would like to get the opportunity to introduce you all to our panelists for today, Danielle Reynais, Doctor. Lauren Campbell, and Doctor. Esper Fishman. So with that, I will turn it over to Danielle who will act as our hosts for today's webinar. Danielle? Thanks, Akina. Hello, everyone. I'll be your host today, as Akina mentioned. My name is Danielle Reynais. I'm a senior research associate, and I've been with NCQA since twenty eighteen. And throughout my time here, I've led several measure development projects, primarily for our risk adjusted measures, but also for measures of respiratory health quality and appropriate antibiotic use. I also lead and coordinate our HEDIS public comment periods, and various other cross cutting projects across the organization. Another speaker you'll hear from today is Lauren Campbell, who brings over ten years of experience managing quantitative work streams and leading analysis on large scale national implementation and evaluation projects. Since joining NCQA in January twenty twenty four, she has been overseeing the risk adjusted utilization, older adults and aging, and long term services and supports measure portfolios. And then last but certainly not least, our final speaker is Ezra Fishman who conducts statistical analyses to estimate NCQA's risk adjustment models. He analyzes large claims datasets to support the development, maintenance, and revision of our risk adjusted measures. And he has been with NCQA since December twenty nineteen. This is the agenda for today's presentation. We'll start with an overview of our risk adjusted measures, and then we'll spend the majority of the presentation on reviewing some potential upcoming changes for measurement year twenty twenty eight. And then we'll close by outlining for all of you what our planned analyses are to assess the changes in addition to our next steps. I will note that the changes we're previewing for you today are not set in stone as we proceed with this work. We will continue to follow our standard process of vetting the changes through our advisory panels and other experts. And then for some of the changes, will leverage other communication and feedback channels such as public comment. And as Akina had mentioned earlier throughout the presentation, please feel free to add questions to the Q and A, and we will try to get to as many of them as possible when we pause for questions at the end. Right. So we're going to start with an overview of our risk adjusted measures. And first, let's do sort of a thirty thousand foot view of risk adjustment and why it matters. Before comparing health plan performance, it's important to recognize that outcomes are influenced by underlying member characteristics and not just quality of care. Some plans care for members who are older, have more comorbidities, or otherwise at higher risk of poor outcomes. So without adjusting for those differences, plans serving more complex populations may appear to perform worse even when providing excellent care. Risk adjustment addresses this issue by controlling for differences in case mix and ensuring fair comparisons across health plans. We use statistical modeling that accounts for member characteristics to predict outcomes and adjust how a plan's performance looks. And the goal here is to isolate the portion of performance attributable to care delivery rather than, patient risk. The risk adjustment process protects providers and health plans that serve complex populations while still allowing us to assess performance. NCQA statistical models produce risk weights that are published in risk adjustment unit yeah. Sorry. Risk adjustment utilization tables. Health plans apply these tables to their member populations to calculate an expected event rate for a given measure such as readmissions or complications. The expected rate represents the level of outcomes that would be anticipated based on the plan's specific case mix, assuming average performance. Once the expected event rate is calculated, it is compared to the observed event rate, which reflects what actually occurred for the plan's members. Dividing the observed rate by the expected rate yields the observed to expected or O to E ratio. This ratio is the central metric for assessing risk adjusted performance. An O to E ratio below one indicates that fewer events occurred than expected given the plan's the population's risk profile, while a ratio above one indicates more events than expected. To support consistent interpretation across organizations, the O to E ratio is further calibrated by dividing the individual plans ratio by the national average ratio. This calibration step standardizes results and allows performance to be interpreted relative to a national benchmark. So a calibrated ratio of less than one signals better than expected performance, and then a calibrated ratio greater than one signals worse than expected performance. And for those of you who would like more information on this topic, we did release a blog post a few months ago. We can drop that link in the chat. It's also available on the, on the slides here. This is a snapshot of our current measure portfolio. We have ten risk adjusted measures in HEDIS, including the four new outpatient surgery measures that assess acute hospitalizations following select procedures. Aside from the outpatient surgery measures, we also have four other measures assessing hospital utilization, and then we have two measures of emergency department utilization. Some of you may, recognize the emergency department utilization measure from our our, recent public comment period where it was proposed for measurement year, twenty twenty seven for Medicaid. So why are we making the changes we're going to discuss in this presentation? Ultimately, the goal is to continue having a quality measurement approach that is fair, meaningful, and focused on improving outcomes for all individuals. We want to measure what matters and capture aspects of care that can be improved through better