Good morning. How are you all this morning? Great. Awesome. I'm Hilda Nugent with the Professional Development Committee, and I have a couple of housekeeping things before we get started. First of all, make sure that you have put all your evaluations in the meeting app, and then you will be able to submit your final evaluation for your continuing education credit. Also, if you would like to speak at next year's conference, we have a call out. It will open August 14th for abstracts. You can find all kind of information on the AANO website about abstracts and abstract submissions, so we'd love to have your work and have you to present at a session. And then, it is my pleasure to introduce our speakers, Dr. Kevin Griffith and Julia Harris. Dr. Griffith is an Assistant Professor in the Department of Health Policy and an Investigator at the Partner Evidence-Based Policy Resource Center at the VA Boston Healthcare System. Ms. Harris is the owner of Vaporworks Nursing Anesthesia and is completing her PhD, yay, Studies in Health Policy at the University of California, San Francisco. Please join me in welcoming our speakers as they present, The Effects of Anesthesia Availability and CRNA Scope of Practice on Veteran Access to Surgery. Welcome. Thank you for that introduction. All right. We have slides. That's great. That's mine. Sorry. All right. I'm Kevin Griffith. I am a doctor, but I'm not an MD anesthesiologist. PhD is Latin for a doctor who cannot help you. I'm here with Julia Harris, who's a rock star PhD student at UC San Francisco. We got the pep crowd here. And then, we also have two co-authors who are not, one's present, one's not. We have Jeff Darna, who's a CRNA and faculty at University of Southern California. And then, we have Dr. Hillary Mull, who's a PhD out of Boston University, and she's also a practicing VA health services researcher. So, we're excited to share this work with you today. I'm going to let Julia get it started on some of the background, and I'll come up and talk about the nerdy method stuff. See you in a little bit. All right. Good morning, everyone. I think I'm mic'd up. All right. Let's do it. I'm so excited to share this work that we've been working on. We, as in mostly Kevin, have been working on in the last couple of years. So, my job is to just give you the background in terms of what the changing landscape is in anesthesia practice in the United States, and why it matters to the population of the United States, but specifically for the veterans that we care for in the Veterans Health Care Administration. All right. So, I think this was Kevin. Anyway, we were funded by the American Association of Nurse Anesthesiology to conduct this research. So, disclosure. We talked about these guys, and our research that we did was very large. It started with 2.7 million observations, but by the time we got rid of some of the cases that were not reflective in all types of practice, meaning only one type of practitioner did those, and we couldn't really compare it to a different type of practitioner, we removed those from our analysis, because our analysis was primarily interested in seeing differences between the provider types. So, we ended up with 1.7 million observations, which is very significant, many of which were reviewed by nurses that go do chart review to an extent that is not seen in private practice. So, this is very robust information. All right. So, my point. Trends in surgery are changing rapidly in the United States. So, from 2007 to 2024, where we are now, the population-based increase in surgery is 16%, which is huge. And then when you take into account the growth in population, it's increased by 30%. And we have not kept up with the number of providers in order to meet that need, and I'll show you in subsequent slides. One of the reasons for this is the aging population, right? So, we have a, you know, steady increase growth in total population, but the aging population is expanding remarkably, as we've been talking about for the last 20 years. We have the silver tsunami. So, it's the baby boomers that are aging into the golden years, and when you get to the golden years, you need to see more healthcare practitioners, and you need more surgery. So, that is increasing the demand of both surgeons and anesthesia providers. The other thing that's happening is there's been a shift to non-operating room anesthesia care. It's estimated that 30% of the anesthetics delivered currently are non-operating room procedures, and it's going to expand to 50% of the total anesthesia work that we do. So, it's huge. And that also is related to the aging population in that previously, older people, certain older people were too frail to undergo procedures, like an open heart surgery to replace a valve. Now we can do that with a non-invasive modality, so the number of things that we're doing to people is rapidly increasing, so we need more providers. Why are there not enough providers? Well, let's start with the physician anesthesiologist. The main thing is that their pipeline was capped in the late 90s, and that had to do with the supply was outpacing demand at that time. So, there's some economic incentives that I won't go into, it's beyond the scope of this lecture, but the AMA was not happy about a surplus of physicians, because what happens when you have a surplus of physicians, you have increase in competition and decrease in reimbursement. So, in the late 1990s, there was a cap place on the number of training slots or residency slots that Medicare would pay for, and that stayed steady up until 2021, where they had more slots increased, however, the population didn't stop growing. So the population is definitely outpacing the supply of physician anesthesiologists. That's like the main thing. The other thing is, physician anesthesiologists are aging out. So about 60% of them are age 55 and older, almost 60% of them. And the mean retirement age of physician anesthesiologists is 62. And the closer you get to retirement, what commonly happens is they decrease work hours. So when they're decreasing work hours and there's not enough providers, that puts additional workload on the existing providers that are there. And that contributes to the burnout that we're seeing in physician anesthesiologists that started primarily with the public health emergency. And that has just continued to get worse as more and more providers are exiting the profession. So my poor eyes, let me goodness. Okay, so we're currently the HRSA, which is like a national workforce evaluator, let's just say. So when you need funding for graduate programs or whatever, you go to the HRSA website, you bring that to Congress to say, please pay for more of our education. Like they're the standard sort of for the United States. They are saying that the physician anesthesiologist workforce is currently 4% under what it should be currently. And by 2036, it's going to be under by 10%. So if you're thinking we have a pinch now, it's going to get two and a half times worse by 2036. So luckily, there's other types of anesthesia providers out there to meet that need. And we'll get into that longer in a little bit. But the main point is that we cannot keep our models the same as they are now because there's just not enough providers. So there's lots of reasons for wanting to change or arguments in favor of changing provider models. There's just not enough providers to keep it the way that we have been doing things in the past. But it's the main thing. So this just goes into a little more detail about our provider status. So there's multiple estimations of numbers of providers, depending on which website you look at. If you look at the American Academy of Medical Colleges, they say one thing. If you look at the