Hello, everybody. Thank you for dropping into our session, which focuses on how to project and prepare for the abstraction burden of follow-ups in the LAAO registry. A little about me. My name is Christy Verschelden. I work in a quality and registry department that serves three hospital campuses. Baylor, Scott, and White, the heart hospitals are located in North Texas, serving the communities surrounding Plano, Denton, and McKinney, all suburbs north of Dallas-Fort Worth. All of our cardiovascular registries are abstracted in-house, and I'm a bit of a jack-of-all-trades. My current role is to support our team by providing progress reports related to registry productivity and data accuracy. I also float between registries and I help with abstraction, process improvement, and the development of reporting solutions. Let's start with an overview of the session. When analyzing our calendar year 2021 volumes, we identified that our LAAO registry volume had doubled. We jumped from 159 cases in calendar year 2020 to 320 cases in 2021, and that's particularly significant in this registry because every base procedure entered is potentially eligible for data collection across four additional time periods. In our department, twice every year we do analyze our registry volume trends. Since we abstract data in-house, the frequency of that type of review allows us to quickly bend and shift responsibilities when needed. So our short-term goals for the project were to identify the current weekly abstraction burden assigned to this registry site manager, and with that knowledge, we could make some informed decisions for workload rebalancing if necessary. And given that base procedure volume increase, we recognize that our follow-up volumes over the next two years could also potentially double. As part of long-term planning, we knew we needed to forecast our future follow-up volumes as well as the monetary impact of outsourcing our abstraction to a vendor should we no longer be able to manage it in-house. We have three learning objectives for this session. We want to identify tools to help you, number one, estimate your future LAO registry follow-up volumes. Number two, calculate the annual abstraction hours required for an internal abstraction model. And then third, to calculate the budgetary needs for outsourcing abstraction to a vendor. We've done all of these things, and we want to make sure that you can replicate everything that we've done here, and all the tools will be made available to you in an Excel workbook format, which will direct you to. This slide shows our trended base procedure volumes. We started participating in the LAO registry in August 2016, and our volumes have consistently trended up, as you can see by the blue columns. The orange line shows the rate of change from the prior year. So, an average between years 2017 to 2019, we saw about a 43% growth in volumes annually. And then our procedure volumes grew at a slower pace in calendar year 2020, only hitting about 24% growth from the prior year. And that's a trend we saw in many of our registries, given the onset of the pandemic, which delayed scheduling our elective procedures. We did expect the rate of change to catch up in calendar year 2021. However, we didn't really realize that it would double. So, our volumes ended up much higher than anticipated, with 101% growth from the prior year. Every patient that does undergo percutaneous left atrial appendage closure is analyzed for inclusion in the LAO registry. Successful, canceled, and aborted procedures that meet inclusion criteria are then abstracted into the base procedure form. In addition, all these cases are eligible for a 45-day follow-up. And after that, every successful procedure is potentially eligible for follow-up data entry at six months, one year, and two years. So, since our base procedure doubled, we did expect the volumes across each of these follow-up time periods could also double. It's fairly easy to do a quick calculation to show your abstraction burden over time. What you would want to do, just as a quick estimate, is take a calendar year, look at your base procedure volume, multiply that by four for the four follow-up time periods, and you'd know that over the next two to three years, you might see a specific number of follow-ups generated. In this example in green, we have calendar year 2020, 159 base procedures. Multiply by four. We knew that we could see about 636 potential follow-ups between calendar years 2020 to 2023. And then likewise, in yellow, calendar year 2021, we could do the same type of calculation. What we don't see by doing it this way is how many specific follow-ups we might do in one particular year. So, using just kind of those quick estimates without additional formulas involved, we wouldn't know, per se, how many follow-ups we might do in calendar year 2023 specifically. And that was the goal of the project, is long-term forecasting of what our follow-up volumes might look like. I do want to introduce you to Amber Lopez. She's our LAAO Registry Site Manager. I've listed her contact information on this page. So, if you'd like to reach out to her for networking purposes, she would be happy to hear from you. At the start of the project, Amber was responsible for abstraction, data cleanup, outcomes reporting, and process improvements in the LAAO Registry and the TBT Registry, as well as a