So, good afternoon. I'm Jerry Lauder, and myself and John Shepard are going to be reporting on the work from the OGS Payment Reform Committee. This is going to be very much a work in progress talk on our work on alternative payment models. So, neither of us have any disclosures. First, I want to recognize all of the volunteers from OGS that have been working with us on this endeavor. This is an all-volunteer group. We have a wide range of folks from private practice to academics, from one fellow all the way up to senior members. And we have several liaisons that have been important from the Executive Board, the Quality Committee, the Coding Committee, and, of course, Jill has been guiding us through this process. So, to give you an idea of kind of where we started, because we're a very young committee, and, once again, this is a work in progress, I want to share kind of what our goals were that were set forth by the Board. So, first, to educate its committee members, because most of us, this was far from being on our radar, on the process for advanced alternative payment model development, as well as to learn from other groups, because there are a few other societies that are ahead of us and have been able to help us with this. And then, to gain expertise in the development and execution of APMs, as defined by CMS. So, our charges, starting back in 2016 to the present, was to establish OGS as a resource for developing APMs for our society. And so, as we learn more to be a goal to educate the membership of the new world of physician payment, this meeting is part of that. Our first charge, as far as an APM, was to create an alternative payment model for stress urinary incontinence. And while that, in some ways, may seem like low-hanging fruit, we've found that it's not as simple as we hoped it would be. Eventually, it will be to teach members how to implement APMs in their community. We're going to need to be constantly monitoring the regulatory environment and ensuring that OGS continues to increase its voice and influence, and then work with other committees to advance these goals. So, that will be the Coding and Reimbursement Committee, the Quality Committee, and ACQUIRE. So, I wanted to give you a little kind of work to date. You know, it started back in 2015 when Ryan led a task force. So, it was the APM Task Force. And, you know, one of the things we found in reaching out to members within the society in a survey that had gone out was that, like a lot of the members of the committee, it was not something that was on our radar but was something that members wanted to learn more about and that we needed to try to get, if not catch up to the curve, try to be a little in front of. After that task force was over, Ryan presented the report to the board. Subsequently, there was a call for volunteers to develop a committee. And there were volunteers. You saw the ones that had volunteered. I was asked to chair the committee. And at the last OGS meeting in Denver, Colorado, we had our first in-person meeting. And some of that was education through Jill. But we also had from that meeting our first kind of assignments for the in-person meeting that we're going to have in November. And what we wanted folks to do was to create all the scenarios they could for the evaluation and treatment of stress urinary incontinence and get those to me before the meeting. And then we were going to work on those at that point. The November meeting in D.C. was when we had a more involved education session, and I'm going to talk about that in a minute, and then small group work on the SUI scenarios. We've also been doing conference calls almost monthly. And recently had a conference call with Medtronic to try to get the claims, Medicare claims data use, because we need this to be data driven. Unfortunately, Medtronic has in the past or recently has made their purchase of the Medicare claims data use available to some other societies for just this type of work. So we are working to get use of this data in kind. So to go over some of the work from the pre-in-person meeting, once again we were asked to create all the scenarios, plausible SUI clinical scenarios, evaluation, treatment. So nine group members at that time created 72 total scenarios submitted. The median was about 10 with a range from 4 to 12. There were 32 unique scenarios. So what I did is I compiled all those and basically grouped them into eight SUI categories based on similarity. Once again, thinking that stress-unit incontinence will be kind of low-hanging fruit, you can see just by similarity came up with a category of simple SUI, which you also may see referenced as straightforward SUI. SUI that does not undergo surgery, stress-predominant mixed, urge-predominant mixed, prolapse with stress incontinence, prolapse with occult stress incontinence, recurrent stress incontinence, and other. Obviously to create one advanced alternative payment model for SUI is not going to work. So I took all that data and kind of mapped each step of the evaluation and management into Excel and calculated frequencies. So this was just from the example scenario of simple stress-unit incontinence. Once again, this is nine members. So if we opened it up, for example, to the whole body of the organization, the variation would be even greater. So just as an example, about 43% would have done an office urinalysis. About 14% would have done a formal or sent it out to the lab. Forty-three percent would have done a urine culture. Fifty percent, a post-void residual. Cough stress test, 71%. Urodynamic, 64%. So once again, just within nine