Hello, everybody. I want to welcome you to the closing general session of the Quality Summit 2021. My name is Ralph Brindis. We're going to have as our closing Quality Summit lecture keynote, Pamela Douglas. As you can see, we've had the Quality Summit keynote lecture for the last seven years. We're still getting buzz from ZDoggMD's and Quinn Caper's lecture, but I'm particularly thrilled to have my close friend Pam Douglas today talking about using quality to improve the evaluation of patients with chest pain. Dr. Pamela Douglas is the Ursula Geller Professor of Research in Cardiovascular Diseases in the Department of Medicine at Duke University and Director of the Multimodality Imaging Program at Duke Clinical Research Institute. During her 30 years of experience, she has become renowned for her scientific and policy work in improving the quality and appropriateness of imaging and clinical care. A leader, leading clinical trialist, her publications utilizing our NCDR registries, and through the development and dissemination of national standards for imaging utilization, informatics, and analysis. She has been among the pioneers in a number of areas, including heart disease in women, sports cardiology, and cardio-oncology. Dr. Douglas's wealth of experience includes authorship of over 500 peer-reviewed manuscripts and 30 cardiovascular practice clinical guidelines. Pam has served as the President of the ACC and as President of the American Society of Echocardiography. She has had key roles as the Chief of Cardiology at both University of Wisconsin and Duke University. She currently serves on the External Advisory Council of the NHLBI and also on the Scientific Advisory Board of the Patient Advocate Foundation. Pam has published numerous research from the NCDR, and these just highlight a few. The President's page on the question of quality, why national benchmarking, the impact of ethnicity and gender differences, and angiographic coronary disease prevalence and in-hospital mortality from the NCDR registry. A key landmark paper on showing the safety of drug-eluting stents occurred from her merging a database from the NCDR-Cath PCI registry and Medicare claims data showing safety and clinical effectiveness of coronary stents in elderly people, looking at over a quarter million Medicare patients. Another seminal paper from New England Journal looked at low diagnostic yield of elective coronary angiography, a topic that she'll be talking about today. Also, she's been a leader in our appropriate use criteria efforts, including that of coronary revascularization. Pam has been one of the early leaders in the arena of diversity, equity, and inclusion, with multiple seminal manuscripts on the topic, including recently co-chairing the Task Force of Diversity, Equity, Inclusion, and Belonging in the 2021 ACCHA Document on Professionalism and Ethics. In 2020, Dr. Douglas was the college's second recipient of the ACC Distinguished Award for Leadership in Diversity and Inclusion, being bestowed further with the honor of now having this award named as the Pamela S. Douglas Distinguished Award for Leadership in Diversity and Inclusion. Thanks, Pam, so much for agreeing to be this year's keynote speaker for our Quality Summit. Thank you, Ralph, for that very kind introduction that would make my mother blush. It's a special honor, of course, to be presenting the Ralph G. Brindis Lecture, which is the title of the keynote for the Quality Summit. And here is a picture of Ralph at leisure. All of you have seen him in his tie and jacket and formal, so it's worth seeing that he can chill as well. So the title of my talk is Using Quality to Improve the Evaluation of Patients with Chest Pain. My relationships with industry are several, but the only one which has any relevance for what I'm talking about today is HeartFlow, that's underlined in the bottom left, from whom I receive a institutional research grant. So the framework of what I'm going to talk about today was developed almost 15 years ago at an ACC conference about achieving quality in cardiovascular imaging. And at that time, we came up with a schema that we start with a patient and then proceed to select that patient. I mean, we need the right patient at the right time, but we need the right test. We need quality in the way the images are acquired and interpreted. We need those results to be communicated to the care team and eventually translate into better patient care and hopefully better outcomes. So how does this affect all of you, most of whom are involved in procedural registries? Well, if you substitute procedure for test, you can see that you need to pick the patient, select the patient, the right patient at the right time. You need to pick the right procedure. Is it a percutaneous procedure or an open procedure? The procedure needs to be performed well and interpreted properly. And this is the sweet spot for most of the procedural registries. It's the central part of it. But unless the patient is selected well and the procedure is selected well, it really doesn't matter how well the procedure is performed. Similarly, unless the results are communicated properly