SANJEEV AGRAWAL: Good morning, everybody on the west coast or in mountain time port central. And good afternoon to everyone who is having lunch. As Marianne said, my name is Sanjeev Agrawal. And over the next 30 to 45 minutes, we’re going to talk about how having more open time in the OR, and not a fully allocated block schedule helps you complete more cases. And talk about how many of our customers at LeanTaas have been achieving success this way. What we thought we’d do in this session is actually start with some academic literature, some science, some theory behind why this is true. And so we’ll spend maybe 10 minutes talking about how the OR resembles other highly expensive assets, like highways, like airline seats, and many others in which the demand is not fully forecastable. And so the logic for why you need to have open time to maximize the number of cases you do makes sense when you look at parallels in other industries.
We’ll spend a little bit of time talking about the fact that you cannot just have all open time. There is value to having allocated time in the OR for all kinds of efficiency reasons, but it’s a balance between how much open time you have, and how much allocated time you have. Then go through how you actually create that open time, so let’s say you buy all this and you believe it, what do you do in there? We’ll show you some of the tools that customers of mentors have been using to create open time. So 30 seconds are nuts, who are we and why do we have any standing and being able to talk about some of this? So we’re a data analytics, predictive analytics company based out of Silicon Valley.
We work with about 46 of the country’s top institutions. New York, Presbyterian,. Stanford Wake Forest, Memorial Sloan Kettering,. M.D. Anderson Hopkins, the whole list of others. And really our claim to fame is that we bring together lean principles with predictive mathematics that’s not based on simple averages. We have a team of physicists, data scientists that are looking at data in a very unique way. We package our tools that create efficiency in assets like infusion chairs in operating rooms, in clinics, and ambulatory environments into products. So we have two products on the market– infusion, that is being used by 108 infusion centers, seven of the top 10, 15 of the top 30 in the country. We have the operating rooms product that we’re going to spend a little bit of time on as we go through this. And we’re building many, many others for clinics, for labs imaging, and other assets.
So just to give you a quick look at who our customers are, the 46 that I mentioned they include 15 of the top 30 cancer centers, seven of the top 10. And then in the perioperative environment, and we work so far with about a dozen institutions across 400 ORs. All of which we’re helping create open time in. So all that said, let’s start with a little thought experiment. And let’s start with some analogs from other industries just as a way to ease into the discussion about the OR. So I know everyone’s been on a highway where there has been a lane reserved, I thought it’s called an. HOV lane in California. It’s called a carpool lane everywhere in the country. And oftentimes, it’s reserved for people who are two or two or three in a car.
So if you’re a family or your carpooling in order to encourage that, we’ve decided as a society to create carpool lanes. Now, if you’ve ever observed the dynamic of a carpool lane, it often happens that a carpool lane is not fully utilized. But the other lanes that are free for any car to go in are much more utilized than you can imagine being in one of those lanes looking over at the carpool lane going, what’s going on. You know what’s going on, but the fact is that if you were looking at it from an efficiency perspective on a per car basis, carpool lanes are actually quite inefficient. Now, let’s do a thought experiment, and say let’s allocate every carpool– every lane on a highway to a special interest group. Let’s say, buy make of vehicles. So Toyota can go in the left lane, Ford can go in the lane after that, Volvo is can go into the lane after that. You can only imagine what would happen to the highway.
From a utilization perspective, you’d have far less utilization. And if you weren’t lucky enough to be driving a car that had been allocated those lanes, you have no place to go. If you added a new car make, you’d have to add a complete new lane. And so if you think about the dynamics of what’s going on. If the society we wanted to maximize utilization of the highway, especially during busy time, we should minimize dedicated lanes. Now, there is social value to creating a carpool lane where we encourage that for environmental reasons, but not for efficiency reasons. And the root cause of this problem in the case of the highway is this model would work allocating lanes by car make or color would work if you could precisely be certain the number of cars of each make or color, and how long those cars would stay on the highway.
If you had that data, you could decide how long to allocate capacity to each car or each make of cars. The problem is you can estimate that, but it’s never precise. You can never precisely say between 8:00 and 8:30, how many cars of this make or color am I going to see, where they’re going to enter the highway, and where they’re going to leave it. Trust me, there is analog to the OR in this, and we’ll come to that. Same thing with if you’ve ever driven around in a paid parking garage, right? And the garage is full, you can’s find parking, but there are these reserved spots for all kinds of people, right? It could be for people that have purchased monthly parking. The point being this is another analogy where you’ve got this capacity that in the moment isn’t being utilized, and how does this relate to the OR?
Well, I’ve spoken to about 100 block committees in the last year and a half, including one where they said that the time it took for them to schedule a case for a new patient-designed clinic was over 60 days. And so I asked them, well, if I walk through your wards today, are you telling me all your OR are full and being used by the folks that you’ve allocated capacity to? And they said that’s not true is always open time and OR being used by other people. We obviously fill it up, but it’s not used by the person before whom the block is allocated. Very similar to paid parking, right? So if you think about again, the problem here is there will always be error in predicting, precisely, when someone will use what spots.
