Hello, my name is Omer Yousef. I want to thank the American College of Cardiology for inviting me to speak at the 2021 ACC Quality Summit. It's truly an honor and privilege to share with you the value of the EP device registry data and how data and metrics can impact patient care. The goal of this presentation is to share with you the role of the American College of Cardiology NCDR EP device registry to discuss what are appropriate use criteria and how are they interpreted. What is the appropriateness of ICD implants in the United States? How are appropriate use criteria incorporated in the NCDR EP device registry and how can these data and metrics impact patient care? What is the role of the ACC NCDR EP device registry? The NCDR EP device registry history dates back to 2005 after ScudHeft was published and CMS expanded the indication for primary prevention ICDs to non ischemic cardiomyopathy patients and they issued a national coverage determination at this time. And as part of this expanded reimbursement, CMS and Medicare required patients to be enrolled in clinical trial or a national registry that would support understanding of outcomes and that led to the inception of the NCDR ICD registry. Over 1800 hospitals participate in the NCDR EP device registry. In 2017, CMS revised the national coverage determination that ended the data collection requirement that was put in place back in 2005. There is, however, a new shared decision-making requirement, which is basically an encounter that involves using a shared decision-making tool whereby the provider and patient have engaged in a conversation to go over the relative risks and benefits of undergoing a primary prevention ICD. Many hospitals continue to report to the NCDR and many find significant value. Data collection occurs in all types of ICDs regardless of indication and irrespective of the payer. The device registry is very rich and a comprehensive source of data. Over 2 million implants have been recorded in the registry and over 80 peer-reviewed scientific articles have been published from all this data that has been collected over the last 16 years. It has had a significant impact and advancement of our understanding on device selection, outcomes, effectiveness, safety, equity, and efficiency of care. The EP device registry has become the bedrock for quality and safety improvement. It supports various measures including FDA post-market requirements, CMS coverage with evidence mandates, and it's become the model for other cardiovascular procedures such as the transcatheter valve therapy registry and the left atrial appendage occlusion registry. And it really serves as a foundation for translating clinical trial knowledge and guidelines into contemporary clinical practice. And it's also a model for professional transparency. It allows for self-accountability and it's a platform to advance science. So what are the appropriate use criteria and how are they interpreted? In order to understand and improve quality and procedural utilization, we have to look at misuse, overuse, identify areas of uncertain indications, and underuse. Misuse is applying treatment to the right patient in a manner that may result in harm. Overuse is applying treatments to patients in whom the risks may outweigh the benefits. And underuse is failure to apply treatment in those likely to potentially benefit from the therapy. The NCDR is a great platform to understand potential misuse of therapies while the appropriate use criteria provides greater insight into potential overuse of therapy and also identifying areas of uncertainties in indications. The traditional process of care involves a patient in whom a potential treatment is offered, which is then either used or not used, and then they are followed for potential outcomes, which may include survival, readmission, patient satisfaction, cost implications, as well as health status outcomes, which may look at symptoms, quality of life, and functional status. We have a host of other tools in our armamentarium that allow us to improve the quality of care we deliver. These include guidelines, performance measures, as well as appropriate use criteria. Guidelines are evidence-based review of what can be done, while performance measures are a subset of guidelines that must be done in order to observe the desired outcome. And appropriate use criteria are evidence-based review of what should be done. This is a conceptual model incorporating cost, quality, and metrics. Guidelines and performance measures have a strong emphasis towards delivering quality care and a much lower emphasis on the cost of care. On the other hand, efficiency measures focus and emphasize delivering cost-effective care without a strong emphasis on the quality of care. The appropriate use criteria may provide the equipoise bridging that gap between guidelines and practical considerations allowing delivery of quality care in a cost-effective manner. There are three categories when we think about appropriateness in the appropriate use criteria. The first is appropriate, in which case the benefits of the therapy outweigh the