Bureau of Labor Specifics, they say another. If you go look at HRSA, it's another. But these are the best guesses that I can come up with based on what's out there. So the number of bodies, right, of nurse anesthesiologists is about 51,000 in the United States currently. And there's about 43,000 physician anesthesiologists. Now this second line has to do with full-time equivalents. So that's 40-hour workweek. So if somebody is working 60 hours a week, they are 1.5 FTE. So that's why there's a difference between the actual number of bodies here and then the actual FTEs. The advantage of this is that they do give projections, which is really helpful. So currently, nurse anesthesiologists are still outnumbering physician anesthesiologists. But as you can see, physician anesthesiologists tend to work more overtime than nurse anesthesiologists, although you can definitely see both provider types are working over 40 hours a week on average, right? So 2036, the projections are that nurse anesthesiologists will be about 71,000 FTEs and physician anesthesiologists will be about the same as what it was. Because even though they increased the number of DME slots or residency slots, it takes a really long time for physicians to get through the program to be fully trained and to help with the workforce. And this, by the way, as far as I know, does not account for the 20-odd schools that are opening in the next year or so. So our workforce numbers are going to really be a strong argument for needing to make regulatory changes so that we can do the work that we were trained to do because we will be the providers available to do so. All right. So the other thing that you can look at in terms of gauging the market need is job postings. So Gas Works is like the national standard for gauging the need. And so if you take the number of missing slots or the number of advertisements and you compare that with the proportion of providers, you can get kind of a ballpark, back of the envelope sort of estimation of what the needs are in the country. So I did that. I took that on my own. So take this with a grain of salt. So the need basically for nurse anesthesiologists is 5.3, and the need currently for physician anesthesiologists is 4.8%, although HRSA says there's more of a need for physician anesthesiologists, and they call it a surplus of nurse anesthesiologists. And the reason for that is that they calculate physician anesthesiologists' need based on a ratio or proportion to surgeons. But they don't do the same for nurse anesthesiologists. They base our numbers off of the number of physician anesthesiologists, which is inherently long because clearly nurse anesthesiologists are being used in a different way than they have been previously. So this indicates that there's a greater demand for nurse anesthesiologists either working independently or working in more loose, restrictive settings. So that's the point of all that stuff. So this is a map, and I'm sorry, it's a little bit old. It's from 2010, but there's not a lot of good publications about anesthesia needs across the United States. But the main thing that I want to keep pointing to on these maps, and I have several because I'm a visual person, and I hope that helps some of you kind of follow along, is this section here. So the dark purplish colors are where there's the greatest need, or at least there was in 2010. I'm sorry we don't have newer stuff. But keep an eye on kind of this area, because this is chronically underserved area, and there has some policy implications for that. Okay, and so this is a map from the Bureau of Labor Statistics, and again, it's not perfect because not everybody that works in healthcare is an employee. So the recordings are not 100% reflective of what's actually happening. But it's what we have, so there you have it. So this is a map of nurse anesthesiologists. And basically, like the darker color there is, the more providers there are. So as you can see, nurse anesthesiologists are represented throughout the country, and especially in areas where there's not other providers, because we work in underserved areas. We work in rural settings, and we work in underserved populations. And this will keep coming up. So if you look at where the physician anesthesiologists are, or are not in this place, they're not here. And this is a really important region. And they're not here. They tend to follow where the wages, right? And so the wages are higher, right, on the coastal regions and where people want to live. There's big facilities and that kind of thing. So that comes into play when we look at this map. This map is a map of migration, basically, of veterans. And they are moving from the golden area, no pun intended, although they are a little senior, tend to be, the ones that access the Veterans Health Care Administration. So they're moving from these regions to the blue regions. And where are they moving? They're moving to this area, where there's a need and where it's affordable to live. So that's where the veterans are. Where are nurse anesthesiologists? They're in this region. And they're able to help if regulations permit. All right. So current challenges, how many of you have heard of operating rooms closing because they don't have enough anesthesia staff? Yeah. Thank you. It's run rapid. I've heard that in my workplace, where I live. And I've heard that to a greater extent in this meeting. And I see that it's reflective of the people that are showing up here this morning. And thank you so much for being here on the last day. So what we're doing is not working. And we need more providers to be able to take care of patients. So one option is to uncouple physician anesthesiologists from supervising nurse anesthesiologists to free them up to take care of people, to be hands-on, deliver anesthesia. And that will increase the anesthesia workforce. And if you increase the anesthesia workforce, you'll be able to do more surgeries. So it's been a big problem on the ground, wherever you are. So we're going to have to change things. But we can't change things if the regulations do not permit. So this is a map, thank you from the ANA, of where there's no supervision requirements in the Board of Nursing and that kind of thing. So all the green is where there's no supervision requirements. However, in the gray areas, and look where they are, this is a very hot topic area, there's a lot of restrictions for nurse anesthesia practice there. And if you have restrictions there, you cannot opt out. Because one of the requirements to opt out is that there be no supervision requirements in the state regs, that kind of thing. So this is a map that I made of opt out. So most of you are probably familiar with the Medicare Part A reimbursement regulation that has been there since I don't know when. But it became contentious around the year 2000, where Clinton got rid of the supervision requirement for facilities to be reimbursed if a nurse anesthesiologist was working independently. So it's not necessarily illegal or wrong for a nurse anesthesiologist to work independently according to this regulation. But if you do, and you don't have any permissions otherwise, the hospital won't be reimbursed for their surgical services, which is, by and large, the largest reimbursement part of the whole surgical process. So it's prohibitive. So anyway, Bush came in and reneged what Clinton had done, which was to get rid of the supervision requirement and deferred it to the state. So every state has the option to opt out of this requirement in order for facilities to be reimbursed. So the progress of the opt out has gone over time as such as the map. I didn't say that very well, but let me just show you because I'm better visual. So early on, the earliest