The demand is volatile as opposed to fixed. There are many, many other examples that won’t deliver you with the idea, but you get it. Imagine if we started allocating capacity on runways to individual airlines. We’d need many more runways. You’d have fewer flights taking off. The utilization of runways would be lower. And so we pool runways in order to make sure that everyone has a shot at getting their flights in or as you have seen a second cases in. Two last examples, and then we’ll get into the OR in detail. Imagine reserving hotel rooms or seats on an airplane for families or business travelers. How often would it happen that there were more business travelers than we thought we’d need, and but there was no room for them and left families that needed hotel space or airline tickets. And you’d have this weird thing where a southwest plane would be taking off with a third of the seats empty because they were reserved for families when there weren’t enough families flying. So net, net, there’s a lot of theory behind this, but net, net, what’s common about all of these examples is three things.
One, whether it’s highways or runways or hotel rooms or airline seats or paid parking, and by the way, this is exactly true of the ORs. This is an extremely expensive asset. Takes a lot of money to build one. And if you have a set, if you don’t use them maximally in your building or OR before you need them, you’re actually doing yourself a disservice, so it’s an expensive asset. Second, it has what we call perishable value. A southwest plane that’s taking off with a third the seats empty is actually losing revenue. It’s opportunity cost of what you could have made, so the capacity is a use it or lose it type asset. And the third, which is probably the most important is, that it’s almost impossible to precisely predict how many cases of each type in the OR or how many families versus business travelers will I get in my airline seats on a given flight. You can estimate a range, but it’s never going to be as precise.
Now, where does this model of allocating capacity work. Think about our Toyota manufacturing plants, why does it work to say the left assembly line is for Toyota Celicas? Because you can pretty much precisely say I’m going to make 400. Toyota Celica today. And they’re going to take between 30 minutes plus or minus 15 seconds to make because of the automation we’ve done. When demand is deterministic, when you can precisely, more or less, determine how much of that asset is going to be used by what car, in this case, it’s the allocation works. And so the big academic point, if you will, is that reserving assets that are perishable, extremely expensive, and where demand cannot be precisely forecasted usually leads to suboptimal system utilization. All of these examples are just proof points for that.
Now, in the OR, what we have done, and a lot of this has to do with the way EHR map works, and the way block allocation has always worked in the past. We’ve done a good thing. We’ve actually said who are our high volume surgeons and service lines, and let’s go ahead and allocate capacity to them. The problem is, once you allocate capacity, it’s very hard to take it away. Imagine trying to get rid of the carpool lane. And so we find that the first problem is when we allocate too much capacity, we find a similar situation that you find in all these other asset types where on the very same day that. I had Dr. Agarwal who’s been allocated the Wednesday block happens to be at him, or happens to be teaching, or just doesn’t have enough cases to fill into the OR. Dr. Smith who is my colleague and who just joined and who wasn’t blessed by the law committee with permanent blocked time, could be doing cases just doesn’t even know that the capacity that I will not be able to use exists.
And so this is becoming more and more a problem for almost all our customers in a world where capital isn’t available as freely as it used to be with reimbursement pressures being what they are. The market forces that are leading us to do more with less are requiring them to examine how to allocate capacity, better in a world where volumes are rising. How do I take my existing lock allocation? How do I alter it to existing surgeons as well as new ones can be accommodated? So just to prove the point, we were talking about earlier, why the OR resembles many of these other assets we were talking about. We all know it’s expensive. We also know that a wartime is perishable. OR time in your institutions that is not being used today, you’ll never get back.
Cases that are not being done, you will never make– patient access will be lower, revenue will go away. And this is just walking into or talking about the third element of these assets, which is the unpredictability. Not entirely unpredictable, but what you’re seeing on this slide, there are two graphs, these are real customer graphs anonymized. But on the left, you’re seeing within the orthopedics department, the amount of cases done in the amount of time used over time. As you can see, it can range from white noise in the case of some surgeons to a reasonably predictable in the case of others, but it’s very edgy. It’s not smooth. And then on the right, you’re seeing this aggregated at the service line level. It’s smoother at the service line level, but even at the service line level.
There is enough volatility for this not to be a Toyota manufacturing plant, and therein lies the problem with allocating capacity in the face of this kind of stochasticity or variability in demand, all right. A couple of other charts to make the point. You know, we hear a lot of our customers saying that if they had better case link estimation, this problem would go away, it would not. And the reason is that surgery is, by definition, something where it’s the nature of the patient, what happens the day of the nature of the procedure, there pre-existing conditions that determine that the best you can do by wave case length estimation is reduce the error. You’re not going to be able to say– and reducing the error just means that Epic says that the median is x hours and the standard deviation is maybe y hours.
To better forecasting we’ve been able to bring case linked estimations to be better by 20% or 25%, but they will always lie on a normal cave or a curve. They’ll never be precise. And this, again, goes to the fact that not only do we have variability in the number of cases, we now have variability in the length of the case, which you can minimize only so much. So this is like saying, not only do I not know how many but are going to be on the road between 8:00 and 8:30. I don’t even know how long they’re going to spend the night away. And so in that world, if you allocate lanes, you’re just asking for some of them to be underutilized, now others too have a lot of pressure on them. And so amongst all our customers, especially the ones who are getting to the point where it feels like they’re trying to stop a 10 pound weight in a five pound bag, meaning they’re getting to the edge of capacity.
