 Great. All right. Good morning, everyone. As you recall, last week, Bruce Korf, his lecture on genomic medicine, gave you a very broad overview of how advances in genome research can be used to change the ways we start to think about the practice of medicine, and this week's lecture will build upon those themes, addressing how the genomic approaches we've discussed throughout this series really can be applied in a real-world context to improving public health. It's my great pleasure to introduce to you today Dr. Colleen McBride, who is a senior investigator and the chief of the Social and Behavioral Research Branch at NHGRI, and it was my pleasure to help recruit Colleen to NHGRI some eight or so years ago now. And in the time that she's joined the institute, she's established from scratch a brand-new research branch that's devoted to using a broad spectrum of approaches to investigate social and behavioral research factors that facilitate the translation of genomic discoveries with the goal of improving human health and well-being. Dr. McBride is a recognized leader in the field of behavioral epidemiology and an expert in the development and evaluation of behavior change interventions for use in public health settings, many of which have been focused on smoking, diet, and physical activity. And one of the hallmarks of Colleen's research program is that she's continually looking for new ways to capitalize on advances in genomic science in order to improve the efficacy of such behavior change interventions. I've had the privilege of collaborating with Colleen on a project called the Multiplex Initiative, which was the first study to inform the public debate about the effects of genetic susceptibility testing amongst healthy populations. It was an absolutely fantastic experience to get to work with her. You'll hear more about multiplex during today's lecture, and I really feel that that project serves as an excellent example of the kinds of translational research efforts that Eric Green talked about way back during the first week of this series, one that helps give us an evidence base for genomic medicine by looking at how we should deliver genomic information to patients, how individuals will in turn interpret those test results, and how they'll ultimately use this knowledge in making their own health care decisions. Of course, all of these things that are highlighted, you'll remember the NHGRI vision document, they're all highlighted in this document, is part of our vision for the field of genomics that are key ways that we can advance the science of medicine. So I'm very pleased that Colleen can join us today and give the first lecture devoted to these sorts of topics and issues since the inception of this course. Please join me in welcoming today's speaker and my good friend, Dr. Colleen McBride. Testing? Yes, great. I'm on. Good morning. Thanks a lot for choosing this out of all the many things you could have been doing this morning. I first off have no financial interest to disclose that would influence any of the comments I'm going to make today. So as Andy said, we're going to be moving from sort of the sort of genome discovery to thinking about ways in which this new sort of new knowledge that we're gaining might be applied in public health settings. I'm going to talk a little bit about why I think it's really important to be doing that now and then sort of set the sort of distinguish for you the differences between, say, basic science discovery genome methods and those that we apply in social and behavioral sciences to sort of translate all of what we're learning. In order to do that, I sort of have to ground you a bit in principles of public health just a little bit so that you can understand sort of why we take the approaches that we take and then I will walk you through three areas of priority that sort of a national representation of social behavioral scientists have sort of outlined for the field and give you just a few examples of some of the research that's been going on in those areas and then I'll just close with a few take home messages. So as Andy pointed out, you know, there's a lot of rhetoric out here these days about applications of genomics that first off there's the tsunami word gets used a lot in terms of discovery and then the sort of a lot of optimism about how this is all going to play out in terms of improving, bringing tangible health benefits to the public. But yet when we look at the research that's ongoing right now, we see that there's actually not a whole lot going on in the area of translation. In fact, what you see here is what you see here is this sort of large body of discovery research, just a very, very small area in which that you could even loosely call translational research and when you blow that up to say what in fact is that translational research, you see that the lion's share of it is epidemiology, which is basically looking at gene disease associations and GWAS and other kinds of applications. And the only loosely the sort of work that Bruce Korf talked about last week and that I'm going to talk about today is really probably way down here in the pink area. So a long way to go in terms of beginning to think about how this is how this is all going to play out. And I think that we have to sort of think carefully about why is that happening and I imagine maybe not so much in this room but probably others who will watch this in the future will feel some tension about how isn't it too early to be doing this and don't we need to wait a little bit and figure things out. And I think that really maps to our sort of assumptions about sort of the timeline of science and how it's supposed to happen. And typically we think that science is supposed to sort of start with basic research, this very sort of genome discovery, move on to sort of developments of treatments, pharmaceuticals or interventions as I'll talk about in a minute, just establish in very tightly controlled settings the efficacy of that intervention. Does it really work under very highly rigorously controlled circumstances and then take it out to sort of run it around the block in a little bit more real world and then way way way downstream do we really start to think about well how do we tweak this and make this really work in the real world. And we see that the funding agencies all sort of agree and that we sort of characterize research ranging from sort of T1 to T4. I think that in order to as we think about applications and the lessons that we've learned we need to sort of challenge that paradigm and start to think about in fact from a public health standpoint we would start probably at the other end of that continuum. We would start by sort of taking a look at what are the existing public health challenges and unmet needs and then we would anticipate this new technology, this new discovery. How might that be applied to help sort of address those problems then sort of in partnership start to work on basic research effectiveness I mean efficacy and treatment development. So slightly a different perspective that you're going to hear today and I'm going to make the assumption that all of what I'm proposing can be happening in tandem with the discovery. So why do that? You know why what do we have to gain from that and I think that you know we really do want to think about anything that we create having sort of optimal application and we want to we're sort of lots of again folks sort of saying we're sort of stuck right here we have this mountain of data that we've we've now unveiled or discovered and then we see at the other end what might be called premature translation things that are coming out that make people a bit uncomfortable that are being directly marketed to the public in the form of products like DC DTC genetic testing and and what we what we hope to do in terms of our research and what I'm going to sort of outline for you today is to sort of that middle ground of sort of bringing science to this debate. So lots of the suppositions that are put out there about how this is all going to play out are are testable research questions testable hypotheses and those that data then can be used to inform appropriate translation. So an example and in point here is a little over a year ago Pathways Genomics proposed to do something quite radical take their their basically their spit kit and put it on the on the shelves at Walgreens and there was a huge amount of uproar about that that that would be harmful although these these currently are available online on the web to put them in the hands in the in in the drugstore seem to be a step too far and in fact there was such an out cry about that that that Pathways was sort of pressure to pull those products off the not not actually even put them on the shelves now I would argue that there was absolutely no data to inform that decision that was a passionate emotional decision that was made perhaps the right decision to make but absolutely no data to support that so many of us in the genome Institute at CDC and across the country and sort of public health genomics enterprises are sort of stepping back and thinking okay what is this program of research what is this area of research and we're sort of calling it now genomics and societies often called LC and other kinds of names to characterize this very very complicated field and I just want to point this out to you because I think it has bearings on what you'll see as I'm talking about some of the examples that we really have the genomics and society is both the area of research includes both disciplines and a lot of disciplines as you'll see here a lot in the liberal arts and then a very wide swath of methodologies that really probably in their grossest characterization could be sort of called quantitative that is sort of you know we're counting things and counting responses and so forth and qualitative which is done in a more grounded way where we're looking at sort of coloring in the lines of what we see with our quantitative approaches and that our intent in all of this is to generate output that informs this full continuum of application from sort of just knowledge basic knowledge all the way through to public policy and I'll try to point that out as I go along today sort of where I see those opportunities so I think the first step before we go very far is to get us all in the same place with respect to terminology and I'm going to use the term intervention a lot because that's pretty much what public health people do is is develop interventions and and I want us to all sort of understand