Advanced Analytics with Dr. Zavaleta & Matt Oldham

It's now 2022. We have the internet, new space programs, electric vehicles...but anesthesia providers still struggle, at times, to get the simplest report like, "How many minutes of anesthesia were performed today?" Still less are they able to benefit from deeper dives into analytics.

On this episode, Dr. Zavaleta and Graphium Health VP of Engineering, Matt Oldham, discuss advanced analytics and their potential impact on the practice of anesthesia. The possibilities are both exciting and endless.

Transcript

[music]

Dr. Zavaleta: Thank you all for joining us today. We're gonna do a little back and forth. Of course, I'm Jeff Zavaleta, I'm the chief medical officer at Graphium Health. And on the call we have Matt Oldham, he's our VP of engineering. And Matt and I spend an awful lot of time each week, hours each day, working through technical problems and finding solutions and educating our clients and potential clients in the realm of analytics. And we talk... We spend a lot of time discussing what kind of reports are available, what kind of reports do you want, why is it that we're different in the marketplace, what makes reports valuable, etcetera. We just spend a lot of time talking about it. We thought it would make an interesting podcast to surface some of those frequently asked questions.

Dr. Zavaleta: And at the end of the day, I think Matt will agree, that reporting is just difficult and there's a lot of challenges for different reasons, some technical, some personal. And we're gonna go through some of those and describe how we solve them, what you need to look for in powerful analytics that can help you run your group. Matt, thank you for joining us.

Matt Oldham: Yeah, glad to be here. It's always fun. I enjoy working with you and like you said, we talk about this almost every day. We get presented with new challenges almost every day from our customers and we have to figure out how to solve these problems, and they're not always easy. But at the end of the day, our customers, anesthesia practices, providers, they need reports to try to run their business. And so we have to figure out how to make the data that we have available to us, available to them in a way they can use it and so that it's actionable. Yeah, it's fun talking about these problems.

Dr. Zavaleta: Yeah, I would say that most groups when they come to us, they seem to have... That the reports they have available are generated by their billing companies. So it... The data source... The source data is gonna come from your claims and those claims get turned in every day, every week, every month, and then depending on your billing software's capabilities, they'll get a report on some generic productivity. How many cases did I do? How many units did I generate? How many codes were used? Are there other types of reports that you feel that groups have come to us with? What would you say the baseline client is able to produce? 

Matt Oldham: Well, I think you hit on a good point. Billing is the area that we know no one's gonna forget about it. No one's gonna be unwilling to document what they need to get paid. And so the billing software has access to a lot of data but, only so much of the data they have access to is needed for a claim and that's really kind of where you can draw a line around what data is available to generate reports from.

Matt Oldham: And like you said, you can get some top-level productivity reports and break downs by CPT codes or diagnosis code or a provider or location, in minutes. But, then it's gonna start to break down not too far beyond that because you don't have a lot of your operational metrics that administration needs or practice administrators need to run their business. You don't have access to a lot of the clinical data that the hospitals are gonna wanna look at, in terms of patient safety and quality and compliance, and so, yeah.

Dr. Zavaleta: Yeah. That's... It's well said. 'Cause we were talking about... We kinda group these reports into three different areas. We have billing reports, 'cause that... You're right, it describes the source data for that. But right next to that is gonna be operational reports and when we talk about operational reports, we think of intervals between... Intervals within a single case. For example, what is your average anesthesia ready time between providers, between surgeons, between ASA classes, between facilities? You can aggregate and slice this information in different ways but the interval between anesthesia start and anesthesia ready is a lot of value. Anesthesia ready to surgery start, surgery start to surgery end is obviously the surgical duration.

Dr. Zavaleta: And then the emergence time from surgery end to anesthesia end time, those are different intervals between a single patient experience that can help answer operational questions, and if you want things like that, then you need the data source to support them. Groups will say, "Well, we don't record anesthesia ready time. We might have surgery start and surgery end, but that's on our anesthesia record, that's not on our claim." So it doesn't... The billing team, even if it is on your claim, the billing software may not have a place or a workflow to pay data entry people to put in surgery start and surgery end. It's just not a part of the ticket.

Dr. Zavaleta: Other types of operations are turnover time, so if you try to compare intervals between two different cases... Well, we all have wheels-out wheels-in, that's pretty standard report from a hospital. There's other ones that's... We want anesthesia turnover, anesthesia end to anesthesia start between case 1 and case 2, or we want the close-to-cut time for the surgeon. That's what's driving their emotional insight into why are things taking so long? It could be a long emergence, it could be a long prep time, it could be a long... You just don't know until you can look at those... At those kinds of endpoints, like anesthesia ready, surgery start, surgery end.

