Hi, my name is Philip Goodney. It's my pleasure today to talk about the undertreatment of peripheral arterial disease as part of the symposium on the plan to develop a PAD treatment plan, including goal-directed medical therapy as well as emerging therapeutics. Thank you. These are my disclosures. I want to make sure that we express our appreciation to both the American Heart Association and the Food and Drug Administration who provided funding support for some of the work that we will share today. I have three goals for the next 20 minutes or so. First, I'd like to summarize some recommendations for annual integrated preventive measures for patients with peripheral arterial disease aimed at limiting the progression of their disease and improving their outcomes. Second, I'd like to outline the adherence to these measures as well as treatments for severe peripheral arterial disease. Then third, I'd like to talk about ways to potentially find some pathways forward to improve the care, even given the limitations in the dissemination of these preventive measures for patients with peripheral arterial disease. I'd like to express my appreciation to the American College of Cardiology for asking us to be part of this work. First, let's talk about the recommendations for annual integrated preventive measures for patients with peripheral arterial disease. As it's well-known, and this is work that we published in the early 2012, that there is significant variation in care for patients with peripheral arterial disease. This is a map of Medicare patients across the United States formed from a cohort of patients with both diabetes and peripheral arterial disease between 2007 and 2011. The colors in the map correspond to amputation rates for those patients with severe peripheral arterial disease and diabetes. The darker colors, such as the red and brown, correspond to rates of amputation from 3-6 per 1,000 Medicare patients, whereas the lighter regions are regions where amputation rates are lower, usually on the order of 1-2 per 1,000 Medicare patients. But as you can see across the country, there's wide variation and quite a few differences in the risk of amputation by different regions of the United States, and there are several factors which influence those differences. When we start to think about those factors, we study them in the context of whether or not patients had diabetes, as well as some other important factors that appear to drive that risk, and this is a figure that derives from that same cohort of Medicare patients. On the x-axis are four different groups of patients that I would ask you to consider, and on the y-axis is the major lower extremity amputation rate, and this is above and below the amputation, not including toe amputations. As you can see, among all Medicare patients, the amputation rate remains relatively high, about 6 per 1,000 Medicare patients. But when we look in patients with diabetes, that risk goes up and then it incrementally rises if patients also have diabetes and prefrontal arterial disease. The risk continues to incrementally rise, however, if other factors are considered, such as if the patients are African-American in race and if they reside in certain high-risk regions of the United States. Adding all these up, we see that while the consideration of prefrontal arterial disease and its prevention is important, it is especially important in certain ethnic groups, certain racial groups, and in certain parts of the United States. There are management strategies that are suggested to limit amputation for patients with diabetes and critical endoschemia or severe prefrontal arterial disease. It's a multidisciplinary problem, and as such, there are multidisciplinary specialties, societies, and regulatory groups that have all provided recommendations about how the care of patients with severe prefrontal arterial disease should be guided, especially in terms of preventive measures which should be provided to these patients over time. You can see from federal organizations like the Health Research Services Association, to professional societies like the American Heart Association, the American Podiatric Medicine Association, and the Society for Vascular Surgery, to other government organizations such as Veterans Healthcare or national initiatives such as Healthy People 2020 and the American Diabetes Association, that there are a variety of program statements and white papers and position documents on how we should try to prevent amputation from prefrontal arterial disease. Many of these preventive measures have certain elements that tend to be in common. Those three key elements that we see as integrative preventive measures, which can be provided at low cost and should be readily available to all patients with prefrontal arterial disease and especially those with prefrontal arterial disease and diabetes seem to fall into three categories. Hemoglobin A1c testing, diabetic foot care, and some measure of vascular assessment. There are small differences within the recommendations across some of these different bodies in terms of how the testing is provided and what it entails. But in general, those three key components, A1c testing, diabetic foot care, and some vascular assessment, all seem to run true irrespective of the organizing body making that recommendation. There are other elements that are often recommended such as statin therapy and other aspects of medical management, which have begun to consistently appear in many of these guidelines statements as well. However, providing these integrative preventive measures in any coordinated fashion has been limited. It's difficult to garner all the puzzle pieces necessary to provide all these