what that means because I'm going to I'm going to use it very specifically a lot but it actually can can mean a lot of things so really it's just basically this idea that we're trying to we're taking efforts to direct them at some specific target group and we're trying to influence some kind of a desired outcome and those outcomes can range range from helping people make informed decisions about whether they should be tested or not whether they should uptake some kind of an intervention whether we're trying to influence individuals to change their behaviors we're trying to influence whole groups in changing their behaviors say worksite settings or so forth we're trying to change attitudes because attitudes from our theoretical conceptualizations of this stuff suggest that that make they play a very important role in what people do or at the larger societal level we're trying to actually change public policy or inform public policy so that's the term that I'll be using a lot throughout the rest of the talk so we talk about interventions their levels at which we can target interventions from a public health standpoint we can target interventions at healthy populations for what we call primary prevention so this is to keep the healthy healthy and genetic corollaries to that are sort of the susceptibility testing sort of trying to find individuals who are who have genes that perhaps put them at higher risk before they ever get these health conditions and intervene upon them secondary prevention which is looking amongst those who are at high risk to try to screen them surveil if you will to sort of identify those who who get the disease very early to present to prevent negative outcomes so that probably that genetics area would fall into the sort of high risk target groups of hereditary cancers those who have colon cancers and breast cancers identifying those and doing sort of accelerated screening tertiary is really those who have the disease so those affected by genetic disorders many many of which are quite rare but but at a public health level represent large numbers of people in large health impact and those we may be doing things like assisting trying to figure out how to help them live lives with these rare conditions and and studying the impact of that what distinguishes I think social and behavioral science from some of the discovery work that you've been hearing about and even from the clinical research to a large extent is this sort of notion that we we think about things theoretically and conceptually we start off with a plan of why how do we think an individual or a group is going to behave in a certain set of situations and we do that across a host of levels so we do that from the sort of individual level how would this individual in this situation if there's a cigarette smoker for instance how are they going to function in this world and where are the intervention points that we would we would try to to get them to change their behavior we can do that at the interpersonal level which is sort of dyads or small groups or families networks again trying to study the how the how their context influences their behaviors we do it at the organizational level or the societal level but what goes with each of those which I think distinguishes this science from others is these conceptual models so there are established conceptual models people who have over the last past decades done a lot of thinking about what how these factors work together to influence health outcomes and we apply those relative there they map to these different levels of of sort of study and we map those models to that and we use that in terms of how we design our studies how we design our questionnaires and so forth so I'll give you a very simple example we often start with some set of outcomes they can be as I said before in terms of the the intervention outcomes I was discussing they can be a behavior they can be some type of an organizational change some type of a policy we look then to see from the theoretical perspective where are what we can call intervenable points things that an intervention can change and then we outline what those are how we think they will influence the the behavioral or whatever outcome it is we're interested in and we consider that in the context of a lot of background typically the background though not always isn't something that we can readily change or intervene upon it tends to be something that is held constant but we know that it's pretty important and in fact it may be so important that it actually affects the outcomes on its own that looks very very simple and in fact when we look at real models it's a lot more complicated and what you can see here this is the risk information seeking and processing model a model that's targeted to individuals and individual behavior change but predominantly focuses on how do people why do people seek information and once they've sought the information how do they process that information and you can obviously see the relevance of that as you think about genetics and genetic information that's coming out and you can see here that there are many many factors that come into play all of which would be put into some sort of an assessment strategy to try to capture so that we could then see where in these pathways we would perhaps want to target our interventions so I think that is a as a sort of difference between the sciences that we have to hold in mind and we'll sort of explain some of the things that I'm talking about as I go so I'm going to turn now to think about sort of where as we think about public health where where should we be trying to think about designing interventions and as I mentioned we pulled together the NHGRI a couple years ago pulled together a group about 50 social and behavioral scientists to spend a day thinking about gee where where are what is the most important thing for public health and of course it it immediately comes up that it's reducing common chronic disease and that's not really just for the for the Americas that's actually a worldwide a growing problem and as we think about that from a public health standpoint we think about well sort of ideally and to reduce the burden of of chronic disease which by the way is is increasing and has been increasing over the last couple of decades and as forecast to continue to increase prevention would be ideal so keep people from getting these chronic diseases anyway because the cost of treating them is so high in order to do that the ticket is risk behaviors we have to get people to change their health habits and but how we go about doing that really has to be amenable to public health dissemination and I'll talk about that in a minute but public health dissemination is typically characterized by being somewhat resource poor and having to reach very far across different target groups and then also as you probably heard from Bruce last week we it has to be amenable to primary care if that's in fact where we're going to be directing these interventions and primary care is characterized by very short visit times high costs so anything that gets in the way of what the doctor would normally be doing isn't going to have very much likelihood of success and we have to keep asking ourselves where does does genomic information add value because if it doesn't add value then it perhaps isn't I mean it's really not something that then should be adopted and then there probably are going to be answers to that question that are yes it does add value in some circumstances and no it doesn't in others and we want to maximize our applications for adding value lastly is that we have to keep in mind that there are already widespread health disparities and those disparities certainly we wouldn't want to exacerbate those disparities but in the ideal if we could reduce them and genomics would be a way of doing that gee we'd all be very very excited about that so when we think about public health applications though as I mentioned it is sort of this whole idea of the efficacious intervention the efficacy point along the continuum in how we do science what we see is that many times the approach and that linear approach is to come up with this terrific intervention that works beautifully under very controlled circumstances that don't in any way reflect the real world scientists have done that for a long time social and behavioral scientists have done that for a long time what we see increasingly is when we take them out to the real world obviously they flop they flop because they didn't anticipate all of the challenges of the real world and what you see then in most public health research and most thinking around translation and application is that we're on this balance being we're trying to find that sweet spot between efficacy so it's been rigorously evaluated we can say with confidence wow this works and effectiveness which is it's actually going to be useful I mean actually can be applied in the real world and usually we're toggling here on this notion of reach that efficacious interventions typically tend to reach smaller numbers because they're harder to implement effective interventions tend to reach larger numbers but they may actually not be as powerful as the efficacious interventions and we can use an example from genetic testing to sort of show that what what has to happen as we think about translation so we take so we take sort of hereditary colon cancer genetic counseling for an example something that we've been is really clinically useful now and is being used and could be used in public health settings as well but if we wanted to take what's being currently done and translate it into what are public health settings we have to sort of look at what is what is suggested for this area so typically the genetic counseling discussions or interventions with with at-risk families are what we would consider from public health standpoint high dose that is that they are they're going to be intensive they're going to take up to two or three hours to do that that means that they're also resource intensive they're typically done by a certified genetic counselor which we know there aren't a lot of those or may and certainly not enough to reach across a broad public health spectrum they're done in face-to-face sessions where individuals have to come in they have to sit down they have to you know be be available during these time frames and because of those things they're sort of demanding to sustain especially as you think about resource poor settings that would be more public health so we don't have