risks. A maybe-appropriate indication would be one where we don't have enough data of the risk-benefit ratio, and perhaps more research is needed. There may be clinical equipoise, however, these indications should still be reimbursed. And rarely-appropriate therapies are ones where an alternative therapy are potentially likely to be better, perhaps one in which the harm may be greater than the potential benefit. It's important to note that limitations in methodology will make mistakes, and all physicians should have some cases across maybe-appropriate and perhaps to a lesser degree to in rarely-appropriate categories. This was a study published in 2011 using data from the NCDR EP device registry. In this study, they looked at key clinical factors that were used in previous clinical trials to determine the rate of non-evidence-based ICD implants. Previous studies have demonstrated that in certain clinical factors ICDs may not be beneficial, and our guidelines generally have recommended against primary prevention ICD implants in patients who have newly diagnosed cardiomyopathy of less than three months duration, or who have had a recent myocardial infarction within 40 days prior to implant, or neocardial association class IV symptoms, or prior coronary revascularization within 90 days. And as such, this study looked at the rate of non-evidence-based ICD implants based on these clinical factors, and what they found was that a substantial number of patients over the course of nearly three and a half years did not meet evidence-based criteria for ICD implants. This was over 22% of implants. And these findings garnered a lot of attention and negative publicity. And they found enormous variability in the non-evidence-based ICD implants across different hospitals. So what you see here is that, you know, over 30, 40, 50% of non-evidence-based implants were occurring at a substantial number of hospital sites across the country. And this led to the publication of the appropriate use criteria. So in 2013, the ACC, in conjunction with Heart Rhythm Society and American Heart Association, amongst other professional societies, published a series of criteria to determine the appropriateness of ICD implants and cardiac resynchronization therapy. So what is the definition of appropriateness? An ICD or CRT implantation is considered appropriate in general when the expected value in terms of survival and or other health benefits such as symptoms, functional status, and or quality of life exceeds the potential adverse health consequences relating to the acute procedural risk and the long-term consequences of living with an implanted device. It's important to note that just because an indication is considered appropriate, it does not mandate an implant, but rather infers that it's a reasonable consideration. Remember, the appropriate use criteria evaluates overall patterns of care rather than adjudicating specific cases. A maybe appropriate indication is one in which individual physician judgment and understanding of the individual patient is necessary to determine the usefulness of device therapy for a particular clinical scenario. It's important to note that this should not exclude the use of ICD or CRT devices for such patients. The ICD appropriate use criteria are very complex and nuanced. There's 235 diverse clinical scenarios incorporating 369 indications. Of these indications, 106 are aimed for primary prevention devices, 71 are for secondary prevention devices, 40 are distinct to cardiac resynchronization therapy. There's a number of other ones that are distinct to dual chamber pacing, as well as a smaller number aimed at the appropriateness of generator changes. And the appropriateness is based upon the potential benefit of ICD as stratified by a number of factors, but namely near-cardio-association functional class, left ventricular ejection fraction, the temporal relationship to a recent myocardial infarction, coronary revascularization, the duration of guideline-directed medical therapy, and also whether or not there is a pacing indication, if there is bundle branch block and what the QRS duration is, and many more. So the appropriate use criteria have 150 distinct indications that are aimed directly at the ICD component of the therapy to prevent sudden cardiac death. In addition to the ICD appropriateness indications, there are distinct indications for dual chamber devices aimed at determining the appropriateness of the atrial lead, both for sensing and pacing, as well as distinct indications for CRT devices that are specifically aimed at determining the appropriateness of left ventricular pacing for cardiac resynchronization therapy. So what is the appropriateness of ICD implants in the US based on the NCDR EP device registry findings? Our group was instrumental in mapping the appropriate use criteria to the NCDR EP device registry data. And so we looked at 15 quarters of data spanning from 2016 to 2019. We were able to map nearly 79% of all device implants. About 21% of cases were not able to be mapped, and we'll get into that a little bit later. And when we look at the