folks, there's a lot of practice variation for a workup of straightforward or simple SUI. And so this is an example, and some of this print maybe is hard to read, but all the different titles for the scenarios down the vertical axis. And these are all the steps in the evaluation and then treatment. And then obviously color-coded by the individual. And you can see that even though these scenarios varied slightly in title, a lot of the workup and evaluation were similar. Now when we got to the meeting, the first half of the day was education. And so Jill, in the beginning, gave a great talk on MACRA and APMs. We then brought in outside guests, Walter Birch from the American Society of Clinical Oncology and Matt Cuffron from the American College of Surgeons, who already have advanced APMs much further down the road. And they gave us their experience, which was invaluable to our very early working group at that point. We also had Dr. Margie Olson from WashU, who's the director of the Center for Advanced Data Research, come. And she was important. Not only does she have the databases of Medicare, MarketScan, and others, but she really gave us a great idea of how to work with claims data. Because once again, we need this to be data-driven, and these are sources that this kind of data is going to come from. In the afternoon, we did a small group work on the SUI scenarios. And that was important because we needed to decide if there's further steps that needed to be elucidated within the scenarios versus could some be consolidated. We wanted to determine if additional testing, patient education, counseling time needed to be added into the steps, because we really wanted it to be all-inclusive, partly because when we go to the data, if we don't ask for it, we're not going to get it. So we needed to identify kind of all the plausible steps that may be able to be coded for. Determine the E&M visit for each level. And then also just begin to start identifying evaluation management steps that are either not currently performed or should be performed that are not billed for. And so these are steps that in a new bundle we may create, quality steps that we already do in our practice, but we don't bill for. And then also to further determine if some of these scenarios could be consolidated or need to remain separate, because there may be some scenarios that an APM just will not fit for. So you saw this similar slide a minute ago. So from the previous scenario could fit in one slide. This is now just the evaluation portion. And what we also wanted to start adding in are potential sites for quality or markers of quality, because as part of the advanced alternative payment model, there have to be quality indicators, quality markers that are very similar to the MIPS. And so kind of in our brainstorming session, people thought about assessment of urethral hypermobility, neurologic exam, assessing for prolapse, also obtaining validated measures. When we look at patient-centered outcome measures, a lot of us collect those, and it takes time to do them, it takes time to review them, the expense of sending them out, et cetera. And that's not something that we get reimbursed for in any fashion. So this was just the evaluation part. This is now just the management section. The other things we thought about was weight loss counseling or tobacco cessation and counseling, things that may improve our outcomes for our patients, but we can't necessarily bill or code for unless we do it under time, except in certain scenarios. So we brought up the idea of could a nurse navigator or care coordinator improve outcomes for our patients. We also expanded the range of treatments. Once again, probably not many people do a Kelly-Kennedy plication, but if we don't ask for it in the data, we're not going to be able to determine if folks are still doing that. We then, with the follow-up, added in postoperative teaching for CISC if needed and multiple other steps to make this more expansive, in part to once again identify the variables that we may need from the databases, but also as we create the model and create bundles in the future for an advanced APM that we would have the most inclusive point to start from. So once again, what was the goal of developing these clinical scenarios? First thing was to clarify, because once again, we wanted to have a general range of the possible E&M steps for the SUI scenarios, and also to try to assess the different SUI scenarios that clinicians face, because once again, a private practitioner in a small town may see different SUI scenarios than someone at a tertiary academic referral center. So as we're creating an APM for the future that's going to be useful to the membership, obviously we have the full spectrum of providers. Once again, we needed to identify the variables that could be found in an administrative claims database that we would use in modeling. Then we also wanted to cite steps in the clinical pathway where existing quality metrics needed to be placed or that may need to be developed. Then once again, to propose clinical care steps that are not currently reimbursed, that may need to be incorporated into a new bundle. Then also to try and identify potential steps of over-utilization. If we think about the value study, for example, use of urodynamics in simple or straightforward SUI is definitely going to be likely low-hanging fruit for over-utilization of a test. How do we go from the clinical scenarios to modeling? What do we need to model? We're going to need to model the actual state of clinical practice. That data is going to come from the