perfect example of this. A lot of our equipment is pushed away in a corner and then we're supposed to develop and deliver excellent care. Well, we have to look strategically how we can change that and make it work best for us and our patients. So, this is an example of implementation science. Go to the bedside, figure out what people are doing and what works best for them within their workflow, and then the technology can be implemented around those workflows or within the workflows. Yeah, and I will say, like many of us in the room here have probably been part of whenever we've been implementing something new. And it's important for us to do this, right, and give feedback. But even taking it back further, I think you said too, whenever companies are developing this technology, they need to see this and they need our input to guide what is important to us and what we need. And so, a lot of times we put up a lot of barriers, not only with implementing on the ground, but also our collaboration with different industries about what's important to us and we put up those barriers so they don't always know what we need. And I think we need to consider that as we go forward, especially in the space of AI. Absolutely, the dialogue needs to be open. Yeah, for sure. All right, so what does implementation have to teach us about adopting emerging technologies? One, we have to leverage qualitative data. This is not just about all the hard numbers that come out of whatever you're pulling at. We have to look at qualitative data to make decisions and guide implementation. Again, workflow is paramount from a conceptual idea of an emerging technology all the way through to implementation for the end user. You have to match implementation strategies to context, right, and I always say this whenever we talk about change management. It's like you have to appeal to the stakeholders that you're talking to. So what I'm gonna say to one of our CRNAs who's an early adopter versus what I say to one of our CRNAs or physician colleagues who is not so much of an early adopter, it's gonna be very different. So you have to consider that. But what is really important that you involve all of those people, all of the stakeholders in the conversation. You can't put anybody out to the side or not think about them or not include them in the conversation. That is an automatic barrier to implementation. And then design implementable and usable technologies. Hey, tech companies, are you listening? Yeah, really important. So last Slido question here, would you support the implementation of AI-driven tools in your anesthesia practice if they demonstrated improved patient safety and outcomes? And overwhelmingly, about 73%, 76% now, say they would strongly support the implementation of AI-driven tools if it demonstrated improved outcomes and safety. No one opposes or strongly opposes it. We have a couple, 2% that says neutral. So very interesting information. I think that's okay. I actually think that's okay. Yeah. Our minds are open to the new technologies. We are, and I think so. All right, so in closing, again, we really wanted to take, it's a very broad overlook. I mean, I've had hours of conversation about AI and technology and utilizing it specifically in anesthesia practice, not just all over healthcare, but we really wanted to focus on what it means to us, how do we implement it, how do we work with it as it is evolving in our own practice and look at the intersection of all those. In closing, Stephen, what would you say about AI and how we're implementing that? I think three, I think it will improve outcomes. I think we will see this inundated in our daily healthcare lives as we are in our daily regular lives. I do think we still need some hurdles and challenges to be overcome. I do think implementation science is extremely important, and particularly in this science of artificial intelligence, I think we need to get people together to figure out how we can make the technology work for our patients and for us. And I think that that's really, really the most important thing, because if we don't couple the technology with the people who are working, there will never be success. And then, Patty, a question for you. I mean, you've seen all kinds of technology over, you've self-proclaimed that it's true. I'm not just saying that. You've seen a lot of technology that was emerging. The pulse oximetry was emerging whenever you were getting out of school. So in your opinion, as someone who's been at this for a little bit longer than some of us, SRNAs in the room, what advice would you give to the CRNAs here today about how do we adopt and how do we embrace these emerging technologies that sometimes are a black box and kind of scary, but end up changing our entire world and care of our patients? You were so kind how you phrased that, so kind. So I would say, and I have been around and seen all of these probably implemented for safety, and I would say, again, that APSF led a lot of this. I think you have to be open. I think as CRNAs, you really have to be involved. Many of us just go into our work and don't have time to be involved, but I think as an organization, AANA should have you involved with industry and people that are developing technologies and AI because they need your voice. And also to be open to, I know when sometimes when I'm doing cases and a company comes in to do it with a new product, you don't want, you're so busy, you don't take the time. I think engage with them because they learn as much from you as you learn from them. And part of the work that I do is bring clinicians into industry because I feel it's very valuable that industry leaders hear the voice, and you are the voice. So I would encourage that very much for you. It's rewarding. You'll find out you do make a difference. You'll find out how smart you are. Coming back, you'll find out the technologies then change your world a little bit, and most importantly, change the patient's world and the focus on safety and quality. All right. So we talked a lot about the emerging medical technologies and AI and patient safety last year at the Consensus Conference. Tell us a little bit more about the upcoming conference here in September. Yeah, so this marks our 40th year anniversary of Consensus Conferences built through the APSF. And I really wanna emphasize the inclusiveness of these conferences. You're all welcome. CRNAs have actually time and time again created these conferences and chaired the conferences, and they have been extraordinarily successful. The one this upcoming September, which is just about a month away, focuses on opioid-related harm and medication safety. We've actually had two other Consensus Conferences in the last 20 years on this topic. So that tells you that we still haven't fixed the problem. So we're going to take another stab at it, so to speak, and hope you can join us virtually or otherwise in Boston. And in the future, in 2025, we'll be focusing on maternal health and obstetric safety in Chicago. And that will obviously be coupling with some of the