Getting Surgeons to Engage With and Believe Their Performance Metrics Webinar Transcript
SANJEEV AGRAWAL: Good morning, everyone on the West Coast or in Mountain Time and good afternoon to everyone on the East Coast. As Sandy said this webinar is the fourth in a five part series for those who have been following the first few webinars, the overall theme of all these webinars is how to get the most of your OR capacity. And there are many pieces to it, many processes that go into enabling that. We covered the first three in the previous three webinars. So if you’re watching this and haven’t seen any of the others, they’re also available on our website at leantaas.com.
But about a month and a half ago, we talked about how do you get– how do you take bloc, time away. So right sizing blocks, using a metric called collectible time instead of using block utilization, why that is more actionable. A few weeks ago, we talked about how time is left on the margin because people don’t release far enough in advance and how you can enable an. OpenTable like capability that allows no block to be left behind and a much smoother release and request process.
Both of these free up a lot of capacity, both the use of collectible time that we talked about on September 13 and the use of OpenTable to release and request time on September 20. And today, we’re going to talk about another key piece of building the culture around enabling some of the key processes that get you to use your OR capacity well. And this has to do with building a data driven culture where administrators, surgeons, staff all understand the data, understand the metrics, understand why they matter, understand their definitions, and are able to use the numbers and access them really easily. And without doing that it’s also difficult to move to a place where we’re making data design decisions and maximizing the use of our OR capacity.
A week from now we’ll address the issue of how you allocate time which is different from how you take away time. All of these webinars are recorded, and you can listen to any of the past ones or this one later on as well. Just a quick reminder that our experience doing this doesn’t come from theory. It doesn’t come from consulting. It comes from building products and deploying them at over 450 ORs across some leading institutions in this country. Many academic, many community, many on Epic, many on Cerner, all of whom have their own reporting capability. Many with very strong reporting NPI. We think we need to be addressed and move to a place where our surgeons, our preoperative leaders have believable, credible, good data that they can make decisions with at the right time and place.
So there are six pieces to today’s agenda. The first is diving a little bit into what we see as the issues with traditional reporting. Basic EHR reporting as well as reporting that systems has created internally through teams that mine the data in the EHR and publish reports. Then we’ll get into some of the ways that you could engage providers and administrators both on mobile and web. We live in a world where almost everything is available on mobile and web. And so it’s surprising that for a lot of health systems, mobile is not even an option. Then we’ll talk about some of the key performance metrics we think matter. Nothing surprising there, I’m sure most of you on this Webex consider these, but we’ll get into some detail as to why we think these matter.
The fourth part of this presentation is actually quite powerful, because really when we do reporting, we are usually trying to answer a set of questions. And how can we answer key questions that crop up both periodically, systematically, as well as ad hoc. So anything from are we using our robot room well to do just robot cases or are we using it to do nonrobot cases? And who are the biggest offenders and what can we do about it? Which are my top surgeons that always run late even if they leave time during the day unoccupied in their block, but they always run late?
So those are questions that keep cropping up, and so we’ll get into some examples of how modern day tools can help you answer these questions significantly better than what traditional reporting has been. We’ll give you a sense for what’s coming, what’s possible, especially in our product which is called iQueue. And then how you might want to get started, if this is something of interest to you. So let’s start with the issues with traditional reporting.
In a phrase, the way. I like to describe the way traditional reporting works is there is a lot of admiring of the problem. So if you think about weighing scales, I’ve gotten on several in my life hoping that the answer would be different, but it’s not. Every weighing scale that I’ve gotten on has basically said, you need to loose a little weight. So making the weighing scale different, and prettier, and more accurate does not take away the fundamental problem. It’s basically me admiring the fact that I have a problem in different ways or a thermometer or a fancy scoreboard at your favorite football team stadium.
Changing the scoreboard is not going to help you make better plays. So what we find, if you dig deeper, and why I say, it’s admiring the problem is there are four basic issues. One is often time reporting is logjam not because there is a lack of reporting– there is in fact sometimes too much. When we first started working with perioperative teams, we found that there was one institution that actually had 150 reports across their 40 ORs. And when we asked the business perioperative manager, what they use the reports for?
And dug deeper into it, we found that in fact, a lot of these were the result of historical reasons– why someone had asked for a report at a certain point of time. But then actually, if you actually looked at 150 reports, they often didn’t agree with each other, because the definitions were different. So there was no single source of truth that anyone could point to. And when there are multiple sources of truths, you can say whatever you want. Who you’re going to believe? If I have five different scoreboards, and they all show me a different score, what do I believe?