enough genetic counselors to do this there isn't currently any reimbursement so nobody's getting paid to do it it's expensive I mean and not necessarily more expensive than many health treatments but still nonetheless expensive but very efficacious so yes it works it's beautiful but probably would reach a very small number of those who might be in need of this service so you contrast this with what would be thinking about what would be a public health approach and it looks quite different so here you'd have to think about if you were thinking about doing this in a clinical setting you're talking about 15 minutes that's the typical visit time so squeezing something like this into a 15-minute visit doesn't seem very likely that's just from the clinical perspective but even thinking about public health delivery systems not likely to be anything more than an hour and even an hour would be ambitious it's going to have to be resource light so it would have to be implemented by the infrastructure that's currently there public health infrastructure are typically masters trained health educators and that's a high high watermark but many of the public health interventions are done by telephone by mail via internet more and more so again very light touch but this would be sustainable in the sense that it would have to employ existing infrastructure and it would have to be inexpensive and then effectiveness would be the goal the goal is that we want to reach as many people we can to achieve some increment of improvement and even if that increment of improvement is quite small and I want to characterize that for you so you see sort of tangibly what the tradeoffs here are in efficacy and effectiveness so when you think about the current approach as I was saying highly efficacious we have people coming out of those sessions feeling really comfortable in making decisions about genetic testing knowing exactly what they need to do so forth and so on so 80% of the population that we reach are improved by this intervention that is a very very efficacious intervention but we can only reach 10% of the people so 10% of what would be an ideal target group yet receive this intervention so if we look at 80 times 10 we get an 80% or an 8% effectiveness quotient okay let's look at the other approach the public health model where in this case maybe we're gonna not be that effective in the sense of it being a very light intervention and maybe we'll achieve some modest outcome that we've set up front for ourselves 20% in smoking cessation we jump up and down if we get 20% of smokers to quit reach though because it's light it's implementable in many structures is 50% so we're gonna reach a lot a lot of people and we're gonna get each of those people or we're gonna give some small benefit to those individuals or that population multiply 0.20 times 0.50 and what do we get low and behold we get a better effectiveness ratio or quotient even though our intervention isn't nearly as efficacious so that's the principle upon which most public health researchers are thinking about the kinds of interventions that we're developing and how we would be thinking about broadening the reach of translation of genomics so I'm gonna turn and talk about the priority areas that we outlined and then just give some examples from some of the research that I've been doing and sort of highlight each of those areas so the first area that came up in terms of thinking about sort of facilitating translation is that we have to improve public understanding and use of genomic information in essence we have to develop a consumer group that is comfortable with this information and can size it up and can be critical consumers we also wanna if we were gonna aim this at sort of prevention of chronic diseases we have to see whether or not this new information this new knowledge can be used to improve behavior change interventions and then lastly more radical I think and probably a bit more downstream is this idea that could we actually use what we're learning in genomics to decide on new intervention targets so new outcomes that we would be looking at that might be epigenetic or some other kinds of outcomes or could we actually much like what we're seeing with pharmacogenomics could we be doing the same thing with behavioral interventions actually customizing them to things that we learn about an individual's genotype and obviously there's cross cutting themes if you wanna read more about this I don't wanna bore you with it you can see this at this publication in the American Journal of Preventive Medicine so as I said before much of the hyperbole and rhetoric out there are really offer testable hypotheses when I came to the genome and nine years ago one of them was that a very strong sense that the public was gonna be very genetically deterministic that they would misunderstand information that was given to them particularly information that was given to them directly without the intermediary without a healthcare intermediary would use that information to maybe justify continuing to do unhealthy behaviors and so and there was also a lot of supposition about how these were going to play out direct to consumer testing didn't exist in 2003 though there was a lot Francis Collins and others were talking about the notion that folks would take in some chip to their healthcare provider and he or she would help them understand it so in anticipation of that my colleagues Andy and Andy mentioned colleagues here at NHGRI at Henry Ford Health System in Detroit and at Group Health in Seattle decided to test these suppositions put them to the test through the multiplex initiative and what we had to do in the multiplex initiative is to develop a prototype test there were no DTC tests actually GWAS wasn't even really on its way yet so we came up with a set of markers that had a sort of evidence base put them through a fairly grueling review process and you can read all about that in the article here by Chris Wade in genetics and met in public health genomics so I'm not just wanted to show you that these diseases are the common chronic diseases that have the biggest public health burden these are markers are gene variants that are associated with very slight increases in risk and in fact this is the very model for what common disease genetics is going to look like is that even though we may have multiple gene variants in a particular category for some small number of people maybe you know four, five, six percent they're going to have a very, very high risk based on those gene variants but for the lion's share people they're going to be still fairly small risks on the order of 15, 20 percent increased risk so how are people going to understand that and how are they going to use that information the other thing that is characteristic of this is that individuals are going to get multiple genetic variants that tell that say that they're at risk for something so in our study with 15 gene variants on average an individual who received that feedback is going to get feedback that they have nine risk variants for various diseases so it's also a little bit different model than the model of a single gene say BRCA or these others that has a high penetrance how this information is going to be conveyed is really also at high consideration these the approaches that we used in multiplex are based on evidence based from conceptual models in which these kinds of different formats have been tested to determine whether or not people can understand them and currently this is the standard this is the recommendation the highest standard for how to communicate risk and what you see here I'm just trying to characterize sort of how this what is the content of this information so it helps you understand what the challenges are for trying to influence health outcomes so what you see here is that this is the KCNJ11 gene associated with increased risk for diabetes there's three different categories that an individual can fall into they can have no what we called risk versions for our intervention they can have one risk version or they can have two risk versions and you can see the incremental increase in risk with each of those which is quite slight so you go from 35 in 100 to 37 in 100 to 43 in 100 so only at the from the lowest to the highest there's only about an 8% increase in risk there so these individuals are are this is this is the kind of information that they're that they're that they receive now what you also see though is that what these variants have in our favor is that they're quite common so a third of the population think about that at the population level are going to get this risk information they're going to have some risk variant they're going to have this they're going to have one version of this this particular gene now across all of the genes the same thing is true we picked the variants because they were quite common so given where the field was in terms of our methodology we really wanted to use an observational an observational design an observational design meaning basically we observe it we ping them with this information so in a sense you could call that an intervention but we're not randomizing them to any different conditions everybody is getting in a sort of longitudinal approach everybody's getting this information at the same time so the observational approach though if you wanted to have any kind of external validity that is that it can be generalized to other populations you need to have a denominator you need to have a population base so that strengthens the the the sort of rigor of an observational design so what we did is we worked with the NCI cancer research network which funds about 15 HMOs around the country we picked one the Henry Ford Health System because it had a very nice demographic profile we had multi ethnic population they were all insured because this was high-risk research so we needed to if we were giving a risk message we needed to make sure that everybody had a place to go to get their health care we wanted to aim at primary prevention primary prevention meaning healthy adults who we were trying to keep from getting these common diseases and so they couldn't have any of the diseases on the battery and then we had to design feedback we had to design a way to tell these individuals which risk variance they had and and you can see that up at the top so a results sheet which outlined each of the diseases for which they had a genetic risk variant for and telling them that you know having these risk variants didn't mean that they were going to get the