appropriate use ratings, what we found is that a majority of them, over 95% of implants, are considered appropriate. Nearly 3.5% are considered maybe appropriate, and only a small fraction, about 1%, are considered rarely appropriate. And this is both for primary and secondary prevention ICD implants. And when we look at the distribution of types of implants, over 3 quarters of ICD implants are for primary prevention, and about 22% are for secondary prevention. And when we look at the distribution across the types of implants, both single chamber devices are nearly equally distributed amongst primary and secondary prevention. In dual chamber devices are twice as more commonly implanted in secondary prevention as compared to primary prevention. And the opposite is true for CRTDs, whereby a much larger number of CRT implants are implanted in primary prevention as opposed to secondary prevention. And when we specifically look at the pacing indications to determine the appropriateness of both dual chamber pacing as well as CRT pacing, we notice that there is a greater distribution across the appropriateness categories as compared to the ICD categories, where a majority of implants were considered appropriate. In the pacing category for dual chamber, nearly 97% are considered either appropriate or maybe appropriate, and a very small number of only 0.6% are considered rarely appropriate. For CRT pacing, over 82% are considered appropriate, with nearly 13% as maybe appropriate, and about 5% of implants deemed rarely appropriate. And this is, again, looking specifically at the pacing component of the therapy. And so the question is, how do we incorporate all these appropriate use criteria indications into the NCDR EP device registry, and how does this impact you as an individual physician or as a health system entity? So each quarter, we receive performance reports if we participate in the NCDR EP device registry. It's a professional-level dashboard that allows us to view individual data about ourselves to gauge performance and quality and outcomes and appropriate use criteria metrics. It allows benchmark of one's individual care against their peers, either within the institution or nationally. These performance reports also provide quarterly risk-adjusted benchmark reports with performance measures and quality metrics, allowing comparison of institutional performance to other peer groups across the country. So this is an example of the dashboard that you may receive in these performance reports. So this is specifically looking at the proportion of ICD-CRTD patients that fulfill Class 1, 2A, or 2B guideline indications. And in this case, my performance was nearly 98% for this quarter. And you can see that you get an assessment over several quarters, and you can compare how you did in this quarter relative to previous quarters. In this second graph here, this is an example of a dashboard that's assessing the proportion of CRTD procedures classified as appropriate according to the appropriate use criteria. And in this quarter, the performance was 100%. However, you can see that in a previous quarter, the performance rating was lower at just over 77%. And so this allows you to compare your performance over a period of time. There's also a summary that you will receive for each metric description, and it allows you to compare how your hospital system performs. So this is a dashboard that's looking at metrics at the hospital level. So how your individual hospital compares to other US hospitals at the 50th percentile. So if we look at metric 52, this is specifically looking at the proportion of CRTD procedures classified as appropriate. And so this gives my hospital an 86.1% rate of CRTD implants that are rated as appropriate. The 50th percentile for all US hospitals, 90.1%. So it really allows you to compare how you do relative to other hospitals in the country. And if we look at metric 54, this is the rate of CRTD procedures that are classified as rarely appropriate. And so my hospital rating is 3.3%, and the 50th percentile for US hospitals, at least for this quarter, was 0%. And so what we're looking at and what may be most helpful is looking at trends. And if you start noticing that there are outliers, then it's worthy to explore this further. What are the indications that are being ranked as rarely appropriate? And it allows us to really dig deeper into the indications and explore potential areas of improvement. And so how can all this data and these metrics impact patient care? Because that's ultimately what we care about. So if we dig deeper to try to really understand what are the common indications of rarely appropriate primary prevention implants, what we see is that across the board, over 50% of rarely appropriate implants are being performed in patients who have advanced heart failure, near-heart association class four, with a reduced ejection fraction of less than or equal to 30% that are not candidates for cardiac transplantation, cardiac resynchronization therapy, or VADs, right? And so it really provides us some insight into a patient population that really has shown