The second big issue is the frequency and cadence. So if you’re always looking at historical data from say a month ago or even a few weeks ago or 90 days ago, worst case, you’re admiring the problem by looking backwards after the horse has left the barn. So to give you another example, back in 1995, when the web was just starting to get to the point where you could do things like stock trading and things like that, I personally had a broker that would send me a report every quarter. In the middle of. April, he would send me a report for the first quarter.
And it would always say that. I lost against the market in beautiful chart format, great pictures. It would never tell me what he was exactly going to do about beating the market going forward, but it was a day late, a dollar short, but very pretty, not very actionable, backwards looking, not very helpful to me, but it was just informing me of something that happened. The third has to do with, in April 2018, if we have to access our bank accounts, we could go on our mobile browser. We could download an app. We could go on the web. And it would show us current information in a user experience that we could easily navigate, understand. They’re very simple to use. By going to Bank of America, I’ll be able to see every check that ever cleared, that I ever ordered for or like for many months, yesterday if it cleared in fact, I would see it and I would see a copy of it. And so if you were presenting data using paper.
If you’re still emailing people, large PDFs and expecting them to dive into it and read their performance numbers that gets a little difficult. No one usually will make that effort, and especially because life outside of the OR in the hospital, it’s so easy to see my performance numbers. In fact the financial world, if you go back to that for a second, you can program Yahoo finance or Google finance to send you a text message every day after the market closes to be able to see how your portfolio did based on what happened that day. You could dive into exactly which stock went up, which went down, why? What happened? And it’s almost real time, fresh data that is meaningful in explaining what went on and very, very easy to access, and use, and understand.
And then finally, this idea of credibility and transparency in the data itself. So when we’re speaking with surgeons and we’re speaking with perioperative leaders, administrators, these are all data driven people, scientific folks that have spent a bunch of their lives looking hard at numbers. For research, they’ve done during their process of becoming surgeons, and when you present them with data where it’s not clear how it’s defined. If you show me my block utilization and don’t show me exactly how it’s defined and don’t show me the raw data underneath it saying, how much time did. I ask you released? Is that part of the denominator or not? I will throw the baby out with the bathwater, even if there is one thing that’s wrong with the report.
So if there are multiple sources of truth, and we don’t really get down to showing explicitly where the data is coming from and transparently making it visible to everyone. Oftentimes, we find that we’re doing reporting, and a lot of effort being put into it, we’re gathering data from multiple systems, but the question is who believes it, who gets to even see it. Because of all of this reporting is coming out of the EHR, and I have to access the EHR or tableau or some other report, then it’s very difficult to expect that all my surgeons, all my staff will get uniform access to data.
Now, obviously the flip side of it is how you solve the problem. We’d all love to live in a world where much like our bank accounts, and our stock trading accounts, and other commonly used capabilities that we have on the web, we understood the metrics. And the metrics we presented were aligned and defined, and there was a single glossary of definitions as to how we’re calculating first case delays, how we’re calculating turnover time, how we’re calculating block [INAUDIBLE] utilization, how is staff room utilization different from prime time utilization?
All of those– because the details matter. The devil is right down to the specifics of the numbers. And really, if that data is being used to highlight ways to get better. Again if you think about homework grades and midterm grades, when we were all in college, if those grades were basically just telling me a fact– I did poorly, or I got a B or C or an A, it’s really not actionable. It’s not saying that the reason I got an A or B or C is that when I work with this particular case team on a Wednesday, I tend to start much later than when I work with others. So diving into the causals and making it really simple to identify anomalies that are happening before bad things happen as opposed to after the fact. It is pretty important, because reporting has to move from being a pejorative judgmental process to one that helps me do better.
The third bullet is about the ease of use. I expect information to come to me in this day and age. If my flight’s running late, then I expect the airline to text me and let me know what happened, if there is a date change, if a check cleared. In every other aspect of life, there is a combination of push and pull. Information is being sent to me that I can easily digest at relevant times, and I have the ability to go seek it out myself and pull it. And then finally, I need to understand what is being reported. And I need to be able to dive deep into understanding why the numbers are the way they are. If I say I thought. I released my block, I want to see whether you’ve taken that into account.
So these are some of the principles of why we’re even having this webinar, where we think that a part of why it’s hard to use data to make decisions is because of these four basic problems. So now, let’s imagine a world where reporting in periop resembled reporting in almost every other aspect of the world whether it’s airlines, or banks, or brokerage firms. So imagine every week your surgeons, your administrators, or service line chiefs, whoever you wanted to see their numbers, imagine if they got a text message. And I’m going to start with how you push quality information and then show you how you pull quality information in a far easier [INAUDIBLE] way than previously seen traditional reporting on paper, on PDF, and through EHRs and tableau allow.