clinical practice and coordination. We also want to support equitable care across all populations and not penalize organizations for caring for more complex populations. Quality measures should discourage behaviors that improve scores without improving patient care. So we wanna promote transparency, fairness, and accountability. And then finally, we want to drive real improvements and better outcomes. So with that, we'll now dive into what the actual changes are that we're considering for measurement year twenty twenty eight. And the first one that I'll discuss is around denied claims. This is something that impacts or would impact all risk adjusted measures, and I'll start with some background. So denials are not isolated events. They're a routine part of the health care financing system. Research has indicated that health plans deny up to twenty percent of claims at initial submission with denial rates varying largely across lines of business, payers, and types of services. The impact of denials is both clinical with necessary care being delayed and treatment interrupted, and financial since the burden of paying for denied services falls largely on the individual. Automation and AI do streamline claims processing, but this has led to concerns about whether there's appropriate clinical review and decision making. In addition, research has suggested that denials are often used to control costs and not as a way of controlling overuse of services or ensuring appropriate care quality. Today, denied claims can be used for risk adjustment and for the hospice exclusion, but they are excluded from being used to identify measure events in the HEDIS risk adjusted measures. So this means care represented by denied claims is effectively invisible to quality measurement. Over time, we recognize that there are potential unintended consequences of excluding denied claims. One concern is that providers caring for more complex patients may appear to have worse performance when important care events are omitted. The exclusion may also disproportionately affect vulnerable populations who are more likely to experience coverage or administrative challenges. And then there's also concern, as I mentioned in the previous slide, that denial decisions are often driven by payment or administrative considerations, rather than clinical appropriateness. So, this really raises the question of whether measurement should depend on payment status or on the care and outcomes that, patients actually experience. And so the proposed change to include denied claims provide provides a more complete picture of health care utilization and quality and reflects a more person centered perspective. From an individual's point of view, a hospitalization, ED visit, or other health care event occurred regardless of whether the claim was ultimately paid. So this new approach reflects what actually happened, not just what was reimbursed. It ensures that organizations are not unfairly penalized or advantaged based on, payer denial practices that may be outside of their control. In addition, including clinically meaningful events can improve the accuracy of quality measurement and capture the true quality signal of care that was delivered. And then finally, another important consideration is that including denied claims can remove possible incentives for plans to improve performance through administrative actions rather than actual improvements in care, which ensures accountability and measurement. So ultimately, we think this potential change, would help preserve the integrity of our measures. So now I actually wanna turn it over to Lauren Campbell to take us through the next topic. Great. Thanks, Danny. Hi, everyone. Again, my name is Lauren Campbell, and I'm going to share some information about how we use clinical groupers at NCQA to support our risk adjustment methodology and what proposed changes could mean for health plans. So to start off, a clinical grouper is a standardized methodology that categorizes clinical data, including diagnoses, procedures, or healthcare encounters, into meaningful clinically similar groups. By grouping members with their patient severity or the complexity or disease burden, we would anticipate these members would have similar care needs or similar expected outcomes. By organizing complex health care data into clinically meaningful categories, clinical groupers support our quality measurement through risk assessment. Clinical groups are critical for birth care as a QA because translate individual diagnoses into categories of complex disease burden. Performance and supporting fair comparisons of health utilization across health plans. The next slide shows a brief illustration of how we use clinical groupers in risk adjustment. Let's say that plan a has much more proportions of members with heart failure, COPD, and diabetes, and that Plan B has smaller proportions of members with these conditions. If Plan A's members are sicker or more complex than Plan B's members, Plan A might also have higher rates of inpatient admissions and emergency department visits than Plan B. And this raw comparison, that is a comparison without risk adjustment, would not be fair to health plans. This raw comparison would be misleading because it ignores differences in member health status between Plan A and Plan B. Instead, we use a clinical grouper to identify variables for risk adjustment. The clinical grouper categorizes diagnoses to explain member complexity and disease burden, enabling fair comparisons. And in our risk adjusted utilization measures, we use risk adjustment to estimate expected utilization based on member complexity, enabling equitable evaluation of plan performance. So after applying a clinical grouper and risk adjusting for member complexity, we would find that Plan A treats a much higher percentage of members with heart failure, COPD, and diabetes than Plan B. By risk adjusting for these conditions, we can evaluate utilization relative to the expected complexity