to not particularly benefit from ICD therapy. And a smaller number, but still significant, about 18% of patients that have a rarely appropriate implant are occurring in patients who are post-acute MI within 40 days, who have coronary anatomy that may not be amenable to revascularization and have not had any non-sustained VT. Even smaller number, about 9% of rarely appropriate implants occur in patients who are not a candidate for advanced therapies and are on continuous inotropic use. And lastly, about 7% of implants that are deemed rarely appropriate are occurring in patients who have a newly diagnosed non-ischemic cardiomyopathy with advanced heart failure, who have not been on optimal guideline-directed medical therapy for at least three months. And similarly, when we look at the top indications of rarely appropriate secondary prevention implants, we learned that there are three common indications which are considered rarely appropriate. So about 35% of patients who are deemed rarely appropriate for a secondary prevention implant are implants that are occurring in patients who have had a life-threatening ventricular arrhythmia, either BF or polymorphic VT, during the acute MI period within 48 hours, okay? And about 32% of rarely appropriate secondary prevention implants are occurring in patients who have had a ventricular arrhythmic event in the acute MI period post-coronary revascularization within 48 hours. And about 28% of rarely appropriate secondary prevention implants are occurring in patients who have had syncope with an ejection fraction of at least 50% or greater without any other structural heart disease or familial sudden cardiac death syndrome and a normal electrocardiogram. And so I think these indications highlight areas for potential improvement if we're implanting, if we're consistently implanting devices that are falling into these categories. And if we explore the most common indication of rarely appropriate dual chamber pacing, we see that 100% of rarely appropriate dual chamber implants occur in the setting of longstanding persistent or permanent atrial fibrillation. And we know that these patients are least likely to benefit from an atrial lead. And what about CRT pacing? So what are the most common indications of rarely appropriate CRT pacing? As we can see here that this distribution is quite varied. And that's because the indications for CRT pacing are quite varied based on the type of cardiomyopathy, the type of bundle branch block, the duration of the QRS, as well as ejection fraction and near-heart association functional class. And I'm gonna try to break this down for you to get a little better understanding of commonly encountered rarely appropriate CRT pacing indications. So we can see that a number of these have an ejection fraction of over 35%. We know that there's very limited to almost no data that suggests benefit of CRT therapy in patients who have an ejection fraction of over 35%. And so it would be reasonable to understand why these are considered rarely appropriate. Moving ahead, we see that there are a number of indications in which CRT devices are implanted in patients who have a QRS duration of less than 120 milliseconds. And we also know that CRT therapy is not deemed to be effective in patients without significant ventricular conduction delay. And lastly, the rest of the indications here relate to non-left bundle branch block with relatively narrow QRS durations ranging between 120 to 149 milliseconds with near-heart association class one to two heart failure. And based on guideline recommendations, these patients are perhaps unlikely to benefit from CRT pacing given presence of non-left bundle branch block and relatively narrower QRS durations with mild degree of heart failure symptoms. Here's an example of an appropriate use criteria, ICD scoring grid. And so I just wanna highlight and show how specific these indications are. So this is a table on non-ischemic cardiomyopathy and it requires several layers of input including the time of diagnosis, whether it's been within three months, what the QRS duration is, what the ejection fraction is and what the near-heart association functional classes. And based on these findings, there is a score that's generated, R for rarely appropriate, M for maybe appropriate and A for appropriate. Just to expand on this, we can see that these indications are very nuanced and specific. So here are some indications from the appropriate use criteria for secondary prevention devices. And this is a category of patients who have coronary disease, who had ventricular fibrillation or hemodynamically unstable VT in the setting of an acute MI within 48 hours. And we can see that based on their ejection fraction, they get different degree of appropriateness categories. And there are a number of these indications based on whether the ventricular event is recurrent, what the timing of the ventricular arrhythmia is relative to the MI, if there's non-sustained VT that occurred four days after MI and subsequently whether an EP study was performed that had inducibility. And based on these findings, you