Next Steps and Action Items

Next Steps and Action Items
    Watch the video

    click to begin

    Youtube

    ELIZABETH: I'm not sure how we ended up ahead of schedule. It seemed like we were in a full-on panic in our working groups there. And this slide reflected that panic, that I just wanted to say to everybody: you still have an opportunity to provide feedback. This request for information entitled "How Do We Better Facilitate Cancer Systems Epidemiology Research?" is out and we're happy to get anybody's input there. I just wanted to really start by saying thank you to everybody. I probably am a little bit biased, but I think the meeting went very well, and I think that it was really fantastic to have you all here and everybody was extremely thoughtful and really engaged, so we really appreciate that very much. I just wanted to especially thank our planning committee who helped us really nail down the agenda. And we tried specifically to get different perspectives on that planning committee, and I really thank you all for participating. And then I also wanted to just acknowledge there were a lot of NCI staff who were also involved in planning this meeting, and I'm not sure we mentioned this yesterday, but in the spirit of trying to think more holistically about cancer and cancer-related outcomes, I just wanted to recognize that this group does reflect several different branches from within the epidemiology and genomics research program. So we're trying to make the gap smaller maybe or start with collaboration from within. And then also just to recognize the people outside of EGRP, who also many people contributed to the logistics today. I just wanted to go back to the workshop goal that Leah articulated yesterday, and that was really for us to facilitate interdisciplinary discussions about the application of systems modeling approaches to epidemiology or to population-based cancer research. Again, maybe I'm biased, but I do think that we've accomplished that goal and really had a fantastic discussion over the last day and a half. So thank you again. We had several different workshop outcomes, and I think that most of the breakout groups were designed to try to address these specific outcomes, so I think we've done a good job thinking about, for example: how do we identify those youth cases for systems modeling and population science? We talked a lot about the potential barriers and facilitators. Then again, another outcome or goal of this meeting was to form new collaborative interdisciplinary relationships between statisticians, various fields, and I think that we've really done a fantastic job of that. And I think you can see that just by the distribution of the different fields that are represented in this room. And I think we may have saved Marilyn from becoming a geography expert and Marta from having to be a chromatin expert, so I think if we can just recognize that it's good to come together and really try to make some connections outside of our normal field of expertise. I think that relates specifically to – and I'm sorry, I made these slides before the four groups reported back, so there's a little bit of redundancy here, but we absolutely did hear that call for need for transdisciplinary teams. And I think Bob highlighted this yesterday, this isn't necessarily new but it is something for us to be mindful of, so how do we encourage people to think outside their lane? And again, this is for the extramural community and those of us here at NIH. How do we put together the right collaborative teams and how do we facilitate doing that at the beginning, at the time that we're starting to think about the question, and not at the end when all the data is in hand and our study is complete, really. Another thing to think about is: how can we nurture or build this systems epidemiology community? And I think that's something that we'll be mindful of in trying to work towards. Bruce, this morning, talked about the need to challenge us to think about how we encourage these different types of trainees. How do we facilitate the hybrids, the connectors and the glue and mortar? I think that's really important. And something that has come up several times, and I think that we at NIH do have a mechanism to explore in thinking how to facilitate that marriage between the modelers and the content experts. We, at NCI, have other groups who have used these innovation labs or sandpit meetings where you bring together different communities in an effort to try to facilitate or foster collaboration. We talked about it a bit yesterday in that idea of bringing people together but having the time to try to nurture those relationships. And I think we can certainly explore that further. Another key thing was just enhancing the training pipeline. And so I just wanted to make sure that everybody was aware of these two funding opportunities. One is already out, and that was Liz Janetzi [phonetic] who supports this RSA and it's specifically focused on predoctoral training and advanced data analytics for behavioral and social sciences research. That's a T32 program and that RSA has already been announced, and NCI does participate in that RSA. The other funding opportunity, which you should keep your eye out for, is a Pathway to Independence Award for Outstanding Early Stage Postdoctoral Fellow. So we talked about really trying to enhance training at the earliest levels. So this may be a very appropriate mechanism and there will be specific focuses on population and behavioral sciences and then also on behavioral science. It will be quite relevant, I think. We talked about perhaps shifting our focus to a more problem-based initiative or training. And maybe with Liz, also who