Matt Oldham: Yeah, that's right. And for a lot of practices where they practice across multiple facilities, they find, many times, that they don't all track the same things and when they do, sometimes they call them different things. We may call... One facility may call anesthesia ready "In OR time," and they're the same thing to them, but another may call anesthesia ready equivalent to anesthesia start. They don't distinguish them and so when you have to kinda zoom out and say, "Well, how do I measure these at different intervals," like you described, "across my entire practice?" I've got to figure out the answer to those questions. How do I make these things consistent across all of my data sets so that I can generate consistent reports? And so, yeah. Therein lies the challenge, many times.

Dr. Zavaleta: And I would say it even gets more complicated when we jump into this third realm of source data: Billing, and then we go to operations, and the third one is this quality and safety. And when you get into the world of quality and payment program, MAC or MIPS measures, QCDR measures, whatever term you wanna use, those measures are typically...

Dr. Zavaleta: Very complex. There are two, three or four pages of definitions describing what type of case is included, what's excluded, what's performance met, performance not met, data incomplete. And you and I spend hundreds of hours interpreting these new measures every year. As they get renewed, new measures come along, existing measures get retired. We come up with what end points we need to appropriately code any anesthetic case, but maybe you can speak to... Once you update your data model, or once you define these new terms and then update your data model, and then make those available in the EMR, it becomes a couple of orders more complicated in getting that semantic layer correct.

Matt Oldham: Yeah. And that's... The compliance is one of those areas where it's uniquely complex because the measures we're given by the QCDR, or the registries, or by CMS, are... They're unique in that the way they measure things according to the measure specification doesn't always make the most sense to the everyday provider in terms of when they're documenting their case. You can boil the technical language down to concepts that they understand and terms that they use frequently in their clinical workflows, but it's not always immediately clear what's being asked to be tracked or what's immediately asked, "Did you perform this?" or "Did you perform that?"

Matt Oldham: So where we spend a lot of our time, to your point, is breaking those down into concepts that can be understood across all of our customers, which means even in a given practice, do all your providers speak the same language? And so words matter, and we wanna make sure that when we're taking a very technical concept that's probably overly complex in the sense that it could be said in a more simple and understandable way, we don't lose meaning in that and we don't change meaning. And so that translation to, "What do we need to satisfy compliance reporting?" versus "What do I need to track? What do I need to have my providers document for every case and make sure they understand what they're documenting, that they're measuring what needs to be measured?" That can be a challenge sometimes, so that's why we spend so much time on it.

Dr. Zavaleta: Yeah. And then once we set those flag poles within our EMR system, it makes it consistent across multiple facilities within a given organization and consistent across multiple organizations across the country. And then we run into the customer that says, "Well, we use one of the major EMR vendors in the marketplace, and we record that information on our EMR system." It's like, "Well, our experience is the devil's in the details. You may think that you record all of the necessary information, but what if you're missing three or four fields? What does it take to update your user interface so that you do collect these new fields as they pertain to the quality payment program or as they apply to... " Look, operations, most people are recording all of those time intervals. That's not... It's really when we get into the quality macro compliance that we see things really come off, the wheels come off, so to speak, in that each organization has implemented terms differently.

Dr. Zavaleta: And what that means is that they say things that may be the same or they may be the exact words with different meaning, and they may have different words with the same meaning, and they surface themselves in different places in the user interface, so the check box is not consistent across different implementations and across different organizations. And then we're tasked to go in and say, "Hey, just grab all of the data from EMR system X and map it to our data model," which... Why, Matt... In three minutes or less, why is that so difficult? Because it's so much easier said. And I would say as a physician, it's hard for me to explain to my colleagues that it is simply not that easy. It's why is it so difficult? Why can it take months and months and months, even years? We have some clients that are still giving us last year's data because they're updating their interface. Why is it so hard, Matt? 

Matt Oldham: I think the answer is multi-faceted, so there's just the one... All the EMR vendors who implement the ability to capture the data required by each of these measures implement them in a different way. And so they're therefore faced with the same challenge we are about, how do we implement them? Some take the approach that we take where they try to simplify it and make it into concepts that are more understandable and that can be applied and be used in multiple ways, meaning if a measure gets retired, well, maybe I still wanna know that when I'm administering multimodal anesthesia. And so I want to be able to... I want that particular measure or data point to live on beyond maybe the life of the measure itself, but other vendors may take the approach, "I'm gonna implement a given measure exactly the way the measure specification from CMS states." And so you've got to deal with the different implementations across different vendors.