get an appropriateness score. So what I just wanna highlight is that it's very important to try to fill out as much detail as possible in the NCDR device registry forms, because all this data is captured from those device registry forms. And so if it is incomplete, or if the data elements are not accurately filled out, then that will reflect on the appropriateness rating for that implant, or it may be deemed unmappable if the various data elements are not captured because they may not be filled. We also looked at hospital variation of appropriateness in ICD implants. And so what we can see here, shown here in the blue are the number of appropriate implants. And number of hospitals is shown here on the X-axis and the appropriateness rate on the Y-axis. And so what we see is that the distribution of rarely appropriate implants as shown in red is really present across increasing number of hospitals. And so while there are a small number of hospitals that have a high appropriateness, a rarely appropriate implant rate, in general, the rate of rarely appropriate implant rate is distributed relatively evenly across the number of hospitals. And the same can be said about maybe appropriate and appropriate implants. And most importantly, a large number of implants are in fact appropriate. And looking at this just in another format, shown here on the X-axis is the hospital rate of rarely appropriate implants. And on the Y-axis are the number of hospitals. And so what we see here is that, over 90% of hospitals have a very low rate of rarely appropriate implants, which is less than 4%. And when we look at the number of hospitals with a very high rate of rarely appropriate implants, over 10% and over 20%, that's a very small number of hospitals, about 0.7% and 0.1% of hospitals. And shown here is the relationship of hospital volume to rarely appropriate ICD implants. And what we see here is that hospitals with the lowest procedural volume have the highest rate of rarely appropriate implants. Specifically, hospitals that are performing less than 100 ICD implants annually have rarely appropriate implant rates of upwards of 10% and as high as 25%. Whereas majority of hospitals that are performing at least 300 implants or greater per year have a rarely appropriate implant rate of less than 5%. The appropriate use criteria is much more than guidelines. It is really a bridge between clinical trial data, guidelines, and commonly encountered scenarios that we have to address for each individual patient. And so if we go back to the study that looked at the rate of non-evidence-based implants and showed that over 22% of implants were deemed to be non-evidence-based based on a number of clinical factors. And if we extrapolate those clinical factors into what the appropriate use criteria say, we get a slightly different picture because the appropriate use criteria is really incorporating much more factors beyond just trial-based inclusion and exclusion criteria. And so specifically, if we look at patients and we look at PCI or CABG having been performed within 90 days prior to an implant, that's considered a maybe appropriate indication. And if the patient had a preexisting cardiomyopathy prior to undergoing a PCI, then that's actually an appropriate indication. And what about near-card association class four symptoms? Well, if the patient has a VAD in place, that is considered a maybe appropriate implant. If they're a candidate for cardiac transplantation or CRT, it would be considered an appropriate implant. And what about if they're on Milbranone while awaiting heart transplantation? That's currently not addressed by the appropriate use criteria, but you can see that there is much more to the clinical context beyond simple exclusion criteria, such as having undergone coronary revascularization in the past 90 days or near-card association class four symptoms or an MI within 40 days prior to implant. And if they had a preexisting cardiomyopathy and underwent an implant within 40 days of an MI, that's considered an appropriate implant. And so perhaps many of these patients that were deemed as non-evidence-based implant perhaps may have either an appropriate or maybe appropriate indication if we apply the appropriate use criteria for their particular indication. And it's important to note that CMS has very specific criteria for reimbursement. They have a national coverage determination that really outlines indications for reimbursement. And a lot of these indications are based on clinical trial inclusion and exclusion criteria. And so specifically, their national coverage determination memo excludes reimbursement for certain implants. Namely, if someone has had a coronary revascularization within the 90 days, an MI within the last 40 days, near-card association class four heart failure, near-card association class one heart failure. And more recently, they expanded their coverage to include patients who may have otherwise fell out based on these indications that I just mentioned. But if they have a need for interim pacing. And so those patients are now considered to be reimbursable under the