Medicare claims data or market scan, which is private insurance claims data, because we need to know what's actually happening. Once again, that's why we needed to identify all the potential steps in the pathway so we could get that data. We're also going to need to model the ideal state of clinical practice and where it may be primarily just evidence-based steps for what we have in the literature. Then we're also going to need to model the clinical practices of a new bundle, and that's where maybe we have the streamlined kind of ideal state, and then we create the new steps in the evaluation management and treatment that are not currently billed for or billable, but we may have to create CPT codes for them and justify that those are steps that improve the quality for our patients. How do we estimate actual state clinical practice and costs? As I mentioned, it's going to be for insurance claims data. One of the nice things about the Medicare claims files is that it provides procedure and diagnosis information, dates of service, payment and charge amounts, beneficiary demographic information, and limited professional provider and facility data. While you may say, well, that's only in folks over 65, which is true, the nice thing about it is we're going to be able to follow those out, because once they're in the Medicare system, we track them until death. So we thought that this would be a great database to start with. Once we test it in the Medicare claims data, then we hope to use it in a private insurance database, which will obviously have younger women in it. How do we move from the scenario to model variables? This is three tables that are actually much longer, and I just cut them off in three different sections, just to give you an idea, because really in a claims database, the only thing that's in there is what's been coded or what's been entered, and that's usually based on the diagnosis code or a procedure code. So if you think about a claims database and the evaluation and treatment of stress urinary incontinence, there are really only three clinical paths, if you will, for which an administrative database is useful and data can be obtainable. And that would folks who came in to see a urogynecologist, urologist, and from that initial visit were evaluated and went on to surgery. So they didn't undergo physical therapy or an incontinence pessary, anything like that. The next group would be some women that were not initially offered surgery but then had surgery. So they were seen, evaluated, maybe tried a pessary for a while, tried physical therapy or both, and then went on to surgery. And then there's going to be the women, for whatever reason, never had surgery. Maybe the PT and incontinence or incontinence pessary were successful, but those are essentially the three groups or how we've kind of grouped the women for stress urinary incontinence that we're going to be able to get data from the data set. Now obviously that data is going to be messy, and we're going to have to clean it up, but with the limitations of an administrative database, this is, we think, the best way to move forward. So from that, we'll enter the data into cost-effectiveness models, and those are models that evaluate the value of interventions, and John's going to talk a little bit about that. So Jerry and I actually worked together when I was a fellow, and I got involved in some cost-effectiveness modeling, and that's, I think, probably why he tapped me to be part of this committee, is to use some of those skills that we can apply to this new regime here. And so what we're trying to, it's more, instead of being a cost-effectiveness model, it's largely the cost and trying to figure out how to make the proper decisions and where can you find value. I think that the goal of the alternative payment models is for practitioners who are doing the right thing to make more money, the payer to pay less money. So it's an everything, everybody wins game. So we're trying to figure out how to make that happen and put some numbers behind it. So we're looking at the way that we can evaluate these different interventions. How many people in the room know anything about cost-effectiveness modeling? So we got a little bit for two to three people, but for the most part, not too much knowledge here. So I'm going to start with something to try to make this as basic as I can to explain how cost-effectiveness modeling works. And so this is kind of just like a very simplified scenario of some of the tests that you might do when somebody comes in the office with stress incontinence. And so they come in with stress incontinence. You could do the urinalysis or not do the urinalysis. And off of each one of these, so if you did the urinalysis, then you could do urodynamics or not do urodynamics. Same thing down here with both of these two options. And so what we can do is we can go in and say what percentage of people are doing these tests. And we say 90% of them. And these numbers are totally made up. There's no true Medicare data or cost data behind them. I just, for example purposes, put them into the slide set. So if we've got 90% that do the urinalysis and we know that 60% do urodynamics, you can then do some simple math and figure out that the 0.9 times the 0.6 gets you to 0.54. For the people who came down this path and ended up at this terminal branch, had urinalysis, had urodynamics. We can then do the same thing and calculate the percentage of people that ended up at each of these terminal branches in the tree. And you should see that these should sum to one. Next thing that we can do is