states to adopt opt out are the ones in lighter color. So as you can see on the map, the lighter colors are really the rural areas, which make sense. And as the states become more purplish, it's been more recent. So you can kind of see how the neighboring states, once it got started, the neighboring states of those states started to adopt the practice of opt out. And again, as you can see, this area of the United States is still really lagging behind. And unfortunately, that's where the greatest need is in our country, as far as I can tell. And this is a total nerdy thing that I did for my theory paper. Poor Kevin is going to have to read this. But anyway, if you guys are familiar with the Rogers diffusion of innovation theory, it goes that there's some really odd people like Tom Berrybold and I with opioid free anesthesia. Nobody else was doing it. I thought we were absolutely nuts. But we went out there and started talking about it. And then slowly, some other people thought, well, their patients don't do too bad, right? Let me try it. And then they try it, and they like it. And they talk to their friends, and so on and so forth. So it really follows a pattern that is in line with the characteristics of the people that adopt, like how willing they are to take risks and that kind of thing. Well, that can also be said of states. So this is kind of a visual representation of that. So the first, oh, the colors are gone, but that's OK. The first state to opt out really kind of took a risk. And they were the innovators. And then the other states are the early adopters. And then these are the early majority, and so on. But it's predicted, according to that theory, that more and more states will continue to adopt until it's 100%. And typically, when it gets to this point, it's a pretty steep upward trend, right? And we had a few things happen in between. We had the COVID-19. Then we had surgeries resumed, and there was a huge backlog. And now we have a great shortage of providers. So these are the other factors that are going to be influencing health policy here. Now how this pertains to the Veterans Health Care Administration, the VHA is huge. It's ginormous. It covers the expanse of the United States. It has like 171 facilities. It has approximately 1,100 nurse anesthesiologists in it. So it really can set the precedence for other systems that are interested in seeing how it can be done on this large scale, let's just say. So we kind of put a focus on this center because of various changes that happened that started in 2014. So there was this Phoenix scandal, wait time scandal. And this really came to light in that the center, the Veterans Health Care Administration in Phoenix, was really trying to make their mark in terms of meeting their quota so that they can show that they look good. Because the way they had their staffing and everything was almost impossible to meet those marks. So they were running kind of behind. So they gamed the wait times. And a whistleblower came out and called them on it and said that there were even patients dying while waiting. So this garnered national attention. It was in the media. There was an independent investigation. Congress passed legislation so that veterans can go outside of the Veterans Health Care Administration and get paid on the cost of the Veterans Health Care Administration. And that has gotten even more liberal, let's say, as the year has gone on. But again, that's beyond the scope of our talk. So this incident really got us to where we are now. So they had an independent investigation that told them to do three things. Hire more physicians so you can see more patients so that we'll get rid of this backlog. Organize independent practice for all APRNs, Advanced Practice Registered Nurses. And then go into virtual care and meet veterans where they are. And this will help with the backlog of cases. Seems simple enough, doesn't it? Well, how can the Veterans Health Care Administration do that? How can they just make a broad statement that a provider can practice to the full extent of their training and education in any state, regardless of state law, was because of this federal supremacy clause, which is part of the Constitution. It basically says that federal regulations trump states, no pun intended, so that whatever the federal government says, that will supersede state law. So currently, it has been, and the way it was reinforced during COVID times was that if you had a license in one state, say I'm in California, and I go to, let's see, Missouri, and I have a California license as a CRNA, and I go work in Missouri in the Veterans Health Care Administration, I don't need to get another CRNA license in Missouri. I can work under my California license. And technically, I could work a full practice authority in Missouri, even though Missouri state laws are restrictive, because my state license allows me to, technically. However, the bylaws at the facility would need to allow for that. So just to give you a preface on what federal supremacy is, and the proposal with this directive was that not only would they grant full practice authority to all APRNs, but they would also mandate that all the facilities change their bylaws to allow for APRNs to practice independently, which also has been a huge deal. All right, so we had some barriers. The AMA and the ASA, so the Physician Anesthesiologist and the American Medical Association, were really upset about this. So they say it's not safe. We went to school longer. We paid a lot more educational bills. We should keep the models the same as they are. Then the ANA said quite the opposite. There's tons of studies that suggest that we are equally safe and we're cost effective. And we are everywhere in the United States, especially in areas that are underserved. So what literature do we have to back it up? Well, there's mixed literature, depending on who funded the study, basically. So all these upper studies, basically, were all funded by the American Association of Nurse Anesthesiologists. And I will say that, by and large, there are PhD-prepared people doing this that some of them are CRNAs and some of them are not. And the other ones that show that there weren't as good outcomes were all published by Physician Anesthesiologists. However, there are systematic reviews, two of them, that show no difference. There's no superiority of one provider over another provider. So even though there's mixed evidence, we're not stacking up bodies behind the facilities where we work independently. All right. So what happened? So because of all the pushback, and I'm telling you, a lot of pushback, they received more comments about this proposal than they had for any other proposal ever combined, as far as Rochelle says. Massive effort. And it wasn't just against APRNs, by and large. It was targeted specifically for nurse anesthesiologists. So they excluded us from that ruling. So all the other APRNs practice a full practice authority within the Veterans Health Care Administration, except for nurse anesthesiologists. And their rationale was not that we weren't safe. They said we were great, that we were safe. But there's no problem. There's no access problem with anesthesia services. And that's not what CRNAs or nurse anesthesiologists in the Veterans Health Care Administration were saying. That's not what the OIG report's saying. They're saying there are shortages across multiple settings. There was even an independent investigation at the very own Phoenix Hospital where this all started. And they showed there was deficiency in anesthesia related to an over-supervision model, overly restrictive model. Anyway, so then the public health emergency happened. And we had another directive to basically encourage facilities