smaller population of our CAF PCI Registry she was working with for only one of our campuses. So, when we did our analysis in quarter one of 2022, at baseline, this is the annualized registry volume that was assigned to Amber. The blue portion of each of these columns represents our base procedures, and then the yellow portion represents our follow-up volumes. On the far right, you can then see the assignment that Amber had on her plate, which equaled about 989 base procedures with 1,299 associated follow-ups. Before we move on to some of the formulas we've used, I do want to point you in the direction of a tool we uploaded to the NCDR Participant Resource Sharing site. If you navigate to that platform, you'll find an Excel workbook titled LAAO Follow-Up Projections. All the calculations we've used are built into that workbook, and we're going to focus on three of the tabs that you can see in those kind of colored arrows. I'll point you in the right direction as we go, and we have aligned our calculations with the participant guide from the LAAO registry related to follow-up timeframes. So, that document can be found in the resources section of the registry. The first thing we needed to do was analyze Amber's current weekly abstraction workload. We knew that that could help us redistribute cases internally if necessary, and you can easily replicate this analysis for any registry. You just need to know your annual base procedure volume and your follow-up volume. That can be pulled from your data quality report or DQR on the submission results page within the NCDR website. That DQR tool will help you see the number of patients submitted for each quarter, and then you simply have to annualize those numbers. I did try to color coordinate this with the tabs on the Excel document, so in orange you'll see our weekly total abstraction hours calculator. If you're abstracting data internally, you can use this to estimate your abstraction hours required for the registry. To do that, you just need to know the average time spent in minutes abstracting each case type, and that can be accomplished through internal time studies. In yellow, you're going to find our annual vendor abstraction cost calculator, and you can use that to estimate some budgetary needs. For that, you're also going to need your vendor cost per base procedure, as well as your vendor cost for follow-up. This page shows our analysis for Amber's role in full at baseline. It does include all three of the registries she was managing. If you're following along in your own workbook, that light orange section is where you're going to enter your annual base procedure volume, annual follow-up volume, and then the average time in minutes that you spend abstracting each case type. The numbers we have on this page are reflective of our own workflows and work generally for the majority of our registries done in-house. We've proven that through some internal time studies, but you're probably going to want to plug in numbers that are more reflective of your practice. The numbers below that are auto-calculated. You don't have to do anything there. Just know that you're going to get those numbers by multiplying your procedure volume by the time spent abstracting your base procedures, and then adding that outcome to your follow-up volume multiplied by the time you spend abstracting your follow-ups. What you end up with is the average weekly hours spent solely on abstraction within a registry or a role. In this example, Amber was spending about 31.5 hours abstracting cases every week. Why that number caught our attention is because we look at an 80% operating capacity in our department since we do all of our data management in-house. In a 40-hour workweek, an 80% operating capacity for abstraction means there's only 32 hours available there, and she's already hit that ceiling. We give the registry site managers another 20% of their week for discretionary and administrative activities. Those are things like department huddles, meetings, any training that they have to do, as well as working on projects, simple things like case finding where the minutes add up over time, inclusion screening, data submission, validation, and trending, reporting data, time away from work. We also have to take into account the number of registries assigned for each registry site manager. So, with this 31.5 hours per week, Amber's already bumped up against that 80% abstraction operating capacity number of 32 hours, and if we're not doing something about that, we're potentially not setting her up for success within her role. So, this is where we knew we needed to take action, especially when paired with the knowledge that her LAAO registry numbers had just doubled. So, this is what we did. We almost immediately reassigned her CAF PCI registry role. I actually partnered with another team member so we could quickly offload that registry from her plate. We then reused that abstraction hours calculator and only looked at the LAO and TVT assignment, and that dropped her abstraction hours per week from just at 32 hours to 25 hours per week, and that created some breathing room to accommodate the continued LAO registry growth. Our focus then moved to forecasting our continued LAO registry based procedure and follow-up volume growth. We've already seen that year after year our base procedure volumes had been on the rise, and we've