Healthy Data Podcast Stephen Estime (UChicago Medicine) & Jordan Cooper (InterSystems)
May 8, 2026, 4:04PM
23m 23s
Jordan Cooper 0:03
with Stephen Estime. He's the Associate Professor of Anesthesia and Critical Care at the University of Chicago Medicine and the Associate Chair of Strategy and Finance in the University of Chicago Medicine Department of Anesthesiology and Critical Care. Stephen, thank you for joining us today.
Stephen Estime, MD 0:18
Thank you, Jordan, for having me.
Jordan Cooper 0:20
As background, the University of Chicago Medicine Health System is headquartered in Chicago, Illinois, has 2,000 beds across 10 hospitals serviced by 3,000 providers. Now, Stephen, today we're going to be speaking about using data to shape staff behaviors and design incentives across a complex work environment. In particular, I understand you've been using operational data to improve staffing and incentives in the anesthesia department. Can you please elaborate on the genesis of this project, what the intervention is, and how it's going?
Click to read the full transcript
Stephen Estime, MD 0:53
Yeah, for sure. So I think it's helpful to sort of provide a little bit of a groundwork in what's happening. So as you mentioned, I'm a clinical anesthesiologist. I also work in the intensive care units at the University of Chicago. I, another hat, I sort of wear several hats, but as a day-to-day clinical operations leader in the operating room, I do that, I do that several times a month, justsort of coordinating cases, scheduling, you know, resources and staff to cover said operational cases. And then on the back end, thinking about high level strategy and finance with respect to the Department of Anesthesia, how we're staffing things and trying to maintain as best as we can operational andand fiscal sustainability and responsibility. So, you know, the one interesting part with the field of anesthesia is that it's sort of unlike other fields in that a lot of our workflow is captured in time. There are other, if you're a surgeon or if you're a proceduralist, a lot of their workflow and the way that they are compensated is going to be based on how many cases they do. Their RBUs, if you will. If you're in a medicine-based specialty, a lot of that is determined by the complexity of patients that you see, the medical, the complexity of the medical decision making,that you're doing for said patients and basically the volume of patients that are coming across your, you know, your workflow. So the fact that anesthesia is time-based, meaning that, you know, we basically, we bill and many times we're incentivized based on how long is the anesthesia team, the anesthesiologist and by extension the nurse anesthetist, in the room during a surgical procedure, start to finish. That's sort of how the anesthesia is sort of done. What this means, and sort of the problem that I wanted to highlight and talk about is, you know, trying to account for the work that's done. by your frontline clinicians. You know, we in the past were using, I'll call it a legacy system where we use Epic, which is our electronic health record. And we were basically reverse extracting data for surgical cases that were being complete. Andlooking then to try to identify how much time our clinicians were spending in the hospital and by extension, you know, how much we should be compensating our faculty based on those numbers. That led to a number of different issues because one of which isyou know, our time as anesthesiologists aren't always cleanly defined by time done in an operating room, which means that things like emergencies, so if you're responding to a trauma in the trauma bay, if you're responding to emergencies throughout the hospital, if there was waits ordelays either between cases or to start cases, that's time that we're in the hospital, but that's not time that's being recognized. And so what that led to was us adapting, well actually what that led to then was multiple systems trying tocapture this time outside of the operating room. And it sort of kind of blew up into having three to four different systems trying to track and account for our time, which led to a pretty heavy lift from an IT perspective and from a data analytics support personnel perspective. And that's difficult in an environment like this. And so instead, we decided to transition and implement a more clinician-facing time and attendance platform to better make some of this information user-generated and hopefully increase some reliability in the data that we wereCapturing, and then, and then also you making decisions on.
Jordan Cooper 4:55
What kind of, you said you're trying to create user-generated data. What kind of platform did you implement in order to enable that user-generated content?
Stephen Estime, MD 5:08
Yeah, you know, it's a third-party platform that we're using, so it's not something that's homegrown. It's not something that, you know, that we created. It was basically a scheduling software that we've been using for years, although the scheduling platform had recently released a time and attendance, meaning thatclinicians could sort of self-document their own time, meaning that they could sort of own that. And so when they were doing work for us with anesthesia, that work that's particularly costly are our night and weekend work. And so when clinicians were doing said night and weekend work, they could capture that time. It's a You know, it's a platform that you can also use both desktop and on a phone, and you can capture the work when it starts and then the work that it that when it finishes. And it's user generated, meaning it's coming directly from our clinicians. It's not something that we're extracting, you know, sort of in the background and then trying to extrapolate. the work that our clinicians are doing. And so I think it does a couple of things. It creates a little bit more integrity in the data because it's the work that is physically being done by our frontline clinicians. And it also helped us because we were able to whittle it down to sort of 1source of truth, if you will. One data source where we could then take that data and actually start looking at things and looking at trends versus trying to parse data out from three to four disparate systems. So it sort of helps to make things a little bit smoother and more efficient.