to utilize CRNAs or nurse anesthesiologists to the fullest extent of their education and training. However, they did not invoke the Federal Supremacy Clause. And they did not mandate facilities to change their bylaws. So here we are. So we're going to take a look under the hood, basically, and see if there were any changes that did happen as a result of the public health emergency and the directive that was ensued. And then we're going to see if there were changes. Were there any differences in outcomes, basically? Were there any harm to patients if more nurse anesthesiologists provided care independently versus their either supervision model or a physician practicing independently? So essentially, that's what it is. Kevin, we'll take over from here. Thank you, Julia. And now as the resident statistics nerd, I come in to take over. I'm a health economist by training. And so I looked at this, which I figured this was going to be a conflict zone when I came into this study. But I had no idea how political it would be. But our goals were really simple. We wanted to look at what were the practice patterns for anesthesia care within the VA as this large influential provider that is kind of seen as potentially a big domino for other federal state health system policies. We wanted to look at whether supervision requirements were associated with patient safety. So if you have an MD doing the anesthesia independently versus CRNA versus team care, whether CRNA is supervised, does it make any difference for patient safety outcomes? We also wanted to look at the supervision ratio. So perhaps if you have a supervised surgery, does it make a difference if it's 1 to 1, 2 to 1, et cetera? And then we also were curious what happened during the public health emergency. So health economists, we never like to waste a good crisis. So you have these changes in practice patterns that happen. And they kind of come in exogenously. It's not someone making a decision based upon what they see on the ground with what's going on with anesthesia care. COVID comes in, it changes practice patterns. And so we also leverage that. We say, is there a difference if your facility changed their anesthesia care patterns during the public health emergency, did that have anything to do with patient outcomes? And so why is the VA a really good setting for this? So one, it's, as Julia mentioned, there's a big national footprint. But also, there's a lot of variation. So even when there are bylaws, they're maybe not followed exactly. So you see a lot of variation in the anesthesia care models and in the supervision ratios used, both within a given facility, and then especially between facilities over time. We also have really detailed medical records. So some of the literature that Julia mentioned, one of the big flaws in it is it uses claims data. So claims is very sort of crude for looking at surgical or post-operative complications. It also doesn't have detailed staffing. You're relying on code modifiers to sort of see who was in the room for doing the anesthesia, were they supervised or not. Within the VA, we have their medical records. We have this large-scale chart review. The VA chart reviews tens of thousands per year. And we also have precise surgical staffing data. So in the VA, I can see everyone who is in the room. I see who was the principal anesthesia provider, secondary, supervising. And there's no coding. There's no incentive to game this. Providers in the VA aren't paid on volume. So they aren't paid more if it's supervised or not. So surgical nurses check in who is in these different roles. And I see that as well for every single surgical case in the VA. And so we have mentioned this VA corporate data warehouse. Awesome. National emergency electronic medical records. I have all the details on every veteran, including all their enrollment data, their demographic data. I see the credentials for every single person involved in the surgical case. For the few people I'm missing their credential, these are usually contract providers within the VA. Use this data set from the Centers for Medicare and Medicaid Services as additional credential information. And then this is the one that's really cool. It's called VASQIP. You know, the VA, it's a militarized organization. We love our acronyms. So we have this large-scale chart review of about almost 500,000 chart reviews, which is insane to me. As a researcher, I'm usually working with 200 chart reviews. We have hundreds of thousands for this study, so it's a pretty amazing resource. And we also see a standardized set of post-operative complications in VASQIP. So this is also really neat that the surgical nurses are going through and they're just coding up if they see anything in the notes or a standardized set across every single surgery in the VA, every single facility. And so we started out looking at 2016 to 2023. We have about 2.6 million surgeries in our data set. We removed, you know, rare types of cases that could complicate our comparisons. For example, cases that were classified as emergency, you know, ASA 5-6, that were either like very low or very high surgical complexity, right? There's no anesthesia base units around. We also removed cases that were, like as Julie mentioned, weren't observed in multiple provider types or multiple care models. For example, I didn't wanna have cases or procedures that were only observed in the VA supervised or only observed in the VA independent. I wanted to focus on the types of procedures that were seen in all the different settings so that I can compare the outcomes between them. So when we did those removals, we ended up with about 1.7 million surgical cases in our data and about 460,000 chart reviews even after those removals. Amazing resource in the VA. And we were interested in two things. So you're, what I call the anesthesia care model, right? So the baseline here is gonna be MD anesthesiologists working independently. And I wanna see, does it make a difference if they were supervising a CRNA or if the CRNA performed the anesthesia independently? And also the supervision ratio. So for that subset of surgeries in the VA that were said supervised anesthesia, I could identify, I have very detailed timestamp data for every surgical case so I could see what's the overlap. How many concurrent surgeries was that MD supervising at the same time? And so I create, we created these supervision ratios. We didn't have to rely on coding. So we could see is it one to one, two to one, three to one, et cetera. And then we identified 14 post-op outcomes that we were looking at. And so some of these we think are more sensitive to anesthesia care. Some we think are less sensitive. Example, going to this, I really didn't believe we would see a difference in mortality. I just didn't believe that that was going to, that CRNAs were killing people. But because people looked at it previously, we wanted to include it. Because if we include it, then we could perhaps either confirm or refute things that people have found in prior studies. And there's others, I mentioned, I'll come back to this at the end, but there's other surgical companies we might have been interested in, certain anesthesia complications, especially more minor ones, but they're just not in this standardized set of complications that's in BASCWP. So this is the set that we have. It's not every single complication that you might want. And then we had two key comparisons. So one, we looked at facility level. So I could say, okay, at a given facility for this specialty in this month, what were the anesthesia care models they were using, what proportion of their cases were done by the CRNA independent, the CRNA supervised, MD independent. And then look at changes within