just learned we have a new heart hospital opening by the year 2025, and that data collection efforts will also fall under the umbrella of our department. So that knowledge really reinforced our desire to better understand where we were going with this registry. This table is reflective of our actual volumes for the LAAO registry. So in 2020, we had 159 base procedures. We actually entered 412 follow-ups that meant our weekly abstraction hours set at about seven and our vendor cost when we use the calculator in the workbook would have been about $25,000. That number doubles, not a surprise, in 2021 to about $48,000 with about 13 and a half hours spent abstracting. Again, the missing piece of the puzzle is what is an accurate number for us to use for follow-up volumes to continue to trend out how much time we're spending or the cost we're spending for the registry. The projection model we created again has been built into that Excel workbook on the participant resource sharing site. It's under the green tab and we'll dive into the data you're going to need to replicate this project as well as some of the calculations we've used over the next several slides. To forecast future follow-up volumes, we started off with a staggered table to capture our base procedure volumes in a monthly format. That allowed us to best estimate when a patient would be eligible for a follow-up period. To replicate the project in the workbook, just use the data extracts feature on your LAO registry website. Enter the date your program started through the current date and that'll allow you to pull an extract and then count your volumes. The registry does count follow-up eligibility by procedure date, so make sure that's the number you're using to count the monthly volume. Don't use the discharge date. From there, you're just going to enter that procedure month, base procedure volume, and the first two columns of the table. Everything else is going to auto-calculate out for you. At a minimum, our department strives to achieve follow-ups in window for each time period at or above the 50th percentile benchmarks and that's what led us to choose the following metrics for our calculations. If you feel that the benchmarks for your volume group or the U.S. hospitals rolling for a quarter is a better fit for your program, those values can easily be plugged into the workbook. You'll just update the values in orange and your formulas will automatically be updated for you. This slide shows the calculations we created to forecast follow-up volumes at 45 days and six months. Please know I am going to show this in a visual format shortly. I did just want to give you a deeper dive into each of the formulas used. So, the first follow-up time period defined by the registry is at 45 days after procedure. They do allow for plus or minus 14 days, so you'd enter that follow-up anywhere between 31 to 59 days post-procedure. And we place this calculation at the second month after procedure to fit that timeline. Successful, canceled, and aborted procedures all require a 45-day follow-up. Thus, the only real adjust we made in this calculation was removing the average number of patients who were deceased at the time of hospital discharge. After six months, only successful procedures require a follow-up, so we remove the average number of patients who had a canceled or aborted procedure, and then we remove the average number of patients who were deceased at the time of the 45-day follow-up. These are the calculations used for one-year and two-year follow-ups, and these calculations just rely on the average number of patients being alive at the prior follow-up time period. And so what we ended up with was a staggered workbook table that helps us to forecast our monthly follow-up volumes. We started our program in August 2016 with a monthly base procedure volume of three patients, and that's seen in column B of the table in the darkest yellow fill represented by three little people icons. If we follow those three patients over time in column D, we see the expected 45-day follow-up volume is still three patients. In column H, the expected six-month volume is 2.8. The one-year volume in column J is 2.7 patients, and the two-year volume in column L is 2.6 patients. So by August 2018, we were fully vested in the registry, abstracting through the end of the two-year time period across all follow-up time periods. And so we look back at column M. Let's take, for example, row 26 in August 2018. We expect to have abstracted 30 follow-ups that month, and that number is derived from all the patients eligible for each follow-up time period for that month. Some tips if you're using this workbook table. You want to make sure that you're entering all your historic data. It'll accommodate up to 10 years. I definitely recommend, again, you doing the full time period so you can see how your volumes grow over time. In addition, you're going to want to enter an estimate of your future-based procedure volumes, or your forecasted follow-up volumes will be falsely low. I suggest that you use an average of your prior 12-month volumes for this value. As an example, our average monthly volume between July 2021 through June 2022 was 34 cases per month. I then used that number for two years into the future to predict our base procedure volumes, and that's seen by red in the screenshot above. To test our formulas, we