Jordan Cooper 6:43
Sure. So what are the data sources for this staffing, scheduling, user-generated data platform?
Stephen Estime, MD 6:48
Yeah.
Jordan Cooper 6:55
Is it just, I mean, to what extent is it integrated with Epic and does it have any other external data sources other than kind of clinicians going directly into this third party staffing system and entering when they are in the hospital and when they aren't?
Stephen Estime, MD 7:13
You know, there are other there are other sort of functionalities that this that this platform does, and so there is some data integration between Epic and between this platform where there's sort of bi-directional communication. Those are other sorts of use cases that there will be some back and forth, you know, but but I think specifically for the time and attendance of better capturing
Jordan Cooper 7:16
Mhm. Mhm.
Stephen Estime, MD 7:37
user generated time, this one, there isn't really bidirectionality. And, you know, part of that was intentional by design, because I think that, again, the data that we're getting from Epic with respect to this particular issue was incomplete. And I think in many instances, probably caused a little moreconfusion and it was probably less helpful than just getting this stuff directly from the user.
Jordan Cooper 8:03
So how is a user-generated data validated?
Stephen Estime, MD 8:08
I mean, we basically do audits. So we'll take a look and we'll do audits on the times. You know, it's helpful because I am a frontline clinician and so I sort of understand the work patterns and I can actually see it in Epic as well too, how many cases were going, what was happening on any particular day.
Jordan Cooper 8:16Yeah.
Stephen Estime, MD 8:27
And it's pretty easy to just sort of take a look at a day by day and a month by month basis to sort of scan and look where these things accurately placed in. There's also a free text option that's in this platform so that if there's discrepancies or if there's areas wherethings don't completely line up, then that can be an explanation that the user can create. And so it makes that process a little easier.
Jordan Cooper 8:55
What percentage of cases are subject to this spot checking manual auditing against Epic?
Stephen Estime, MD 9:03
I don't know if I can, I don't know if I can peg down a specific number, but I would say that, you know, me and my team, we will review every single every single data point. You know, so we'll take a look at a high level and just and and and look through what those time punches are, how the hours were spent, and really we sort of flag areas where there seems to be an outlier.
Jordan Cooper 9:12
Mhm.
Stephen Estime, MD 9:26
And if there seems to be an outlier where there seems to be a time where, wow, this certain clinician spent significant amounts of time either in this case or in the hospital this day, we can dig more into that. So we're really flagging for outliers, you know, when we're doing these audit checks.
Jordan Cooper 9:41
Mhm. It sounds like when you're flagging for outliers, you're trying to identify trends or as you say, outliers, that seems like a perfect opportunity to leverage Gen AI. Have you thought about doing so? Mhm.
Stephen Estime, MD 10:14
I think we're going to need more time with this data before we do it. And then also, you know, the questions with respect to data security. You know, we want to make sure that if we're feeding this into some kind of, you know, AI system, that we're doing it with a lot of responsibility with respect to confidential data.
Jordan Cooper 10:17
Hmm. Just to go to back up at a high level for our listeners who don't work closely with anesthesia, you did mention it's a different kind of specialty. It's measured and reimbursed differently than the other specialties. So just at a high level, is all time spent in the hospital or another care setting by an anesthesiologist or nurse anesthetist or someone involved in delivery of anesthesia care? equated with reimbursable time delivering care? Or is there any time where the provider is at risk and is not meant to be reimbursed for that time?
Stephen Estime, MD 11:06
That's a great, you know, Jordan, that is a, that is a very perceptive question. You know, so no, I think that the way by and large that anesthesiologists are reimbursed physically by their time, by third party payers, et cetera, is by that time that's spent in and operating. when you're actually doing said procedure. You know, when that makes it tricky because, you know, I think, as I alluded to before, while that time is the reimbursable time, and it's really the highest value time for
Jordan Cooper 11:43
Hmm.