facilities over time. So again, I'm at Vanderbilt, so look at the Nashville VA. What's the Nashville, if it actually sort of fluctuates from month to month, are those fluctuations associated with changes in patient safety? And then once we do the facility level look, we move down to the individual surgery level, and I look at, okay, the staffing for that particular surgical case, the anesthesia case, doesn't matter whether one person did the anesthesia or not, or another. But we have this big problem. So case assignment in the VA and probably most systems is non-random. So this chart shows anesthesia base units on the horizontal axis, and the proportion of cases in the VA that are handled by red is supervised, blue is MD independent practice for the anesthesia, the golden rod is CRNA independent. See, the dominant paradigm in the VA is supervised surgeries but you do see this sort of trend where at higher base units, it's less common for the CRNAs to do it independently, more common for the MDs to do it independently. So if we just sort of naively study this, that might be a knock on some of the prior literature, we see a potential pro-CRNA bias. Does that mean we think if we're looking, if we're just comparing outcomes, well if CRNAs working independently are doing things with less base units, we might find a positive benefit. But on the flip side, supervision ratio goes the other way. So the supervision ratios tend to be higher, so less intensive supervision within the VA when you have higher base units. So if we just look at that naively, that would bias the analysis against CRNAs. And this makes sense because where you have higher supervision ratios, these tend to be the VA facilities that are larger, handle more complex cases, and they just hire a lot more CRNAs. And so we have to control for that. And so we try to control for a wide variety of things. So we have patient characteristics, I mentioned I have all that VA enrollment data, so I have the patient's age, sex, race, ethnicity, all their comorbidities and their pre-op workup. For the care setting, I actually control for the facility you're at, the anesthesia provider you had, the surgeon who treated you. So I'm really controlling for provider quality. So there's, think about a regression model, this has tens of thousands of different variables because every single provider and their quality is controlled for. And then we have procedural complexity. So we have the ASA class in our models, we control for base units, the surgical complexity category, mean work RBU, CPT code, month and year. So really everything that we think might be associated with these quality outcomes we put in our models. And some might say that might not be enough. So I'm not going into the sticky details of this, but one cool thing we also do is what's called instrumental variables. So you wanna know, we wanna know the likelihood of post-operative complications and whether it's associated with the anesthesia provider type. And I mentioned before, we're trying to control for a bunch of other things. So we control for patient risk, we control for procedural complexity and the quality of the surgeon and the anesthesia provider. But the case assignment's non-random. So someone might say, well, OR or VA facilities are looking at the risk and they're choosing to assign cases to certain provider types based upon the likelihood of complications. And thus your analysis is worthless. But what we do is we say, well, anesthesia staffing patterns within this, so the larger patterns within that facility, what were they running with on that day or that month? That affects your likelihood of getting a certain provider type, but it doesn't directly affect your likelihood of post-op complications. So it's just another, sounds like statistical voodoo, but we do this to try to reduce that bias. So we do the look where we look at, okay, just your normal regression analysis, does the provider type affect outcomes? Then we also do this instrumental variables, which is a really complex econometric method to say, okay, we're not even looking at the care you received. We're looking at what was the care being given at that facility on that day? Was that associated with your outcomes? It's just another way of looking at it. All right, so getting into some of the results. So these are the trends in anesthesia care over time, going from 2016 to 2023. These are the number of surgeries. So the purple line is where CRNA is leading the anesthesia, so it's supervised by a physician. The, I'll call it teal at the bottom is CRNA independent, and then the yellow is physician only. So of course, they all kind of, something weird happened in about March 2020. I'm not exactly sure what, but there was a big drop-off, but it didn't really affect the actual practice patterns. So this is the same chart, but now it's proportions on the vertical axis instead of frequencies. And you see, if I add up these, it's kind of hard to tell, but if I add up the top two lines, where the CRNA is supervised and the CRNA independent line, it increased by a few percent during the, like after the public health emergency was declared, but it wasn't a large change. So the VA, basically that directive that Julia mentioned from the VA that during the public health emergency, you should really embrace full scope of practice and use, it's all hands on deck, use people as best you can. Not a whole lot happened in the VA in response to that directive. And this within the VA, these three histograms show the proportion of cases on the horizontal axis, and then the number of facilities in that bin. So for example, in the top left, yeah, audience left, not stage left, we see most cases, most facilities have a small number being done by a physician independently, but there's this long tail where some or almost all the cases are being done by a physician independently. But most facilities for most cases, it's done in the middle with CRNA being supervised, a large portion of their cases. And there's really just a few VA facilities that do a lot of CRNA independent practice. So it's not none at a lot of facilities, but that first bucket is zero to 10%. So there's a lot of facilities that have, it's very rare where CRNA is operating independently, and there's maybe 20 or less that have a higher number. And some of these are the trends in the supervision ratio over time. So the dominant paradigm in the VA for supervision ratio is one to one, one to two or one to three. I think one to one is most common. And the red line is at the surgery level. So when we look at our data and we say, when you have this surgery, how many other cases is that MD anesthesiologist supervising? And it's usually under two, and it dropped a little post-pandemic. And then for the top level is at the day. So at the clinic day level, how many unique CRNAs did that surgeon supervise just over the entire day? Maybe not all at the same time. They weren't all working at the same time. And that's obviously a little higher, but still not very high within the VA. And we see the pretty wide variation within the VA in terms of supervision ratio. So on the horizontal axis here, this is average supervision ratio for that facility. These are these 171 medical centers. And the vertical axis is the number of facilities. So in the plurality is one to one or close to it, but there is this tail where some will have two to one and as much smaller number actually will reach three to one. It's now for, those are the orders for the main course. These are our main results. And I'll walk you through this for, is there a quality difference between CRNAs and MDs? So let me orient you to this for a bit. So these are regression results. The one side is CRNAs supervised compared to MD