 Our next speaker is Dr. Simmer Singh, MD-PhD, internal medicine resident at University of Arizona College of Medicine, and a member of my research team. He'll be speaking on epidemiology of cardiac amyloidosis in veterans. Thank you, Dr. Dave. So take a little time to talk about some of the research we're doing on cardiac amyloidosis. And then I want to talk a little bit about my journey into data science and how it's been navigating that foray. So cardiac amyloidosis is actually a pretty rare disease. And getting accurate data on cardiac amyloidosis is somewhat hard. It's cardiomyopathy. And oftentimes, symptoms are actually quite not specific. And this makes diagnosis quite difficult. And the median time in diagnosis is actually 39 months. Oftentimes, patients wait years before a diagnosis. This can be nerve-wracking, not only for the patients, but very difficult for the health care system, which is caring for them. And it's unknown, actually, the prevalence and incidence of cardiac amyloidosis in the veteran population. So that's something we sought to address. So number one, we wanted to really identify what the burden of cardiac amyloidosis was in the veteran population and guide the development of screening programs that would help benefit those patients. And along the way, identify a cohort of patients that we can use for training AI models and also looking at other clinical outcomes. And so the way we approached this was to conduct a retrospective cohort study using the VA data and the MD Cloned Atoms Platform. And we surveyed all VA patients with an inpatient or outpatient encounter in 2012 and 2021. And then we used MD Cloned to actually create our case definition. And this is something that went through a rigorous iterative process where we would initially create a definition of cardiac amyloidosis. And then doing a chart survey actually go in and validate those findings. And through this process, we actually found that traditional case definitions relying on ICD codes were actually not very informative. In fact, we lost, I think, a lot of patients or misclassified a lot of patients. So this was a nice iterative process, which was kind of very quick, actually. And then we were able to take that data and provide some statistical analyses and take a look at, in fact, geographic variation, which we'll take a look at right now. So the first thing I want to point out here is that we've essentially assembled the largest cardiac amyloidosis cohort in the world. And not only that, we think it's the most accurate cohort. And so right off the bat, we have data from 2012 and 2021, which shows that we have an increase in incidence. And this is likely to be an increase in detection. So now that more people are looking for cardiac amyloidosis, we expect that we should pick up more cases. Interestingly, cardiac amyloidosis is actually a disease that has a higher prevalence in people of African descent. And when we look at the geographic variation, the data doesn't really show that we're detecting cardiac amyloidosis in places where we expect a higher prevalence. And so that was one of the insights that we really found that we have areas in this country where we need to do a little bit better job of identifying these patients and get them the care that they deserve. So in total, we identified 5,200 patients, which I think would be the largest cohort that we can use for training and further analyses. And so we were able to take a look at the geographic variation here on a state level. And we see that the incidence is higher in the Northeast. And this is possibly due to the fact that there's more cardiologists up there. But also, they just may be better at detecting cardiac amyloidosis. Whereas in the South, where you'd expect, given the higher proportion of patients with African descent, you would expect more to pick up more cases. Prevalence kind of echoes the same findings. And so this is just kind of talking about what we discussed here. So let's skip that. So what we can take from this is that there is a significant regional variation in the incidence and prevalence of cardiac amyloidosis. And this may be a difference in the detection rate, the ability of these separate VA health systems to find cardiac amyloidosis cases. And so this is something that we'd like to use to develop better screening programs for at-risk populations. So with that being said, I'd like to switch over to kind of talking about the user experience, how to kind of get involved in the sort of research. So for me, this is kind of a new type of research. So I've always been a traditional bench scientist, a basic scientist. And then specifically, I looked at epigenomics, metabolomics, and kind of integrating those type of data sets. Then I got to residency and found I didn't have too much time to take part in actual bench work. So I look for different ways of engaging in research. And what I did have a lot of time to do was interact with CPRS in the course of my clinical duties. So I have my continuity clinic at the VA, where it's been an honor to take care of veterans. And my attending on my first day told me, she goes back 10 years of imaging and data. So the first thing we do is we look at our chart. We open it up so it has 2,000, the last 2,000 notes. I mean, that's an incredible amount of data to kind of sift through. So I realized very early on the potential of this sort of data. It's a goal in mind, as many people have said before. And I wanted to start exploring some research questions. So I tried doing that myself at first. And I found this very difficult, honestly. A lot of the challenges and questions we've seen brought up today. I also encountered those. And I'm someone with, well, I consider myself literate. I have three postgraduate degrees. And I still had a tremendous time trying to navigate this. And not only that, as someone who traditionally is not at the VA every day, I found it difficult to kind of do hypothesis testing where I have to email my query using a VA email. And then I have to try to log on somewhere. I had a card reader. And it was very difficult. And then not being there during business hours all the time. That was also another kind of setback there. So it really wasn't until I was introduced to Dr. Dave and Dr. Agarwal. And I was able to join their data science team and kind of collaborate with them. They kind of championed my, I guess, entrance and credentialing process. That really the breakthroughs I think started occur. So we'll talk a little bit about what my advice for future researchers is. But I really think that finding someone who can really get you involved, who has a knowledge of the process, because it is still a difficult process to navigate. And really when I was introduced to MD Clone, I was surprised at how quickly we could go from a question or an idea to having some preliminary data. And this is really the power I was taken back by, honestly. And what it really allowed us to do was to work collaboratively, generate data, look at it, and then keep refining that process very quickly. This was a process that may, I don't think would have been possible if we'd gone through the traditional route emailing a query, getting a feasibility study done. It really was almost instantaneous. And what it's allowed us to do was actually accurately phenotype these patients. For cardiac amyloidosis, there's a lot of noise in the diagnostic, in the charts and the diagnosis. So it's really allowed us to accurately phenotype those patients. And we've created the largest study cohort for cardiac amyloidosis, which I think is no small measure. I think it's something that I'm really excited to do a lot more with. And as someone who's enjoyed being a part of impactful research for quite a while, I feel like this is really a high value, a high impact research. We're working with patient data. We're working with veterans, which I think makes it all the more important. And it's actually not too difficult to learn, use and teach. And in the short time I've been able to work with this data team, we've been able to generate several abstracts, presentations, got two manuscripts in preparation and I hope many more. And really become part of a data science team and it's been the ability to draw on all their different expertise that has really made this an exceptional experience. And we really hope that it's gonna really take care of patients, really improve the way that we care for them. So advice for our future researchers. First of all, it's a great opportunity. I don't think that there's a more rich and more rich and accessible data set than this. And so I would really think of trying to find a PI that can really support you, that can get you involved and help get you credentialed, honestly. And that includes getting a VA laptop. Honestly, if you can get yourself a VA laptop, that's gonna make streamline the entire process. And start this credentialing process early, as you heard, it does take some time. And then get hands-on with the data, start practicing, start building queries in MD clone. I think this is really the way to get involved. And I think I just wanna thank the Southern Arizona VA Health System, Dr. Dave, Dr. Agarwal and our entire data science team thank you for listening.