Stephen Estime, MD 11:44
what you're doing and how we're going to be paid. It's difficult when you're trying to sort of build a department where the reality is that anesthesiologists have to do work that is beyond just work that's in the operating room. And so how do you account for that time? How do you show value for that time? how do you sort of build commitment with the other things that just have to happen in a hospital, like responding to emergencies that, you know, when you're responding to an emergency. So, you know, part of my happen, you know, working as an anesthesiologist, I got to respond to codes in the hospital. Technically, those codes,You know, being reimbursed very, very little for those for those codes, if if at all.
Jordan Cooper 12:29
Why wouldn't a claims payment be issued for care provided in response to an urgency, to an emergency in another part of the hospital? Or would it, and it's just not translating?
Stephen Estime, MD 12:40
You know. I mean, I think sometimes, sometimes you can, but it's pennies on a dollar. And so it's incredibly, it's incredibly low. Plus, you know, that would assume that everything is documented in the right way to actually get that going to claims data. So like we do, I think we do generate bill, we do generate revenue from
Jordan Cooper 13:02
Mhm.
Stephen Estime, MD 13:02
set emergencies. It's just, you know, it can be difficult to capture. And it certainly doesn't capture the entirety of the preparatory work that happens when, you know, I'm thinking sort of doing an emergency airway, for example. It doesn't account for that preparatory work. Sure, maybe thethe intubation or the airway itself might be accounted for where, yes, said procedure was done, but all the time that sort of is associated with that isn't always neatly captured. And, you know, one of the struggles of trying to manage a department where it's a sprawling enterprise or a lot of moving parts is how do you keep
Jordan Cooper 13:28
Mm.
Stephen Estime, MD 13:43
How do you retain your faculty, you know? And I think sort of thinking about all the important work that they do, trying to capture the important work that they do, and then trying to figure out how you reimburse in a responsible way for that work sort of goes in line withhow to keep a department functioning, especially at a place as complex as the University of Chicago. Because, you know, in a lot of ways, this, you know, I think there's a morale component to it because, you know, when you're working with hardworking clinicians frontline, there's this, at times it feels this insatiable demand.
Jordan Cooper 14:23
Yeah.
Stephen Estime, MD 14:23
for expanding service lines and doing procedures, making sure that you've got the, you know, making sure that you can actually get a supply of providers there to keep pace with that demand is really important. And I think recognizing the time and effort that clinicians are putting in for their workis an important part. And that's where this tool has really, I think, helped us in some regards with that.
Jordan Cooper 14:42
Yeah. So where is additional revenue being drawn from when more reimbursable hours are identified?
Stephen Estime, MD 14:55
Well, I mean, I think that's a multifaceted question. Some of that is a little bit beyond me, but I know that, you know, I think the reality is increasing surgical revenue and providing and expanding surgical services is that pop-off lever. I think certainly trying to bereally responsible with these high cost personnel, the high cost services that you have. You've got to be responsible on that end. But I think it's really led by health system growth and trying to make sure that you
Jordan Cooper 15:28
So it's coming from payers, essentially. You're saying that you were going to get reimbursed $30,000 for this procedure, but now that you're able to better account for anesthesia care delivered, maybe it's $34,000. I'm just making up numbers. Right. Okay, great. So you mentioned, oh, actually one last thing on this topic.
Stephen Estime, MD 15:42
Sure. Yeah.
Jordan Cooper 15:48
Admin time. I'm just trying to wonder where pushback could be. Obviously, you have other providers across the health system, not anesthesiologists, I wouldn't imagine too much, but other kinds of specialists who have pajama time. And they often are doing in-basket messages after their kids go to bed at night at home. And that isn't directly reimbursable. There's no charge associated with that. And as you're seeing more, as you're seeing anesthesiology being able to capture more time that is reimbursable, are you seeing other departments say, hey, I want my time to be also reimbursable when I'm not doing an actual procedure associated with the CPT code?
Stephen Estime, MD 16:31
You know, I'm going to pause with respect to the other departments, just because I don't work in a lot of these other departments. I will say that my friends that, you know, are gastroenterologists and, you know, even in primary care who, you know, a lot of their time, you're right, is now spent with doing things like in baskets and doing times that aren't
Jordan Cooper 16:37
Good.
Stephen Estime, MD 16:50
directly reimbursable. I think those questions are coming up more and more. And I think in a lot of ways, it's probably contributing to some of the morale and wellness and perhaps lack thereof, the burnout that we're starting to see across medicine because of these tasks that are being done aren't
Jordan Cooper 17:05
Mm-hmm.
Stephen Estime, MD 17:09
directly reimbursable. It's hard to figure out how you account for that time and how that translates into compensation. I think it highlights a real struggle and a real challenge with, you know, how do we account for that, you know, and how do we do it in a responsible way?