independent practice. And then it's CRNA independent practice compared to MDs. The zero line, that means there's no effect. And then the rows are each of our surgical complications. And I mentioned those three models I talked about. We have the estimate is a little dot and then the bar is the confidence interval. That are not, if you were to like a 99% confidence interval margin of error, if the margin of error includes zero, then we say statistically there's no difference. And it should jump out right to you. Almost all these bars overlap zero. There's a few I'll draw your attention to. So for example, so for stroke outcomes in the facility level models only, there actually is, it says CRNAs have fewer stroke outcomes compared to MDs operating independently when they're supervised. There's no difference when CRNAs are supervised. There's a tiny, tiny difference in PACU time. And I'm talking, when I say tiny, I mentioned I have this really large sample. So sometimes it's like using a giant magnifying glass like frying an ant. I can find like really tiny differences as statistically significant. This is a one to three minute difference in PACU time. So clinically meaningless. But that's another thing that did come up as significant. And that also came up looking at, when we looked at independent practice, that was the other one that came up was really just PACU time. So very, very small differences and kind of perhaps not clinically relevant. One thing I should also mention is when you run a lot of statistical tests, sometimes you just get, come up, you get some that come up as significant by chance. So if you're doing the math at home, these are 28 different analyses we're running. You might expect one or two to just come up by chance. So really I look at this and I say, I'm not finding any evidence. In fact, I'm finding consistent evidence there is not quality differences between CRNAs and MDs. The next part is, this is the supervision ratio. So now taking that 1.7 million, looking at just that 60-ish percent of cases that are supervised, looking at those separately. And so again, you see all these estimates are really close to zero. All these confidence intervals, these margins of error include zero. So for most of it, say there's not a different, in fact, the only difference that we found was a very small difference in PACU times. So if you were, had a higher supervision ratio, is that's less intensive supervision rate, you had a slightly higher PACU time of just a couple minutes. So clinically meaningless. None of the other outcomes had any differences. So again, very convincing evidence of equivalence across levels of supervision. And so what happened during the pandemic? So this chart, each dot is a medical center. On the horizontal axis, it's the proportion of cases that were led by a CRNA during the public health emergency. And on the horizontal axis, it's the pre-pandemic proportion. And so we say 60 medical centers increased their proportion of anesthesia cases led by CRNAs during the pandemic. 48 medical centers decreased it. A lot of these changes were really small. You see, if they're close to that 45 degree line, that means it's just a very tiny change. If you're further away from the 45 degree line, it was a bigger change. Some facilities really increased their proportions of CRNAs. But the kicker, again, I'm not gonna show the bars for this because I don't wanna be too redundant, but none of these changes were associated with differences in surgical outcomes. So facilities that really pumped up their numbers of CRNAs made no difference for their rates of surgical complications. Facilities that decreased their number of CRNAs also didn't make any difference for their rates of surgical complications. So our key takeaway number one is the rates of surgical complications are very similar across these anesthesia care models within the VA. So the dominant paradigm in the VA is supervised surgery, but we find no consistent differences between CRNA or MD independent practice, team care. We're getting very equivalent surgical outcomes. Then we also find that these pandemic-related changes and practice patterns did not change rates of surgical complications. Second big takeaway is it's unrelated to supervision intensity as well. So the dominant paradigm in the VA is one-to-one to three-to-one, but lower supervision ratios, more intense supervision by an anesthesiologist was not associated with changes in patient safety. And then similar to the prior slide, changes in supervision during the public health emergency were not associated with change in outcomes as well. So we saw that we went to lower supervision ratios post-pandemic, I showed you. We didn't see a gain in patient safety because of that. Now, a few things I wanna say this we do not show, or at least not yet. We don't say there's no differences between CRNAs and MDs, right? Our data don't include, let's say like minor complications, things like nausea. We don't include all anesthesia complications, just not in VASCWP. But one thing I will say is that the results I've shown you today should sort of temper our expectations for what we would find if we even had those. Because if we, if there's some complication that we were missing, that was really important, and there was a huge difference between CRNAs and MDs, it would show up in some of our larger measures, right? We're not seeing differences in the stuff that really matters. We're not seeing differences in readmissions, mortality, all these other complications. So we're not saying there's no differences. There's not in anything we can observe and not in anything major. And we also don't claim yet, this is planned, so I will be, you guys are coming to my way next year. You guys are in Nashville. So I'll get you guys set up with a, I don't know, what's the Nashville thing? Country music and pedal taverns. But we, these are overall results. So the analyses you've seen today, we're looking at all those 1.7 million, or excuse me, surgical cases. But there's questions like, does it differ by specialty? Does it differ by ASA class? So that's planned. We're gonna look at that, actually, in the next few weeks. And we're gonna see, do these results hold? My expectation, again, is that if there are differences in some of these subgroups, they're going to be small. Because if they were large, we would see it in the overall analysis. So if there are differences, they're not enough to affect that overall result. But that's something we're gonna do. So we're gonna look at ASA 1, 2, 3, 4. Surgical, the VA has a standardized matrix of surgical complexities for every single CPT code. So it doesn't matter if the surgery's more complex. Doesn't matter if your patient has more comorbidities or is older, or there's certain specialties where this seems to matter. We wanna be very thorough and try to leave no stone unturned. Because we know this is going to be pretty heavily criticized and examined. So we wanna be as transparent and thorough as possible. And I'd say one thing we also, we're just looking at safety. So we're not actually looking at workforce here in terms of like, we're showing that there's clinical equipoise. So if you wanted to swap out CRNAs for MDs, if you wanted to run an RCT, this is justification for doing so. But we don't actually do a full cost-effective analysis. We're not looking at productivity of provider types. We're not looking at your salaries. And as Julia mentioned, there's this huge growing demand. So the data suggests both are needed. Both are important. Both bring something to the table. And with that, that's the end of our presentation. I'll take any questions that you have. And I know we have mics here, or if you can project. I'm happy to answer your