Some of the symptoms that indicate that they’re getting there are, like I mentioned, first, takes me a long time to actually book someone into the OR, but if I walk around my class today and tomorrow, there’s always room. And I’m filling it up, room utilization maybe high but block utilization is low, so that’s the first two points are related. The third is if I can take a lot of add-ons despite having a fully blocked out schedule, how does that compute? OK. Things like I have full morning, but I’m still working late into the night, so my anesthesiologists are unhappy. Almost every– all my staff is unhappy. But it’s because I’ve allocated these lanes that no one else can use, and so. I need to get cases done later in the night. And then, of course, if I have a lot of staff towards some of which should’ve been closed,. I just have allocated them it’s very hard for me to take that time away. So these are some of the symptoms we see a lot, OK.
Now, the flip side of the argument is so why not just have all open time? And again, I’m preaching to the choir here because all of your practitioners of this on a daily basis, clearly, you can’t have all open time because there are constraints like when surgeons do clinic. And so they’re only certain days they can operate. If I’m a surgeon there’s a lot of efficiency in combining similar cases and doing doing doing them back to back. I might need the robotics room. Or I might not might need an alarm to do a spine. And so combining cases is useful, and so some degree of allocated capacity makes sense. And of course, from an administrative perspective, it simplifies your life– you know, who’s doing cases, what staffing levels you need, nursing, anesthesia equipment, all of those things are simplified.
So the answer is not as drastic as let’s make everything open, but what we’re positing is the answer certainly isn’t to allocate everything, all right? So what is the answer? So we’ve taken a very high level view to start with, and then we’ll dig deeper into– say, let’s take all OR time across your ORs and break it into three types of capacity. So over on the right, you see the red box that it’s time for trauma. We work with a lot of customers that have that our trauma centers. And if you’re not one of this, may not be relevant for you. But if you are one, clearly you need to reserve some capacity for trauma. The map can help you figure out how much. If you have 50 ORs, should you keep one room, two rooms, of five rooms? All depends on historical volume of trauma, and this is reasonably predictable in terms of how much time you can– you should set aside. Over and the left, is allocated time. This is what we all do.
We’ve seen at places like university of Colorado Health where we first started working– 98% of their time was completely allocated. Non-trauma time, 98% of that was allocated. There was very little open time. And if we need to hire new faculty member, a new surgeon, we have to guarantee them time. And they certainly got to the point where there was no more time to guarantee. And so they had to start looking at how do we get more open time because we’re leading time on the table, the same time that’s been allocated. So let me just cut to the answer, and then show you the how. For trauma time, reserving time for trauma cases absolutely makes sense. For open time, this is what we’re going to spend a lot of time on. If you can find a way to open up some chunk of allocated time, and we’ll show you two ways and how to do it, magic will happen.
This is like lanes on the highway magically opening up when you need them because you have extra– it’s rush hour, and by the way, it’s gridlock. If you could open up a lane, how would you do it? So the gridlock that happens and allocated time whether it’s used or not, how do you simplify that? And the third element of the strategy is to look at allocated time differently. Almost all our customers and everyone we’ve spoken to, looks at allocated time and makes decisions on block policy, who to take time away from, who to give time to based on a metric called block utilization, which, with, all due respect is just broken math. Block utilization, as a metric for block policy, is broken because days in which I am 10 minutes late to start my cases are five minutes late for turnover or leave 30 minutes at the end of the day because I was efficient as a surgeon. Those are not repurposing chunks of time in which you can put a case in, but you ding me as a surgeon because my block utilization goes down because of those small grains of sand that you couldn’t have done anything with anyway.
So we’ll introduce a concept that says apply the right block policy to the reserved portion of time, and it’s a concept called Collectable Time. So the rest of this is going to focus on how do you create these extra green slices called open time, and do it in two ways. One, by making it really simple to release and request allocated time that we know is not going to be used, so take some of the blue bars or blue blocks and move it to the right and be more of an OpenTable-like situation where you can ask for that time. And then, second, look at the fundamental basis of allocating reserve time, and reexamine the metrics we use to do them, which will also help you create more open time. Because unlike block utilization, Colletctable Time is a far more implementable metric, far more surgeon centric metric that you’ll find easier to implement.
So punch line is open up more time by doing two things. 1, help create larger pools of time by providing visibility and transparency, and encouraging ways to predict who is not going to use block time well, and getting them to release it. So we spend a lot of time on how do you get a culture where releases work. And then how do you get those releases in front of everybody else that could be doing cases, and fill the whole on a daily basis. The second is collect. Instead of using block utilization, let’s use Colletctable. Time to take time away from the right block owners who are leaving large enough chunks of time on the table.
So let’s start with the first tool. How do you create open time using what we call a day trading metaphor almost, so time’s been fully allocated. There are 50 of us that have a block each a month. But life doesn’t work according to a block schedule. I take two weeks of vacation. I teach. I have clinic. I go to him, then I speak. When that happens, how do I release my time? And how do you make it available to everybody? So imagine if you lived in a world where all your clinic schedulers, your surgeons as well and maybe if you wanted access to it you as well. Your clinic scheduler that are, by and large, in control of your surgeon schedules. And some surgeons had access to a mobile browser-based tool that allowed them to see open time, and request chunks of open time based on cases they needed to do.