of each plan's population, adjusting for disease burden and member complexity, and enabling fair comparison of plan performance. Each measure in the Risk Adjusted Utilization Measure Portfolio includes a combination of risk adjustment variables, such as age, sex, comorbidities, discharge conditions, or surgeries. The clinical grouper in particular is used to categorize comorbidities for all risk adjusted utilization measures. The comorbidity variables derived from the clinical grouper that are selected for risk adjustment can vary by measure and product line indicator. This is because the comorbidities used in risk adjustment are selected through statistical methods in order to optimize model fit. This means that the comorbidities included in risk adjustment for each product line indicator are those that are meaningfully predictive of measure event. It also means that when comparing variables across product lines for the same measure, those product line indicators could have differences in the stats of comorbidities included in risk adjustment and would have different weights for the same variables. We identify comorbidities or risk adjustment using CMS hierarchical condition categories, or HCCs, as our clinical grouper. The CMS HCCs were developed for use in Medicare Advantage, meaning that they are accessible, publicly available, and updated regularly. Through HEDIS measurement year twenty twenty seven, we are using version twenty four of the CMS HCCs as our clinical grouper. However, version twenty eight was implemented in payment year twenty twenty four and phased in through payment year twenty twenty six. This signals to us that plans have been or are in the process of implementing and transitioning to version twenty eight. It's important to us to align our measure requirements with plan processes. And with that, we are proposing to use version twenty eight as our clinical grouper starting with HEDIS measurement year twenty twenty eight. Our intent here is twofold, really. We want to avoid potentially unnecessary plan burden with maintaining two different systems while ensuring that any planned transition activities to version twenty eight are mostly complete so that plans are ready to use version twenty eight with our measures in measurement year twenty twenty eight. We also want to ensure that we stay current in the field. We recognize that as utilization patterns, characteristics change, If we implement this proposed update, there would be some corresponding non substantive changes you would observe in measure documentation. For the measure specifications, we do reference some CCs and HCCs as examples with. And there are some changes in the diagnosis code mapping to CCs and the HCC number in version twenty eight. So if we move forward with this proposed update to version twenty examples, specifications tool with version twenty eight. Similarly, these differences between version twenty four and version twenty eight would also be reflected in our risk tables and the risk weights themselves. Now I'm going to turn it over to my colleague, Doctor. Ezra Fishman, to introduce the next topic. Thanks, Lauren. Hi, everyone. I am going to talk about our outlier strategy. The outliers are a important part of four of our risk adjusted measures, which are listed here, AHU, EDU, HPC, and PCR. The first, I'll explain our current strategy, which is for these measures, we exclude from the denominator, outliers, which are especially high utilizing individuals. And we do this to avoid, distortions in the measured utilization patterns at the plan level. And I'll show you, what I mean by that in a moment. We use third party administrative data to estimate the distribution of measure events. And based on that distribution, establish a threshold of events per person above which a person would be classified as an outlier and excluded from the measure denominator. As you'll see, this approach enables us to view the rest of the plan that is the vast majority of enrollees more accurately, but at the cost of excluding a particularly high utilizer or group of particularly high utilizers. And these are people who disproportionately have high needs, and we would like ideally to include. So the alternative that we're exploring is called Windsorization. And this approach keep would keep these especially high utilizing individuals in the measure denominator but only count their events up to a specified threshold. And the goal is to include these especially high utilizers in the measure denominator without the distorting effects of their extreme event counts. So let me show you what I mean by all that. If you imagine this is a health plan, and the numbers you see on the seesaw are the number of events measure events per person. So for example, this could be emergency department visits. What we see is that the vast majority of the plan's enrollment has between zero and three events, but there's perhaps a small number of people who have ten ED visits. And those people, make the plan appear to have a very high utilization rate represented by the seesaw tilting towards the ten. And this would obscure the fact that the vast majority of the plan actually has three or fewer events. This is what a health plan's data might look like before we consider outliers. If we then exclude outliers, which is the current method in HIE's volume two for for the form, measures that I mentioned, then the plan's utilization looks like this. Everyone who is in the measure denominator has between zero and three ED visits. Now this is a better view of the bulk of the plan's enrollment, but there's a cost. Right? The cost is that the person who was on the far right end, is no longer in the measure denominator at all. And as a result, the measure does not hold the plan accountable for that person's management, or coordination because that person is not in the denominator. And this is especially unfortunate because the people on that extreme right end are disproportionately a high needs group. So