use the DQR to trend our actual follow-up volumes entered into the registry. So the green line represents our actual follow-up volumes. We then trended the quarterly volumes identified by the projection model, and that's reflected by the blue line. That's our potential follow-up volumes. So we interpreted this graph as proving that the model enhances the ability to forecast future follow-up volumes. Of note, in late 2020, we did revamp our follow-up abstraction processes. So at that point, you can see that we were performing more follow-ups than forecasted, which is a really positive sign that we were meeting or surpassing the benchmarks that we used for our calculations. At the time of this recording, we're nearing the September 1 data deadline, and our registry site manager is still actively entering follow-ups for that deadline, and that does account for the lines starting to separate again. We then went back to our workbook calculators and plugged in the projected base procedure and follow-up volumes for the current calendar year, 2022, and we can do this with some confidence based on these projection models, even though we're only at the mid-year point. Again, these are projections only for the LAAO registry. We anticipate that we will perform 424 base procedures this year. We also expect to abstract 1,226 follow-ups by the end of the year, and that equals about 20 hours of abstraction time per week, which is up from 13 hours in the prior year. Our vendor cost in 2021 was estimated to be about $48,000. As you can see, that calculation jumps by $23,000 by the end of the year, and we're now projected to surpass a cost of $71,000 annually should we choose to outsource our abstraction. Putting it all together, we can now see our historic and forecasted base procedure volumes in the blue columns, our follow-up volumes in the yellow columns, and the upswing in vendor abstraction cost over time represented by the green line. This graph highlights the ratio of base procedures to follow-up data entries for the LAAO registry. By 2025, we anticipate performing 408 procedures annually, but predict we're going to be entering over 1,500 follow-ups for the same registry. Putting the data together in this format could help your program determine what's the best model for abstraction, be it in-house staff versus outsourcing to a vendor. Rounding back to the table where we attempted to predict the future for the LAO registry, you can see that we've successfully been able to do that. We can now look to what our future volumes might be as well as our costs. If Amber, our registry site manager, only managed this one registry, this is a totally reasonable assignment with room left over to support other registry assignments. However, when we combine the LAO projections with a similar analysis for the TVT registry, we find that offloading those CAF PCI responsibilities earlier this year was just a temporary solution. By the end of this calendar year, we do expect Amber to again be maxed out on abstraction hours, spending more than 80 percent of her time abstracting and leaving less time for those discretionary and administrative role responsibilities for two registries. Knowing that our volumes continue to rise across both registries and we have that new hospital opening by 2025, we've begun the discussion for creating a new full-time employee within the department. And so, the data created by this project was used as part of the justification to write up the request for that new employee. In summary, we did achieve our stated goals. We've been able to accurately estimate our future volumes, identify the abstraction burden assigned to one of our registry site managers, and to assess the monetary impact of outsourcing to a vendor should it become necessary. The data we've gathered has been actionable. We were able to quickly respond to volume trends and reduce this RSM's workload that enhanced her ability to be successful in her role. We also used the data to advocate for the department's future needs, creating a justification for an additional abstractor in the near future. And finally, we were able to create a workbook that our NCDR peers can easily use to replicate this project. The biggest takeaway that I want you to walk away with from this presentation is that you need to analyze your volume trends routinely. Once you've completed two years of participation in the LAAO registry, you're going to be capturing follow-ups in all eligible time periods. And as your base procedures volumes grow, your follow-up volumes are going to grow exponentially. It's really imperative that you project and prepare for the abstraction burden that that creates internally and or the financial burden created for programs using an abstraction vendor. Thanks again for your time. If you have any questions about the workbook or this presentation, please reach out and let me know. If you're interested in networking with Amber Lopez, our registry site manager for the LAO registry and the TBT registry, again, I know she'd be happy to hear from you. And we both wanted to give a special thanks to our Director of Healthcare Improvement, Sue Dorval, without whom none of the work that we do here would be possible. Thanks so much for your time and looking forward to hearing from you. Bye. Bye.

LAAO registry

follow-ups

abstraction burden

volume trends

registry department

Texas