 Okay, good morning everyone and welcome to week six. As we've touched upon several times during this series and as we'll continue to do so in future lectures, one of the major drivers of genomic research really involves changing the way we think about the practice of medicine. And so this week's lecture will speak directly to that theme, addressing how complex genomic approaches can be applied in a real world context to improving public health. So it's my great pleasure to introduce to you today's lecture Dr. Colleen McBride. She is a senior investigator and the chief of the behavioral and social research branch at NHGRI. When Dr. McBride joined our institute some 10 years ago she created from the ground up a brand new research branch, SBRB, and this new branch very quickly became one of the premier research programs that's dedicated to investigating social and behavioral research factors that facilitate the translation of genomic discoveries with the goal of improving 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. And one of the hallmarks of her research program is that she is continually looking for new ways to capitalize on advances in genomic science in order to improve the efficacy of behavior change interventions. I personally have the privilege of working 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 and that was an absolutely wonderful experience to have that opportunity to work with Colleen on this project. You'll hear more about the multiplex initiative during the course of today's lecture and I really feel that this project is an excellent example of the kind of translational research that Eric Green spoke about during the very first week of this series. One that provides an evidence base for genomic medicine by looking at how we should deliver genomic information to patients, how individuals interpret their genetic test results, and how they'll ultimately use this knowledge in making their own healthcare decisions. All of these, of course, themes that are highlighted in NHGRI's current vision document for the field of genomics that are key to advancing the science of medicine. So with that I'm very pleased that Colleen can join all of us today. Please join me in welcoming today's speaker and my good friend Dr. Colleen McBride. Thanks, Andy, and thanks to the course organizers for having me today. I have no financial interest to disclose, unfortunately, but what I have organized today is to sort of walk through in five parts talking about how we might be applying what we're learning in genomics for public health improvements. I'm first going to make a case, because I presume I need to make a case, for why we should be doing translation research now before we have it all figured out. And I'm going to give you then an overview of the kinds of social and behavioral research approaches that we use. It's a broad spectrum just to get us all on the same page about terminology and what we're talking about in terms of methods. I'm also going to just briefly touch on how the philosophy of public health might influence how we do translation research and what our ultimate aim for is in the products. I'm going to give you a few examples of translation research, some of my own, some where I'm looking kind of forward, thinking about ways that maybe aren't quite ready yet, but could be in the near future. And then I'll leave you with some deep thoughts at the end. So we hear a lot about sort of the tsunami of genomic discovery and the fact that we're going to be overwhelmed with new knowledge, and at the same time we're hearing a lot about a lot of optimism about how this is going to be used to directly benefit the public. I think probably Eric talked a lot more about how that might be integrated into medicine. I'm going to focus a lot more today on how it can be integrated into public health, and you'll see the distinction of the two as we move along. But when we think about what has actually happened in this time frame of discovery the last decade and change, what has happened in terms of real applications that are tangible benefits to the public, what you see here is the publications that have been occurring over since from 1995 to about 2007. That same trajectory has not changed at all in the last seven years. Is that we see that the lion's share of what's finding its way into the literature at least is discovery research, and what would even under very loose liberal senses be considered to be translation research is really quite small. Now, when you blow that up, what you see too is that the lion's share of that is really epidemiologic studies like GWAS to look at gene disease associations, and anything that moves us into the clinic or into practice guidelines or anything really even medical oriented is really quite small. And I think that the reason for that really comes down to our assumptions about how science is supposed to be done. And really what we think about is that the path to translation is a linear one. And it starts from the very beginning with basic research. That basic research is the discovery research we're talking about with respect to genomics. That then goes into sort of a treatment development phase that is again tightly controlled. We do very, very tightly controlled experiments to figure out if those treatments are efficacious. And it's only in stage four and stage five that we start to do anything that starts to bring these new things into the real world. Now when you talk about drug development, that can be two decades. So we could be thinking today that given if we're not doing any translation research that we could well be at least 20 years away from having any tangible benefits. I don't think any of us in this room or anybody in the genome project or in the institute believes that that's really the way we should be going. And that's actually reified by the way funding is done at NIH and most of our extramural colleagues, how they're deciding what their ideas are and how they're going to get them funded is this T1 to T4 sort of timeline. So what I'm going to ask you to do today is sort of shake that up a little bit and start to think about what if instead of that model, we started out with more of a problem based sensibility. What are the unmet needs, the big challenges that where genomics might have some