So imagine a simple text message was pushed out to the right folks, where let’s say, it was a surgeon who wanted to get a text message. And the text message says, here are your five key performance indicators for this week. Your room utilization, your minutes used, your block utilization, your turnover time, your first case delays, your case length back here, so you keep saying that the case is going to take three hours, in fact, it takes 4.5. And then underneath it, imagine if there was a link that allowed them on their mobile phone, because everybody has a mobile phone and everybody has a mobile phone with at least a browser. And imagine if they clicked on that, they accessed on their browser in 90 seconds or less all the detail behind those key performance indicators.
How I stack client against every surgeon in my service line. How I have used every chunk of time that I used in the OR last week? When did I start late? When did I end early? When did I go into overtime? When did I use two rooms, and how well were the two rooms used? How does the pie of my use of time break down? How much time I used in my block? How much time did I use in certifying or a surgeon group block or outside time that I used that as open? If there were delays, why those delays happened? How many of those happened? Were there trends? Do they happen more on. Wednesdays and Thursdays, when I’m working with a particular team?
So now when this information comes to me, it’s irrelevant to me. It’s something that I care about and I get it on a periodicity much like my bank sending me texts messages, informing me of things that had happened. I don’t have to do anything. I don’t have to go look for information. I don’t have to wait for the. OR committees to come tell me, how I’m doing? And basically right before they take time away or try and take time away from me. And that’s it. Once a week, a 30 second overview of my numbers as simply as on the palm of my hand.
Now, the flip side I’m going to turn to an actual application. And for those who have been following this webinar series, we’ve covered other parts of this application before. But the sister to these mobile capabilities, and if you remember, we talked about how to release and request time through exchange. We talked about how to right size blocks using collect or collectible time. So now I’m going to focus on the portion of this application called Analyze that has three sections to it, Dashboard, Explorer, Surgeons. And the goal of this application is to say, there is a lot of wonderful information hidden in our EHR that sometimes surfaces through these reports direct.
What if we could just automate a majority of the reports and put them in three buckets. The first bucket is called the Dashboard that takes all the standardized reporting that you would expect to get, all the things that everybody cares about, and provides you a very simple way of navigating it. So let’s take say these nine key metrics. And this is a demo application. It’s not real data. In the demo application, let’s say you as an institution decide that prime time utilization, total OR usage, block utilization, first case delays, first case delay length.
How many add-ons you are having? Your cancellation ratio, the number of cases you’re doing, and your turnover length, what are your key metrics? And if you had the ability in one, once you signed in, which you can stay signed-in forever, to be able to dig into just about any metric you care about at a great degree of depth. So let’s just take one as an example. I want to know how my block utilization has been doing, and when I jump into this application, I want to be able to drill from the high level numbers all the way down. So at the highest level,. I want to know things like is the trend of my usage of blocked time trending up or down? And just at a very, very high level.
Then I want to know, by month perhaps, how this utilization is trending. Am I seeing seasonal patterns? Each of these numbers I wanted– I expect when I click on them, I should be able to download the raw data that shows me why that number is 70%. And it should open up exactly who the surgeons are? And how they’re using their time?
When I go into this case called the leader board, I should expect, and you can share this leader board selectively with different surgeons, or you can make sure that everyone gets access to everything. That’s an institutional priority which you can decide. So then I want to know something other than a little deeper. I have a mix of surgeon groups, certifying and individual block owners here. So if I’m having a discussion with general surgery in how they’re using their block time, in two clicks, I expect to be able to see first of all, what is that utilization number?
Much more importantly, how it was defined? Even more importantly, where these numbers actually came from? So if you’re telling me that utilization is total little block divided by net allocated that breaks into in block plus turnover divided by allocated minus release. Then I want to see, exactly how did I lose every block or how the general surgery used their block. Sometimes they had two rooms, they use them well. Sometimes when they had two room, they didn’t. And all of this should surface where there is no– this is incontrovertible evidence.
This is right here, being pulled from your EHR on a daily basis, and being refreshed. So yeah, you can argue against facts, but then I could say the sun is not going to rise tomorrow. The point being, by making it really so simple to both aggregate that a lot of work that goes into making this, this simple. But if you do, you can move from a world where all your standardized reporting, all the efforts you’re making around standardized reporting, could actually become a continuously available stream of good data, for those of you who use applications like Fitbit. When you go into the. Fitbit dashboard, essentially it takes every health metric and tears it apart and says, let me show you how much you walked, what time of the day you walked, all the things you care about. OK?