and translate into better patient care, the quality of the procedure becomes less important. So you can see this framework can be adapted for what you all do in the NCDR registries. And of course, this is particularly relevant from those of you from the chest pain accreditation group here today, which actually encompasses the entire range of this quality sequence where you need to figure out what patients need to enter into a chest pain evaluation process. You need to select a test, whether it's diagnostic angiography for ST elevation or high sensitivity troponin for somebody with less significant EKG findings and clinicals. And then, of course, whatever image or test you select needs to be performed well and interpreted properly with the results communicated to be able to translate into better patient care. Here you see in chest pain version 7, I've highlighted some of the phrases that are particularly relevant for what we'll be talking about today. Appropriate risk stratification. Appropriate noninvasive testing strategy. Early access to care. Transitions of care. Cross-department processes with results communication. And of course, in any quality initiative, identifying gaps, measuring results, and revising current processes to meet goals. So let's get back to imaging and the imaging schema here. And let's start by talking about a patient. Here's a typical patient that was enrolled in the PROMIS trial, which was a 10,000 patient trial for diagnostic evaluation strategies for patients presenting with new onset but stable as opposed to unstable chest pain. The average patient in this trial was a 61-year-old woman with new onset chest pain, had multiple risk factors, obesity, hypertension, hyperlipidemia, smoking, sedentary lifestyle while present in more than half the patients. Symptoms were generally atypical and occurred at rest as well as with exertion. Using the Diamond and Forrester algorithm, the likelihood of significant obstructive CAD was greater than 50%. And using a Framingham risk score, the risk of a cardiac event was 17% in the next 10 years. So not particularly low risk. And as you think about care here, here's the patient. What do you do next? What do you need to know about this patient? So to understand that, let's look at stable suspected heart disease in 2021. First of all, I don't have to tell this group, cardiovascular disease is the number one killer and often presents with chest pain. But the volume of new stable chest pain patients presenting every single year in the United States is staggering, greater than 10 million. As it turns out, obstructive coronary disease is unlikely, only in about 10 or 20% of the patients. And the outcomes are excellent with CV death or MI about 1% per year. This raises concern for possible over-testing and also what, which is a kind of quality concern. And also you'll note as being quality experts, we have many evaluation approaches with no consensus strategy, which is a flag for a potential quality issue. So let's look at patient selection. Let's just determine who needs testing. And all the guidelines, the US, the European guidelines all agree that estimating pretest probability or PTP is a first step. As it turns out, the Diamond and Forster algorithms that are in our current guidelines and have been recommended for years, actually overestimate pretest probability, which has been corrected now in other algorithms that use contemporary cohorts and CTA data. As it turns out, adding calcium score or just simply risk factors further improves prediction. So let's look at this. Here's data from the PROMIS trial comparing Diamond and Forster risk with actual anatomy in 4,400 patients of those who had obstructive disease. And you can see that the pretest probability in this plot on the right in the gray bars is Diamond and Forster. And for every category of angina, the three columns, and for men above and women below, the gray bars are higher than the black bars, which are the actual CAD presence in these patients. As a result, pretest probability is decreased by 50 to 70% over the old modified Diamond and Forster. As an example, women under 60 years old, all women under 60 years old, all are less than 15% risk, which is often a cut point for testing, versus greater than 15% with typical angina in the old guidelines. And in the new, men with atypical angina are rated as 15% likelihood of coronary disease versus 34% in Diamond and Forster. So you can see there's a dramatic change in who might be eligible for testing by risk. Here's a paper that we published last year in the fall. Could we improve this revised pretest probability even further? And does adding risk factors or calcium score actually improve CAD prediction? Because Diamond and Forster simply includes age, sex, and quality of chest pain. And so we combined a number of data sets to develop a machine learning model in 41,000 patients with validation in 15,000 patients. And as it turns out, I wouldn't be presenting to you if it didn't improve accuracy. And you can see in the bottom right the increase in the area under the curve with subsequent additions of pretest probability. And if we look at a cut