extra traction that could help us make some maybe little B breakthroughs and anticipate that and direct our research in that way. So part of the reason that we want to do that is that there's been a lot of, I think, well justified concerns that if we're not careful, we're not going to wind up with optimal applications. So we hear a lot about this notion that we're going to be lost in translation, that we're going to wind up just with a pile of information that we really don't know how to apply. At the same time, all the optimism that's been well publicized about the promise of genomics is having this pull effect, which is bringing new and maybe untested new products into the market that people are concerned about. And those are sort of the personalized genomics, things that we hear so much about, the 23 and me. And you see a lot of the public discourse that says this is really of great concern. When I think about translation research, and I want to keep that notion, translation research, this isn't translation, this is research to guide translation. What I think we're trying to do is find something in the middle, the sweet spot, the optimal ways that we could do this. And that takes, in my mind, social and behavioral science to do that. So when we, I don't know if any of you are familiar with the pathway genomics controversy, which happened about three years ago. Pathways genomics was going to put their test kit, which was literally just a buckle scraping kit on the wall green shelves. And the FDA said, no, you're not. They pulled it off and said it was not something that was safe enough for the public. Now, mind you, this decision was based on no data. This decision was based on sort of over, sort of, I wouldn't even say over, a protectiveness, a sensibility that the public needs to be protected from this information until we have it figured out. So when we think about social and behavioral research a few years ago, the Genome Institute convened a large group of social and behavioral scientists to sort of get clearer about what do we mean. I mean, many of you are probably familiar with the ethical, legal, social implications or ELSI program. That's a component of social and behavioral research, but it isn't the full umbrella. Let's define a broader tent. What is in that? And what you see is that there are both disciplines and methods that come into play and doing this kind of science. And the disciplines are a bit staggering. Some of my friends, I think Andy in particular, sort of likens us to the plates coming on Mount Sinai of, here's the commandments. But anyway, so these are the disciplines that you can say really fall into the tent of social and behavioral research and they're quite broad. And likewise, the methodologies that get employed are also quite broad. They include both quantitative and qualitative methodologies. And I'm going to try to give you some examples of those. I've come from a background in behavioral epidemiology, as Andy pointed out, but have roots in social science as a sociologist. But you can see there's quite a few things. In addition, what this pursuit, this translation pursuit, I think we're at least social and behavioral sciences involved, is really inherently transdisciplinary. So this really requires us to have teams that don't necessarily speak the same language, even if we all sort of are roughly social scientists. And the challenges of doing that are large. And we have a range of outputs that we generate from these exercises. So everything from the basic knowledge that would be equivalent to discovery research, all the way through tools and methods, interventions that I'm going to talk a lot about today, all the way to public policy. So when we think about public health and what we're trying to do in public health, much of what we talk about are interventions. And so when I say intervention, I'm going to use that term a lot in this talk. And I mean a very, very broad swath of different approaches that we could do. We can, basically what we're trying to do is we're trying to influence a target group to some desired outcome. And in this case, in public health, obviously, it's a health outcome that we're looking for, or something related to a health outcome. And those things can range from informed decision making. We can focus on individual behavior change, or group behavior change. We can also look at changing how people think, because how people think we know has a lot to do with what they do. And then it can also, we can go all the way to the organizational level, or societal level in terms of trying to change public policies, that also really influence how people behave. So that's what I mean when I say intervention. And we can think about interventions following sort of a whole span here as well. So we can talk about primary prevention interventions, which are primarily targeted to healthy individuals to keep them healthy. And in the genetics, or genomics domain, what we see happening more often than not right now is sort of what we hear susceptibility testing, telling people that they're at higher, telling healthy people that they're at higher risk for disease as a means of intervening on them. Secondary prevention is to, is really focused on early detection among high risk groups. So in this case, in genomics, we're talking about high risk families. What we hear a lot about are families that are at high risk for inherited cancers, colon cancer, breast cancer, targeting those individuals to undergo screening and maybe a more frequent pattern so that we can detect the disease and when it's curable. Then we have tertiary interventions, which are really targeted to sick people, and in the, and trying to help them manage disease, manage their psychological well-being while they have those diseases. And in this case, in the genomics field, that really focuses on families that have diseases that are genetic influenced and trying to help them deal with those diseases. So also in terms of interventions, one of the things that's, I should say, social behavioral science and translation research that I think separates it from basics, maybe basic biological sciences, is that we rely on what we call conceptual models. So we pose our questions and how we think things are going to, how our interventions are going to influence