Matt Oldham: Another reason that it's complex is because, even though a given clinical workflow may capture a certain data point, and this really pertains, this particular challenge pertains to kind of...

Matt Oldham: Or we see it more commonly in like pre-op workflows, or certain parts of the clinical workflow where maybe different staff are responsible for capturing certain data, meaning the nursing staff will capture it in one phase of care, or the anesthesia provider will capture it in a different phase of care, in a different facility. And so reconciling all of that becomes difficult because in one phase of care it may be documented in one way, maybe in a clinical note, in another phase of care it may be a checkbox, like you said, or it may be a drop-down, or some discreet value, and you have to figure out how to reconcile all of those, and not only do you have to figure out how to reconcile them, and as a consumer of the data, like Graphium, as if we get that data in interface, the people who are building that interface for you on the vendor side, and trying to get that data out have to go find it. Just because the system supports it doesn't mean it's easy to find when it's documented, and where it's documented, and what phase of care, and in what form are you gonna find that data. So there are other challenges too, but those are just the ones that come to mind.

Dr. Zavaleta: Yeah, it seems like there should be a big easy button where you just push it and that happens, but...

Matt Oldham: Right.

Dr. Zavaleta: It's just not the reality. After doing this for 10 years, and setting up hundreds of interfaces, it's very... It's always surprising how no interface seems to be the same, and there's always... When you get into the world of customization, there just seems we spend a lot of time mapping words, to say, "This is our Boolean field in our data model, how does that map to what they have?" And usually there's a group call, four or five people, and they say these are the fields we have, this is where we think that field is, and then they implement the interface, and then they send us the results, and then the clinicians on that side say, "Hey, this doesn't make sense." But it turns out that the mapping was wrong, that we picked the wrong field, even though we thought it was the right... And so Matt... Yeah, like you say, it's multifaceted, right? It's not only do you have the actual same model property that we have in our data model, but is it accessible, and then, is it the same data model across multiple facilities within your organization? 

Dr. Zavaleta: And I know that might sound strange. What I mean by that is if you're one... Like if you're the chairman, or the director at one hospital and you're trying to get these reports, that can be complicated. I can appreciate that, I understand it, and I live it. I will say that it's an order of magnitude more complicated when you say, "Not only do we want reports out of your system, but we wanna compare apples to apples between that report at your hospital, in your organization to a completely different healthcare system." That may be using the same EMR vendor, or they may be using a different vendor, but even if they're using the same vendor to try to get reports that are apples to apples, to try to get any sense of national benchmarking on what is the average Anesthesia Ready Time across the country look like, I doubt there's any report that would show that consistently across the country. And it's one of the values that we offer, in that regardless of who uses our service line, they're going to be powered by the exact same data model. And that, really it's... You can't overstate the value of that, I think.

Dr. Zavaleta: And it's lost, I think, on most clinicians, 'cause they don't understand what a data model is, which I can understand, but we can take our data model, and because we spend so much time making it so detailed, and intentional in its design, there's a lot of consequence to doing that when you go from one organization to the next, and within each organization across multiple facilities, regardless if you're just using us for charge capture, or you're using our EMR, or you want us to create custom interfaces with existing EMR tools. That's fine. At the end of the day, what you're trying to do is get all of that data into our data model. And once it's in our data model, then our software can understand from an analytics standpoint exactly what you need, right? So it's like we finally realized the full potential of being electronic once you have a consistent data model. And I feel like it's a lost point on many people, that they just don't understand the value of the data model, or they're clinicians, and they have to interface with hospital IT, and sometimes the hospital IT doesn't even understand the data model because of its complexities, and it's not easy. It's not a fault per se, it's just it's not easy. And so our specialty is really focusing on what makes an anesthesia data model complete, and then, what can you do with that, and so what are some examples of data reports that we can provide because of our consistent data model? 

