 Okay good morning everyone. I'm really excited to see you all here in person and online. If you have any questions for those in the online audience please go ahead and post them in the zoom chat and we'll get started. So my name is Sunday Shadeve and I want to welcome you to this workshop. I am the director at Institute for Future Health and cardiovascular projects. I'm a cardiologist at the Southern Arizona VA healthcare system. These are my disclosures. Why is this workshop important? I think just looking at the scale of Veterans Health Care Administration it's the largest integrated health care system in the United States and there's over 1,000 health care facilities, 172 medical centers, 1,000 plus outpatient sites of care, serving you know roughly 9 million Veterans per year. So a couple sort of general comments and perspectives I wanted to share with you and then we'll get into our talks. So you're going to hear about working with VA, collaboration opportunities and entry points. You're going to hear about VA data resources, Vinci. You're going to learn about the arches and MD clone data platform. You're going to hear directly from investigators who are working with VA data from University of Arizona, ASU and VA. And you're going to hear from our VA's about how to partner with them. You're going to hear more about other types of VA data resources and we're also going to hear from our universities about their interests and needs and how this could potentially go forward. You know from a clinician perspective, clinicians I think are naturally curious like all of us and you know by being able to access data themselves they can rapidly, we can rapidly answer questions and generate hypotheses. Last night as I was preparing this talk I thought well you know let me just try something myself and see if this is really true. And so I went into the system and I said how many VA patients nationally have been treated for hereditary ATTR polyneuropathy and I found the answer in less than two minutes. Now that wasn't a full study obviously but that was very powerful. You know that data might be available to let's say a national pharmacy program manager or local pharmacy manager but that data is typically not available to clinicians themselves. So I think this is really an important self-service model which is what we're seeing right across society and it's I think the difference between an ATM and a bank teller. And so from a clinician perspective you know this can support quality improvement and program building right. So if you want to design a program for patients with high cholesterol or aortic valve stenosis you can start to generate some numbers and begin thinking about this immediately. And also some of your ideas may be percolating and you may not you know get an answer with a single ticket into informatics. You kind of want to generate your ideas iteratively. It's not a simple request. From a researcher's perspective I think some of the advantages of the platforms you're going to hear about today is that they do not require studying extensive data dictionaries. That's not to say that everything is you know immediately interpretable to someone who's not familiar with VA data but one can get started relatively quickly. And you know obviously understanding health care data is still important. But there's no database programming skills that are immediately required. There is I think a strong user support system with some of the newer platforms that you're going to hear about today. The functionalities are constantly improving and I personally can't keep pace with them but you know things are constantly in flux. So something I see that's not possible might be possible later. The approval process to access data is pretty fast. Once you have in my case as a researcher IRB approval the data access is almost instantaneous within a matter of days. However you know not all types of data are necessarily accessible by some of the platforms you're going to hear about today. So you know things like Medicare data, progress notes. They may not be currently available. You're going to hear more about structured data today. And currently in the VA arches environment there's limited software for statistical analysis. I believe R and other packages may be available but things like SAS it's sort of a bring your own license approach and I think that can be a challenge. So in terms of my interest as a researcher in cardiology a lot of the areas that our group is interested in can be answered with the CDW access corporate data warehouse access. So issues around clinical epidemiology, outcomes research, disease prediction, quality of care, quality improvement interventions or at least planning for them compared to effectiveness costs of care. So I think this is a very intriguing platform that the VA is providing. I want to just briefly mention VA data challenges. This is not something that I'm particularly knowledgeable about but this came to my attention the VA in collaboration with FDA and UK sponsored cardiac health and AI model prediction challenge VCHAMPs. And I pass this on to one of our cardiology fellows at University of Arizona who himself is a data scientist Jeffrey Tran and we connected him with a data scientist in VA and another data site budding data scientist and this small team of three people competed for this international challenge and as I was looking at the website approximately 16 countries are represented 123 people were part of those teams and submitted to that challenge and most of the winners seem to be with an industry and our team actually did not win but the point is that they competed and they learned a lot of important lessons and I think that this could be very powerful for universities and data scientists if they're interested in getting involved with VA data. So from an education perspective this is something that I think we need to really creatively think about and and this is something that I think needs to be discussed more. So I think that students you know so medical students graduate students residents fellows undergraduates they have shorter time horizons they have limited data analysis skills in some cases they generally are you know a team of one or two they have less experience with IRBs and they have less experience with VA IRBs naturally. So with VA arches I believe they can conduct impactful research and quality improvement if they have proper IRB support if they have some statistical support depending on their level if they have credentialing support with with VA if they have some training on the data platform. So I don't think that they can do this themselves but with a little bit of support they can be off to the races and from a perspective of industry I think that there's a lot of real-world evidence here and you know if someone is interested in the entire veteran experience in many cases they will be interested in Medicare data and so that that's an important linkage that might not be accessible through some of the platforms we're talking about today. However that's not always the case with every research question and you know veterans are primarily males and so you know the research has that perspective but you know many diseases are prevalent in males and females you know take the case of prostate cancer heart failure ischemic heart disease hypertension etc. And I think I haven't really seen a lot of research coming out of VA that that that partners with industry for this type of work. So a couple other interesting tidbits as we were doing some of our large-scale data work to answer heart failure questions you know there were questions around the validity of some of the underlying data what is the meaning of this ICD code does it really relate to the disease of interest things like that. So we were looking at ViRec website and and we found VOOGLE which is Google-like resource in which you type in the patient's name or data identifier and you type in the concept of interest. So in this case the example is diabetes and so it pulls up all of the mentions of diabetes within that patient's record and so one can then systematically go through this longitudinal patient record and find out all mentions of that particular disease and then one can actually go into the actual the note and so one can sort of validate these these concepts that we're trying to develop and this was something kind of unknown to me I'd never really heard about this before and so my point is that there's a lot of tools out there you know if people know you know if people are trying to answer specific types of questions. Similarly we're looking to do a national chart review and the Capri application which was actually created around compensation and pension records it actually supports now research functionalities and so we're hoping to do a national chart review using Capri and you know full disclosure it's taken us roughly four months to get approval to use this platform so it's not been a rapid process but we were approved. So and I think I just give you an example of where you can do large scale data analysis but you can also do really individual level you know patient record extraction so it's it's something that you wouldn't necessarily get from Medicare data set right you cannot directly link to individual patient records you only have you only have codes and claims. Now in terms of barriers and opportunities I think that you know if we want to scale from individual researchers individual students there will be some sort of investment needed to centralize some of the resources to accomplish these things. Regulatory approvals are challenging you know they are possible but they really require people with experience VA experience understanding of VA regulations and then I think you know VA credentials and training must be kept active you know I have to confess I brought on a researcher once from a university and I don't think that we ever did a project together and I'm pretty sure their credentials lapsed and that's really you know not a good use of their time or the VA's time and so I you know I think I learned a lesson there but you know in terms of opportunities I think this is a goldmine of data and there's really no cost to access to data it's a public good you know but we need more investigators you know we could there's enough here to support a hundred investigators or a thousand in Phoenix and a similar number in Tucson and other universities in Arizona and other entities in Arizona. This fills a genuine need for researchers and educators in my opinion and there's a massive scale I mean I believe we have access to 25 million longitudinal records with the complete EHR that's very powerful and it can be utilized for non VA research so NIH research industry research there's really no limit as to how you can use the data if you have proper approvals and so