So I have Wednesdays, but in addition to my Wednesdays, I have a backlog of cases if I could go in. And I could look at a calendar of open time and request a piece of time that I know I’ll be able to use because my next allocated block is two weeks from now, I have three cases that can get done. Can I find some open time in your life quickly? When I request time,. I can put in a request for special equipment. I can put in request for a robotics room or something that I need. It gets better. What if you could live in a world where in addition to being able to request open time, you could say, if there is no time open on a day that I want to do cases, I put my name on an Amazon wish list. So if time opens up, you can let me know.
On the flip side, if I’m in Hawaii on vacation and realize, oops, I forgot to release my block. Myself or my clinic scheduler could go in. And in two clicks, released that time and to an open pool that would then go on the other side of the equation and be seen by others at open time. This is essentially. OpenTable for open time. And the reason this works so well is on the margin, this is not taking block away from people. It’s saying, I am only asking you to release time that you wouldn’t use, otherwise you still have that Wednesday block Dr. Agrawal. That’s one Wednesday when you’re teaching or on vacation that you’re putting it into the open pool, so no one loses. Access to the or especially for those not blessed by the or our committee goes up a lot.
In fact, the. University of Colorado health, who I mentioned earlier, they were able to hire 11 new surgeons and absorb them because when they moved from here, I don’t have anymore time to allocate because I’m fully blocked out too. I’ll give you a tool where you’ll always find open time. And by the way, as you build your book of business you prove to me that you need permanent block. Are you please willing to work with this tool, and your office willing to work with this tool. The whole process of request and released by the way leaned out. So I don’t know if this is true in your world, any of your world today, but I’ve sat in a clinic clinic schedulers offices and our schedulers offices. And the number of phone calls and emails and text messages and doctors walking in demanding time, and post it notes that I see on their computers. Frankly, it’s ridiculous. That in 18, we’re living in a world where all this is happening by paper and pen, and by people chasing each other to release time and request time.
So this idea of creating a much more transparent visible over the air process where I could request and release time. Now, truth be told, a majority of the time this what is used is the web version of this mobile tool. I’m not going to go through this entire tool, but my point is make this available to clinic schedulers who are supporting multiple surgeons to be able to look out and ask for open time just like I showed you on mobile, this is the exact same thing on the web. Now, that’s just stage 1 of this tool called exchange. Stage 2 is far more interesting, frankly, from someone with our backgrounds who trying to do predictive analytics to help you release capacity. We know with your data and you know this, that when you look at the lead time between when different service lines and surgeons do cases, the difference between when the case is done and when the case was booked, that’s the lead time, can be very small, it can be held both cases regularly 2 to 4 days out. But there are many, many service lines.
I’m just highlighting an example, in this particular case of one of our customers, anonymized where orthopedics tends to book 71% of their cases way before seven days. And a lot of our customers, when we first walk in, typically have a seven day, 14 day, four day auto release within their EHR or whatever their EHR is. Frankly, that’s a blunt instrument, you’re taking a baseball bat to something that actually needs a scalpel. Because if you look at the fine grained patterns of the lead times, what you can do is you can be far more precise about trying to recover time and getting people to release it. Here’s what I mean. Imagine that 80% of my case their book 14 to 20 days in advance. If I have block time 13 days from now, it’s completely appropriate, much like a dentist’s reminder, for me or my office to get a little text message saying, hey, you know Dr. Agrawal has signed 13 days from now. It looks at odds with his behavior, so on the Grid and Epic. I don’t see any cases. You know, would you consider filling this block up if you accidentally forgot to put cases on it, or consider releasing this time?
Sounds like this is not time he’s going to be able to use way before your seven day order release or whatever the generic assumption is for auto release. Now, I’m Dr. Agrawal and I’m so important. I can say, I can ignore this. I can say, you can keep sending me texts messages, I’m not going to release anything. Fact of the matter is after 10, 15, 20 weeks, now you have a lot of data. Within the application, you’re able to see, OK, I sent you 15 text messages. That’s fine you ignored all of them, but on four of those days, your block went completely under utilized you know the. OR did all kinds of gymnastics to fill it up, and yet so room utilization is OK, but block utilization was essentially zero. Or in five of those days, by the way, Epic auto-relased it. And when you auto-released it, it forced many service lines, it’s too late.
Having a seven day visibility into being able to get my patient in given how long it takes to get them to agree to come in all the processing and the paperwork I need to do in the insurance work that I need to do. Seven days is just not enough time. So the goal of this tool is to create a marketplace on top of your EHR, create more inventory for you to be able to provide on the margin to folks that need extra time, without hurting anybody’s practice because they were going to use that time anyway. Now, if release reminder capability, like that dentist’s reminder here, you’re going to show up, works really well hand in hand with this other little capability called the Wish List. Meaning if I want time, 10 days from now outside my block, and none is available, but. I know I can use it well. What if I just allowed you to put your hand up as you saw in the demo, and say I’ll be able to use it if you give it to me. It’s very, very powerful.
OK, because this sometimes unearths latent demand that wouldn’t surface otherwise. Because if you’ve just given me Wednesday block, when I’m in clinic and. I’m seeing patients, I think I only have. Wednesday capacity. So even if I could see more patients, what’s the point? I couldn’t get them in the OR for months. Now, when I know that in addition to my Wednesday time I have some more flexibility. And I can actually get some time maybe two or three hours on a Friday or a Monday that other people won’t use without them losing their block, but the meaning being able to sort of ride on the lane if you will. That’s tremendous. And so this idea of unearthing the demand signal and unearthing supply.