preventive measures. As such, there have been few initiatives which have tried to determine the effect of providing integrative preventive measures. The LEAP program, which was built off of the PAVE program and provided at a Veterans Health Center in New Jersey, did try to put together patient education, screening initiatives, and provider evaluation at an attempt to reduce amputation rates. While this increased some of the educational scores among patients and providers, it was not found to have any effect on amputation rates. Similarly, the Healthy People 2020 initiative also provided online patient initiatives for education and national tools for education. We have seen a diminution in amputation rates over time, but this relationship is likely is not a causal and really just cross-sectional information that's been collected over time. It's hard to know exactly what's happened with the provision of preventive measures and whether or not it's had an effect on amputation rates. Initiatives like this are especially important, as was mentioned earlier, for certain high-risk groups. This is data from the Dartmouth Atlas of Healthcare, which shows that stark differences that arise for African-American versus non-African-American patients in the United States. Each of these dots represents one of the 306 hospital referral regions in the United States. On the y-axis is the amputation rate per 1,000 Medicare beneficiaries in those regions with diabetes and peripheral arterial disease, and on the x-axis is race divided into two groups, Black and non-Black. As you can see, there is extensive variation in the rate of amputation among Black Medicare patients across the United States, where there is much less variation among non-Black patients in the United States, and even the highest-risk regions are much lower than a majority of those regions caring for African-American patients. While this is an important initiative, it has an utmost important for some of the highest-risk regions across the United States. In summary, in terms of which things matter most in terms of preventive measures, most guidelines, society statements, and white papers can be called down and distilled to hemoglobin A1C testing, some manner of vascular testing, and foot exams as universally recommended, low-cost, and seemingly achievable peripheral arterial disease care. Now that we've summarized the recommendations for annual integrated preventive measures for patients with peripheral arterial disease, we now will test and outline the adherence to these measures, as well as invasive treatments for patients with severe peripheral arterial disease. In other words, we'll do some experiments to see how well these treatments are provided and what their effect is on vascular care. We did this work as part of American Heart Association funded strategically focused research network. This is a collaborative effort between the Dartmouth-Hitchcock Medical Center and the Brigham Women's Hospital in Boston led by Dr. Mark Binocco. In this effort, we took those same three universally recommended preventive measures, hemoglobin A1C testing, vascular studies, and diabetic foot exams. We looked to see how the utilization of these preventive treatments varied across different regions and different racial groups across the United States, both by time and by racial group and by region. Our objective in this study was to evaluate the timing to receipt of these preventive treatment strategies among patients with concomitant diabetes and PAD. We studied this within a large national cohort of Medicare patients with diabetes and PAD. We wanted to examine the effect of race and region on adherence and timing of these preventive treatments among that same cohort of patients and we did it in the following manner. We started with all fee-for-service Medicare patients between 2003 and 2016. Patients were required to have diagnosis codes for both peripheral arterial disease and diabetes. Ultimately, our cohort was formed from nearly 10 million patients who met these criteria. Once we had identified these patients Medicare claims, we followed them forward from their time of their diagnosis to assess some key exposures and outcomes. The key exposures were that we wanted to record the timing of the administration of hemoglobin A1C test, a diabetic foot exam, and a vascular imaging study using the CPT codes that are put into place for those billing events. We then examine the results of these exposures by race and region of the United States. These are the demographics of our cohort. Most patients were between the ages of 65 and 75, but there was a small proportion of patients who were under age 65. Those are patients typically on dialysis and Medicare claims, and also a small proportion of patients, approximately 15 percent, who were over the age of 85. Approximately half of patients were female sex, as expected. Finally, 77 percent of patients were white, 14 percent were black, 6 percent were Hispanic. Even though it was a relatively small proportion of patients who are Native American or Alaska Native, this still represented nearly 55,000 patients with both diabetes and peripheral arterial disease. We had large sample sizes, even in some of our smallest cohort, to study some of these differences. The comorbidities we witnessed in this cohort were typical for populations of patients with peripheral arterial disease. Approximately 20 percent had COPD, approximately 17 percent had congestive heart failure, approximately 12 percent had cerebrovascular disease, and approximately 10 percent had a prior myocardial infarction. We looked at the period prevalence of diabetes and peripheral arterial disease by state during the same