a particular health outcome based on theories. And those that you can see here are smattering. This is really a subset of the number of theories that are out there. And these theories actually map to different levels of influence that are different levels of interventions that we're trying to deliver. So we have specific theories that are specific to when we're targeting just an individual. We're trying to change how an individual thinks or how an individual behaves. Likewise, if we're moving to an interpersonal level where we're trying to influence either small groups or dyads, we have theories that match to that, so on through organizational levels and societal levels. And those become the sort of cookbook, if you will, for how we're going to shape an intervention and how we're going to design the surveys and assessments to evaluate whether, in fact, we've done what we thought we did or if we didn't wind up affecting the outcome, why didn't we? We can unpack that. So when we think about research questions and hypotheses at a very simplistic level, we're thinking about some outcome, whether that's an individual or a group's behavior, an organizational change we're trying to influence or a policy. We think about mediating mechanisms, processes that we need to influence if we want to change that outcome or have a positive impact on that outcome. And those, again, can range from how people think, what skill sets groups or individuals have, and system incentives at an organizational level or, in the case of tobacco, changing tobacco taxes so that we get people to smoke less. And then we also consider background variables that influence those. Demographics of an individual, whether they live in poverty, and if it's an organization the size. And those we also can hypothesize that the background variables will influence whether or not our outcomes achieve. So that looks very simplistic, and I think roughly that's kind of what all of our models do. But when we look at an actual model of behavior change that we use to guide how we develop our assessments and how we develop our interventions, you can see here that it gets quite complicated quickly. So here are a wide array of background variables and mediating mechanisms. There may be multiple levels of mediating mechanisms that we have to consider to get to our outcome, which in this case is something that probably, to many of you in the audience, seems really simplistic. Something as simple as going out and seeking information about genetics, and yet what goes into play in sort of developing these assessments and interventions is actually quite complicated. So now I'm going to turn to thinking about public health and the principles of public health and what we do with public health interventions that I think is somewhat different than even the medical model of what we do in clinical practice. But in generally in public health, what modern problems that we have today are our chronic disease. And chronic disease, really if we're going to save money and lessen the negative health impact on the population, we're going to have to try to prevent these because it's way more costly and way more difficult to get to treat these conditions when someone has them. And when we think about preventing disease, we go right to health behaviors, which account for most of chronic diseases and there's like five health behaviors that do that, diet, exercise, smoking, sexual behaviors, I think. So these are the key underpinnings that most of our public health interventions are trying to influence. And when we think about these interventions, we want them to be what we call public health friendly or primary care friendly. That is that they need to be low in cost so that they can be widely disseminated and they need to have time, be time sensitive because primary care is a very fast-paced situation and health care providers don't have a lot of time. And the question then becomes as we think about translation, is there, what benefit does genomic information bring to add value to interventions that we're currently doing that meet some of these requirements because there is a fair amount of skepticism about new technology, not just genomics, any new technology, increasing costs when we implement it into these different settings. And the other thing that I think generates concerns, and again, doesn't matter what the technology is, is that we already have pretty serious disparities in health outcomes. And we certainly would want any new technology that comes in to exacerbate those disparities, make them worse. In the ideal, this technology could be brought to bear to decrease those disparities. And that becomes one of the challenges and one of the important research questions. But when we think about public health applications, unfortunately, when you think about that timeline that you saw before, the stage, the step one where we don't even test out our interventions in the real world until they're pretty far along and then we throw them out into the public. And what you see is that these very efficacious interventions are just, I have no real chance against what are real world challenges of trying to implement something that's been conducted under highly controlled circumstances. And that creates this tension, this ongoing tension that we face in public health between sort of this notion of having an intervention that's really efficacious, that really works, but doesn't have much reach versus an intervention that's not so effective. But has really broad reach and maybe meets other requirements in terms of cost savings or time savings, the things that are really of concern in terms of trying to put these into place. And so I think probably most of you would say, well, gee, wouldn't you just take the most efficacious intervention? And that's what you would use. That just makes the most sense. You want something that really, really works. Well, when you take sort of apply the public health principles to that, maybe you come up with a different conclusion. So I'm just using here the circumstance of hereditary non-polyposis, colorectal cancer genetic counseling. And what is the current standard of care for how that's done? And what you see here is that it's a high dose. Typically, these are two to three hour sessions with an individual or