Think about the Wish List that unearthing the latent demand signal amongst your surgeon base and the release reminder that’s unearthing the corresponding capacity because what happens if you put your name on a wish list two weeks from now, the system not John, or Jane, or me, or a consultant, the system goes back and says OK there’s demand for time on Friday. Who has time on Friday? And should have put cases into the grid by now, they haven’t. Let me send just those five surgeons in their offices a release reminder and try and unearth that demand to make the demand supply match happen. Now, you don’t want to lose control of granting time to the clinics.
Obviously, if I asked for an OR, they need to make sure that there is no equipment conflict. If I ask for the robotics room, you need to make sure that happens. So the only difference between this an OpenTable is you don’t consummate the transaction at the point of the demand signal. You let the OR manager determine whether to approve or deny the request, but their life changes from all these phone calls, emails, text messages, post-it notes to one streamline portal where every release and request is sorted for them, and they get to examine each one of them saying, oh,. Dr. Agrawal wants the alarm. Eight days from now, is the are going to be open? Is anesthesia available? Are we stopped for it? Is this in alarm that we can make available.
Once I do the checks, my life is that simple as approving or denying that request. And if I deny it, I deny it for a very specific reason. Hey, that could be the third alarm we only have two. So I don’t feel like you’re beating me up. I know that it’s a fair, egalitarian, transparent, visible process for everyone. It’s not as if you know Marianne gets the table for when she goes to OpenTable, and. I’m refused the table before because the. OpenTable doesn’t like me. So a lot of the cultural issues we talk about, frankly end up being the circle of trust. Do we trust that the intent for a right down from the OR committee to a service line cheap, to an administrator, to a scheduler. When you create this depersonalized capability of anybody can again ask for time, and anybody can get a release reminder, magic happens. Because the same highly argumentative extremely data driven people like me will never ask a human being for directions.
But when I get in my car, and my GPS says take a left, I take a left. Because I believe this is a tool that is egalitarian and fair and knows what it’s doing, and it’s not pick singling me out and sending me left. So the point of this tool is to create a small open marketplace on top of your existing EHR without disrupting your block schedule, one way to remember this tool called exchanges, no block left behind. OK. And so if you’re having situations where people have a lot of blocked aren’t using it well all the time because seasonality has a big part to play, just to give you an example. Nothing I’m sharing nothing with you that I haven’t asked our customers to approve of.
At Ohio health which is a customer of ours, phenomenal experience they had with this where we got their clinic schedulers around the table and said, do you know your surgeons’ summer schedules? And they said we do. In about an hour, more block time was released using this tool then in two years put together. So this idea of we know our surgeon schedules way out into the future. The point isn’t to try and release capacity the day off or two days from now. The fact is there is no open transparent tool built into the system to be used for scheduling today. So last slide on this to show you some results of this tool at one of our customers and I can repeat this for about five or six others that are using it.
First is the in Metro Denver across 38 ORs at University of Colorado Health you’re seeing two graphs here. The one on the left at the top is the amount of locks released over 18 months, and on the right the number of locks requested in that same period. The inflection point where releases really took off was this release reminder capability. And then on the right, is the number of locks requested. This wish list was magical in unearthing demand. And so if you think about the numbers that they have published with us. OR utilization going up about 4 percentage points being attributed to this tool. Like I said earlier, 11 new surgeons weren’t given permanent block time upfront, but had used this tool to get to the point where they could ask for block.
Surgeon happiness is tremendous. They may or may not even know this tool exists, even when they talked to their clinic schedulers. In a world where their clinic scheduler used to tell them, in order to get your time, I’m going to call John or Betty or Joe and see if I can get it versus, hey, I’ll go into an open tool called iQueue exchange, and I’m going to request time. And I know for a fact, that if that time’s available, I’ll find it or I’ll put my name on a wish list. This last metric is huge and I know you know it. But in addition to all the financial benefits, this idea of release sleep time being almost 20 days. Imagine having chunks of time available to you that you can ask for three weeks in advance as opposed to just the one week of auto-release. So that’s the first tool. Remember we’re talking about two ways to increase open time.
One is no block left behind by extracting the capacity that is not going to be used by people that own it and making it available to people that need it. Think about that as a day trading in the financial context. So going back. Just to ground us, going back to this slide. Remember when you said, you can make open time on the margin by exchange. And then now, we’re going to talk about item number two which is your block policy. So block policies that are based on block utilization are completely broken. And I don’t know if you found this, but we found many, many folks that when you try and implement a block policy based on block utilization, here’s the problem you run into. And there is math behind this. It’s fairly simple to understand, but it’s nuanced, and so let’s spend a minute on it.