disease trying to address any of the misconceptions that we imagined from the literature would suggest would come up and also the suppositions about the misconceptions that would come up we could test those but we also had to spend a lot of time talking about the caveats of the information what the information couldn't tell us and how important all of these health behaviors were in terms of influencing these health outcomes so it's not just genes it's basically the message so the observational design as I said the importance here of having a denominator having a denominator of individuals that we know a lot about we can do that in a quantitative quantitative way via a survey so we identified this sample of patients at Henry Ford it was about 6,000 patients that we approached we tried to conduct a baseline survey with them ended up conducting that with about 2,000 individuals then we offered them web access to make a decision about whether or not they wanted to be tested those 2,000 individuals they then if they decided after going through this online system that they wanted to be tested then they had to schedule an appointment with a to go into a clinic to have a blood draw and then their results were provided by mail with a health a research educator calling them up about 2 weeks within 2 weeks of their receipt of their results and then we followed them all up follow those who got tested we followed up again at 3 months follow up so as you can see here there's this you'll see this cascade of public health applied approaches light quite extraordinary for the field of genomics and yet absolutely essential if what we're going to try to do is actually put this out in the real world and this is from the baseline survey and I just wanted to show that sort of from an observational standpoint how we can bring some data to this discussion one of the concerns about determinism and here we have a population of about 2,000 healthy adults who by all accounts have their I'm not showing you their demographics but they were a very representative sample of insured adults so nice representation there and what you can see here is we're trying to bring data to this discussion of is the population in the absence now this is before they've gotten any kind of education about genetics they don't even know the studies about genetics at this point they're just completing a baseline survey and what you can see here is that we asked them for these these eight health conditions that we these eight common chronic diseases we asked them the extent to which they thought behavior contributed to those health conditions and the extent to which genetics contributed to those health conditions if you go across the bottom here and all the diseases red bars being behavior contributes blue bars being genetics contributes you can see that that behavior trumps genetics across the board that people believe the behavior causes diseases more so than genetics yet they do also believe that genetics plays a role in these diseases as well and there's nuance in this there's nuance in say lung cancer for example where we have huge public health campaigns about tobacco use and so you can see here that individuals are far more likely to endorse behavioral contributors than genetic contributors so it says to me this data would say in bringing some data to that debate that actually adults healthy adults are pretty capable of and they they understand that there are both genetic and behavioral contributors and perhaps our concerns about sort of hyper determinism may be may be not not the right concerns to have there may be other concerns but this may not be the one likewise we also heard a lot about uptake of these tests that there would be such such great interest in these tests that it would overwhelm healthcare delivery systems it would overwhelm public health systems because there would be too many people opting in for these tests so what you see here is this is the beauty of having a denominator we had 1,959 who were surveyed who went to the website less than half about a third actually went to the website to even consider testing so two-thirds weren't even interested in considering it and of those who considered testing 350 said yeah I think I want to be tested but only 266 showed up for the blood draw so overall at a population level only about 14% of those who completed our survey actually ever showed up to be tested and wanted to be tested that seems like a small number I mean I would imagine you all think that's a small number it's far less than the 50, 60, 70, 80% that might have been estimated or guesstimated but at a public health level again 14% given the current state of affairs in terms of our healthcare delivery system and our public health system would be a tsunami and would be far on above what our resources are capable of managing so even though it's a small number if you generalize that across the population it can have a big impact so the other things that we can learn about is from these sort of observational studies is where would be the intervention points so we're looking as you as you remember in those models I was pointing out we're looking for that middle point those those intervenable factors and so here we looked at individuals how they went online to consider the decisions how many pages they read so we could do that electronically via the web count the amount of time they spent on it and the number of pages they read and and and look at other kinds of factors as well that might be intervenable so I'm going to just draw your attention to three of these the first being the number of pages viewed so what we see is that individuals who viewed more pages were more likely to decide to be tested that might suggest that the content of the material was perhaps a little more positive than it should have been except for the fact that we had about 50% of those who showed up to decide to be tested didn't get tested so it we didn't see an exorbitant number of people deciding after that but looks like reading more meant that that either they were they started out being more interested or it was more convincing to them to get tested but also what we see is that it increased the more pages they read it increased regardless of what their decision was whether to be tested or not it increased their ease of making a decision so this kind of this sort of says to us our intervention is perhaps usable in a target if it had said that reading more pages made them think that the decision was less easy than we would have realized that we needed to to reframe that intervention and do something differently but you can see also is what affected some of these outcomes is the individual's perceptions of their ability to understand the information we call this jargon and lease self-efficacy and that too was associated with making a decision to be tested and with ease of the decision so they're bolstering individuals a target an intervention target would be bolstering individuals confidence that they could understand the information likewise that those who thought the information was important were also more likely to think the decision was easy to make so again an intervention might target ways in which making this information the salience of this information what they're going to gain or at least emphasizing that in the intervention you're not going to get much out of this this isn't particularly important to you is going to help help them make a better decision and lastly we can influence communication patterns in larger groups either with the healthcare delivery system or we can understand how that's happening in this short run that would be very useful in planning and public policy so what you see here is after those individuals those 266 that got tested we asked them at follow up who did they who did they share this information with and again the concern was that this was going to be largely something that happened between the healthcare provider and the patient with the healthcare providers in this study not being involved at all and really not knowing much of anything about common disease genetics we saw that only 11 percent of those individuals I think it's only actually that's a number I think it's only 11 of the individuals actually showed up to their healthcare provider with the test results more often they were telling their family members about it what they were telling their family members and whether they were accurate and what they were telling their family members again could be an intervention target it could be something that we would explore in more detail but it's clearly that's where the action is in terms of communication about these results and and they were as likely to tell their friends as their healthcare providers healthcare utilization so again this notion this rhetoric that it's going to really overwhelm that those who get tested are going to disproportionately use healthcare services and those could be those could be not just healthcare delivery service could be public health services and what we see here again is the data doesn't support those concerns as we look at these are a little bit hard to see but you've got this solid line is those who underwent the genetic testing the multiplex testing this dotted line that's very very faint in the middle is those who just showed up to think about to consider the testing on the web but didn't get tested and then the dashed line at the bottom is those who just completed a baseline survey didn't go online and didn't and didn't also didn't get tested and what you see is that there's absolutely no effect of the test which happened at about this point in the quarter there was no increase in use of healthcare services but mind you those who showed up to get multiplex testing were higher users to begin with perhaps the worried well in this instance we know they weren't the worried well because the target sample that we got were not an overly healthy population they didn't have any of those health risks but 30% of that but they didn't have any of the diseases mind you but 30% of them were overweight 28% of them smoked cigarettes so they looked like the general population in terms of their health behaviors so I'm going to move on to the second priority area that was the sort of notion of how do we communicate how do we build and inform consumer how that what that research might look like how it could be informative to different public policy or or or engaging new lines of research now I'm going to move to how can we how might we improve behavior change interventions so the big hope