we can go through and say how much each of these things cost. And if you say, assume that urinalysis costs $20 and urodynamics costs $1,000, the person who ends up in this branch of the final branch of the tree, their cost was $1,020. And you can figure out the cost for each of the four branches with this one being the most expensive, this one having zero cost. And then we can take this information and what we do is we go back and we say, okay, let's find the weighted cost for everybody. This person, we knew that from the previous example, they were occurring 0.54% of the time or 54% of the time, and we knew that their cost was $1,020 because they had the urinalysis and the urodynamics. This person had their relative cost and their relative percentage, same thing for all four branches. And that gets you to an average weighted cost in this example of $618. And so we can take the – if we build these trees with the clinical scenarios, we build them for the codes that got coded in, for the visits that come in, for the tests that get run, for the procedures that get done, you can figure out what kind of like right now where are we at, what's the average cost for treating stress incontinence. But then we've got to break it down by simple stress incontinence, so, you know, mixed incontinence, recurrent, with prolapse, with no prolapse. And so there's a bunch of different – you've got to do a bunch of different scenarios to come up with these. And once again, this is very simplified. This part is easy. Like with that simplified tree, is there anybody that couldn't follow what was going on with that? And I think that that's very simple. But the trouble is that these trees get very large. The tree that we presented had four terminal nodes. Jerry and I worked on a model one time that had over 100,000 terminal nodes. And so that's where the computers come in because I can't do that math in my head anymore. The other thing that we can do with the models is we can model change. So let's say that we have the people who had urinalysis was 90 percent. Let's say we're going to take that up to 100 percent. And let's say we just talked about the value study with urodynamics, and maybe not everybody needs urodynamics. Let's say if we want to take that 60 percent, take it down to 40 percent. And then you can figure out what happens. Your cost, if we went through the stage that we call that rolling back the tree, you can figure out what the average weighted cost is. And you say that, okay, we increased one test, we decreased another test, but the test that we increased was a much cheaper test, and the test that we decreased was a more expensive test. And so there's cost savings. So we can find the areas where we can reduce cost, maybe cut the total reimbursement a little bit, but overall make the amount that we're getting when you're doing the right things more. So everybody wins. And I think that's kind of the goal for the alternative payment models. And when I initially got involved in this, I was, like, very skeptical, like, how is this going to help me? I think that they're just trying to figure out a way to cut costs. But if you're doing the right thing and we're using evidence-based practices, we can potentially decrease overall cost, provide better care for our patients, and at the end of the day, as physicians, get paid a little bit more for doing it. So this is another tree looking at treatment strategies. And you can see that these trees can kind of get built onto each other. So you could take this tree and just stick it onto the very end, like here, like the tree that I'm showing here. You could take the entire tree and just kind of copy and paste it and put it here, put it here, here, and here. So that's how the trees get bigger and bigger and get very, we say, and as a cost-effectiveness joke, instead of being trees, they turn into bushes because they get very large and complicated. So this is, once again, for stress incontinence. If we know that right now maybe only 10% of our people are getting physical therapy before they have their sling, and we know that there's a chance that if they have physical therapy, they're going to be successful and they'll be done with treatment or they're failures and may have to have that sling anyways. So what happens to our, in this model, I calculated the average cost to be a little under $2,000. But if you could just increase that to 20% getting physical therapy, you save your cost, you save, you know, 5% to 10% of your total cost. So, I mean, there's a ton of different ways where we can look at this to try to figure out ways to save cost, ways to provide more evidence-based care and ways to get buy-in from everybody on a system that's going to radically change what all of us have used in the past. But that's kind of as basic of a rundown of cost-effectiveness modeling as I can make it, I think. All right. So, you know, as we discussed earlier, there's going to be a lot of or several different potential scenarios that we can create APMs for. Once again, as we boiled them down a little more, straightforward SUI is the one we're working on now. Once again, you all may ask yourself, well, how many women have straightforward SUI in your practice? You know, as we look at the data in the future, once again, you know, there's a code for mixed urinary incontinence, but there's not one for stress-predominant mixed or urge-predominant mixed. So that likely may be something that we just create one for mixed. But part of the process, you know, there's an art and a science to the modeling. And as part of the modeling, we're going to