When we look at block utilization, block utilization is an average. And the way averages work is if my head is in the freezer, my feet are in the oven, on average, my block the mind my body temperature is OK. And this is what happens a lot. So imagine that I, in Surgeon. A, who has 10 blocks of quarter. This is my epic grid and I use my 10 blocks with wheeled into wheels out being the dark blue. Surgeon B is that let’s say a neurosurgeon, and they also you have 10 blocks a quarter and that’s how they use it. Now, imagine on the left. I’m a surgeon who does eyes, or hands, or I’m the type of surgeon that does fairly predictable length of cases. I do for two, four and six hour cases because I do knees, arms, fingers. OK. And so I end up using 7 and 1/2 days of my 10 days really, really well. I packed the award for 7 and 1/2 days.
What is my block utilization? 75%. Imagine I’m Surgeon B, remember that variability chart that I showed you of length of cases, I happen to be one. That on January the 3rd, the same case that I did six to seven hours long. For the same patient type when. I did that case on February the 1st, it took five hours. Why? Because surgery is fundamentally not like making Toyota cars, you know this better than I do. It varies a lot. And there is no way anyone can predict a priori how long it’s going to take. It’s after the fact that when. I look at Surgeon B’s grid they use all 10 days. And on average, they use six out of eight hours each day. What is their block utilization? 75%. So now, think about what we’ve just done. We’ve said they’re too extreme surgeons, head is in the freezer, feet are in the oven. Both of them have the same lock utilization. The basic problem with block utilization as a metric is that 1 minus block utilization means absolutely nothing. What does it mean to have 76% block elevation. Does that mean you can take 24% time away? No. And now, when I have an OR committee meeting and I have Surgeon A or if I did try and take say two blocks a quarter away from them, wouldn’t actually hurt them much.
This is their regular pattern of behavior. Using the same blunt instrument called block utilization, then I go to. Surgeon B and say, I need to take two blocks away from you. Good luck with that, there is no way they can fit their elective practice in. And now, it’s blocked utilization is what makes me an MVP. What would I do as Surgeon A? I would consistently release two blocks. What would my block utilization be? 75% over 80% over 90%,. I’m going to be a hero. I get that parking spot everybody wants. So the problem with block utilization is many. Instead, let’s do a thought experiment and introduce you to this concept called Collectable Time.
Think about what is the problem we’re trying to solve with block policy. The problem we’re trying to solve with block policy is to identify repurposing chunks of time that could actually be used to put a case in. So the smallest unit of time in an a OR is not a minute. The smallest unit of time in an OR is the smallest length of case that you can get in there. So I’m going to ask you for a minute to ignore small grains of sand. So on January the 3rd, if I started my case 30 minutes late shame on me. It’s terrible for the patient experience. Just don’t believe that those 30 minutes are going to give you extra access to fit a case in because all that would have happened had I started my case exacting on time is that I would have had 30 minutes left at the end of the day. What exactly were we going to do with it.
So instead of focusing on these little blue triangles that are small grains of sand, that even if you collected you could do not much with, you should try and collect them for the patient experience. It’s just not something that will increase access. Instead of that, why not focus on truly repurposing chunks of time that you could fit a case in. What are they? There is a large red diamonds. Number one, if I am leaving large contiguous chunks of time on the table, whether it’s in the middle of the day like on one 10, the end of the day like on two 22, or the beginning of the day on three 1. There are many surgeons and service lines and surgeon groups that use the morning really well not the afternoon so much, but still have the entire block. That would be one example.
The second is, if I’m abandoning days, that’s a travesty. This is a sign of this plane that’s not even taking off, forget about carrying few passengers. And the third is, this is probably the most nuanced. How much block release should you allow? So if you’ve given me three blocks a quarter or two blocks a week, let’s say. And I have 156 blocks a year, and I’m releasing 60 of them. In theory, my block utilization could look great like Surgeon A. But the fact is you’ve just given me too much time to begin with. So for a moment if you start looking at the world as, forget about this metric that we’ve taught ourselves as being relevant. Block utilization, as an input metric, leads to that Surgeon A,. Surgeon B situation a lot. Colletctable Time, I don’t care if this was Surgeon A, B, or C, or what practice this is.
If you’re abandoning days, if you’re leaving large enough chunks of time someone else could have used, or if you’re abandoning, or, if you’re releasing too much block how does it matter what service line I belong to or who I am or what my specialty is? Time is time is time, OK? Then what the tool allows you to do with. It allows you to configure these characteristics like how much contiguous time could you put a case in. So orthopedics might be 2 and 1/2 hours is enough to put a case in. Neurosurgery might be five hours. And what we give you through this tool this tool called Collect is show you the lowest hanging fruit by day or week by block owner, by block owner, by the week, by location, how much of their allocated capacity is collectable? So questions like, I’m hiring a new surgeon who can only do cases on Wednesday, where should I find time?
Look at the collectable table, find the Wednesday, find the folks that have a lot of collectible time on Wednesday. How much time can I make available in this location? Add up all your collectible time by day a week, maybe you don’t need to run or 27 hours, maybe you should only be running 23 if there is no demand for it, right? But that discussion with surgeons, when I go to Dr. Dyer, and say Dr. Dyer, you have 27 allocated blocks and we think we can take six away without hurting your practice. That is no longer based on a fight about why I start my cases late for the 15 minutes you’re trying to recapture from me as if they will make a difference. Instead, you have a far more airtight argument to make with surgeons, so when I go to Dr. Dyer and say you’ve got 27 blocks and I can take six away. I’m showing him not block utilization is 68% in our policy 70%. It’s saying, look, let me show you a total pie of allocated time.