for genetics and genetic has been really genetic risk feedback we've focused a lot of attention on that with the idea being that and this is just an amalgamation of many conceptual models around how people respond to information risk information but the but the thinking has been that the way that this would work is that we would increase individuals we'd personalize individuals perceive susceptibility this is your genes and your genes are saying you're at higher risk as opposed to saying in the population people who sort of look like you have this percent risk and that that would increase their motivational relevance wow this is about me okay I got to get serious the other side has been very concerned about the fact that if we talk about genetics then people are going to think oh there's nothing I can do it's my genetic makeup so I might as well give up and that that would decrease their confidence that anything they did would change and that there'd be a host of other kinds of individual or group level characteristics like you know how capable they were how confident they were how motivated they were when they came in for testing and their attitudes and dispositional factors and so forth and that that would all happen in some context and the context might frame and make certain kinds of risk communications more impactful say a family member has just been diagnosed with a health care with a disease and now you're getting genetic testing that would be a contextual factor but all of this the active ingredient is motivation is that somehow this information is going to be more motivational than any of the other many amounts of risk information that we provide so before actually the completion of the the sequence of the human genome there was already research going on in this area anticipating that we were going to find out about genetics and it was largely starting in the area of smoking cessation and there were two clinical trials I did one and Karen Lairman at Georgetown at the time did another that basically just took genetic risk information and plunked it into a smoking cessation program one group got it all the other groups didn't and then looked at its impact on on smoking cessation and these the genes that were used were two different genes one was sort of a processing gene a sort of metabolic processing gene the other was well they were both they were both in different parts of that downstream system of metabolizing carcinogens and what you see here is that didn't have any effect no effect on cessation outcomes they're modest again as I said these 20 percent success rates that you see but no difference at any of the follow-ups between different between different risk groups those who got a high risk genetic message versus those who didn't and that has really that was a very dampening effect on the field that gee this isn't going to have any real benefit but I would argue I get to because I did one of the studies that that actually there were these studies were very flawed they were flawed because they targeted highly motivated smokers motivation was supposed to be the active ingredient they also did not they basically plunked in these genetic tests they really couldn't disentangle the separate effect of genetic testing in fact there there were other things that were confounded with it and so and it was single genes and so forth so there was lots of limitations to these studies and we would never in any field a study do two studies and say sorry we've got the answer and it's not working so from that point we stepped back when I first got here and 2003 to look at whether in a more public health-friendly model these were all clinic-based studies as well could we influence things other than smoking cessation because mind you smoking is an addictive behavior many of these behaviors have addictive properties and genetic underpinnings that suggest there's an addictive property to them to think that somehow telling somebody that they're at a 15 or 20 or 30 or even 50 percent increased risk to think that that would prompt smoking cessation is a bit optimistic so let's think about different kinds of behavioral outcomes that could be important from a public health standpoint one of the big challenges we have national hotlines around the country a smoker can call up today and say ring ring help me quit smoking and do that for free not many people do it so perhaps this kind of information could be used to get people to make them avail themselves of these available services so what we did is I'm sorry we used a web-based program developed by some folks in this room to actually invite family members of recently diagnosed lung cancer patients to undergo genetic susceptibility testing for susceptibility to lung cancer and we did this online because families of lung cancer patients don't all live near their lung cancer patient family member they live all over the country so reach is an issue here broadened the reach they decided online whether or not they wanted testing they they submitted a a spit sample by mail we analyzed it we directed we sent them a ping and said come back to the website and we gave them their test result online radical again in terms of thinking the way we are currently thinking about genomic applications but from a public health dissemination model probably the only way this kind of thing could really happen but what we found was that even in this modality even though we thought we would be able to reach lower motivated smokers what we found is that we didn't we still saw ceiling effects we still saw that those individuals who showed up to consider testing which were almost identical to the people who then logged on it was a very small difference in those two groups were highly motivated at baseline to quit smoking so what is genetic risk information going to do for them if if if the if what we think the active ingredient here is is motivation and so in that regard again it strikes we begin to see based on the multiplex in this that people who are already very motivated to change their behaviors are probably the people who are most interested in genetic susceptibility information that really refrains thinking about how we would even apply this information you can already see that it challenges that large-scale supposition about wow this is going to be a great motivational tool and in fact when we our other outcome was okay we're going to give them free smoking cessation materials so maybe that will be the distinguishing factor those individuals who get a high-risk message will be more likely to take up these free services that we're offering to them nicotine replacement therapy self-help booklets which this is all standard of care for smoking cessation active support meaning six telephone counseling calls for free that these individuals could use what you see regardless of their test result no difference in the use of pharmacologic therapies no difference in the self-help selection small but not significant difference in active support suggesting that maybe those who got the high-risk message were a little bit more engaged a little bit more serious but overall my guess I'll put it out there is that actually they were expecting more from these genetic tests that this would tell them a little bit more about how to make it easier to quit not just that they needed to quit but I'll come back to that in a minute so that sends us back to the drawing board what do we do okay so let's rethink this whole model so in from a public health standpoint we've got this variability of dose right we're trying to map that we're trying to find what's the optimal target group for delivering or for testing out some of these sort of uses of genomics so we've got these two sort of bars the low intervention dose to high we know that public health we're aiming for to move lower on that continuum and we're trying to do primary prevention so we're trying to find those individuals who probably the idea of changing their behaviors is a little less salient because they're not suffering really any of the health effects yet and that we're looking along that that curve for where is the optimal place to target these kinds of interventions and so ideally what we would be trying to do is to by using genetic information enable ourselves to lower the intervention dose so make make a lower dose of intervention which will be cheaper more effective more potent so where do we go from here then well that sort of directed us to thinking about young smokers and so thinking about target groups like college students that are in the transition from living at home to a stressful first year college where they learn quickly that nicotine has some benefits for helping them concentrate for helping them manage anxiety and so these are this has been widely identified as an important target group very difficult to get their attention very difficult to make any kind of health risk salient to them and there's lots of conceptual models that can guide our thinking about that we know that that from our previous work that and this is where we see these biases that individuals in these kinds of groups might be inclined or motivated to seek testing to find out a good news message I'm not in that high-risk category therefore I can keep smoking so there are lots of conceptual models about why do people do that how do they go about protecting themselves distancing themselves from information and so forth that we can apply in this area and it turns out that just in our preliminary data surveying about 640 college students freshman college students in North Carolina that you can see here that most of them believe that they can quit smoking any time or that they score relatively high on that and that they don't think that the health harms of smoking are going to happen anytime soon now that's a survey methodology quantitative approach that we use over and over again and it's only a survey methodology is only as good as the questions that you put in the survey and if you serve if your question doesn't have a lot of variability or it doesn't or the results seem uninterpretable you really do need some other methodologies to help you understand that especially when you're in hypothesis development and intervention development so we supported this with qualitative assessments meaning we called in college students we put the test in front of them we gave them sort of a booklet of information about what they would see to sort of see in real time what their reactions to that are and what you see is that is that these individuals sort of flesh out this