Matt Oldham: Yeah. Well, I think one approach that we've taken that I think really helps is to categorize the type of data that we get. And we do that in a couple of ways. One, we classify these groups of reports and do the areas that people will want to measure, or different parties, and maybe an anesthesia practice wanna measure. Maybe those in the finance area wanna measure the billing segment of the data that's being collected, and aggregated across the practice, that maybe those who are responsible for operations wanna look at the operational metrics, and then you have, obviously, the clinical metrics, and you have compliance metrics for those who are responsible for compliance, and safety. So those are the kind of the broad categories that we most commonly see people, being interested in, and so that's kinda how we broadly categorize the data reports that we make available...

Matt Oldham: Beyond that, I think, that's where we kinda get into the technical aspect of how we implement data accessibility, and that's building the semantic model on top of the technical layer of our data model, and that's just really casting it into business... Casting these technical components into business concepts that users understand.

Dr. Zavaleta: So, talk a little bit about... Yeah, talk a little bit about semantic layer, what does that mean to a clinician? What in the world are you talking about? I just need to know... I just need to know turnover times. What is the semantic layer? 

Matt Oldham: Right. So, turnover time, as an example, is a calculation of your cases over a certain day, over certain locations and times, and so you've got these different components, these different data points that make up, go into the calculation of a turnover time. And so, those individual components are valuable because they may be measured... If I have a start time of case A and a start time of case B, those are independently valuable, because maybe I need to measure the anesthesia ready for case A and compared to anesthesia ready for case B, that's super valuable. But if I need to calculate turnover time, those two data points are used in different ways. So, I wanna make anesthesia start time something that's available and accessible to be able to report on and extract and look at and compare and aggregate, but I also need to surface a concept called turnover time, so the semantic layer really allows us to hide the complexities of things like a turnover time calculation into just a single measure called turnover time.

Matt Oldham: I don't need to necessarily know how to calculate that by picking all the individual components of anesthesia start of case A and the anesthesia start of case B. I hide all that in the semantic layer and surface something called the turnover time. It also allows us to hide other technical parts of the data model, because there's not just one big bucket where all this data lives, right? It's split out into this model, if you will, that's comprised of different entities. And so, the tools, the reporting tools that we make available to our users have to know how to make that data work together, how to join that data, how to link it together, how to aggregate it, how to navigate the data structures to deliver any particular piece of data and give the right value, right? Not calculate something incorrectly or associate it to the wrong patient or case. And so, the semantic layer allows us to kinda do that behind the scenes, hide all of those technical complexities and surface simple business concepts, like patient, or anesthesia staff, or start time, or more complex topics like turnover time, or wheels-out, wheels-in that are more aggregate calculations, that's really all the semantic layer is doing.

Dr. Zavaleta: That's all it's doing. That's very well, [chuckle] that's quite humble of you. I will say that, in my mind, so... When I see you create the semantic layer and when you're writing the sequel queries to get us what we think we should be seeing, I can't give a better example of the devil is in the details. And without having experts understanding what is it you're actually looking for, and you and I would consider... I would consider you and I experts in this space, 'cause we've spent so many... So much time, over so many years focusing on these same things that the devil really is in the details, and he and I will go back and forth with, "I understand that we think we got it there, but let me tell you why I think it's not quite right." And that kinda conversation happens once, and then once the change is made and implemented, it goes across the entire platform, so everybody is updated consistently, so it's always comparing apples to apples.

Dr. Zavaleta: One of my favorite examples is looking at Heat Map utilizations, right? I mean, this is... This, to me, is kind of a pinnacle of what can you do when you have a standard data model, what can you do when you have an accurate semantic layer that you can put on top of all of that, visualization tools that let you create things like heat map utilizations? I mean, heat map utilization, we look at each day of the week for any given month, and we wanna tell you for every 30-minute interval on that day, for all 24 hours, from midnight to midnight, for each 30-minute interval, how many ORs are you actually running, how many FTEs do you have allocated? And then put that in the heat map visualization, so you can see where the hot spots are. And then we can take that same data and put it into a max location utilization, where we can look at, throughout the day, when are you running the most locations, right? So what does three o'clock look like? What does seven o'clock look like? What does nine...

Dr. Zavaleta: And I will say that those kinds of report... Those types of reports, I don't think that the anesthesia market even knows those are possible. They are still stuck, I feel like, and it's very frustrating for me that they are still stuck on trying to get consistent case counts, right? They're still like, "I just wanna know how many cases I did." And they don't have the insight to understand it's even a possibility. If you, as an operations person, can look entire organization or a single facility and say, "I know that we're contracted to run eight locations and I have eight FTEs on the ground... How many times in a quarter am I actually using nine people? Or eight people?" Whatever the FTE is. "How many 30-minute increments…am I actually using that many people?"