The second part of this application is the Explorer. It looks blank for a reason, because this is the section which is the ad hoc role, your own set of metrics, where these are the metrics where I as a surgeon or you as a service line chief or someone else as the owner of a set of ORs in one location. There’s a very specific ad hoc report you want to run. So this is essentially tableau built into an application like this, where you can pretty much say, I can choose whatever filter I want. And there are dozens of these filters.
Because as you know, it’s day of week, it’s location, it’s surgeon, it’s you want to only see the high volume ones. Is it the robot case? What kind of patient class we want? And if it’s a patient class, do we want emergencies, et cetera? And you can build your own report, save it, run it, and it’ll be sent to you whenever you want it. But not everybody is interested in this report, so it’s something that’s individual to you. And I’m giving you a high level overview, because there’s a lot we need to cover.
But imagine a third piece of it, where we have surgeons score boards or surgeon dashboards, so simple for both the surgeon, the OR committee, as well as surgeon service line chiefs to be able to use. So if I’m sitting down with Dr. [INAUDIBLE] instead of dozens of pieces of paper, what if I could have all prior relevant information in one place. Dr. [INAUDIBLE],, these are the cases you’ve performed in this period of time, and I can change that period of time. I can see how long it took. I can see wheels-in, the wheels-outs. I can see if it was a robot case or not. I can see what the distribution is in the case length estimate versus how much time it actually took.
But I showed you block utilization already, but imagine if that utilization was broken down, if she did have multiple types of blocks, individual blocks, service lines block, et cetera. Here the minutes you used, and the minutes where you were primary on the panel. Because sometimes you were primary and sometimes you were secondary, on some of these battles. How many of them were outpatient versus inpatient? Is your volume of inpatient or outpatient surgery rising?
So by making it really simple to have the surgeon conversation based on very simple to use web tools much like we use everywhere in life, but several things happen. You get the ability for people to actually engage in care. You’re making it so simple that it’s hard to argue why you wouldn’t use it. You eliminate a bunch of process and paperwork and rework and back and forth, because you’re drilling down all the way to the data itself. So the idea is that most companies, most systems now are moving to the point where reporting in and of itself should be largely automated, so people who are doing reporting today can free up their time to have the harder conversations.
How should we right size block? Who should we get time to? How can we build a relationship with each of the service lines to be able to get them to use this? Instead of spending a majority of our time fretting over pulling reports and doing pivot tables, I expect this to be something that continues to improve where institutional knowledge is not lost if someone leaves the hospital. Lots of benefits of being to be able to doing this, this way. Now, when you do this, and you clean your data up in the process of doing this, obviously the mode in which you distribute the data is less relevant. It can be made and customized to whoever wants to receive information in whichever way.
So I showed you mobile. I showed you web, but the same information being put into a. PDF to the extent that there might be certain surgeons or certain service lines or administrators that still prefer the model of a push based PDF come into my email. That’s not hard to do at all. In fact, at the other end of the spectrum are institutions that are so forward looking that are saying, hey, you know what, I’m able to order an Uber by speaking to my phone, by talking to Alexa. What if I could get all of these metrics sent to me by asking for my utilization through a chatbot.
So let’s not confuse the modality of consuming this information with the cleanliness and the ability to see the right information at the right time. So that’s a very, very high level flyby of some of the ways in which you can make it really simple, elegant, easy to consume information that does contribute hugely to engagement from a physician and a service line, and an administrative perspective. So when I show up to OR committee, I’m no longer surprised. I’ve been getting B and C all semester, so I show up at my finals, and at the end of the semester, if I get a B minus, it makes sense to me.
Let’s move into the section about key metrics. None of these are surprising to anyone. But when these kinds of dashboards and these kinds of tools are built, we obviously need to make sure that at least these four buckets are captured. The first is volume, both in number of cases and total OR usage minutes. How it’s attributed? Which is inblock? What’s out of block? How many to in open time? What am I doing add on cases that are running late?
All of that being shown as part of the volume piece of these reporting capabilities. Then of course utilization, which is the backbone of what we do, are we making sure we’re being good stewards of our time and utilizing it well. We want to know when delays are happening and to the extent that we can improve both towards case delays and turnover time. We can improve the patient experience as well as the surgeon and nursing experience. And we need to be able to dive deep and detect trends, and be able to see what’s acceptable or not very, very simply. And then more metrics like, are there certain service lines or surgeons where the add on case ratio is very high?
Is that truly all urgent emergent or is it something that could be done on a different day? That’s another topic we discussed when we talked about reducing and requesting time. Are a lot of my case has been canceled last minute? How is my turnover trending? These are other metrics we are working on. Of course, you can add however many metrics you want, quality metrics, financial metrics, cost metrics, if that information is available. Really the ability to use these metrics comes from being able to dive deep into certain major dimensions and being able to filter it. So the obvious ones are, I want to be able to focus on certain locations and runs. I want to be able to focus on key expensive pieces of equipment, like my robots. If I have multiple robots then. I want to know Si versus Xi. I want to know whether it was an inpatient case? Whether it was a trauma case? Whether it was an urgent emergent case?