point of 5% for testing individuals, you can see that using pretest probability alone, either revised pretest probability from the European Stable Chest Pain Guidelines would recommend at 5% would recommend that 89% of patients receive testing. If we add risk factors, now that we reduce that substantially, 27%, to only 62% need testing. If we add calcium score, only 46 under half need testing, further testing, of course, beyond the calcium score. So let's look at this concern for possible overuse that has and gave rise again about 15 years ago to the ACC Appropriate Use Criteria, which was largely based on expert consensus. But we do have some data suggesting harm from overuse. After all, it's a non-invasive test. It's not that big a deal, perhaps, than an invasive procedure. Let's look at those data, which were generated from the CAHPS PCI Registry. So important for you all to know how some of the data are used. So we took the CAHPS PCI Registry and linked it to Medicare Part A hospital and Part B physician billing data and identified a quality issue here and tested a previously identified quality issue, which was that routine surveillance non-invasive testing after PCI is one of the most common, rarely appropriate indications for stress testing. This was identified in numerous data sets outside of NCDR. So our approach to examining this was to identify PCIs at that time, 656 NCDR sites, and we looked for non-invasive testing in the Medicare, these are greater 65-year-olds, in Medicare bills within 15 months after their PCI. And our hypothesis was that higher frequencies of testing, those on the right, would indicate surveillance testing, whereas lower frequencies there on the left would indicate symptom-driven or appropriate testing. And so that is the bars there are the sites and the incidence of testing post-PCI. So you can see the middle, the 50% bar is about 40% of patients being tested. As you know, the results were the mean here is that 32% of patients had stress testing overall, but there is high variability. You can see there's a lot of spread in those bars with sites ranging from testing in under 10% of their PCIs to almost two-thirds of the PCIs. We further examined this to look at the timing of testing. And here we had the symptom-driven group with the lower prevalence of testing at the bottom, which is pretty much a flat line over the time since PCI. But you can see in Q4, there's really distinct peaks at 6 and 12 months that increase with the increasing testing at those sites. And there's even a little blip at three months early on. But clearly, stress testing is driven by the months since PCI supporting the hypothesis that this is surveillance testing. And then when we look at these quartiles of testing use, we can see no difference in mortality, no difference in acute MI, but repeat revascularization is higher significantly in all the quartiles of use above that symptom-driven quartile, indicating that there was potentially unnecessary revascularization driven by excess use of non-invasive testing, which is deemed really appropriate, so supports the AUC guidelines. So finally, we do need to identify those then with limited benefit from testing, who are these people? And using PROMIS, we identified a lower score and found out it could occur in any pretest probability group. So here's the PROMIS risk score. We took 4,600 patients in PROMIS, right? These are patients with new onset symptoms with unknown cardiovascular disease. As it turned out, 27% of them had what we would call no risk. They had no calcium, no stenosis, no plaque, and no events. And we could predict this low risk group using 10 clinical variables, and we validated that in both the Scott Hart and the Dan Nykad cohorts, and the C statistic was 0.76, which is pretty robust and consistently better than both the Framingham risk score and the Diamond and Forster risk score shown in the AUC curves in the upper right. The bottom right horizontal bar graph shows a significant prevalence of minimum risk, of course, in the 0 to 5%, which is the top bar, but also in the 5 to 15%, which might be tested. And there's still a smattering in those above 15%, suggesting that there's some value for using the PROMIS risk score in conjunction with the pretest probability estimates to triage patients for testing. Now let's turn to test selection. We have many choices in these patients. Current stress testing performance, the dominant choice, is about 80% sensitive and 80% specific. However, there's both a low prevalence of positive stress tests and obstructive CAD, both about 10 to 15% in multiple populations, suggesting that positive stress tests are poor indicators of obstructive CAD, and in fact that using CT may improve cath yield. Let's look at these data. So here's a closer look on a patient basis to help us with understanding the impact of those numbers. One in five patients having stress tests, current generation stress tests, will have erroneous test results, that's what 80% sensitivity and specificity means, may lead to additional tests that aren't needed, time costs, and even a misdiagnosis. The prevalence of 10% of obstructive disease