mentioned this is an example of a training program at Arizona state, which is focused specifically on solving – it's a really problem space-focused training. And then, perhaps, obviously, but really thinking about the need for effective communication and really recognizing that we often speak different languages and we need to really be aware of that and try to think about how we can facilitate communications across these really disperse groups of individuals. A couple of other key themes: these were also highlighted by Bruce in his presentation so we don't want to let the perfect be the enemy of the good. We talked about there not being one perfect model or one perfect approach. And that we really need to be thinking about this more iteratively, thinking about incorporating simulated and real data and really to think about it all as an iterative process and that there isn't any simple recipe. The mind the gap picture there was just to remind to mention – maybe it was Marilyn who made this point, just that we're not necessarily gonna solve this overnight and we can't possibly think we're going to fill every gap that exists. But if we could just takes steps towards just making the gaps smaller, that's at least progress. We started yesterday's discussion with this working definition of cancer systems epidemiology, and I know, I think it was Jonas [phonetic] who recognized that we don't actually use the word "data", but I just wanted to put this up and see if anybody else had any thoughts about things that might be missing. I don't think we want to get into any wordsmithing at this point, but just conceptually, is there anything that's missing? We said that it's an approach to study cancer risk and outcomes that incorporates high dimensional measurements from multiple domains, so for example, environmental, sociodemographic or clinical. It considers the interrelationships between those risk factors and changes over time. And Leah showed the results from our survey and there was not necessarily that focus about changing over time, but I think we, in our discussions, came to a consensus that that needed to be there. MALE: People had a discussion in our group about a broader community that we're engaged in and health impacts that we want to make in doing this research in the first place, so there's nothing there about why are we doing this and how do we get this back into the community for action? ELIZABETH: Yeah, that's great. I don't know if you were here for our very first presentation yesterday, but Kathy Holtz's hour sort of started us off by talking about the definition of epidemiology, which certainly and explicitly includes that point, so I think that's a good one. Thanks. Anybody else have any thoughts about this? We're always welcome to have your feedback because I think we want to come to a consensus about this definition. That's another place in the RFI, right. Do we have a question about that? Then just email Linda. We intend to write a paper about this meeting, and I think we'd like to include that definition to try to get it out there and get feedback. I think we'd like to discuss some of the challenges or common misconceptions we've talked about in the last two days and really what we might do to address these. And then also talking about the feedback we just heard from all of your discussions this morning about facilitators, barriers, opportunities for problems-based studies? What needs do we have in terms of approaches and methods. And what data availability needs do we have? And I just thought if you could ever imagine giving a talk in which you can have Anakin Skywalker in your slides, that's pretty good, right? Just the last one, just a couple final thoughts, something that came up yesterday which seemed fairly straightforward and probably really isn't, but: thinking about how we develop an online catalog of relevant resources. For example, Rocky mentioned the CISNET smoking simulator, and you could imagine that's something that lots of people would be interested in using, and also the idea touches, again, both on thinking about access to resources, access to code and access to data and how do we enhance that? Also, related is this idea of an online catalogue of available data sets. Yesterday we heard a lot about the RTI synthetic population data and it was with a lot of enthusiasm about how you might overlay additional data layers on top of that. So I think if we can help to catalogue some of those resources, that would be helpful. We also talked about somehow creating a venue for model buy-in. That was the idea of getting feedback on your models before you've gone through the entire process and perhaps that relates to sort of the development of the systems epidemiology community and how you could have a place to do that. And then just, finally, I was going to mention another point that's come up quite a bit, which is the idea, maybe the challenges and limitations in terms of NIH peer review or also publication peer review. And for NIH in particular, the idea that systems epidemiology, ideally the aims aren't linear in our traditional sense, and perhaps we have a need for a dedicated review panel in this space. And this figure just shows a resource that Rocky mentioned yesterday, which is a catalogue of genetic simulation resources. And Jill actually shared another catalogue resource, that I wasn't aware of, this morning, so I think we have really an opportunity to try to share some information about what resources already exist. I guess with that I'll just say thank you and thank you in particular to Leah who really did the lion's share of work planning for this and just remind everybody that we're happy to get feedback through the RFI. END OF FILE