Obviously, by surgeon, and service line, day of week. So many of these are obvious. Some of the newer emerging ones we’re seeing is, let’s also see this with the lens of the anesthesia providers. Let’s see from the lens of the circulating nurse. Let’s see from the lens of the payer mix. And all of these are interesting lenses we’re building into this capability, but really what you want is one single source of truth where the data has been cleaned up and aligned and that’s socialized with everyone. But then the ability to go in and look at very similar to my bank account.
If I went into my bank account and wasn’t clear what was a check versus an. ATM withdrawals versus a cash payment or something that are connected to a credit card or PayPal, well, I’d be very confused. And there would be no way of actually attributing the right dollars to the right source of funds or use of funds. Very similarly, cleaning the underlying data and making it attributable to the right metrics and filters once.
That’s the hard work we do with you once in order to then have a scalable platform that you can share with 500 surgeons and 200 administrators should you choose to do so. Now, let’s move on to potentially, at least for us, the most fun part of this webinar. Reporting has to, at the end of the day, be the answer to questions. It has to be so simple that if I was similar to going into my bank account and saying, you know I’m getting a little bit of work done in my yard, and I paid the landscaper three checks. Which three checks? And for how much? That’s what I’m focused on. And I need to be able to get to that answer really, really easily. Some banks make it harder than others.
In the OR world, these are 10 examples. Obviously this is not a comprehensive list, but oftentimes we have questions like, does a particular surgeon run over his or her block very often? And if he does, does he do it on days where he’s also leading time on the table during the day when he has walk, which is kind of a travesty, because I’m not letting anybody use my time, and then I’m running late, which is a great source of dissatisfaction to my staff. And sometimes, it can be avoided. Things like how busy is the. OR across my 15 OR on Monday morning. Am I working weekends, and am I working weekday nights and how busy.
They’re very simple questions, but having the capability to access the answers very, very easily instead of having to run a report that takes 30 minutes to set up and then cannot be repeated, because it’s a painful process. Who is contributing to the fact that in our first case on time starts are actually decreasing? Who’s being a good steward of their time and releasing time early? Is the robot being used or the robot room being used for too many nonrobotic cases, et cetera, et cetera, et cetera, you get the idea.
So applications, like the one you saw, allow you to dive deep, and for example, if you follow these instructions, they come with kind of an instruction manual. Very if you say, I’m trying to figure out if surgeon A frequently runs over his or her block. And so if I go into the application, I can follow these instructions. I go to be analyze dashboard, and I’ll just give you one example. And I go into the dashboard here, and it says Click On The. Block Utilization Details. So let me go to block utilization, and I hit Details. And then when I hit the. Details page, I get Trends, I get the Leader Board, and I get the Visualizer. Let me hit the Visualizer. And it says click. View utilization for a given block owner. So let’s take a particular block owner and view their utilization.
Now, this is one of the most powerful visualizations to answer questions like this, because it says, if I walk in, and I go into his. Tuesday, well, Tuesday is a bad example for him, but let’s go into Wednesday and say, show details. You can go in and you were able to see for that particular surgeon, for that block owner, how they’ve used every Tuesday, and how often they frequently run, in this case, it’s once in a while, not very often. And this actually is then using a second run. So they’re actually not using their block. They’re using two or three simultaneous runs on Sunday.
So the point being, if you can get to hard questions without running massive reports, but a single dashboard becomes the way for you to find answers very quickly because you have the recipe of getting answers to questions like these. For example, how busy is the OR on Monday morning, on weekends and weeknights, very similar analysis by going into the dashboard and saying let’s look at prime time utilization, individualized or be able to hover over any portion of the day and be able to see how many rooms were occupied by day of week.
Things like first case ontime starts are decreasing. Who is the biggest offender? Well, let’s go into the Analyze dashboard. Let’s go into the First. Case Ontime Starts. And let’s click on Details, click on the Leaderboard. We pick a time range. Let’s just keep it the way it is for now. Set a minimum number of case. So let’s forget all of that for a second. But if we go into and see the biggest offenders are the ones who are at the bottom of this sort of a list. And of course, you could say, I only want to see the ones that have high volume. And you only want to see the big offenders. And be able to say, OK, when I’m having a conversation with Dr. [INAUDIBLE],, being able to look very quickly about why her cases are running late all the time. And what are the reasons for it?