have to make estimates and we're going to have to make assumptions. And, you know, if we think that there's going to be significant different costs associated potentially with urge-predominant versus stress-predominant, that may be something that we, you know, try to model. We're just going to have to see. And then once again, when we get into prolapse in the future, you know, are we going to create an SUI APM off of prolapse or is it going to be an SUI, you know, with prolapse? And then, you know, the occult is, once again, there's not a code for that. And so in the recurrent SUI, and like I said, some of these we may not be able to develop APMs for, and that's fine. They'll just continue to go through the fee-for-service type model. You know, other potential model considerations, once again, with the evaluation is we look through the claims data and, you know, we look at evidence, clinical practices, best clinical practices and what's in the literature. You know, we're going to look and see whether there's some tests that we think are required in the evaluation and management, and that will be part of the, you know, proposed clinical pathway. There's more than, you know, one correct way to do things at times, so within the model we may have optional tests, ones that are reasonable. Now, granted, if some people, you know, providers decide to perform those tests, their costs may be a little higher, but that's something we may need to build into the model because, once again, this needs to be usable to the general membership. And then we're likely going to find some tests that should not be performed. And, once again, if it's straightforward SUI, we already know that there's literature to support not doing urodynamics. And in treatment, you know, there may be a best option versus all options, and we may need to model that. You know, there's nothing wrong with doing a BIRCWH, you know, procedure, even though the literature seems to favor, you know, the synthetic mid-urtho sling. And then, once again, you know, in the model we build it such that it incentivizes less expensive options with better outcomes. So these are all model considerations we're going to need to think about as we're building it. And, once again, some aspects that are not currently reimbursed is – and so a lot of you may be thinking kind of like John was saying earlier, well, if we cut out UDs, for example, because, I mean, let's be honest, there are a lot of – I mean, when I was a fellow many, many years ago, I would hear about some practices that new patient, systo UDs on their new patient visit, whether they needed it or not. And, you know, that was definitely a moneymaker for a lot of folks. So if you cut out the urodynamics, people would think, well, that's going to drop our reimbursement. I mean, that was a cash cow, if you want to be honest. So what we're hoping to do is when we identify steps that we do, such as, for example, nurse education preoperatively or nurse phone calls afterwards that may decrease – and we talk about this a little bit later in the slide – presentations to the emergency department. So if we can justify that those are quality or steps that improve our outcomes and the quality of care, you know, we may be able to generate CPT codes for that so that then we can bill for that. So once again, as John was saying, we can decrease costs on other areas, justify work that we're already doing or our office staff is already doing, but we're just not getting paid for. So that may be how we can justify the same reimbursement for what is being done. So once again, preoperative screening and treatment of other components that we know can improve outcomes, post-op phone calls, and then certain aspects of the post-op visits that are not currently covered. And there may be others that we identify. So what are the next steps? As far as the model type, you know, there are about seven different types that are proposed by CMS and AMA. We've chosen episode of care payment. We're going to need to determine how long that episode is. As we've talked about, we're going to create the cost-effectiveness models and the run analytics using the cost data. There's some other things we're going to have to define of the methodology and characteristics and requirements of the APM, which is the payment methodology, all the quality measures. And then, you know, really look at the practice capabilities. You know, what is going to be useful for all members of OGS? And then I had this from earlier, obtain funding. Like I said, hopefully we're going to get through an in-kind grant from Medtronics, both use of their clinical or the Medicare claims database, but also some of their health policy and economics experts. And if there's any questions about that, I'll be happy to answer it. And so once their models are finalized, once again, as John mentioned, we're going to share with the relevant committees and SIGs for input. Ideally, we will have some forum to distribute it to the entire membership if possible, let them look at the steps, and once again, help identify if we've missed something, some things that need to be included, and just look for overall input. Then we'll submit it to the board for review and comments. Once the APM is finalized, we'll share the white paper with the OGS membership once again for review and comments, and then finally look to start meeting with payers and submit to CMS. Thank you.

alternative payment models

OGS Payment Reform Committee

physician payment reform

stress urinary incontinence

cost-effectiveness modeling

evidence-based care

payer and CMS submission