 Now we're going to go into a 20 minute Q&A. I'd like to invite Posca, Janet, Bonnie, Travis, and Michael. I want to start off with the first question. So let's say I am, let's say Bonnie's perspective. I have 30 master students. This is directed to Posca. I have 30 master students in public health or a similar discipline who are interested in research and they have to graduate in two years. Do we have the capacity to bring them up to speed on that type of request? Well, if they're dedicated, we will help you. So yes and no. Meaning, yes, we definitely want to help. OK, these master students need to be associated with a VA researcher. I think that's perhaps the easiest way to actually help to be included in already existing research projects. That's perhaps the fastest way to do this. And the existing research projects that are ongoing can always have additional strategic aim, right? An amendment to the IRB can be done to just add that. So that will be perhaps the fastest way. While these students get the VA appointments that are needed. For example, without compensation or walk appointment, it's not paid, but it gives the students access to VA data. Does that make sense? Hopefully, Bonnie. Yes, we love we love to we love to collaborate. In fact, I know Dr. Migrina has left. But I guess you're there in there. But we definitely will connect you with Ray. Yes, he's a great researcher. So yes. Travis, go ahead. Unless Bonnie, you were about to say something. I was, but go ahead. Go for it. Then I'll ask my question. Yes, so I understand that. And I at some level, I really agree with you because it would be better for students anyway to be affiliated with research where that especially if we're going to talk about masters. And I believe that Sandish is using masters because of the timing of like the two years of that that these guys have and they don't have the four years to get the, you know, set their own research questions up, which PhDs and and postdocs would. So I agree that probably for master students, they would necessarily need to be associated with a research project anyway. So because of the timing. So that makes great sense. I also would like to again, stress that the idea of the challenges are also a way that we can get our students interacting not with the full data, like a VA data health record, but more of the subset that I understand last year's challenge. So more of those and having more eyes on those opportunities probably in about three months prior to them coming out is something that would be really helpful for for us. And those, so those are the two points that I wanted to make about that question, but I think it was a great one, Sandish. Thank you. And I think just to emphasize, it is good to then, you know, for sort of our different programs to collaborate first to make sure that we are prepared, right? There may be some, what we call umbrella projects, okay? That can be ready to actually incorporate incoming students into the projects. So we're working on that. So yeah, we're more than happy to discuss this more. Yes, Travis. Maybe that's a good place for me to throw in my question. My impression is that association with VA researcher or active research programs means that there needs to understandably be some sort of oversight or engagement from the VA side of things. And I think it's important to get a sense of what kind of capacity you collectively have to be able to provide that, right? If we said, here are a thousand students and let's get projects on them. The answer would be no way, but do you really have the capacity to take on meaningful projects for dozens of students? Thousands of students right now at the same time, maybe not. But 30, we can talk about it. So let me add that if the university has a resource to help, you know, navigate the paperwork, the approvals, the PI leadership, there's no limitation on that. And so I was sort of, you know, when I said we want to link with the VA investigator, that was assuming like a one trainee to one faculty model or one, a few trainees to one faculty linking with VA. But I don't think, you know, creatively thinking we could scale it, but there would have to be resources brought in to manage all of those people. That is correct, yes. So definitely things could happen if the university can help manage those trainees. Yes, that is also, yes. And I don't think there's a restriction that there has to be an existing VA PI, like you could be the VA PI. If you've got VA credentials. Exactly. So what I'm saying is that's like, you know, you don't have to find someone in the VA. You are, you can become the VA. Yes, a student cannot be a PI. That is VA regulations. So yes, there has to be a VA PI. But a VA PI doesn't have to be paid by the VA. At the very least, a VA PI should have what is called a without compensation or walk appointment. I hope that makes sense. Michael, did you want to add something? You know, all my questions are answered now. Thank you, Narkosh. So I had emailed Jeff Tran, who had participated with a few other colleagues in the challenge. And I did not have time to incorporate some of his thoughts. But since the challenges were mentioned, he said, you know, be on the lookout for opportunities by registering at precision.va.gov challenges. The challenges take a huge amount of time, considering everyone has full-time day jobs and every day, every week counts. Not only will the team have to spend time learning the data, cleaning the data, training models, but actually clarify the nature of the challenge. So sometimes asking for clarification on what is being asked. Because sometimes the ask is not always as clear as we would like. And have pre-arranged teams, so kind of alluding to what was mentioned, like the teams have to be essentially ready to go with including a technical analyst and a clinician. And there's a big learning curve in terms of just how to communicate to each other. And so, you know, in this case, they assembled a team on the fly, which already put them behind. And being clear about how much time there was to devote to the challenge, he said they are, quote, super time-consuming on both sides, clinicians and data scientists. And, you know, the teams need to learn each other's domains. You know, obviously these are all just good practices for research teams in general. And then he said compute can be an issue if not provided