And sometimes, this brings up both data issues and process issues. Data issues in that, we recorded this delay. Maybe the delay wasn’t recorded right, and maybe the attribution wasn’t right, but this is where garbage-in, garbage-out is going to hurt. But the fact is, how do you even highlight that the data isn’t clean or not believable or credible? So these are some of the things that you can do. Many other that you saw in there being able to dive right into looking at block utilization, being able to pick a particular surgeon, being able to dive deep and look at their specific numbers, and be able to show them precisely by days, week, why we’re saying what they’re saying?
There was a day which we know they released, it was marked as release, but when we’re having a discussion with the surgeon, they might think that it was released, and if it wasn’t because there was a process issue. Often we find that clinics get so busy in OR schedule and get so busy that surgeons may think they’ve asked for time to be released. Whether it was actually released in the EHR, we find out after the fact. And that’s one contributing reason for why I as a surgeon don’t believe the data anymore, because I thought I released it. And somewhere along the process of releasing it in the EHR may not happen.
So surfacing and shining light on the details of these numbers goes a long way. One very, very interesting recent metric we introduced was, if I’m looking at my robot room, how many cases in my robot room were truly robot versus nonrobot? And then do I both anecdotally and strategically hear about people asking for time on the robot and not getting it, in which, case there is demand for that time. There’s a balance between the robot [INAUDIBLE] in the room, and if I’m a robot surgeon, and I have multiple cases on a single day, you really don’t want to having to oscillate between rooms. But at the same time, if. I’m holding other people up, I need to know who I’m holding up and whether I should get an entire block where I should try and fit in as many real life cases or not.
These are things that gives you the ability to have that discussion openly with data, with physicians. Some of the additional metrics that we’re working on that we talked about are specifically focused on robot time and looking at average case length for the same procedure type. If I’m doing an open versus a robot case, how does that differ? Of course, you want to marry that with quality and [INAUDIBLE], et cetera. Things like what percentage of cases are actually running over a certain amount of time depending on your average length of case. Things like cut to close, things like skin-to-skin. There’s a bunch of other metrics.
The point behind all of this is why do these metrics matter and what questions are you trying to answer with them, so metrics for their own sake really won’t get me as far as we’ve talked about. Some of what’s coming in the future through these capabilities that you should be aware of. When I was flying in an airplane and airplane control systems have gotten to the point where, let’s say, they hit an air pocket. I’m sure we’ve all experienced times when they hit an air pocket, and unexpectedly there is a pitcher or a role that the pilot has to correct. In the case of airplanes, they get an almost immediate signal back saying, you know you need to veer right or left or fly back up or descend a little bit to avoid air pockets.
Now imagine, if it took nine or 10 seconds for the airplane control system to actually tell the pilot what to do. Well, we’d all be in trouble. Because in nine or 10 seconds, a lot of bad things can happen to an airplane. And oftentimes when we do reporting, it really is like the pilot being informed nine seconds after the fact, because we look back after a three month period, and we say, oh! Our cancellation ratio suddenly went up. Our turnover time suddenly went up. Wouldn’t it be much nicer if much like airplane control systems do, we get to be informed early when bad things were happening that bad things were happening.
So you see two charts here where statistically if there is a metric that is going outside a range, whether it’s positive or negative, it’s my turn over time, and suddenly going beyond a certain range, or my cancellation ratio is rising, it would really help me if you told me during the quarter, almost on a weekly basis that these are the metrics that are trending badly, and these are the metrics that are trending well. If you patted me on the back for work that I’ve done with my case team to get my turn over and my first case delay is down, and giving me feedback more in the moment, so instead of again the periodicity of reporting being a month or a quarter, if it got a little more frequent and much like an alert.
You know, you’re going to stop paying attention to alerts. For example, airlines never send me alerts saying my plane is on time. Airlines only send me alerts when my plane is not on time, because that’s what I’m expecting. So you have to be careful about how many of these alerts you send in, when you send them, but when it’s powerful enough like say, you might want to do something about Wednesdays, say you might want to do something about the fact that you know your cancellation ratio is going up before it really becomes a problem, we call this anomaly detection. It helps surgeons and teams do better by letting them know during the time that if I read chapter six and seven a little more thoroughly then I’ll do better on the midterm.
Similarly, this notion of looking at the data we have, and every hospital obviously does a lot to clean their data up especially for billing purposes. But when we get feeds from our customers, sometimes 1 to 2% of the time, which still can be a lot if you have tens of thousands and millions of timestamps in your EHR. Sometimes we find issues which are fairly obvious when you think about them, but are a result of expected human error once in a while. So if the timestamp for the wheels-out of a case actually is before the timestamp of a wheels-in of a case. If the timestamp says the case started at 7:45 in the morning on the 23rd and ended at 12:30 PM on the 27th, well, that’s going to cause us many, many billing issues. Hopefully, no case takes 100 hours.