Let me show you how much time you’ve used well, and I also give you credit for the little blue triangles. Forget about those, I won’t deny you for small grains of sand, I’m only showing you the egregious amounts of time you’re leaving on the table. So I’m only penalizing you, if you will, by showing your repeated behavior over the last month, the last quarter, the last year where you’ve either not used a full block or you’ve released more than the 20% capacity that we decided in our. Block policy was, OK. Or you’re leaving more than 3 and 1/2 hours in this case of time on the table. You don’t believe me? Let’s double click. Let’s go into each of these chunks, and show you exactly down to silicon. Down to the Epic Grid or the Cerner Table to show you exactly when you’ve left this time on the table.
So your ability to take back this time and put it into open time goes up a lot when you use this kind of an argument versus a Surgeon A, Surgeon B block utilization argument. So the funny thing is that if you want to increase block utilization, don’t focus on block utilization. Focus uncollectible time and reducing the holes in the Swiss cheese at the right holds, not the small holes, the big holes. Those are collectable holes. OK. And so I’ll end on one slide, which really is a piece of totally mine and a lot of people in this field. We’ve gotten used to looking at data historically and making decisions based on, with all due respect, the open metrics. OK. It’s sort of like getting up on a weighing scale and realizing, yes, this one says I’m 10 pounds overweight, the other one says it’s 9.9 pounds overweight.
Fact is I’m overweight, what am I going to do about it? So when we look at the analytics, we typically see our customers use the boat they usually fall in one of the first $2 here. They’re either descriptive, I call that admiring the problem. Oh, I got on a scale again, and it says I’m overweight. Well, great I knew that yesterday too. So I have a lot of data on block utilization, minute use, case volume, first case delays, turnover time. In some cases, we see accurate diagnostics why it happened, but that again, is kind of admiring the y. What you really need to move to is prescriptive analytics based on predicting what’s going to happen. So if you think about exchange, predicting who is not going to use time well and getting them to release it.
Looking at historical usage of time as collectable time, and turning the page from landscape to portrait mode and saying there’s a different way to look at the same numbers that is far more acknowledging of the fact that surgery is not like building Toyota cars. The problem with metrics like block utilization is we believe we’re building. Toyota cars when in fact we’re opening people’s brains up and putting them back together. The two are completely not in sync. So that’s how you increase open time. You make time available on the margin by getting people to release it and others to requested, and open transparent tool called Exchange. And you start looking at your basic method for block allocation. And looking at your block policy in ways that are a little more sophisticated than block utilization.
All right, so it looks like there are a few questions here or at least one. Can you share any benchmark targets for the amount of open time an organization holds available, and what you believe are the determined variables? That’s a phenomenal question. So let me go back to one extreme of a Toyota car factory. How much open time should they leave on the table? Imagine that they got all their cases of cards exactly in 30 minutes or 45 minutes or an hour. They don’t need to leave much open time, so the big determinant of how much open time you should leave, it’s not a straightforward answer, it depends on this. It depends on your data. If you were, for example, if you’re doing just eyes, or if you’re doing just orthopedics with fairly spiky case lengths, which can be pretty much assumed at the beginning.
You can get away with far less open time. But in general, if you are a multidisciplinary hospital, I don’t want to throw numbers out without any basis for them. We’ve been able to uncover 20% to 30% of open time. I mean if you think about the map for a second, everybody goes on vacation two weeks a year. 2 divided by 50 to the bare minimum amount of open time you should have because that someone’s out. Now, add, just that the surgeon’s not being available for a chunk of time, it’s not hard to come up with a number like 10% or 15%. Depending on demand, can you actually use that time. How you use the open time is also interesting. You could hire new faculty and new surgeons or staff.
You could try and take business away from competition by saying, I’m the only community facility for example or academic facility that’s showing you availability of time on a mobile app. So there’s a number of variables, these don’t take 20% to 30% as an official answer. This does depend on the types of cases, you do in the mix of cases, you do, and how predictable they are.
MARIANNE BISKUP: And I want to remind everybody, if you have a question, we love the engagement. You can utilize the Q&A icon in the upper right-hand corner of your screen. It looks like we have at least one more question. How can we convince surgeons to look at their OR time in this way, it seems like a change culturally.
SANJEEV AGRAWAL: Yeah, so what we’ve discovered is that trying to implement these tools without surgeon by end is hard. So whenever we go in and do this, we usually like to have physician champions as part of the team early on in the process. The good news is that these are very sophisticated human beings that have studied for many, many years to become who they become. And when you show them data. And you show them the logic. And you show them we’re increasingly OR access. And you show them analysis like the lead time to booking a case from a new data than versus when you did it to versus when you booked it.
They believe numbers, they believe facts. And so we’ve never run into situations where once we’ve gotten in front of the committee, what we’ve gotten in front of surgeons that this hasn’t flown because it’s just makes sense. We live in a world where we booked restaurant reservations through OpenTable. Why in God’s name are we calling people and e-mailing them and walking through their doors to try and get access to time. So honestly, we have not found it that difficult to convince surgeons this is a better way of doing things rather than just doing more of the same. How do you initially create the open time? I think the process likely works well once the time is– How do you create stakeholders put the trust into this?