we would really like to find some way to include them. That leads us to the approach we're exploring, which is Windsor rising. So what we do here is the people who were on the far right end stay in the measure denominator, but we only count their first few events up to the prespecified threshold. And the view that we get now of the plan's utilization as a whole is not tilted toward those extreme events. So we remove that distorting effect that, we saw on the previous slides. And at the same time, these high utilizers are still included in the measure denominator. And they're thus, included in the measures account application of accountability for those members' management and coordination, albeit in an imprecise manner. Right? We only, we only count a certain number of events. So what we get about those, extremely high utilizers is imprecise. The I'd like to I'd like to show you what this looks like for a health plan preparing its data for submission to give you another perspective on on how this works. If you consider our plan all cause readmissions measure, the denominator counts index hospital stays. Suppose that we have prespecified a threshold of four events per person. So that, that gives us let's say we have these two patients. We have patient a and patient b. Patient a has five index hospital stays. Patient b has two, and their respective discharge dates are shown there in the third column. Under our current approach, patient a is classified as an outlier because they have five measure events and is thus excluded from the denominator. So the denominator data looks like this on the right, consists just of patient b. Under the new or proposed method of Windsorization, this is what this is what the plan's data looks like. Patient a stays in the measure denominator. Patient b does as well, of course. And the only thing that gets dropped is that fifth event because it's above the prespecified threshold. So this would be a change in kind of how plans, process the data, but, it also shows, you know, how someone like patient a would be included in this measure. So that's another another view into how how this would work in practice for a plan. I'd like to give some thoughts on how we're gonna analyze these changes and what we'll be looking for in our analyses of, you know, what this will do to, performance measurement. So our basic approach will involve comparing our status quo methodology, which is embodied in HIE tos volume two for measurement year twenty twenty six, comparing that to results that are based on each of the three changes that we've discussed, the change to denied claims, the change to clinical grouper, and this change to outliers one by one, which will enable us to see how each of those changes would affect the results. What we're looking for is the measured plan performance, so especially the observed to expected ratios. And what we would like to see is that the distribution of observed to expected ratios does not change all that much, which would suggest that, the measure, remains reliable and valid because of that consistency. We are also looking at diagnostic statistics of our risk adjustment models when each of these changes are implemented compared to the status quo. The diagnostic statistics look at how well does our model differentiate or discriminate between, those who are likely to experience a measure event like an ED visit versus, those who are not likely to experience a measure event. And it also looks at how well is the model predicting overall the the correct number of measure events. And again, what we would like to see is that implementing one of these changes such as, Windsor rising outliers, does not greatly change or perhaps even improves these diagnostic statistics. Those are the major, that's sort of the major analytic framework that we're taking to see what these changes would would look like. On that note, I'm gonna send it back to Danielle to talk about next steps. Great. Thanks, Hosbrara. Yeah. So the last thing we wanna review with all of you is the next steps in our process. I mentioned earlier that that we'll be gathering feedback from a variety of sources throughout this process to ensure our measures remain feasible and relevant, as well as to validate our methodology and vet our approach and our findings. And then we also intend to use other channels to solicit feedback, including but not limited to, public public comment. And then finally, the changes that have substantive impacts on our measures will also be reviewed for approval by our measure governance body before implementation in HEDIS. So I do wanna move into the questions. I do wanna just acknowledge briefly that I know we had some audio issues a little bit ago, which I apologize for. We will be sending the recording out. I know we got a couple questions about this. We will be sending the recording, but also the transcript. So hopefully that will help. And then if any of you have any qualifying or any clarifying questions, of course, we're, you know, definitely open to answering them during the the question and answer period. So I'm just gonna monitor the chat here and see see if we've got anything. Just give me a moment. Just make sure you are you are using the q and a function. There's a question about the outlier threshold. The question is what's the methodology that we're looking at to, determine the outlier threshold. So I can give a few more details on that. What we do is, we have in our testing data, we take the measure denominator prior to any exclusion or winsorization. So the measure denominator but with outliers included. And we look at the distribution of measure events among those with any. So we get the mean, standard deviation. We look at quantiles like ninety fifth percentile, ninety seven point fifth percentile, ninety ninth percentile, etcetera. And, typically, we, we take a threshold that is inclusive of about ninety nine percent. In other words, we don't wanna we don't want the outlier threshold to exclude too many