questions. Craig Adkins, Colorado. Where's your next step? Do you take the research from here? That's a really, so this is part of a multi-year study. I mentioned we're doing a lot of those subgroup analyses. And the reason why we're focused on this really thoroughly is because we're looking at the data. We're doing a lot of those subgroup analyses. And the reason why we're focused on this really thoroughly first is safety is the most important thing. If we don't find equivalence in safety, nothing else we could look at would possibly matter. But our next steps, we're gonna look at access. So we also have really detailed data on all the, we mentioned we have all these surgical cases. We have all the workforce data. So we can look at, does changes in the workforce, the number of FTEs actually matter for how many cases the VA can handle? We've seen from a lot of VA medical centers that we're having a lot of cancellations or delays. There's also a big report in Becker's Healthcare a few weeks ago that also mentioned ORs canceling a lot of surgeries. So we're gonna look at, if you hire more CRNAs, how's that affect your ability, your surgical volume compared to MDs? We're also gonna look at wait times. So we have wait time, I see timestamp data from when you were referred to surgery to when you actually get the surgery. I wanna see, my assumption is, less staffing, fewer CRNAs. You're gonna see longer wait times. We also have data on the VA's community care utilization. So since 2014, since that big Phoenix wait list scandal, the VA has really liberalized veterans' access to the private sector. In 2018, during the Trump administration, this was further liberalized under the Mission Act. And again, to VA, we love our acronyms. So we wanna look at, if you have less providers, less CRNAs, or more restrictive care models, does that perhaps have spillover effects? Is the VA spending more money than sending people out to the private sector because it can't meet the demand in-house? And the last thing we're planning to look at is patient satisfaction. So the VA does a large-scale survey, about 400,000 veterans each year. A much smaller amount of those are actually for surgeries that have anesthesia, but we wanna see, is there any difference in patient satisfaction, depending on who did the anesthesia? Hi, Pam Chambers from anesthesia, from anesthesia, from Arizona. Is it possible to get your slides, because I was wanting to read along as you were going, kind of just to look at it a little more closely. They weren't in the app. Yeah, you shoot me an email, my contact information's there. Okay, I'll email you. My Twitter, it's SuperNerd, it's a stats joke, you know, assume normality, because I'm a giant nerd, but I know exactly what I am, and I'm okay with it. The reason, I mean, they're not online, it's not being recorded for later viewing, and part of that is because this is so controversial, and also because this is a work in progress. One of the things we're concerned about, and we wanna be fully transparent about, is these results are likely to change. We don't think they're gonna be, I certainly don't expect them to be big changes, but you know, those bars might shift just ever so slightly, and we just don't want someone to be like, well, what are you doing, are you doing something with the data, and that's why it's different from this time you presented it before. It's just because it's a work in progress, so it's gonna be evolving, and that's why it's not posted yet. And Michelle, hold on one second. We have a few questions online, so I'll address a couple of them, and then we'll get back to you. So the first one is, did you try to compare quality of care provided by anesthesiologists and care by licensed independent CRNA practitioners in terms of patient safety and access? And yes, we did do that, and I hope that that was made clear that there are no definitive differences in the information that we found to date. And then another question was, what is the purpose of one-to-one? Why not have an MD do their own case? And maybe you can talk to some of the VA providers here that might provide more information beyond my pay grade here. And then we'll go to Michelle, and there's a couple other comments, but. Hi, I'm Michelle Zou. I am a VACRNA. Thank you for doing this. My question is, when would this be published? We need this out there and everywhere. Hopefully, it will be, I mean, this will determine, you know, if I find what the editors of the journals think, what the reviewers think, and how persnickety they are. We think we're going to submit the quality. I mean, we're going to submit the quality part separately and try to get this published sooner rather than later, just because that's, like I mentioned, that's kind of the first thing. So nothing else matters if you don't show clinical equipoise. So I think that'll probably go under review at a peer-reviewed journal within the next six months. And then hopefully, so hopefully by the time you guys are in Nashville, this will be published and out and will be freely available. Well, maybe I will hope next year by mid-year assembly, we can take this with us to the legislator's office and take it to the Hill. Right? Yes, definitely on our radar. We are interested in having as many people know about this as possible once it's published safely. So one other online comment before we move to the next in-person one is that it addresses anesthesiologists' assistance. And in this analysis, we didn't include them in our analysis for safety outcomes. And I didn't address it in the introduction section either. There are anesthesiologists' assistance in the Veterans Health Care Administration. However, to the best of my knowledge, there are not very many at all. The numbers are very low. And I believe that is a product of the low salary they have for anesthesiologists' assistance in the Veterans Health Care Administration. And I can't comment on how many are actually in the data. I just know that I removed them. So... Christy Hoke from Arizona. I just want to thank both of you for your time and effort and work on this. This is something that we've all wanted to look out for a very long time. So kudos to you. My question is, looking at the supervision ratios and we're seeing that CRNAs are being supervised one-to-one, that is an incredible cost. Is that in your plan to review and look at and bring that data out? As the VA is federally regulated, you know that that is going to be huge. I mean, thoughts on that? It'll definitely be in the report. We are not planning to do a full cost-benefit analysis, but we are going to produce some important data for folks. We're going to be able to show what the supervision ratios are. And we're going to be able to show what the FTEs are. And so we'll be able to see... Someone could do kind of a back-of-the-envelope calculation and see how productive are these different provider arrangements. And then if we wanted to make a change, they could perhaps use some of what we're producing as justification. But we're not planning to do an actual cost-effectiveness analysis for what if we swapped out X percent of folks for... Okay. Thank you. And just to add to that, the slide that Kevin showed reflected that about 60 percent, I would say, of cases done in the Veterans Health Care, between 50 and 60 percent are one-to-one over the duration of our study, which is not standard of care in the private sector or anywhere in the United States that cares about keeping their doors open. So that's me only. That does not represent Kevin. So I'll just say that. And also, I will be looking into additional finance information, the financial implication of anesthesia care models in the Veterans Health Care Administration as part of my