Now all of these are simple errors, and they can be avoided, but we’re all human, and as we enter these timestamps, sometimes we don’t enter them right. We had one instance where we saw that there was a tissue sample that was used or a tissue that was used as part of a case that was worth hundreds of thousands of dollars. All that had happened was that the unit in which that the volume was measured and entered was the wrong unit. It’s like saying instead of 1 centimeter, I said it was 1 kilometer. And so the unit field was wrong and that ended up showing a fairly big anomaly. And just giving you some examples of where without anyone meaning to do the wrong thing, mistakes happen, and being able to audit that mistake and send it back to you, and say, you may want to look at these timestamps.
There are the cases where the billing, the compliance issues, and some decisions we might make. Because imagine those kinds of cases going into your block utilization table, going into your room utilization table. It would seem like on a few days, you worked all night. So all of these get surfaced, when you take the data into EHR and run it through cleansing algorithms, run it through a process where teams like our teams work with your teams to truly align the metrics and make sense out of the data that’s in there. So why is this approach different?
Believe it or not, this is actually a chart that one of our customers created with us. And in their words, it’s different for five reasons. The first is meaningfulness. Actually show me information that matter, that I can do something with, where I can easily get answers to the questions I need. And not just need, but I can actually democratize this information and this capability across hundreds of people. The second is comprehensiveness, where we track all the major metrics that matter to all of us in that thing called the dashboard. We allow the capability to have ad hoc reporting. We focus on things that are really important to us like robot rooms. We are able to have these certain scorecards to have effective meaningful conversations with our surgeons.
So most of the use cases where we need to report numbers and engage surgeons are covered. The third has to do with timeliness. The fact that we get a push message every week. The fact that we are not a day late and a dollar short like my first stockbroker, pretending that they really care and sending me yet another report that I really cannot do much with. So how often this data is refreshed? How current is it? And what I can do with it when you tell me the anomalies that are happening much like steering an airplane back into flying normally?
Credibility, how these metrics are being calculated. Don’t ever show me a number that says, my role utilization or rather my block utilization is 62%. On my first case delays are 15 minutes, 25% of the time. I want to see how it was calculated, I want to see every block of time I used, I want to see if you considered the releases I made. I want to dive all the way down to timestamps in the EHR, and if you can show it to me in an elegant, easy to use, easy to digest way, I may start believing the numbers.
And then of course, there’s accessibility. If it takes an act of Congress for me to get access to Epic, and then I have to know how to use the Workbench, and then I have to be able to pull five reports, I’m just never going to do it. If you’re going to send me a 200 page report and expect me to leave through it or understand all that’s in there, I’m just not going to do it. If you give me an extremely lightweight way of accessing this capability on mobile, on web, you send it to me with just the right number of alerts, I might actually start engaging. So if we want our surgeons, if we want our leaders to use data, to understand it, to believe it, we’ve got to make it very easy, very accessible, the metrics right, credible, and comprehensive.
So just to give you a sense, what it takes to do something like this? What it takes to do something like this is four major pieces. The first is to get data out of the EHR, and the right tables out of the EHR, where we understand, for example, what each field means, how you’ve changed the definition perhaps, what rules you’ve put in for how you’re defining minutes and delays and rooms, and what is primetime for you, how many rooms do you operate, how many room are you staffed for, which one is the robot room, all of those the mapping, if you will of what the data actually means.
Now we’ve worked with most major EHRs to the point where we actually have scripts that we can provide your. IT teams to very simply be able to pull this data. If it’s Epic or Cerner or MEDITECH or Allscripts or Paragon, how we worked with most major EHRs to be able to get a pull of this data. And once we have that pull actually aligning the definitions, understanding the constraints, understanding how you operate is a pretty important phase. How your service lines are built, how you allocate block, where do you capture that block allocation information.
From a timestamp perspective, what are your rules for turnover, if you have multiple locations, inpatient or outpatient, do they have different rules for their targets? Similarly, for first date delays. All of this happens in the alignment phase. Once you do these two pieces though, magic happens, because we’re able to create this infrastructure for you that you saw, very lightweight for mobile and web, and ongoing for as long as you want. The push and the pull becomes easy every day of the week. No one has to do much work beyond it on paper.
You can stop doing a bunch of reporting and make this your one single source of truth. And the maintenance that happens really is just the daily ongoing feed of data that comes from your data warehouse, if it’s Clarity or if it’s Millennium or whatever data or Oracle, whatever data warehouse you’re using. The ability to sync that data once every night through an upload to an FTP server. That’s what was used, and once in a while sanity check, whether the original definitions that you told us are still accurate or are we finding that maybe you changed something, maybe you changed the block structure without telling us and that the data and the anomaly detection will identify.