means that nine out of 10 will not have significant CAD, and 95% will not need surgery or stent, and only one or 99 patients will be free of an event in the next year. Pretty good prognosis. So let's look at what happens in the cath lab. These are data from the cath PCI registry on the diagnostic yield of elective diagnostic caths to rule out CAD. This is about 10 years ago. We looked at almost 400,000 patients, and what we found really surprised us, and the paper had quite an impact, only 38% of these patients had obstructive coronary disease, defined with the usual cut points, and fully 62%, or almost two-thirds, had no obstructive disease, or no what we might think of as actionable disease, since most of these people were being cath to do a PCI or CABG, or to make sure they didn't need a PCI or CABG. While this seems very different from what we expect, after all, we don't tell our patients when they're going to the cath lab that their chances of having the disease we're looking for is worse than a coin flip, but they are consistent with Bayesian principles. Given the very low prevalence of CAD in this group, any positive test is more likely to be a false positive than a true positive, so we really need to understand the source of the problem. Is it the test not being accurate enough, or is it that the patient selection for testing is not very good, or is it both? We went back and checked this again. The results created so much of a stir that we were asked to go back and re-look at the data more recently. Five years later, we went and looked at more current data, this time almost 700,000 patients. We had very similar findings. You'll record it was 62% in the earlier study, and now 58%, so really hadn't moved the needle very much. It turns out non-invasive testing was only done in about two-thirds of patients, and the other one-third went direct to the cath lab. Of those non-invasive tests, 52% or half of them were abnormal, but just under 10% were high-risk. Only the high-risk results were helpful in selecting patients for obstructive coronary disease. The other non-invasive test results, either intermediate or low positivity, had no value beyond clinical factors. You can see that in the graph on the right. The little dotted line of 42% was the overall prevalence of obstructive disease. You can see the high-risk bar is quite high, but the other positives, negatives, and no tests were not significantly different from each other. More recently, there's been some confirmation about the difficulty of stress tests in identifying patients with obstructive disease. Here are data from the ischemia trial. And as you all know, this primary question in the trial was about whether patients had a benefit to cath and revascularization over guideline-directed medical therapy. But on the way to getting into the trial, patients had a safe DCT to ensure the absence of left main or triple vessel disease so that there'd be equipoise about not revascularizing these patients. So of the thousands of patients who had core lab adjudicated moderate to severe ischemia on stress testing, 32% of these were excluded from the trial by CTA. The vast majority of these are roughly 20% of the patients with positive stress tests, very positive stress tests, had no obstructive disease, and about a quarter had an unprotected left main, and then there were a smattering of other reasons. But clearly, the CTA added tremendous value in diagnosing a presence or absence of coronary disease over and above conventional stress testing. So let's look at that CT issue. And as I've said, standard testing has frequent false positive and negative results. And if we use CT, perhaps we can avoid cath without actionable disease and also increase the conversion rate of a diagnostic cath to PCI. My interventionalist friends say they have no interest in doing caths in patients who don't have coronary disease, who aren't gonna have the angioplasties because the angioplasties are the real fun stuff. So you'll see the two NCDR data bars on the left, 38 and 42%. Soon after that, the VA group published slightly better at 52%. And then we had two large randomized trials that randomized people to usual care with some form of functional testing or CTA. Scott Hart had a similar prevalence of obstructive disease, 44% in the functional group, but 80% prevalence of obstructive disease in the CT group. Promise was 48% in the functional testing group and 72% in the CT group. So CT seems to improve use of cath lab over and above stress testing. So how about event? So predicting CAD is one thing, but what we really wanna do is predict events. And unfortunately, it turns out that's really suboptimal using both stress testing and CT. Part of that is because of the large negative test cohort, the majority of events occur in this group, those without obstructive CAD. And also plaque is very common, but events are very rare. So here, I'll walk you through this slide. On the left, you see from Promise, a data showing their positivity rate with CT and stress, very similar at about 12% of the population. So most patients don't have positive tests. In the middle graph, you can see the