Jordan Cooper 17:25
This. So you mentioned that now you're able to more accurately capture. time that is spent by anesthesiologists delivering care, you want to avoid reinforcing the wrong behaviors and incentivize the correct behaviors in a complex work environment. Can you speak to which, how the intervention, the scheduling system and the user-generated datahas been reinforcing the behaviors that you want to see and incentivizing them to work in the way that you want them to work and how have they it moved providers away from the wrong behaviors.
Stephen Estime, MD 18:03
Yeah, and I think I think that's, you know, one of the things because OR time is the time that is reimbursable, I think, you know, we make it pretty clear that documenting time that can be collaborated with what you're doing in the operating room is where we want the majority of that of that time spent. You know, and so if you are doing procedures or you're doing work that is quote unquote unreimbursable by third party payers, then that stuff has to be documented, adjudicated, and discussed to see whether that's the best way to spend your, you know, set time. So I think that's probably one way of aligning some of that stuff. I think also just, honestly, a lot of this data has been nice just from a conversation starter. And I think having data isn't magic. It doesn't make everything better. But I know that one of the issues that oftentimes comes up in our group is this concept of fairness, you know, and is fairness being distributed? across with respect to work with individuals or with certain subsections within departments. I think people want fairness. That's something that also builds morale. And I think by having tangible data, it helps us to better identify whether these are just sort ofgut feelings, like if we've got, if we have clinicians that are coming in and saying, hey, I feel like I'm working too much or one section is working harder than another section. I think being able to have this helps us in these conversations and, you know, sort of ground level and ground sets it.
Jordan Cooper 19:39
From a data integration perspective, you mentioned that there's just one platform, the scheduling platform with the user-generated data, and that's where you can capture more areas where anesthesiology, anesthesiologists are performing reimbursable work, or you said sometimes non-reimbursable work.
Stephen Estime, MD 19:52
The.
Jordan Cooper 19:58
You also mentioned that you manually review the Epic, the clinical records in Epic to see what care actually was delivered and you kind of try to reconcile those. Is there any other data that's being integrated or reconciled? Are you pulling in claims data? Are you using Epic Payer Platform? Are you using external data sources to accomplish any of this work? Or is it really just there's this one standalone solution and there's periodic manual review between that and Epic and side-by-side screens, but there's no data integration. What are you looking at in terms of data sources?
Stephen Estime, MD 20:39
Yeah, I think those are I think those are important next steps. I think right now, again, because we launched this platform about a year ago, and we've got six months worth of worth of data right now, it's sort of manual side-by-side integration or side-by-side comparison. But I think certainly the future, you know, the future to try to figure out how to better integrateYeah, some of this payer information is there. We haven't implemented that yet.
Jordan Cooper 21:03
Got it. And what would be the business justification for finding a way to better integrate claims data and clinicals data and the time entry data together in one cohesive, normalized, aggregated, and deduplicated data stream?
Stephen Estime, MD 21:21
Yeah, I mean, I think that's, you know, tying all these things to reimbursable behavior and making sure that that you have a better understanding of sort of how time is spent that will that will sort of lead to and increase revenue and where time that doesn't reflect that. you know, sort of how to account for that time. I think I think that can, you know, I think that there's opportunities for that.
Jordan Cooper 21:48
So we've approached the end of this podcast episode. I'd like to pose a final question to you, Stephen, which is, what advice would you have for yourself in May 2025, one year ago when you were starting this program?
Stephen Estime, MD 21:56
The. Yeah, I think, you know, one of the things that I think I learned is that I think data and systems are great. And I think that using more and more powerful data is the wave of the future. It's something that we've got to increasingly recognize. But I think at the end of the day, it's people first. You know, and one of the things I don't think we talked much about here, but one of the things that I really, that I struggled with and that I learned was this stuff doesn't work without buy-in from your people. And I think taking the time to empathize and sort of understand where your users are coming from, and that helped A lot. for myself as a clinician in this, you know, working in this complex environment, I think starting with trust and empathy can ultimately, I think it does a lot of the hard work, the hard lifting for you.
Jordan Cooper 22:56
All right. Well, thank you, Stephen. I appreciate you joining us for this episode of Healthy Data Podcast. For our listeners, this has been Dr. Stephen Estime, an Associate Professor of Anesthesia and Critical Care, and Associate Chair of Strategy and Finance in the Department of Anesthesiology and Critical Care. at UChicago Medicine. Stephen, thank you so much for joining us today.
Stephen Estime, MD 23:20
Thanks for having me, Jordan.