dissertation work. And that will be coming up in the next year. I'm primarily... Thank you. I'm primarily interested in the surgical productivity as it relates to anesthesia staffing. As you saw in the audience, many, many, many people raised their hands and said that they're having surgeries canceled at their facilities because they don't have enough anesthesia providers. So I'm going to try to bring that to light and show that, you know, based on all the information I'm seeing, you know, there's a huge anesthesia shortage. And I can't imagine that is not the same in the Veterans Health Care Administration. And if there is a great anesthesia shortage, surgeries can't get done. And if surgeries can't get done in 28 days, I think, for specialty care, that veteran has the ability to go outside to community care to get their services. And all of that expense is paid by the Veterans Health Care Administration in addition to all their baseline costs. So that's kind of where I'm headed for my dissertation work, is to just kind of highlight productivity, efficiency, and the costs associated with that. The VA has really good data. Obviously, we have all the salary data. Also, something Julie is going to look at, separate from this project, is how much do we spend on community care, which I think will be really interesting to a lot of folks. Hi, my name's Allison, and I'm from Philadelphia. You may want to reiterate that last portion that you just said because it wasn't catching on the microphone. But I do have a couple questions for you. One is in regards to the AAs. I know that you said that it was a small number, maybe not statistically significant, in order to be able to look and actually do any kind of comparative analysis there. I guess my concern with the big push from the ASA pushing AAs is would that be seen as bias for not looking at it and addressing and giving a number to, pardon me, the amount of AAs that are currently in the system, especially if they make less than us, which I was not aware of within the VA, and then pitching the fact that maybe that would be a better use of the VA's funds to have more AAs, and expanding their ratios that way instead of just with independent CRNAs. My other question for you is how... My memory is... I'm losing my memory already, so let me jump in on that. One thing we didn't look at for the AAs is also I mentioned we really wanted to focus on where we have procedural overlap and facility overlap. So, also, there's not only the AAs. It's a smaller portion or number than the VA, but it would really restrict the type of comparisons we could make because, like, an AA is not doing liver resection independently. So it's like we just would have to throw out all of those, and then we'd be looking primarily at the lower-complexity procedures in the VA, and so that's why we just wanted to leave that out. But it certainly could be a future project. Yeah, for here, it would just have changed the research design a lot. Currently, there's no AA in the VA. Okay, yeah, no AA right now is in the VA, and then there's one person is a PA slash AA. She is hired as a PA functioning in pre-op clinic as a PA. All right, well, that's why... Thank you, Rochelle. That makes that move. All right, I think we are... Oh, do we have one more question? I'm not going anywhere, so I know it's... We hit 9 o'clock, but I'm happy to keep going. I'm Roger from D.C. I thought this whole thing was interesting because I started doing civilian contract work on the military bases up in the D.C. area. We're completely independent. We work alongside anesthesiologists. You guys have access to that data, and I was initially perplexed working there because I'm like, wow, who cosigns? I'm like, you sign everything. You have access to that information, and it was also interesting knowing that we require supervision in VA facilities, but on military bases, we operate completely independent. Will you get access to that data? Possibly. Despite it seeming like everyone should be pulling in the same direction, the data have been siloed for a long time. There has just recently been data sharing between DOD and the VA. I have access to it, but I haven't looked at it yet, so I don't know. I could probably certainly see utilization. I don't know that I'd have the same level of outcomes. I'm not sure they're doing the same chart review, at least for some of the bigger outcomes, probably like mortality, readmissions, maybe PACU time. Might have that in the DOD data as well. You want to talk about the Indian Health Service policy? Between active military service and, yeah, yeah. So to your point, in every branch of the military, CRNAs work independently. We're at the front lines, but in the Veterans Health Care Administration, unfortunately, that does not carry over, and the argument is or has been, even if it's not written down in paper, certainly in meetings and in Congress and on physician professional websites, that question our safety, you know, when we're in a controlled environment in the United States, when we're not at war, but when we are, apparently, we're very safe and effective and frontline providers. So I cannot speak as to, you know, who made that decision and whether or not that decision was justified, but it is the way that the programs are running, and we hope to bring some light on our safety efficacy in the Veterans Health Care Administration. Certainly, veterans are a different population than active military service. They tend to be older. The people that qualify for Veterans Health Care Administration said our vets are a subset of the veterans out there, so they typically do longer service. They tend to have lower economic status, and they live in rural areas. They do make it a different population than the general population at large and the active military service, so we're focusing right now on that population, and those are the best rationale for why we are. Thank you. My name is Barbara, and I'm from Pennsylvania. I just wondered, this may not be the right forum for this, but if there were any updates on the bills that are before Congress, the Senate Bill 2070, and HR, I can't remember the number, for denying full practice authority in the VA. I don't know. I think the government affairs team took off yesterday, but they'd be the, I know, probably the ANA folks would be ones for that. I'm not aware. No, and in fact, it would behoove us not to address that since we are doing our best to be, do everything above bar and to be as objective as possible, so we are doing our best to carve out political influence, although I'm sorry I slipped a couple of times. I'm so sorry. Not reflective of Kevin's work, but we are trying to avoid as much as possible the political talk and just talk about the needs of healthcare in the United States and particularly those of our veterans that we serve. On behalf of the ACRNAs, I want to thank you for everything you have done and everything you are going to do for all of us, so thank you because guess what? I work in Windows 121 supervision doing cataracts, and I can pull my hair out sometimes, so thank you very much, and then it's really, we are there for the veterans. We are there for most for the patients because I take pride in what I do every day. I get to take care of American heroes. What a job. Yes, I have to put up with supervision, but I'm gladly to do that, so thank you. I think I'll stick around. We now have a little frowny face that says time's up right here, so we're going to step down, but I'll be around the chat afterwards. Thank you, everyone.

Hilda Nugent

Kevin Griffith

Julia Harris

CRNA practice scope

anesthesia availability

veteran surgery access

VA system

patient outcomes

surgical complications

anesthesia care models