hazard ratio for an event by test results, means that both CT and stress were highly predictive, three and a half to four fold greater risk of having an event if you had a positive test. But on the right, you see that the green population, those with negative tests had two thirds of the events, whether it was a CT or a stress test that they had, suggesting that neither one is a great event predictor, at least a straightforward positive stress or obstructive CAD by CTA. If we look, however, at non-obstructive disease, which of course is silent on a stress test, we see that there's a significant increase in event rate with lumpy bumpy disease that can be detected by CT over those with clean coronaries. And that's circled there at 2.9%. The 7.7 and 10.1 are single and multiple vessel disease. Now, of course, stress tests for ischemia rarely detect this kind of disease and therefore not able to triage patients at risk. And the confirmed CTA data registry, about 30,000 patients, you can see non-invasive disease in these survival curves, no diseases atop, and then one, two or three vessel non-obstructive disease are the curves going down from there. Now, I'm not gonna say too much about the image acquisition interpretation because you all are not imagers, but I am gonna say a little bit here about non-invasive hemodynamics and what we might help to understand the significance of coronary lesions. We do know from the FAME study 20 years ago and other studies that coronary anatomy is not the same as coronary physiology. And regular CT angiography is excellent in detecting stenosis, but that doesn't help us since many patients with seemingly obstructive lesions, certainly over 50% and even some over 70% do not have hemodynamic plaque, significantly hemodynamically significant plaque. And that's important since angiography guided by fractional flow reserve is superior to PCI guided by stenosis. It turns out using machine learning, there is a technique called fractional flow reserve by CT, which creates a finite element, a model of the coronary tree using conventional CT images, and then uses a big data artificial intelligence process to simulate flow under various conditions and can create an FFR data very similar to that obtained by an invasive flow wire. In testing this against all the other forms of non-invasive testing, which the Pacific One trial did, it was a small trial, 200 patients, but the strength is that every test was done in every patient. And the gold standard was the invasive FFR rather than stenosis. And as you can see, the best test for predicting invasive FFR was CT plus FFR-CT or FFR-CT alone, far superior to any of the PET or SPECT or anatomic CT. So what do we know about using this FFR-CT in practice? The platform study was done about five years ago, looked at sequential observational cohorts, people plan to go to the cath lab. And in the intervention arm, we added a drive-by CTA plus CFFR-CT on the way to the cath lab. And as it turned out, when the physicians saw those results, they canceled 61% of the cath. As a result, the rate of cath without obstructive CAD fell from 73%, little higher than what I've shown you before, down to 12%. With no impact on safety or CAD detection, you can see that the obstructive CAD was 27% in each cohort. So we didn't miss any CAD by canceling those caths. And of course, we saved a lot of money, over 3,000 per patient by not doing the invasive procedure. So finally, let's look at patient care and clinical risk stratification by plaque detection and preventive treatment of non-obstructive CAD. Of course, we think of patient care and imaging as being on opposite sides of the spectrum, but there is a continuum from the technical capabilities of an imaging test through its diagnostic performance, how it changes your diagnostic thinking, your therapeutic thinking, and your strategy. And all of that can truly change clinical outcomes. So here's a data, again, from the CONFIRM study, looking at statin use in patients with non-obstructive disease, which, as we recall, is only found by CTA. And this is 10,000 patients with a follow-up level just over two years, and about 5,000 of those patients had only non-obstructive disease, and we're comparing them to the 1,000 who had no plaque. And as you can see on the left, those with no plaque, and this is an observational study, so it was people who had beyond statins versus those who weren't. In those with no plaque, there was no difference in outcomes between statins and no statins. But in those with non-obstructive disease, only non-obstructive disease, we see a reduction, a 68% reduction in death, in mortality, amongst those using statins versus not. Now, this is observational data. There has never been a randomized trial of treatment of non-obstructive coronary artery disease, but clearly we need one. Let's turn now to the randomized controlled trial data that we do have about CTA and how that might improve outcomes. So the first thing we do from both PROMIS and from Scott Hart is that it increases use of preventive medication. So detecting that obstruction, actually helps us in writing prescriptions for aspirin, statins, beta-blockers, and