notion of how our public health messages have backfired and how we might be able to use genomics to help sort of reduce that backfire effect so most kids have bought into this notion that you can quit anytime what we tried to do with public health was to say it's never too late to quit because we didn't want people to give up well in fact the young adults now are using that information to say it's never too late to quit I'm going to smoke through college I'm going to use it to my advantage and then I'll quit later and of course by that time they get addicted they've underestimated the addictive properties and there you can see a traction point perhaps underestimating the oh sorry perhaps underestimating the potential for addiction is an angle for genetics to be coming in to talk about the genetics of addiction and helping people see the susceptibility that they may have to addiction not to health outcomes and the other thing that we see here is that they misunderstand sort of the susceptibility exposure I guess dialectic here that actually for less exposure those who are susceptible are harmed by less exposure so this waiting until I can wait notion really isn't accurate it's a misconception and that misconception is revealed and that is addressable and intervenable and that is what we would do either we would address the misconception and or we would try to give feedback on perhaps on addiction and we'd have to think about how do you do that how do you do that in a cost effective low dose manner well thinking about using these infrastructures of social networking might be a very viable way to sort to have these kinds of implementation of this sort of genetic information so I want to turn now to the other public health epidemic and thinking about how we might bring because this one is getting just huge amounts of attention rightfully so that we have an epidemic of our kids getting overweight of our adults getting overweight of basically everybody getting overweight and we if we're thinking about primary prevention and moving down the continuum as we did with the young smokers we're going to want to move down the continuum with kids too because a lot of the eating patterns and so forth are established very young and if we want to try to influence those behaviors we've got to start young but as you can imagine the notion of doing genetic testing in its early phases where we are now with kids not very not a very popular idea so we have to think sort of creatively and methodologically about how can we start to ask these questions again before not waiting for the answer so if we can't do genetic testing perhaps we can use tools like family history so family history though probably much inflated over what we're going to learn from genetics in terms of risk levels here you can look at a child's risk of becoming overweight if one or one of their parents is overweight is a double so having a mother or father who's overweight doubles a child's risk of becoming overweight and if they have both of their parents overweight it almost it's at least three times almost four times no maybe it's five times actually increased risk of becoming overweight so family history is powerful but with young kids of course there's some downsides to family history it hasn't fully established and so forth but it may be a good surrogate to start to use to anticipate sort of how this genetic information might play out the other sort of challenge though for and I say clinical integration here but it's actually public health integration the other challenges is that we as a social behavioral scientist so often have to rely on what people tell us so we ask them a question did you change they say yes they say no and we we may do some things to try to figure out if we can trust that answer ask them to give us a biological sample of some sort but very very difficult to validate what people tell us so if we want to really start understanding the effect on actual behavior we have to get creative and that is really sort of keeping in mind that this nature of genomic technology is changing so we have to be anticipating it it's not here yet that some of these situations as I said with the genetic testing are either unpalatable from a scientific standpoint we can't do them or they're just very difficult to envision and that where these things are happening is typically in complex context where there's a lot of variables that are moving and so it's difficult to see where the genetic information can be sorted out as the factor so we developed in our social behavioral research branch a lab called the immersive virtual environment testing area to begin to bring sort of some rigor move away from hypothetical vignettes of imagine you get a test and imagine you have to do something for your child what do you think you would do you know a lot of us have trouble with that but that's not very rigorous so this immersive virtual environment allows us to put an individual in a headset and in that headset there are cameras that are following these this individual as they're moving around in this virtual world and what the individual is seeing in the camera in the headset is some virtual world in this case it's a cafeteria putting a mother in a cafeteria so she's going to be making food choices for her child is the sort of context and we can measure we can observe everything that this individual is doing in this in this cafeteria we can track their eye movements we can figure out what they're putting on their plate and so forth so it gives us a lot of behavioral outcomes that we can measure after we've given them some information so in this case we decided to take on this notion of giving genetic information to mothers one of the great concerns from an ethical standpoint is that if parents are given information risk information they get scared on behalf of their kids that they will they will over parent that they will over control eating habits for instance and we know from our conceptual models that over controlling over regulating food intake in children backfires and makes them want to eat more makes them sneak food and so forth so here we have an individual the mother is in this virtual reality and we're going to ask we've randomized them to so this is an experimental design now we're moving away from the observational and the quantitative qualitative so we identify individuals we're looking for overweight mothers because we want all of the kids to at least have that want that have some increased risk some hereditary increased risk and we're looking at young children who really are sort of pre-obesity typically and we identify these mothers they come in they volunteer we do a baseline survey we screen them in the lab to make sure that they're not they don't get nauseous when they wear a headset we have them practice the behavior so that we can get out any noise practice so that it's not that I couldn't manage this plate in the in the virtual world but they can actually manage a plate in the virtual world and then we randomize them to one of three conditions where we're asking one food safety we're not telling them anything about we're just telling them about how food gets spoiled and can cause you know different kinds of food poisoning here we have behavioral risk information where we're just saying that this is the health behaviors that are associated with kids becoming overweight these are the eating patterns and here we add this assessment of family history so we tell the mom your child your index child is at this level of risk based on you and your the birth father of this individual and then we do a post survey we put them into the buffet so we see if we see if they understood this information so we can have this internal validity that the information actually contributed to the the behaviors that happen in the buffet and then we do a post test survey now I don't have the movie I'm sorry I didn't have the appropriate software but this is just to give you a sense of the mothers what what these visual worlds look like they're very immersive and they are and the kinds of foods you would have seen the mother selecting foods if I'd had the movie but she's filling a plate we can measure everything about the plate in terms of the calories the calories from different sources protein fat etc and all I'm going to present to you today is just to give you a feel that one of the things that we try to do in order to establish rigor is to make sure that our experimental manipulation worked so did the did the mothers believe this did they feel that they were really in a buffet and what we see over and over again in these in these virtual worlds is that in fact they are quite immersive and that people do feel that they are there and that they are that they are filling a plate for their child so last example in this area is the is public health is public health is in the trenches and it happens typically in resource poor settings so as we think about genomics we really want to make sure that it has some promise for improving public health in the in the in the to the lowest common denominator and an example of this is a health condition called podoconiosis that is endemic in rural Ethiopia this is a really ideal kind of situation for evaluating whether genomics could have any promise for these kinds of fields because these are resource poor settings and podoconiosis let me step back first podoconiosis is an elephantiasis that's caused by exposure to silica particles in the clay soils of the highland Ethiopia and it is absolutely preventable if those who are at high risk and genetically at high risk wear shoes to protect their feet from the clay soil 50% of the population of Ethiopia are under the age of 15 so if we're going to do primary prevention of podoconiosis that is a huge number to try to deliver services to even if we take a small proportion of that who might be at risk a 5% proportion that would still overpower most of the public health settings this on average these individuals these families make a dollar a week so this is an opportunity where a moving query and others might say that genomic information might be used to target to stratify risk groups and target resources where they would be most useful but you can imagine when you start to talk about targeting limited resources to special groups that it raises all kinds of concerns about how that would be played out and the backlash on that and stigma so what we did here again working with a community in a community kind of endeavor which is another one of the methodologies I talked about in large