time period, and I've shown this on the left-hand side of the figure. Those areas in the lighter colors had a lower prevalence of diabetes and peripheral arterial disease, typically less than 14 per 1,000 Medicare beneficiaries. Whereas those regions in the darker colors typically had rates of diabetes and peripheral arterial disease which range from 26-36. We can see places like Billings, Montana had particularly low rates with 11 concomitant diagnoses of diabetes and peripheral arterial disease per 1,000 patients. Whereas other places such as New Jersey had much higher rates, 37 diagnoses per 1,000 Medicare patients. What about adherence to preventive treatments within six months of your diagnosis and within those states? Again, here in a similar map, we show the rate of receiving each of those three key measures within six months of when the patient was diagnosed with both peripheral arterial disease and diabetes. Again, the lighter colors indicate approximately 2-5 percent adherence to preventive treatments, whereas the darker colors indicate almost 10 percent adherence to all three preventive treatments. You'll notice that all the ranges are low with only a few regions even approaching double-digit adherence to all three measures within the first six months after diagnosis with diabetes and peripheral arterial disease. Within those same regions as we discussed previously, in Montana, overall three percent adherence to all three preventive measures within the first six months of diagnosis after peripheral arterial disease and diabetes. Whereas in New Jersey, it was slightly better, 6.7 percent adherence to all three measures within the first six months. But again, all of these rates are relatively low. When we look at these two maps side-by-side, we do see some correlation that those regions that had the highest burden of diabetes and peripheral arterial disease also tended to be those regions that had the highest rates of adherence to those three integrated treatment strategies. Although again, many of those appeared overall low, especially within the first six months of treatment. We then look more carefully at each of the individual preventive treatment strategies to see what happened over time. On the figure shown here on the x-axis, we see the time to the first preventive treatment measure, and it's laid out in 2.5 year increments, and our x-axis ranges from 0-12 years overall. On the y-axis is the incidence curve, which shows the probability of actually receiving that one integrated measure. When we look to see how long it took an individual patient to receive a hemoglobin A1C test after being diagnosed with peripheral arterial disease and diabetes, we see it's a little bit like doing your homework. Initially, a lot of people seem to get their testing, and then we reach the flat of a curve. Most testing events happened for diabetes A1C testing. Most testing events happened within the first year and happened fairly rapidly up to a rate of approximately 80 percent of patients receiving a hemoglobin A1C test. But once it reached 80 percent, the remaining patients often were not tested. Second, patients receiving vascular imaging had a similar shape of the curve, although somewhat muted in terms of its uptick. At approximately 2.5 years, approximately 60 percent of the population had received some vascular examination, which is slightly lower than the hemoglobin A1C testing dissemination at that same time period. Again, after approximately five years, nearly 80 percent of patients had received some vascular testing, but it took a significant amount of time to reach that endpoint. Finally, diabetic foot exams were less reliably coded or less reliably administered during this same time period. Within the first two years, approximately 40 percent of patients had a billing code evident for a diabetic foot exam. But this only reached approximately 60 percent by the end of the time in the cohort, suggesting that perhaps diabetic foot exams were not reliably coded or were not reliably administered to patients with diabetes and prefrontal disease. Probably the former is what our supposition is based on our review of some of the billing codes themselves. When we look at all three measures in their entirety, we see that it's likely driven by the low rate of coding for diabetic foot exams, that at approximately two and a half years, only approximately 30 percent of the cohort had received all three treatments. Again, likely limited by poor coding of diabetic foot exams overall. Adherence to each preventive treatment measure also varied by race. Here I'll show you a series of several bar graphs to illustrate these findings. When we look at hemoglobin A1C testing within the first six months, approximately 51 percent of the cohort had received this. Rates were similar roughly between white patients shown in blue and black patients shown in orange, with a slight decline for African-American patients relative to non-African-American patients. Vascular imaging studies were similar, although rates were slightly higher in black patients. But again, only approximately 40 percent overall at six months within their entry into the cohort. Finally, those patients receiving a diabetic foot exam, again, we found this coding to likely be unreliable, but there were not stark differences in the provision of preventive measures in this particular outcome either by African-American or non-African-American race. What about interventions? Well, we've seen before that variation exists in the intensity of vascular care provided to patients with CLI. This is work that we published in Circulation Cardiovascular Quality and Outcomes back in 2012. When we look at the intensity of vascular