ACE inhibitors. And in the Scott Hart trial on the right, you see new antiplatelet prescriptions above and new statin treatments below, substantially greater in the CT group compared to the functional testing group. And in Scott Hart, we see a dramatic increase And in Scott Hart, we see a dramatic reduction in death and MI over five years, a 40% reduction in CV death MI and non-fatal MI on the left. And then in the far spot on the right, this is true for virtually every single subgroup. So what's next in this area? I think we need a head-to-head randomized trial of functional testing versus CTFFR, but we also need to know what to do with those very low-risk patients. Do we stress, do we test them, or do we just do watchful waiting? So the PRECISE trial, which just completed enrollment a few months ago and should be available for results next fall, is a prospective, pragmatic, randomized clinical trial, which will determine the effectiveness of precision-guided diagnostic strategy for stable chest pain versus usual care. The approach care will be physician discretion, either a stress test or direct-to-cath, whatever the doc wants to do. And the precision pathway will be assigned by the promised minimal risk score. The lowest risk, 20%, will have no testing, and the rest will have CT with selective FFR or CT. And the end point, primary end point, will be an effective chest pain evaluation, adverse events, of course, but cath without obstructive disease or vertusia with secondary end points of other clinical outcomes, symptoms, quality of life, and cost. And here's the schematic for that. You can see the usual care on the left with the range of risk, getting functional stress testing or direct-to-cath. And on the right, the precision evaluation, guideline medical care is the green box, and the higher-risk people will get CT plus FFR, including those with known non-obstructive plaque or anybody whose pain is poorly controlled in the watchful waiting group. So let's summary. How do we apply quality principles to imaging to improve the evaluation of patients with chest pain? First, I hope you can appreciate that these stable chest pain patients are complex, and there are many opportunities for quality improvement at every step of that quality framework, particularly in the areas of patient selection and test selection. I also hope you've seen today that real-world data from NCDR has helped informed patient selection and the appropriate use criteria, and identified a better need for CAD risk stratification and for testing with CT. As always, careful attention to quality principles and quality domains will, we hope, reduce unneeded testing, unnecessary casts, and costs while improving preventive care and clinical outcomes. Thank you. Pam, thank you for that brilliant presentation. It truly highlights the challenges of non-invasive assessments of potential coronary artery disease in determining the best cost-effective strategies for decisions when to pursue diagnostic catheterization and probable needs for coronary revascularization. Your work with PROMIS and PERFORM clinical trials really offer a great argument for the use of CTA and CTA combined with non-invasive FFR as a non-invasive strategy of choice. When we look at the national utilization of CTA in patients undergoing catheterization or PCI in the NCDR-Cath PCI registry, the actual utilization rate was only approximately 3%. From your perspective, what are the present barriers that are hampering CTA plus or minus non-invasive FFR utilization in the U.S. for diagnosis of potential coronary artery disease? Thank you, Ralph. It's such a pleasure to give this talk and be part of the session today. I think CT is growing substantially. We have learned to use it in the emergency room because of several trials there that I didn't show, but pretty definitive trials. And of course, now we're moving on to high-sensitivity tronins. But I think we're also learning to use it in patients with stable chest pain who actually make up a small minority of the NCDR population. So the numbers may be a little misleading. In Europe, the guidelines really rank CT on a par, if not higher than conventional stress testing. And of course, we've got new guidelines coming out this fall from ACC AHA, which of course, we don't know yet what they will say. My guess would be that it's gonna elevate CT to be on a par with other tests, which will really be the nudge that's needed for people to learn to use this. It exists in radiology departments often, cardiologists aren't as familiar with it, but we're very capable of change. And if it provides better patient care, I'm sure that our physicians and our care teams will learn to use CT side-by-side with the tremendous information that we get from perfusion and other forms of functional testing. I'm always concerned about issues that everybody, the hospitals and offices own nuclear machines, and maybe there's bricks and mortar issues related to how available CTA is a possible barrier. What are your thoughts along that line? Well, there's some evidence that groups that own their own nuclear cameras test at a higher rate than those