on the side of the challenge organizers. And so I think there was even a question about computing power in the challenge that they participated in. So if Jeff's here, you know, feel free to unmute yourself, but otherwise I think that kind of captures his recommendations. Yeah, and I can see that. I've never participated yet with the VA challenge, but I'm happy to. I have done the Kegel challenges before and like having the teams available. That's kind of where I believe that, I can't talk for Michael, but I think we're kind of very similarly set up. I mean, we do have students that could rapidly be deployed to this kind of experience and it would be awesome for them to see it. I think it's like, it is a lot harder if you're an active clinician that your side hustle is playing around with data. Like Jeff, I mean, he's fantastic, but as he said, it's not his, it's not something he's devoted to where students really would be. So I liked also hearing some of the perhaps challenges with making sure that we understand what the ask is from the challenge. That is also something that we're familiar with. And my understanding was that the data set, and this is kind of a question, it's gonna sound like a comment, but it's a question. My understanding was because the data set was kind of blinded and almost, it was more like a synthetic data set that we could do the computing anywhere, right? So if we have large scale computing at our fingertips due to the kind of work we do all the time, then for some of us, that wouldn't be such a challenge. I think that's correct. They were on such a time crunch that even 10 days to start up something at his normal hosting site would have totally eliminated their competitiveness. So I mean, they were under extreme time pressure. So clearly things could definitely be better the second time around. And again, going back to our talk and also what I heard like, sorry, Michael, I'm gonna pick on you and your upper right hand corner. I mean, also this is where the collaborations with data science type folks who do this on a regular basis. If anyone else were to want to participate in these challenges, we likely could facilitate expedited getting those kinds of infrastructure ready for teams in the future. I wasn't part of the challenge, but I believe that it was first a synthetic data set and then those models were tested against real data. So there was a sort of further phase to that. Sorry, Michael, I think I interrupted you. No, it's all right. Just to kind of continue on Bonnie's comment that that's a lot of our thinking about trying to at least have some visible higher or more aggregate organizations that deal with health and data science at ASU. Having done a couple of challenges in the past or kind of coordinated a little bit, some teams that did challenges, just echoing some of the things that have been said already, that planning a time to plan or an opportunity to do it again with a little bit of a heads up is really valuable. And I think sometimes individual faculty labs encountering those opportunities right in time can be very difficult. Even the communications get out there on time. So, you know, part of our role is to try to be a central location where some of those opportunities can go and then we can get those to the right people because yeah, you can end up in a situation where there are people who are just straining to do work on the project, but across the street, you've got people who are ready to go and have no idea it's even happening. So I think one exciting thing about moving forward here is when you get, you know, you can envision a student over the course of their career participating in more than one challenge. At this point, I cannot give them the way that challenges have gone, right? But the idea that they could actually have some experience doing it more than once, I think it's very exciting. I agree, and one thing, a kind of piggybacking on this, I mean, I think there's two things that we're talking about here. One is possibly infrastructure support for other faculty or other people that we wouldn't necessarily be even participating in the challenge unless you wanted us to participate. But we could actually help people find the resources they need in the infrastructure they need to do their own project. And then there's the second us helping, having our own teams that we might like to participate in challenges. I think both of those are really interesting opportunities. And that's the one where I wanna talk and add just one more thing on. When I say workforce development, I don't just mean workforce development for all of our students who would love to get as much exposure as they can to data science and large-scale computing and predictive modeling, but also the workforce development for people who could consider careers in the VA Health umbrella. And I think that that is something that would be of great interest to some of us. So if there's some way that we could think of either providing training to existing employees and some of the things that both Michael and Travis and I talk about, that's one opportunity there. Another is for us to actually have a pipeline where we would be training the next generation of VA data scientists. And at least from my perspective, those are all kind of separate but overlapping then diagrams. Just want to make a comment on that. That's a great point, Bonnie. In fact, I think a few years ago, the VA central office of research and development actually wrote out an initiative similar to that basically trying to get university affiliates to partner with VA's to actually train clinicians or whoever in data science. I don't think I have seen more in the last two years of how that is going, but I can tell you that the, again, the central office is interested in developing VA workforce as you mentioned. It doesn't have to be just solely VA, right? I mean, again, in partnership with affiliates. So we'll keep our eyes open too and we'll definitely look into that more. So sometimes we get requests for even more short-term trainees. So medical students looking for a summer research program. That seems to be very challenging but the term umbrella research was mentioned. So I could imagine a situation where you have a cadre of 30 or 40 students just looking for a three-month experience. They join some sort of umbrella project and they're able to get