care or how commonly patients receive an invasive procedure in the year prior to amputation, we see it varied almost twofold across the country, from 33 percent of patients receiving some invasive treatment in the year prior to amputation to nearly 58 percent in some of the more invasive treatments, and these differences were highly significant. Naturally, of course, we wanted to know if that intensity of vascular care was related to their overall risk of amputation. We simply transpose this map where the regional intensity of vascular care during that time period onto the regional amputation rate as shown in the map on the right, to see if those regions that had the most invasive vascular care also had the lowest rates of amputation. When we plotted this in a scatterplot with the intensity of vascular care on the x-axis and the regional rate of major amputation on the y-axis, and each individual dot here represented by a hospital referral region, we see that the relationship was inverse and direct. The places that had fairly invasive vascular care like Santa Cruz, California had relatively low amputation rates, whereas other places that did not have intense vascular care, like Florence, South Carolina, seemed to have high amputation rates. This difference had an R-value of 0.9 and was highly significant. It seemed that more invasive vascular care was associated with fewer amputations. Even though this may have been provided in the absence of preventive measures. Now that we've outlined what happens with adherence, as well as invasive treatments for severe peripheral arterial disease, and seeing that those three preventive measures are consistently under-delivered within the exception of diabetic foot exams, which we suspect are not reliably coded, but the receipt of hemoglobin A1C testing and vascular testing appear to take a long time after the diagnosis of peripheral arterial disease and diabetes has been made. Similarly, invasive measures are inconsistently provided for those patients with severe PAD. We want to try to offer some potential pathways forward that might help to improve some of these limitations. In determining our pathways forward, we divided them into two groups, either patient-level pathways or population-level pathways. I'll speak briefly about some of our work that's looking at the patient-level pathways first. To try to better understand why these seemingly low cost and seemingly widely available treatments may not be taken up by patients or their providers, we've undertaken a qualitative assessment of why patients do not receive preventive measures and simply have started a project around the country where we're talking to patients in a variety of different regions of the United States and a variety of different care settings to try to better understand the patient and provider challenges to the receipt of these seemingly simple measures. Using a Wagner chronic care model, we're looking to understand the community, health system, patient and patient care team interactions which can try to affect these outcomes with the hopes that better understanding these relationships will result in improved health outcomes in diabetes and peripheral arterial disease. This qualitative work has allowed us to enroll 25 patients so far who have all participated in qualitative interviews to help both the patients and the providers better understand why they may or may not have received some of these preventive measures. By better understanding these gaps, we hope to make informed recommendations at a population scale. At the population level, when we do have those recommendations to make, we now know where we should target them. This figure here shows the amputations per 1,000 Medicare patients with diabetes and peripheral arterial disease over time. You can see by different racial groups that the racial groups which need these recommendations the most are clearly Native American patients, who even though they have had a decline in amputation rates, have an amputation risk that is nearly threefold higher than white patients. Even though these rates have declined over time for both groups, they remain significantly higher for Native American patients, African American patients, and Hispanic patients. When those recommendations as to how to better deliver preventive measures arrive, we know we need to focus them on the most at-risk populations. Even though progress has been made in all groups, the most progress remains to be made for Native American, African American, and Hispanic patients as shown here. We look forward to this focus on delivering preventive measures where they are needed most. In summary, it's been my pleasure to summarize some recommendations for annual preventive measures for patients with peripheral arterial disease and diabetes. These are simple, low-cost, and effective treatments. Things like vascular testing, hemoglobin A1C testing, and a diabetic foot exam. We've also outlined the adherence to these measures, as well as the treatments for invasive PAD, and we've seen that adherence is low, especially in high-risk populations, and there's definitely room for improvement. Finally, we've offered some potential pathways forward both at the patient level and the population level in efforts to improve these limitations. When these pathways arrive, we hope they'll focus on groups at the highest risk. I'd like to thank the American College of Cardiology for inviting me to be part of this work, and I'd like to thank the American Heart Association for funding to support this effort. Thank you very much.

peripheral arterial disease

undertreatment

amputation rates

preventive measures

high-risk populations

delivery of preventive measures