Another great question. So what we find is you could be in one of two situations where there is a little bit of open time because you don’t have enough demand and you could start there. That’s the easy answer. I typically have a few hours a week where you can start. If you go back to any marketplace. How do you fill an empty bar? How do you create a demand supply situation? Believe it or not, you don’t need 100 percent of clinics, and 100% of surgeons to start playing ball. On the margin, if you can convince a small number of physicians who get the fact that liquidity in a marketplace is their friend, if you can get 10% to 20%. Let alone 10, if you can get a few people to start releasing and requesting time.
If you can get people to maybe put a rule like you can only request time if you release it. The experiment I talked to you about where we got a bunch of clinic schedulers together. There’s another experiment we did at a hospital which allowed this where we had a little contest. We had a little $50 Starbucks card I was going to say, but really it was more like a prize-based on contest saying you know when you’re your surgeons are going to be on vacation why not just release it. And we’ll put your name on a leaderboard. The incentives that it takes for a few physicians to play ball isn’t that high. Usually it’s the ones that are actually hurting for time, and need more time. And their desire to play ball is much higher.
So I agree with you that in order to kickstart this, you need a little bit of fodder, but you’d be surprised at how little that fodder is to ignite this. More questions. What are the costs– sorry there’s a few others up there. What are the costs to the institution to implement this process? We can certainly talk about the costs of the tool. I think the more important thing is– are a few of the things that have been addressed already. One is your clinic scheduler your or schedulers and your surgeons need to get the fact that this is a far more efficient way of doing things. Once you start using something like a collect as your only way to release and request block time, magic happens. It’s sort of like. OpenTable the only way I make restaurant bookings anymore.
Costs are dependent on the number of our costs. The costs you pay us, are dependent on the number of awards you have. They’re dependent on whether you do both exchange and collect, or just exchange or just collect. And there are other modules we can talk to you about. Certainly, this is not meant to be a sales pitch, so I’d certainly not want to go down that road. Happy to give you a demo. Fact is the ROI on this is about 10 to 20x because 1 percentage point improvement in block utilization is worth hundreds of thousands of dollars per OR to you. There are other questions here. After implementing this product and educating each stakeholder, how long did it take for them to trust this process on the data? There’s a method to this madness.
If you start with exchange first, and everyone gets the fact that I’m not– I’m giving you something before. I take anything away from you, it takes far less time. How can I disagree with a much more streamlined process to request and release time? That doesn’t actually take a lot of trust, that just takes looking at an app. And seeing, hey, I got tons of time available. And so enrolling this out. We always try and get exchange implemented before collect, and start educating each stakeholder. Typically, implementation of these tools takes about 90 days depending on IT readiness and your availability of clean data. And then within three months after that, you should be able to see some significant results. Because of which, in fact for me, the cost issue came up earlier.
We actually have a six month money back guarantee for any of our customers because we’d like to take the risk with you. We understand this is not something you can implement by like putting a Cisco router in the closet. This does require going in and creating capabilities that don’t exist today in your environment. And that’s why from a pricing perspective, we certainly make this as risk free as possible for you. Do you recommend any rules around which cases can be booked into open time first come, first serve as there for some urgent shortly-type cases? Yeah, great question again. I’m assuming that if you have urgent emergent cases, you have trauma capacity or other capacity reserving anyway.
What we have found is because there is– we found first come, first serve as a great place to start. And the reason, first come, first serve as a great place to start is, as soon as you go down the path of saying, hey, a neurosurgeon case is more important than an orthopedics case. You start to get into this situation where people trust the system less. You’ve just told them you’re giving a fair egalitarian, simple, transparent way of accessing time. And then you’re telling them, no, but if you’re in orthopedics, that doesn’t apply for you. And so first come, first serve, from a cultural perspective, is the easiest one to adopt. The fun part for us as mathematicians is now you can build all kinds of rules into it. If you really wanted to, you could end up building an Amazon buyer and seller score, or an eBay buyer and seller score.
You observe all kinds of patterns in nature. How many people cancel their. OpenTable reservations? How many people do people don’t even show up after making the reservation? The beauty of the system is it’s all logged. If I’m the type of dog that hoards all my time and never plays over the team, you’ll see it. I never released my time it’s always auto released. If Marianne is the type of dog that’s far more willing to play and release of their time early, you could put rules in that set people who release early get preferential access. Because you know technology, at the end of the day, is pretty dumb. It just does what you ask it to do.
But instead of complicating it too much, if you keep it simple, first come, first serve, it tends to get most of the impact you need. All right, so just to wrap up folks. This has been delightful for us to host. Hopefully, this added some value to your thinking. As Marianne mentioned earlier, if there are other questions that you’d like us to follow up with you on, and/or you want to dive deeper into any of the tools or see an actual demo, then you could text 6308845492, just send us an email at firstname.lastname@example.org and we’d be happy to set up time with you.
MARIANNE BISKUP: Yeah, so thanks everybody for tuning in today. And a big thanks to Sanjeev for spending some time with us. I want to remind you, keep an eye on your inbox we’ll be sending out a link to a recording of this webinar shortly. And again, as Sanjeev mentioned, if you still have questions, reach out to us. We strongly encourage the engagement. Thanks again for tuning in.
SANJEEV AGRAWAL:. Thanks, everyone.