people, even under the current method. So we would pay particular attention to the ninety ninth percentile. We also pay attention to, what is about two standard deviations from the mean, and we also recognize that we need to pick a a whole number. So, usually, those those numbers that I just gave you, like two standard deviations above the mean and the ninety ninth percentile, are not gonna be whole numbers, but actual measure events only come in whole numbers. Right? You don't have three point seven four six ED visits in the measurement year. So we would typically take the nearest whole number above what those, what those calculations would give us, and that gives us the outlier threshold. The methodology for determining what that threshold is, is not we're not looking to change that. We're looking to change what happens to the people whose measure event count is above that threshold. Thanks, Ezra. There was another there was another question that I can take. So someone asked, have these changes these suggested changes been reviewed with major HIE to stakeholders like CMS, etcetera? Yes. We have started those conversations. Some of them have they were included also in the advanced notice earlier this year. So we are definitely, you know, making sure to communicate any potential changes with our, you know, major stakeholders such as as CMS or, you know, any other programs that that might use the the measures. So thank you for that question. Let's see. And then, Lauren, there was a question. Does NCQA have a recommendation on social risk adjustment? Great. Thanks, Danny. I understand there were some audio issues. I do just wanna check if folks can hear me okay before I try to answer the question. Great. Yes. We're right. Can. Great. Alright. So, yeah, so does NCQA have a recommendation on social risk adjustment? So thank you for raising this question. You know, some stakeholders, for example, from our public comment, from our measure advisory panels, have expressed concerns, in the last several years with NCQA's decision to not risk adjust TDIS measures for nonclinical factors such as income or education level. While we recognize that these factors do have an impact on an individual's health, NCQA's position is that we do not want to give health plans a pass, and that plans should be providing the same level of care to all members regardless of these nonclinical factors. This issue has come up. For example, I'll I'll highlight, our emergency department utilization measure that Danielle mentioned earlier in the presentation, that given the dynamics of emergency department use among the Medicaid population. So, for example, recognizing that issues with access to primary care may be a driver of using the emergency department among Medicaid enrollees. Other related, you know, feedback we've had around this, especially with emergency department utilization and the Medicaid product line, included the different Medicaid coverage by state and the concern about comparing health plan performance across states. So I would also just kind of reiterate here as well that with the risk adjusted utilization measures, you know, the intent is never to have, you know, zero ED visits or zero hospitalizations. The goal here is really to provide a a means for health plans to compare themselves to other plans that are similar, that are serving similar populations, and be able to identify ways that they can improve care coordination and care management for those populations that they are serving. Thanks, Lauren. There was a question that came in. The proposal to include denied claims, would this apply only to RAU measures or all HEDIS measures? So there are twenty one measures total in HEDIS that currently do not include denied claims for certain parts of the measure. The for the risk adjusted measures, as I mentioned during the presentation, we allow the use of denied claims for determining risk adjustment variables and for the hospice exclusion, but we don't allow them for measure events. For some other measures, denied claims are not allowed. Like, it's it's really just for the, primarily the overuse appropriateness measures in in that domain. They are not allowed for some of those measures for the, numerator and I think for others for the initial population. This proposal to include denied claims would include, both the risk adjusted measures and those, eleven other measures. So it would be, the the twenty one in total. Whatever measures currently do not allow denied claims, this this change would mean that that any of those measures would, would now allow them. So it would be across the whole volume, across the twenty one measures. But since this presentation was really focused on on the RE measures, we we kept it to that, but, it does extend beyond. Let's see. Okay. And then there was another question. What are the time frames of these for these changes? So all of these changes that we proposed here are would be implemented for measurement year twenty twenty eight. We are doing a a preview, you know, earlier in the process to, you know, really start vetting the changes early and, you know, make sure we're communicating them as soon as possible for organizations so that they're aware that they're coming. Let me see. I don't see other questions yet. We did oh, we got another question. We'll be sharing the the deck. Just just as a reminder, we will we will share the recording, and then also the transcript since we did have some audio issues. Okay. Yep. There was another question that came in. Can you share the the list of all twenty one measures affected by the denied claim rule change? Yep. So if that, if that does happen, so it's the the ten I believe it's ten measures in the overuse appropriateness domain. So all ten of those. And then also the antibiotic utilization for respiratory conditions measure and then the ten risk adjusted, utilization measures, which I can bring up the list of the the measures and the appropriateness domain just so I can read them out