So lots more we can talk about here, but these are the four major steps to making it happen. I want to leave a few minutes at the end for questions. So I want to stop here, and open it up to anybody that has questions. And Sandy, if you’re able to please moderate that. Just a quick reminder that next week, I think at the same time same day, we’ll talk about an emerging methodology you’ve created around allocating time, which is different from right sizing or taking away time or requesting and releasing open time or the webinar today.
Do you see any question, Sandy? We do have a couple of questions that came in. One is, how open do you think that the system should be? Do we allow everyone to see everyone else’s numbers, or just see their own? That’s a really good question. We’ve worked with all four types of systems. So if you mentally imagine a grid where on one axis, if they were not open before or open before, and then on the other axis what happened after. They were not open after. And they were open after, we’ve worked with all four sorts of system.
Systems that believe that really it’s the surgeon’s data and only they should get access to it, and that is fine. And others that have actually moved, and said, we used to think that we shouldn’t allow people to see each other’s metrics, because that’ll create an artificial kind of fight that we don’t want. We are big believers in open markets, as you saw in the release of request time OpenTable tool and in shining light openly and having open discussions. Because part of the reason where a lot of the cultural norms that are set in institutions– I’m going to sit on my time. I am not going to release it. I’m going to pass it on as legacy to my kids before I give it back to the OR.
A lot of these behaviors come from a lack of trust and come from a lack of belief that if I release my time someone else will as well. And the reason I’m bringing up releasing time, it’s just one manifestation of lack of openness in the system allowing free access to having hard conversations based on fact. So while we certainly would be big proponents of sharing more rather than less, obviously this tool can be customized to share whatever you want with whoever you want it, because it’s a role based tool. Are there other questions, Sandy? Yes, so another one is, similar or related, do you allow surgeons to write back and dispute their numbers? Yes absolutely.
So one of the big ways in which there is a continuous learning loop and that helps people be more confident in their numbers is, we sometimes get good feedback from surgeons saying, wow! I love this, because. I now know my numbers, and they’re coming to me, and it’s so easy to use. And sometimes we get,. I don’t believe this, because I released that date. The example I was giving before, that I was getting before. The beauty of that, and you can write within the mobile app as well as on the tool itself you can send feedback back. We love receiving that feedback, because it makes us go check the data in the process. And often we find, it’s both, it could be a data problem or it could be a process problem.
In that example, like. I said, the surgeon told their clinic scheduler to release the block, the clinic scheduler called the OR scheduler, couldn’t find them, maybe left a voicemail, maybe called them back, maybe didn’t, maybe emailed, maybe texted, and meanwhile, the OR scheduler was really busy too, so maybe they received the message, put it on a sticky note and on their computer, and it may or may not for whatever reason not made its way into the EHR.
So the point being, if the ability to let surgeons write back and let administrators write back is actually key to making sure that if there is broken process or data that’s inaccurate, we’re finding a way to fix it going forward. Anything else, Sandy? And one last question, and if anybody has any other questions, please use the Q&A widget to send that in. One last question that’s come in so far is, what tools can you provide us with to deal with the surgeons who don’t have faith in the performance metrics we are showing them?
Right. So I think the biggest tool is within an application like this. So remember the fact that we are, when you go to a page like this, and showing Dr. [INAUDIBLE] exactly what is going on in a very simple way, and allowing her to see this for herself. Or if I go back into the dashboard, if you remember, when we went in to say, any of these numbers, being able to see the details with block utilization for a particular surgeon. How it’s defined? Really what we find is reasonable, smart, data driven people react very well when you tell them reasonable, smart, data driven things. That if you– it’s not a question of I said, you said, he said, she said, but if you pull the basis of any calculation, and you show the numbers in a deep way, that’s frankly the biggest tool.
Some amount of surgeon education can be done where we can do Webexes, we can do training sessions, but at the end of the day, if you show people irrefutable numbers, you know the other metaphor I use is, I’m a fairly argumentative data driven person. But I don’t argue with my GPS. I don’t argue with my bank account. I don’t argue with a lot of tools that make sense to me, because when. I look at them, they’re not different rules for different people. I know I can trust these numbers. And I can trust them because all the raw data is right there for me to look at.
Anything else, Sandy, or was that the last one. I think that was the last one. And we’re just about out of time at any rate. OK, so at this point our webinar is completed. And thanks again for joining us for today’s webinar. Please join us next week same day of the week and same time, and we’ll see you then.