artificial intelligence discussions that we had today in that domain. So a lot of really interesting conference stuff, and hope you can, again, all be involved with the APSF. So, Desiree, just one comment. All of these conferences are on the website, APSF.org. You can go in and search and download and get the information from all the different conferences that have been held. And it is a free conference, by the way. There's no tuition that's charged for these conferences. So if you can, try. We need the input from CRNAs. And it is a great opportunity for your voice to be heard. And that's what I love about the Stolten Conferences. No matter who you are, where you hail from, what your title is, your voice will be heard. And it is an opportunity for us to actually impact patient safety, which I think is absolutely phenomenal. So, with that being said, we want to thank everybody today. Don't leave. I just love to thank everybody profusely. But we do want to take some questions. So please, if you haven't already, and I haven't checked in the last couple minutes, I'm going to check the Q&A feed on the AANA app. Ask questions there. The microphone is yours to have. Have your voice be heard today, if you'd like. A couple places you can find APSF. Steven, why don't you walk us through those? Yep, so we have podcasts. We're on all the social media outlets. You can look at our channels. Our website is www.apsf.org. So there are a variety of ways you can get to our organizational activities. And again, the newsletter is free, too. So you can actually subscribe through our website, and you'll get the quarterly newsletter for absolutely no money whatsoever. And it's a great opportunity to figure out what's trendy in patient safety and perioperative care. Who here receives the APSF newsletter right now? Uh-oh, raise those hands high. Oh, only about 50% of people. So you can give it. We have some work to do. We do, so please, go on the website, subscribe, and then also tell your friends and your colleagues at work to subscribe as well. And you can have patients. We have patient-facing material as well. So all kinds of great information there. So please do subscribe. All right, we've got some questions in. And take it back to your schools, those that are students or educators here. Please take it back to the schools. It's great for those in training. Yeah, all right, a couple questions. So first of all, Lynn Reed, we miss you, we love you. Thank you so much for all your contribution to the APSF, all the work that you have done here in the AANA as well. So we love you, take care. Taylor Merritt, AI algorithms are only as good as the data that is used in their development. How do we ensure that equitable data sets, there are equitable data sets that are representative of the entire population, not just those who consent to data use? I think we talked about health equity. Yeah, I think that's a hugely important discussion and comment you made. And again, that's one of the inherent limitations. It's only as good as the people making the algorithms and opening that wide net to a variety of different groups that we take care of. And so there has to be some checks and balances with some of these things. And that's why these data and these algorithms have to be transparent to the consumer and the people using them, i.e. all of us. I can address that in my work with Medtronic. We address equity, we're focused on it since COVID, as you brought out, the patients were being sent out of the ICU, not knowing if SATs were accurate or not accurate. So it's very, that came back to all device companies. So we are doing that. We look at the scales, but the independent research is the critical research, like the UCSF research. That's the kind of research you wanna focus on to see what products you're using and how they work and what the independent researchers are saying as far as accuracy, it's critical. Brandon Sully, the website says in-person registration is closed for the conference in 2024. So in-person fills up pretty quickly. We have about 150 total people who can come in person. However, if you're interested in doing virtual, that's really open to a much larger scope and group of people. So yeah, it's first sort of first come first serve with the in-person, so sorry about that. But certainly you can get the virtual registration. Thanks for letting us know, Brandon. We'll make sure to talk more about that. Jose Castillo, where do you think we are at? Where do y'all, I love that. Where do y'all think we're at with the use of AI in the prehabilitation arena? We're behind in this space. Are we behind in the space of patient optimization? Pre what? Prehabilitation. Prehabilitation, I think you've mentioned that just. Yeah, I think the wearables are prehabilitation. I don't think we're there. I would guess probably a 0.00% of patients monitor the prehab part, but Steve made it different. No, I mean, the prehabilitation era has also started about a decade ago. And however, things are still needing to be ramped up. There are a variety of, as Patty said, wearables to increase mobility. There are initiatives of playing word games preoperatively. But again, the artificial intelligence aspect of it still needs validation, and it's not yet totally there for prime time. Yeah, all right. And our last question, I know the last session went over a little bit, and our clock was set for an hour, so I think we're a little bit over. But our anesthesia program, this is Alexa Altana. Our anesthesia program started incorporating virtual reality. Have you heard of other programs using this to improve care, any benefits or challenges? I've heard quite a bit of utilizing virtual reality. We can ask Maria, actually. She might be able to answer that question. So Maria Van Pelt, Tiffany, do you know much about virtual reality? In the anesthesia programs today. Yeah, it's definitely being used for intubations and things like that. Here's the mic. Yeah, so lots of it being used. I think you'll probably see a lot more of that. It's great for simulation. And it allows flexibility for the learning experience. So people don't have to necessarily go to a simulation center to get that virtual experience. They can actually do this at home with their own technology and learn on their own time, which I think with all of us worrying about our over workload and burnout, virtual reality training might be helpful to that in that aspect. Yeah, I think it can be good and bad. Yeah. So again, thank you everyone for joining us today on this great conversation by AI. This is evolved. The conversation is evolving. I'm sure we'll be here again and we'll see a lot more of this at the AA&A in future years. So get involved, participate, get your voice out there and be heard. And lastly, all of the questions were brought to you by chat GPT. All right. All right, thanks guys. Thank you everybody.

AI in anesthesia

patient safety

clinical decision-making

post-induction hypotension

deep learning

implementation science

stakeholder involvement

CRNA community

healthcare technology