groups is that we started with qualitative work first to engage the community in understanding what it is that are the barriers to these kinds of to the uptake of the shoe wear and what you can see here is that we we did a very large effort in this case for sites these are the number of cases of podoconiosis in these communities and we did 28 focus groups 38 individual interviews and some case studies and really in our sampling pool so these qualitative efforts can be quite sizable we had 307 participants just to give you an example of what we found is that we all have seen in the literature and we hear again and again this notion that heredity and talking about heredity in terms of conditions where we're still harking back to this determinism concern is really associated with concerns about increasing stigma how might that happen and in our case we did see this when we surveyed our participants they said if they thought that that this disease was solely hereditary that they were less inclined to think that they should wear shoes makes sense and that they were more inclined to think that they should not marry any individual that would have this condition because that would increase their likelihood of their children offspring having it so they did social distancing behaviors putting them away or keeping them at bay they chose their partners based on this and then they also if they had the conditions stigmatized themselves and took themselves out of social situations but what we also found which is interesting is that if they believed that it was solely not hereditary that that too had a whole host of implications for stigma that they might endorse the importance of wearing shoes and they might be more empathetic to patients but they feared contagion and so they wanted to keep social distance from these individuals because they were afraid they would catch it so you can see here that the the fact of this being a gene environment interaction where this is just an increased risk that any intervention we do is going to have to help them understand the nuances of gene by environment interactions this is a I think 10 percent 50 percent of the of the men go to school less than 50 percent of kids go to school a literate population low literacy at best challenge of talking about gene environment interactions to low literacy populations but we have to start trying to do that because that's in fact if we're going to deploy these things in public health settings we're going to have to do that so we have set about now to do a community trial that is just getting started where we will have you know baseline assessments we're using quasi-experimental design first time I've used that term quasi-experimental design we often have to do with large settings real-world settings where we have to we can't randomize communities per se we can't randomize within communities we can't say these kids are going to get this and these kids are going to get that because they're all together in a very small community so we have to use different approaches often what we do is multiple baselines so we see we have sort of a time series then we introduce interventions and then we do multiple time series afterwards so we don't have randomization but we do have some credibility that what we did if we see a consistent blip was in fact what caused that outcome so here we're going to talk about three groups I won't go into them but it's basically the same kind of thing of adding genetics education targeting affected households and unaffected households differently trying to influence the understanding amongst the unaffected households as a means of decreasing stigma and in this case stigmas are our outcome it's not a health behavior change per se it's can we change these people's beliefs about this disease in ways that will foster the behavior so that's the mediating factor we're trying to influence on our way to getting shoe wearing this is the public health this is what public health campaigns look like in terms of the background of trying to promote the behavior but in terms of the sort of metaphors that we have to use to educate illiterate populations and I just use this as an example what we see in the Ethiopian communities is that it's a high mesa so lots of sun not many trees and you'll see a lot of variability individual variation in response to the sun some people are out in the sun it's beating down on them doesn't seem to bother them at all wearing no headwear no protective anything other people are walking around with various contraptions that clearly show they're highly sensitive to the sun and what we see is so a lot of those contraptions are are umbrellas but what we see is that that metaphor resonates that doesn't have a value judgment to it no one's saying gee it's bad to be sun sensitive and so we can use that as a way of talking about genetic susceptibility and then we can transfer that to this footwear where we see the environment in the soil being an exposure we haven't refined this yet this is still a work in progress versus the extent to which so everybody's at risk but those who are genetically susceptible are more at risk trying to so I'm moving into the last section here hang in there with me the last section is this idea of can we change behavior change interventions could we maybe using a pharmacologic model or pharmacoepidemiology model do the same thing with behavior change interventions first have to just make a case that actually the interventions that we're doing right now are based on what most people are getting from a public health standpoint are genetic generic recommendations about what they should be doing and what we see is that overall a large percent of the population are not responding to these generic intervention or genetic generic messages like eat five fruits and vegetables a day don't smoke keep your weight down that sort of thing and in fact what we see from the Hain study is that it's getting worse that we're seeing decline in adherence to these generic recommendations over time so we got a problem here and when we look at the interventions that we currently have these are just a meta analyses of interventions that have targeted low income groups for a variety of behaviors we see that I mean the takeaway message not to see all of these is that the light colored bars are where it's effective where the intervention strategy has been effective dark colored bars is where they haven't been effective and what you can see here is that most of our interventions are as likely to be ineffective as they are to be effective and so room for improvement we should be giving this more time and attention again just taking that back that sort of what's the problem and then how do we bring genomics to it when we look at those who do successfully change behaviors smoking cessation up here weight gain down weight loss and regain down here we see that we start out with the success and we see very steep relapse curves so most people who quit smoking relapse within six months after quitting most people who lose weight this is the weight loss regain it very quickly and sometimes gain more than what they lost so again relapse problem fairly significant when we try to understand what's going on here why are these interventions not working why are people not adhering to them we can we can look at sort of the standard randomized trials this is just a randomized a clinical trial just to give you a sense of how much loss to follow we get from the beginning when an individual is recruited through to completion of the study by six months a third of the sample have withdrawn and by the end of by 12 months half of them have withdrawn so we lose a lot of people and even at the end of it we show fairly modest changes in behavior an average weight loss program may may lose 10 pounds across the population and in these cases there's very little difference across these these different programs in weight loss and then we think the people regain it so a lot of this is the most motivated of the motivated that get to the end and they lose maybe 10 pounds and then when we look at the reasons for withdrawal we see that they didn't like the group they got into or they couldn't tolerate the diet or they were dissatisfied with the weight loss they were doing what they said they what they were told to do but they weren't losing weight and and that they just couldn't be compliant compliance adherence huge problem all right so when we look at this it you know we hear over and over again it's simple stupid energy and energy out but in reality there's huge amounts of variation individual variation and how into how people respond to calorie restriction and physical activity and and lots of genes that are being discovered that point to those what's underpinning those individual variations in fact it's also very complicated because we have an environment out there and we have different levels of exposure to that environment and probably likely individual variation in response to that environment so as we think about going forward we need to start thinking about models of how can we use this how might where might these points be that this would influence behavior change interventions and what you see here is just one model of that genetic factors could influence the physiological response whether someone overheats or not thermal regulation could also affect the subjective response to exercise do they get do they get that euphoric buzz or after a dolphin rush after exercise that could influence motivation which in turn could influence exercise and in fact what we're seeing more and more is that exercise also will affect epigenetic or sort of turning genes on and off so so thinking about those models we could then apply this kind of thinking to improving the interventions so I'm conscious of the time here I want to make sure we have plenty of time for questions these next slides are really just to support that there are genes out there that do show in populations where we've rigorously controlled their exercise and dietary intake that we do see variation in response to the intervention so this is a Japanese population and what you see here is it's a very difficult slide this is calorie this is calorie restriction and changes in weight so we see a very strong correlation about two-thirds of the population if they restrict their calories they lose weight if they increase their steps they lose weight but for a third of the population we see no association that carry this gene variant so they're doing this and they're not getting the benefit so in