loud. Okay. So it would be the non recommended PSA based screening in older men or PSA, appropriate treatment for upper respiratory infection or URI, avoidance of antibiotic treatment for acute bronchitis or bronchiolitis or AAB, use of imaging studies for low back pain or LBP, potentially harmful drug disease interactions in older adults or DDE, use of high risk medications in older adults or DAE, deprescribing of benzodiazepines in older adults or DBO, use of opioids at high dosage or HGO, use of opioids from multiple providers or UOP, and then risk of of continued opioid use or COU. So there was a question about another question about the outlier thresholds, which is that the established thresholds may change with demographic and population data. How do we adjust for these changes? So, the main thing that we do is that we calculate the thresholds separately for the product line and, large age groups. So for example, we calculate a separate threshold for Medicaid, versus commercial. We calculate a separate threshold for, Medicare age eighteen to sixty four compared to, Medicare sixty five and above. So depending on the the product line and age group, there is a different outlier threshold. Let's see other questions. I think there was another one, Ezra, that maybe as a follow-up. How will you adjust for, for risk for SNP plans, which do not fall under normal distribution? So SNP plans don't have their own, risk adjustment model in our, risk adjustment portfolio, but we do have the separate reporting stratified reporting for, duals, dual enrolling Medicare, Medicaid, which is typically which is in the for for our risk adjusted measures, particularly PCR, plan all cause readmission, is, under the Medicare umbrella. So for the so there there isn't a a separately calculated outlier threshold that applies exclusively to SNP plans, but we address the fact that, they have a different patient mix, in those other ways. Other questions? Oh, here's another here's another question about establishing the outlier thresholds. So let me see if I can take that, since we're on the topic. What data please say again, what data are you using to establish the thresholds by product? So we have, third party claims enrollment data, that come from all product lines, so commercial, Medicaid, Medicare Advantage, and cover all age ranges. And from you know, it covers all regions of the country. And those claims databases we use to, look at the distribution of, of measure events, and that's how that's the data that we that we're looking at when we calculate the the outlier threshold. And there is another question that came in, Lauren. How should provider organizations adapt their documentation and coding workflows to align with the updated CMS HCC model and evolving HEDIS utilization measures for measurement year twenty twenty eight? Great. Thanks for this question. So for the risk adjusted utilization measures, we do publish, our shared tables, our risk weight tables, which are available in the NCQA store. This explains every single variable that is used for every single product line indicator and the risk weights that are used for those variables as well as how the ICD ten codes map to each CC and applying the hierarchies to get the HCCs. So those tables are updated every year. As we approach this this proposed update in our upcoming re estimation, we do plan on sharing some insights about those tables and changes in advance of their release so that plans do have a little bit of a head start in how to go about making those changes. Thanks, Lauren. Right. Ezra, there's another question about outliers. Can you consider doing an outlier threshold for SNPs, particularly c SNPs? I don't know if you wanted to respond to that. Interesting. I would be interested to hear what our advisory panels think of that suggestion, and our committees. I don't I don't have a great, I don't have a great answer at the moment, but Yep. It's Yeah. Definitely something we can, like, do yep. Because I remembered or I I mentioned earlier with the next steps. I know we, you know, we're gonna be, having discussions with our our advisory panels and other experts. So that's something we can definitely bring up with them. Any other questions? Right. So I think I'm not seeing other questions coming through. So I think we can actually just wrap up, I'll turn it back over to Akina. Thank you, everybody, so much. And thank you to the presenters, you know, to Lauren and Ezra. And, yeah, I'll turn it back over to Akina. Wonderful. Well, thank you team for a wonderful webinar. I wanted to share with everyone information about our upcoming Health Innovation Summit. We are celebrating our fifth iteration of the Health Innovation Summit that will be held this year in Atlanta, Georgia. This year, we will be at the Georgia World Congress Center right in downtown Atlanta adjacent to Centennial Park, and we hope to have all of you with us in Atlanta. So as a thank you for everyone joining us on this webinar, we have dropped in the link for you to register, and we also would like to provide you all a discount of a hundred dollars off of our current rate. So I have dropped in the chat for you all a discount code to use with your registration that will get you a discount to join us while we're there, and we will be talking more things on HEDIS while we're at the summit. So it's coming up quickly, and we hope that all of you all will find the time to join us in Atlanta. There will be a lot of really good content and even more amazing people to talk with about all the great things that are happening in quality. So with that, I wanna thank you all for joining us for a wonderful webinar. And as Danielle mentioned, we will be sharing with you all materials post webinar in just a few short days. If you have any questions for the team, you guys are more than welcome to connect with them via email. Otherwise, we do look forward to seeing all of you all with us in Atlanta in the fall. Thank you all, and make it a wonderful day.