access and start making insights. Yes, we are really looking into developing this kind of umbrella projects. Again, with a short term of, you know, that these medical students have three months, it's really not much, but they can do necessarily, but there can be a lot of things that they can do with data. As long as, again, they have the VA appointment. If they are health professional trainees at the VA, then they already have that appointment, right? So. So one question I have for Bonnie and Travis is, what, how do sort of university investigators view accessing VA data in terms of what is, is there anything holding them back or are there other sort of competing opportunities? Yeah, what is your perspective on that? Well, I don't even think it's in the purview because I will say that I think there's, at least the people I work with in terms on the clinical research side, not on our student side and not on the Michael, Travis and I side because we've been doing this kind of stuff and we've seen this kind of two year kind of lag time in just about everything we do. So it's not just the VA that has this sort of issue and all of us as another one of these kinds of takes a lot of time to get to know how to use the research hub effectively and actually use the data. So we're kind of used to that, but I will say from clinical research or perspective, I have, what I have heard most frequently is it's not even on their radar just because they don't even have the time or energy to even collectively think about how they would get a project started. Or, and I myself have learned more today than I knew walking in. So all I would say is, oh, yes, I know, poor you. I've heard that from other people, it's not just you. So I think that in my mind, maybe at the first step, it's more the things that we talked about in the first hour or two, spreading the word a little bit more widely to our fellowship programs in particular at the start of the fellowship program. And I'm talking right now about clinicians, Sandesh. I think that would be really valuable. A lot of us have, like I usually give talks in those very first months that the fellows start for like internal medicine and for oncology. And having somebody who spoke today, either the first or second sessions would be greatly helpful just to expose people to what's possible. And then also the maybe fast-tracked way of getting those people access. Those are the two, I feel that they're very easy barriers to overcome. The harder barriers are the ones that Michael and Travis and I kind of talked more about, which is how do we provide those same people once they've overcome the first fairly, maybe lower hurdles. The higher hurdle is perhaps the actual overwhelmingness of pulling together the statistics and data science that's necessary to actually execute the project. And that's I feel where some of us and some of the earlier speakers could all come together to provide that level of support. I think we need to maybe consider including the VA education in the conversation to see if they have any resources that can be brought to bear. We can talk with Dr. Raymond. Oh, please go ahead Travis. Sure. So I appreciate Bonnie's mention of barriers, right? Barriers to entry I think influence almost every kind of project adoption, right? So one of the obvious ones is you just have to be able to get into the data at all. And that means that a person can't say to themselves, oh, I think I'll tinker with this data for a week and figure it out, right? You have to go through a lot of effort just to get access to data. And then once you get access to it, it's in a walled garden of sorts and it's under a particular, perhaps peculiar kind of a computing environment. And that's another barrier to entry or to continue to access. And those are things that you as a person who is willing to do data analysis, you overcome but it's with time. There's another barrier to entry for any kind of a research direction, right? So one is from a clinician's perspective, right? Or someone who's interested in learning how to do data analysis. There's another perspective which is from people who are very adept at doing data analysis but don't know the VA data particularly well and may not be clear on what are kind of open questions or open problems that are effectively low hanging fruit given the data resources that are available but only if you know that those questions exist. And I think helping to knock that barrier down would be really helpful too, right? So this is like, I gave my pitch of we do a bunch of generalist data analysis and I don't really know what are gonna be the open questions that folks within the VA system say, boy, there's gotta be enough data here to answer that question. We just need someone who understands how to turn the crank. Knowing that makes it much more worthwhile to put in the energy to get past the technical hurdles of just like how do you get into the data system and work within the computing environment. But if you don't get past one of the two hurdles then nothing happens. So a clarification on what are the problems to be solved? Is that what you're saying? Yeah, I think from a VA, if the VA is seeking external aid in answering questions, solving problems, identifying ways to take advantage of the data, one of the things that would be really helpful would be to have some kind of a compendium of questions that we think could be answered but don't know how to answer. Problems that we think could be solved but don't have exactly the right tools to be able to solve that makes it a lot easier to kind of, to marry up people on both sides of the conversation. I hope that helped. I think there will be a lot of such questions. So yes, I think just as you said, matching the skills and questions and the people, that's really what we're here for. Yes. And just a comment about audience or who we're appealing to and making these kinds of transparencies. A lot of times, we're gonna be asking to work on a challenge. No one is being compensated to work on a challenge, right? And so it's gonna be a passion project for 90% of the people who are gonna be coming to the challenge today. It is the echo what Travis mentioned. The last time we did a challenge, we had a 50% drop off at least of students who couldn't get through the paperwork just to figure out how to get access to the employee, right? And I know we