national coverage determination. But if we look specifically at certain indications that are excluded by CMS for reimbursement, namely non-schemic cardiomyopathy with an EF of less than 30%, who has been on optimal guideline directed medical therapy with near-card association class one heart failure, this patient would be considered appropriate by the appropriate use criteria. Similarly, non-schemic cardiomyopathy with an EF of 36 to 40% with near-card association class two to three heart failure is considered maybe appropriate. And newly diagnosed cardiomyopathy, so in the absence of having been on optimal guideline directed medical therapy for three months with advanced heart failure, class two to three, is considered a maybe appropriate implant according to the appropriate use criteria. PCI or CABG within three months in the context of a newly diagnosed cardiomyopathy with a reduced LV ejection fraction of less than or equal to 35% is considered a maybe appropriate implant. And lastly, pre-existing ischemic cardiomyopathy with coronary revascularization within last three months is considered an appropriate indication. So you can see that there are a number of conditions that are considered appropriate or maybe appropriate, but are deemed to be excluded in the national coverage determination, which may impact reimbursement of these devices. And so the appropriate use criteria and adoption of these in our quality metrics can potentially have future impact on coverage determination and hopefully expansion of reimbursement for certain types of indications. Since the inception of the NCDR device registry, this study looked at over the following four years, whether there were any changes in adverse events or adoption of optimal medical therapy or CRT therapy. And what they found was that adverse event rates declined after the adoption of the device registry and CRT therapy, as well as optimal medical therapy increased over the course of the following years. And so in conclusion, I just wanna highlight the potential impact of appropriate use criteria on quality. The appropriate use criteria examines and reduces practice patterns that deviate significantly from standard norms of care, provides quality improvement guide that applies clinical trial evidence and recommendations of practice guidelines to commonly encountered clinical scenarios, provides health systems and providers to understand their own performance benchmarked against national standards and provides a unique insight into patient selection. And it's possible that AUCs are likely to be incorporated by payers to encourage high value care delivery that improves outcomes at lower cost. And it's important to note that the appropriate use criteria is ranked procedures by the expected benefit exceeding the risks and costs of the procedures, and therefore are intended to serve as a guide and not a mandate. And in summary, the ICD-AUCs are really a new step forward in improving the value of EP procedures and good evidence-based care really should result in a mix of both appropriate and maybe appropriate implants for most physicians. And while the ACC has really empowered physicians with scientific data and tools with the aim of improving the quality of cardiovascular care delivery, they're also empowering patients in the era of digitalization that we're living through. So patients have the ability to search which quality improvement efforts are being employed in their own healthcare systems. And they're able to go to ACC's CardioSmart.org website and really it allows them to access each hospital profile and learn about the different quality measurement efforts that are being undertaken. And this program really provides unique opportunities to publicly showcase their commitment to continuous quality improvement. This is an example of what patients can see when they search an individual health system and they can specifically see which NCDR registry the health system is participating in. They're also able to review publicly reported outcomes from the various registries of individual physicians. I'd like to thank you for your attention. And I'd also like to thank my colleagues at the Mid-America Heart Institute, specifically Dr. John Spertus, who's been an inspiring leader and mentor and really a force in health status outcomes and quality of care work. Kevin Kennedy, our Senior Vice Statistician and Tracy Breeding. I would also like to congratulate and thank the American College of Cardiology on their tireless efforts and contribution to enhancing cardiac care delivery. I can think of no other professional society within medicine who has invested this much into improving the quality of care that we strive to deliver and improving the lives of our patients. I'd like to specifically thank Beth Denton, Christina Kudros and all the ACC staff for all your support and dedications towards these quality programs. Thank you for your attention.

EP device registry data

patient care

appropriate use criteria

ICD implants

comprehensive source of data

quality and safety improvement

accurate data collection

cardiovascular care