fact what they're telling us in these clinical trials may in fact be true and likewise what we're seeing is that in terms of measuring physical activity response in a very tightly controlled settings that we see that some genotypes are experiencing direct or linear increases virtually linear increases in positive affect when they exercise others not so much but those others that are not experiencing this linear improvement in positive affect are feeling that they're exerting themselves more so they're working harder to get less benefit so in fact what people are telling us about the interventions may have some basis in reality and I think that we also have to consider that perhaps one of the suppositions we may have here is that we this motivation effect which is so powerful that gets people into interventions and what gets people to lose weight and quit smoking and do all of these things may not be where our traction point is for targeting genetic information because that motivation trumps everything but where we may see real advantages is in relapse prevention so here we see this is a neutrogenomics study neutrogenomics very controversial but something that is certainly being talked about as a wave of the future and what you see here is that in individuals who had neutrogenomic we would all sort of question the validity of these kinds of recommendations but people who were told that their diet was being customized to their genetic makeup weren't more likely to lose weight in a short run but were far more likely to maintain that weight in the long run so it may be that either the perception that the diet was tailored or the realities of what they were being told made that easier for them to adhere to in the long run something to continue to consider so take home messages I want to just I think you've probably gotten the point here that I think that translation research is really important and that it shouldn't be we shouldn't be waiting we should be doing it now it can be done in partnership with our basic genomic science colleagues that there we need to look for these ways that it might be used to improve health and I think there are many possible ways that deserve are deserving of us sort of spending some time we have lots of tools and research possibilities for how to do that what will separate social and behavioral science from say GWAS and other kinds of things where we throw everything into the mix and we need whoppingly huge sample sizes is that our work is conceptually and hypothesis driven so we are looking to influence mechanisms we are trying to understand mechanisms within a specific set of behaviors and that as I said before we have lots and lots and lots of methods to use that can be adaptable to the different settings and that keep in mind this balance of efficacy and effectiveness and most importantly that anything that we are doing in this domain and public health is inherently interdisciplinary and that we really do need to have basically multi sort of discipline teams engaging in both setting the hypotheses and actually developing the methodologies and with that I will say if you need to contact me because I sped through this you can and thank you very much for your attention and for staying a little long we'll take questions at the podium thank you for coming today please ask me some questions so I don't feel guilty for talking so long no questions oh they're coming down sorry there you go hi I'll ask a question first of all that was fantastic thank you so much it's and it's a lot to absorb it is I'm sorry I wanted to follow up on sort of reactions that the public get when they hear things and in the context of what you do when you hear that you're susceptible to something more than the general population I like your little pictures and then what do you do you tell them you should eat better you should exercise you're gonna every one of those common disorders pretty much the message has been to the public right do those things yeah watch your weight watch your fat so if that's the case but on what you've been able to show what's the value added of the information if it's deemed genetic if in fact we know that cognitive dissonance sets in and different people react differently sometimes it motivates them sometimes it paralyzes them so on a public health model short of sort of knowing people's personalities which maybe are genetic I don't know yeah they are what difference then you know from the public's perspective would it make well I don't mean to be so depressing no no no that's exactly the point I was making though initially in saying that that are our notions of where where this kind of information was going to be the most useful I think we're off track I think that our early data are showing that that is not and I think we've gotten far at least public health has gotten far too caught up in this whole risk communication is the is the panacea it's not and I think there may be groups where risk communication might be advantageous and that's why I'm saying going after folks that really aren't aware of risk and personalizing it in ways but you're right it's still going to create sort of cognitive responses to that that we have to address but they're all intervenable and we do it every day with lots of other kinds of health outcomes so I don't know exactly I mean I think this last area of can we tell people something different based on their genetics so that we're not giving them a generic message might be a really really useful tool can we do that and how do we go about doing that and in some ways it goes back almost to one of those very first slides you showed about the distinction between clinical and public health where you've shown that in that motivated one-on-one clinical context you can really be effective and then when you move it into this public health approach it's much more difficult well but also from a public for it to have public health impact it doesn't have to be as effective and I think that's a really important thing to not lose track of we don't have to have 80 percent efficacious interventions going out to the far reaches of we don't we don't need that in order to make a public health impact and so I think that's that's where the model we have to keep the public health model in mind great thanks I also want to thank you for such a great talk you had shown some very interesting findings concerning race in your multiplex study and the decision to test so I was wondering if you could talk about that as well as the role of public health genomics in decreasing health disparities as well as potentially increasing health disparities yeah well I think the multiplex project what I didn't show was that actually the population-based sample enabled us to show that African-Americans were far less likely at every stage of the game to complete the survey to show up at the website and to opt to be tested and that is consistent with other findings in the literature but there hasn't been a denominator in the past so it's been who shows up at the door and who shows up at the door largely white affluent women and so what we are take on that actually if you looked at the decision aid and you thought about and you carefully considered genetic testing and believed everything that was said online you would have opted not to be tested because it really wasn't that this test was going to tell you anything that you didn't already know so our take on it is that African-Americans in this instance were far better consumers when they did show up now the issues around survey research and who they're you know different subgroups reticence to be involved in survey research is a separate issue but if you think about it from the genetic testing issue they were far better consumers of that test and that that's the kind of healthy skepticism that the general public needs to have and we need to understand that especially at least to be able to sort the wheat from the chaff what are the good what are the good applications of what aren't with respect to disparities that's why I'm so interested in this photo issue because I think that in this case the disparities are caused and these folks that have this condition are socially isolated and really pretty much you know where they have a completely preventable condition with shoots I mean this is this is not hard or expensive and um but the but the the misconceptions about genetics is actually contributing to the stigma it's actually it's creating exacerbating the the the disparity whereas if if we brought in if we could do it in a way that was really you know acceptable to those target populations we actually could reduce the stigma and so I think that's the imagination they have to bring to these questions is you know get inside of it maybe using some qualitative methodologies and some you know quantitative and then sort of get a good sense of because it's not going to be generic we're not going to be able to find one approach and just plunk it into every different group it's going to have to be contextualized yes so it was curious as far as motivation you know for things such as smoking to prevent people from relapsing and also weight loss how do social networks sort of come into play do they actually motivate you I know there's several social network for exercise loss I mean do they motivate you and I'm wondering as far as you know talking about shoes and so on are there ways to sort of I mean I guess not you know a web internet social network but some other type of social network to actually motivate yeah well you hit on a really good point I mean not much of what I'm talking about has a sort of individual it's done in group but it takes a very individual kind of approach the POTO project does is households and it is families because we do have to talk about this affecting whole families but networks I think is really the rub here for genetics I mean that genetic information although when you get into these common genetic variants it gets a little muddier but that the families really are what they have in common here and that we should really be taking advantage of networks and very amazingly very few public health researchers are doing that you probably know Laura Kaley she's doing that quite ably and looking at family history and how family history information disseminates through the family and where there are interruptions and that information that could impair their adherence to different health recommendations so yes I didn't give that near enough time but that is absolutely a central area thank you for thanks sorry sorry sorry