can't solve that problem today but what's gonna drive them to stick with it is knowing that there is a problem or a set of problems in this data once they're initiated into it that they might deeply care about. And I think being able to couple that, the fact that there's gonna be a learning curve for working in this environment but that getting into that environment is gonna pay a dividend. And that was a big lesson learned from us last time was really to be able to sell it to the students about what a dividend it is to be able to learn and do research in a secure enclave for doing health data. I think sometimes it's easy to believe that that's just a self-evident good but I think for students when they're volunteering to get into this and it's not necessarily part of their program or this is their first time working with clinical data with their data scientists, they have lots of calls on their attention. This is a really invaluable tool I think for bringing more people to the table. Emphasizing clinical relevance basically, right? Clinical relevance but then there's a number of other dividends that I could pay for data scientists, I believe. That's our job. So look, thank you. We're gonna say something. Well, and I... So go ahead, Bonnie. Oh, I wanted to just like, I wanted to, I don't wanna finish. I mean, we've been talking a lot about the students and the workforce development and maybe even workforce development from clinical investigators but when we're done with that I just had one other item but bringing it back to actually research opportunities like that we might have outside of the training aspects. So, but I'll wait until we're done with the training part. Vinod, I had a question because there's clearly huge opportunity here and everyone wants to collaborate but is there a way, this question for Michael and Bonnie Travis as well, is there a possibility that a person can be assigned to run this program of opportunity, right? And possibly have a next step where we can bring all of us together and sort of organize this collection of opportunities to actually take some concrete next steps because clearly I think that state exists right now, Michael and Bonnie. And so you're talking more about the, perhaps having a Travis, Bonnie and Michael sitting together thinking about all of the potential educational, more the educational opportunities, I guess what I would say, both the educational where the educators are the data scientists here who have varying expertise and then the educators are going to both be our own students but potentially also be more clinically relevant or clinically clinical or translational scientists who might need education in data science. Am I understanding that? Ask properly? I'm talking about all the streams of work and activity and opportunity that we've talked about and can we have somebody from the university side to organize this and bring all the parties together and then execute on it? Well, and that's where I was like trying to put, there are a few bite-sized chunks. I think they ask is right now pretty broad because going back to the research collaborations, I think a lot of us have our own research interests that we would like to collaborate with people who are already embedded in the VA system. And I think that's slightly different than the other collaboration that we've been talking about where we're bringing our experience and we're trying to figure out a way that we can help facilitate other collaborations that might not be around our own research interests but might be more around our expertise. And so from that perspective, I think that that would be an ask that perhaps the three of us could consider. And then we would need to work with somebody at the VA to see how we could integrate that to the current VA system that we heard about today where there already are these people that have these umbrella projects and there was already the, I guess I'll say the infrastructure is already set up where we would just be plugging into those and then by doing that, that would give us exemplars for new things that might come about. So that is a thing that I think that we can talk about further and would potentially be an action item for the three of us to figure out how we might be able to do this but then we would need somebody on the VA side to help us figure out how exactly we bring these kind of like multidisciplinary data science expertise to facilitate the research that is necessary to improve the health of VA patients basically. Because that is an important aspect. I just wanna make a quick comment. I think we'll definitely have to unpack this. There's a lot of ideas and great discussion. I think from my perspective, there are clearly some VA investigators who benefit from collaborations with university investigators, whether for data science or other types of expertise. But I think it's actually a dearth of VA investigators. And so I think there's different opportunities for a VA entry point. University investigators become VA investigators or the VA investigators accommodate university trainees or larger teams that wanna form, which I think both are equally valid. But I think we have about 30 seconds for any quick comments. Or so let me actually ask our speakers. Anything else you would like to say before we end? Just want to add that this other person- Well, just that I think the three of us will meet. And so that is already an action item. I think we're probably gonna do that anyway. That's great. I just want to say that the December project that I mentioned, it is still in development. But I think again, with input from maybe the three of you once you discuss this, we can actually further develop that. Right now, as Dr. Dev said, there are some VA investigators with existing projects. You know, maybe they can take one or two students within their projects, right? To add another strategic aim or whatever. But if you want to talk about, you know, something that is really a little bit bigger to that can accommodate more students at the same time than that is in development. Okay, but we're happy to collaborate. Any final comments, Travis, Michael, Bonnie, Janet? Thanks for the chat. Go ahead. Just thanks for letting us chat. Thank you. Yes, thank you. All right, thank you everyone. This concludes. We'll have lunch here and please follow up with our survey that we'll send you soon.