that don't, but it's not, it may be an access issue, right? As well as a familiarity issue and not necessarily a financial issue. So we need to be very cautious about drawing conclusions in cases like that. Access acutely in the need with a chest pain patient, it's not always in the CT kind of mindset. They're used to doing a test a week from now, not necessarily in the next half hour. And that is something that again, we can change, can be done. And faster, newer scanners are much faster and require less patient manipulations like a beta blockers or to get good scans. You know, I wanted to get your thoughts as since you're a truly a visionary leader for the college in the area of diversity, equity and inclusion, and also in CV quality, how can the NCDR registries and our accreditation programs influence patient population health disparities to actually accomplish health equity? I'm sure you've given a lot of thought to this. Yeah, I think NCDR is front and center in health equity improvement and reduction in disparities. And I think that starts with the staff and physicians and the care team that's involved to make sure that we are creating data collection forms and working in hospitals that are diverse, that have diverse patient populations. If we're only looking at wealthy private hospitals as opposed to safety net hospitals, it's gonna be very hard to get the data that we need to do improvements. So diversity is important from a representation standpoint of what hospitals we have. It's also important from a talent standpoint and it's important from a connection to patients where patients of the same race have greater trust and are more likely to agree to procedures and to do preventive care and so on if there's the trust that comes from a same gender or same race ethnicity pair. But once we get past that workforce type diversity we get into disparities themselves. And anytime we have data, if we should be able to look for disparities that patients particularly maybe of black patients are less well-represented, for example, in peripheral aortic valve procedures. And is this because the hospitals that are predominantly black populations are not represented? Is it because the disease of calcific aortic stenosis is different in black people than it is in white people, which it is. And it's less prevalent, similar to osteoporosis or is there some disparity there that we need to fix? But by doing the universal data collection that NCDR does, if we analyze it through a diversity lens, through really looking at issues that we might not normally look at to identify differences in representations, differences in quality, differences in referrals, that will give us the clues as to where those health equity efforts need to be directed and then we'll be able to measure their success. Wow, that's a terrific answer. And our final question, do you have any words or advice or encouragement to our dedicated Quality Summit audience here today in these very trying pandemic times in maintaining our own vigilance for quality assessment and pursuing CV quality? Gosh, you all are the best. You're the ones that build the platform and foundation for people like me to come and do research to be able to give talks like this, but it all starts with what you're doing on the ground floor level. And we need to, even though COVID is top of mind for everybody right now, and it's difficult to be in a work environment, a hospital environment, without worrying about infection as opposed to simply coronary disease, it's even more critical now that we do take unbiased measure of what we are doing and how well we're doing it because these care patterns are changing. People are not presenting for procedures in the same way. And so we can't assume that the data that we collected a few years ago really holds true today. And so only with good ongoing quality improvement, quality detection analysis and improvement, can we ensure that we're providing good care today and setting ourselves up for good care tomorrow. Pam, thanks so much for sharing all your expertise and thoughts with our group here in our closing summit, closing session of the Quality Summit. You're so welcome. Thank you. Ty, what a great way to wrap up Quality Summit with the Ralph Brindis Lecture delivered by Dr. Pam Douglas. I'm sure all of us have been inspired by her thoughtful presentation. I couldn't agree more, Barb. And I know it wasn't the same as being together in person, but we've had amazing sessions the past few days. And if you didn't get the chance to attend all the ones you wanted to, they are now available for the next 90 days. Also, there are 18 on-demand sessions, which you won't want to miss. These are also available for 90 days. Continuing education credits are given for all sessions. So get those credits if you need them. And finally, we want you to save the date, September 14th to 16th, 2022, when we really do hope to see you in Los Angeles, California. So until then, thank you and stay safe. Thank you.

Dr. Pamela Douglas

Quality Summit 2021

chest pain

CT angiography (CTA)

coronary artery disease

preventive care

barriers to utilization

data collection