Dr. Zavaleta: And so a lot of surgeons always say, "We need more people, we need more coverage, we need more time," yeah, but are you using that time well? You just wanna open up a new OR so that everything can be closed by lunch, right? And then they can get to clinic? Well, that's not how you're gonna run an anesthesia group. So when you look at... What are some examples of reports that people typically want, I think they come to us asking for very basic reports and their whole frame of reference is around their billing data set, and when they talk to us, we're like, "Hey, we have these other data sources, operations, quality and compliance that are a little bit more complicated, but man, if you can stick with us and get us those data, source... Get us those individual discrete datas into our data model, the power of really answering questions that are meaningful to your operations is real."

Dr. Zavaleta: And it's really cool Matt, when I get to be with you on a phone call, where you see the light bulb go off in a president's head or medical director's head. And it's really neat to me, it's kind of the purpose [chuckle] of why we started this effort. It's been a joy, 'cause they finally see it and they say, "This is incredible. This is what I've always wanted." And it's just really neat to get that emotional experience about something that, at the end of the day really is important. It's how you keep the lights on, how you keep the money going, how you keep patients safe, how you keep them happy, and surgeons as well, so...

Matt Oldham: Yeah. And it is funny too, because at that moment it's also... While it's so cool to see it all culminate, and see the light bulb go off, it's almost anti-climactic because we're really taking all the work that we've done in sourcing the data, and integrating it, and storing it, and building the semantic layer on top of it, and we turn it into something that's a simple concept where, you drag it onto a screen, and you have this amazing visualization, and it takes 30 seconds. It's almost anti-climactic because they're like... It's hard to appreciate everything that went into that.

Dr. Zavaleta: Right.

Matt Oldham: Because it just seems like... That this is easy. Yeah, this is simple. Why can't everybody do this? 

Dr. Zavaleta: Why can't everybody...

Matt Oldham: Why can't everybody... Put it on my other systems. Right? 

Dr. Zavaleta: That's right. That's kind of the downside. You make it look so simple, that then they... It's easy to overlook all of the work that goes into it to making it possible. There is a reason that these billion dollar companies on the market have such a challenge in generating meaningful reports that are comparative across the country. There's good technical reasons. And that's what I'm kinda trying to get at, it's not easy.

Matt Oldham: It's not.

Dr. Zavaleta: But man, when you make it easy, it sure is fun. And I'll say, too, that one of the other big, I think, advantages that I've enjoyed of the way that we've implemented this, is that today's question that a group is asking, okay, can be answered, but typically the answer will generate an additional four or five follow-up questions. We see this all the time.

Matt Oldham: Always happens.

Dr. Zavaleta: Yeah, it's... "How many cases do we do?" That's fine, we'll answer it. Or "What is our turnover time?" We'll answer that. "Well, what about on Thursdays?" Okay, that's a different question, because before you just said, "What's our average turnover time?" And now you're saying, "Well, I wanna know what the average turnover time is in the spring compared to the summer." Okay, different question. Let's do it. And when you can set filters and generate reports, like you said, in 30 seconds, you start to see these other questions get answered in very quick fashion, which is really neat.

Dr. Zavaleta: In a different sense, sometimes we are surprised by the questions that get brought to us, right? We don't know tomorrow's question. And so part of the fun in the way we've set this up, is that we get the opportunity... Once that light bulb goes off, and I will tell you, they are few and far between... When someone who gets access to the reports, and is in the position to effect change. Suddenly, their wheels start turning and then they start firing off requests to us. Or, we get groups that will say, "We're at a committee meeting and our surgeons are complaining that the regional blocks are taking too long. How can we help?" It's not one single question, it's this sentiment that regional anesthesia is bad, what can you provide to us to change that perception? It's not an individual number, it's not a single PDF that you just send out. That's not how you solve that, that's not how you change that sentiment.

Dr. Zavaleta: Instead, what we do is set up a comparison between all cases at that facility, between the same surgeons with a block, without a block, or between the same surgical cases with block and without a block, and look at how do the anesthesia ready times compare, how do the surgical emergence times compare. Let's actually define how many extra minutes regional block takes over a consistent duration, over a month, or over three months. How many times does that regional block really take over 20 minutes? What percentage of total blocks is that? How does that connect to patient satisfaction? Does it matter? Is it worth it? I don't know if you have... And then once you set that up, they can run the report whenever they want, it's not a one-shot snapshot. It's the tool available to allow you to now, on your own, run that whenever you want, across whatever facilities in your organization that you want, for any data range, for any surgeons, for any providers. Suddenly. You're answering very complex questions, but like you said, you just made it super easy because you put the power in their hands not ours.

Matt Oldham: Yeah. To your point about it not being just a single number, I think that's really where... When you're able to leverage the data for those types of conversations, the credibility really comes when you can show not only this number that proves a point, one way or another, but you can answer the very next question of, what made that number up? It's different having a green metric on a PowerPoint slide. It's another thing altogether to have a green metric on a report that you can click in and just show immediately a breakdown by provider, or a breakdown by month, or a week, that demonstrate that, "Hey, not only can we deliver metrics on the data we collect, we're demonstrating that we're collecting this data consistently across our practice. This is a standard that we're setting. We're setting a bar for how we value the data. We're taking stewardship of the data across our practice." And I think that, that in and of itself, demonstrates a lot of credibility and builds credibility for our practice.

Dr. Zavaleta: Yeah. It's a great point 'cause we'll get feedback from customers that talk about, "The hospital has their data, and we have our data." And now what you're doing is, you have an issue of trust, right? And so, whose data do you trust? And yeah, you're absolutely right. When you can dig down or drill down into a single metric and say, "Hey, here are the 18 cases that made up that average, and you can go look up each one of these 18 cases on a sheet of paper and write down the individual time intervals or put it in a spread... You can validate on your own exactly where these numbers came from." The ability to do that is kind of the ace... I mean it's your trump card. It's incredibly powerful and you notice the power by silence. So suddenly, nobody else is raising their hand to challenge, because this is all live data, this is not a PowerPoint. That this is a web-based interactive, clickable set where, keep asking questions and challenge us, and if you see something that's a question, well, of course, then you might want to get into the semantic layer. Right? "Hey, I need to better understand some of those technical details of why was this case dropped or why is that number different?" And that's great, our documentation we can help show you that, but you're on another level of understanding at that point, right? 

Matt Oldham: Yeah, yeah. You've changed the whole tone of the conversation because of what you've been able to demonstrate. And then, at that point, you've earned credibility, you've earned trust, and then if issues are still surfaced, because not everybody... I mean it doesn't mean just because you have all the data, doesn't mean the data's perfect. Maybe things are recorded incorrectly, or maybe a problem did happen in the clinical workflow or in the documentation data that is legitimate, it's a legitimate concern or a complaint. At least you've taken the conversation from a position of blame to, "Hey, I've established credibility. I've showed you that we have... Our intention is to capture the data and report it accurately." It's gonna be a much different conversation when you've demonstrated capabilities that you have when it comes to finding those issues because you're not gonna be... It's not gonna be blaming anymore, it's gonna be an earnest conversation of, "Okay, there obviously is an issue here, let's dig down, we can...

Dr. Zavaleta: Find it.

Matt Oldham: Access all the data we need, the detail we need to root out the cause." So... Yeah.

Dr. Zavaleta: That's right, that's right. No one's perfect, and the reason we run around through these reports... The reason we want to make this kind of analytic ability possible is not to show perfection. It's to show truth, and when you have reality in front of you, well, you can make a lot better decisions. It's not emotional decision-making, it's data-driven decision-making. And that all starts with, "What is your source data? How are you storing that data? What is your data model? What does your semantic layer look like?" And then, "What kind of visualization tools is on top of that?" And I appreciate getting to have these conversations, we talk every day about different implementations and consequences, and a lot of our time is spent just sending out emails and educating and clarifying about how these things work. The reality is they're not simple. There's a reason that getting reports in the anesthesia industry is so difficult, very good technical reasons.

Dr. Zavaleta: And it's always a thrill, Matt, to talk with you. Thank you for spending this time with me and hopefully, our audience has a little more insight into how it works. We'd love some comments and questions if anybody has any. We'd love if this is of interest to folks to go into, maybe into deeper dives into very specific areas. I feel like we could speak... We could talk hours about different aspects and go through different examples and illustrations of how we solved unique problems to make it work.

Matt Oldham: Yeah. It's always fun, Jeff. We learn new things every day, so yeah, I look forward to... Looking forward to your feedback and let's do more of these.

Dr. Zavaleta: Awesome, yeah. Okay, thank you.

Matt Oldham: Okay, thanks.