a family. They're resource intensive. They are usually conducted by a certified genetic counselor who does that more often than not in a face-to-face session. So they're relatively for a healthcare delivery system or for a public health department would be fairly difficult to sustain because most of these settings don't have genetic counselors. There isn't consistent reimbursement for these services. And so they're expensive. But when we look at the literature, they work. They're really efficacious. People come out, they can make informed decisions. They feel more knowledgeable than they went in. So it's a highly efficacious intervention. When we look at a public health approach by comparison, what do we see in a standard public health intervention that you would see conducted in any kind of public health setting or even in a primary care setting? What you see is that most interventions, let's say a smoking cessation intervention, that's for an addictive behavior, is a very low dose intervention, less than an hour typically, usually about 10 or 15 minutes. You have, they're usually implemented by whatever staff they have available in these settings, which can be and often isn't a nurse. It's a health educator-like person or a nursing assistant. Most of the kinds of serialized contact that is required would happen via the telephone, the mail and more and more the internet. And the sustainability of this is that you have to use what you have. You have to use the infrastructure you have, and it has to be cheap. And so effectiveness here is the goal because if you can achieve those objectives, you're going to have much broader reach. So when you think about that and you cost it out, this efficacy effectiveness trade-off, and you say, well, the current approach has an efficacy of 80%. 80% of the people come out better in terms of whatever the desired outcome is than 80% come out better. The reach, though, is very, very limited. You can only reach those people in high-risk clinics who have insurance, who can get these resources. And when you look at that, the effectiveness quotient of that is about 0.08. Now think about the public health model, the effectiveness, where if you think about the intervention being really only 20% of the people come out doing really a lot better than when they came in. But the reach is 50% because it's much broader, much more easily implementable. You get a higher overall effectiveness ratio. And that has been the underlying principle around thinking that low-impact, high-reach interventions are really the way that we should be thinking about public health outcomes. And so that's the model that we need to bring genomics into in talking with our public health colleagues and our policymakers about these kinds of new interventions. So now I'm going to turn to some examples of research that I think embodies some of both the methods and the principles of public health and sort of walk you through some of those and we do that in sort of three areas. First is public understanding and use of genomic information. There's been a lot, a lot, a lot of concerns about public isn't going to get this, they're going to misuse it, it's going to be really bad for them. And I'm going to talk a little bit about risk communications as a way to motivate behavior change and sort of the evolution of that thought process and where we are with that. And then lastly, I'm just going to talk just to sort of tantalize you with some future areas of new intervention approaches that we might use in the near future. So one of the things that I think translation research could be really, really great at is we see a lot of hyperbole and rhetoric in the literature about genomics and what it's going to do and how it's going to be harmful and how it's going to be beneficial. And I think we can bring data to those debates and we haven't always but this one was one that Andy was referring to with our multiplex study. This was started in about 2005, so there was no 23andMe, there was no multiplex test, but there was a lot of concern that it was coming, that premature translation sort of foreshadowing. And so what the big concern about that was is that the FDA should not allow this, that the public was going to exaggerate this, they were going to bypass the healthcare delivery system, they were going to do things that were really, really bad for them. Or there was going to be this, stampede to healthcare delivery because all of these folks were going to be fretting about the fact that they'd gotten these results, they were going to overuse, misuse health services. So this was a major problem. And so we launched the multiplex initiative to address these very fundamental questions, not to influence any behavior change, but simply to observe if this kind of test was offered and it was done in the context of a primary care delivery system to healthy people, what would happen. And so we had to develop our own test. I'm not going to walk you through that. This was before GWAS, so this was a sort of a consensus process of experts weighing in on where the evidence was strongest for preventable health conditions. We wanted them to be, any feedback that an individual got, there were things they could do to make it better. We didn't want any genes that were pleotropic for things that they couldn't do to make better. So cardiovascular genes that were also associated with Alzheimer's, we didn't want to do that because we didn't want to raise those kinds of concerns. So this was purely a prevention-oriented genetic test. And what you see here are the eight health conditions and the 15 genes that met the quality review of our expert panels. And this is all described in this article if you want to read more about it. Since GWAS has come out, all of these genes have been sort of determined or gene variants have been determined to meet GWAS criteria. So they still sustain sort of credibility, if you will, in terms of giving feedback to an individual. And one of the unique things about this is that individuals are going to get results on multiple, everybody's going to have some genetic variant that's associated with increased risk for one of these health conditions. And you can see here the curve that on it, most of them are going to get at least nine variants, nine of the 15 variants that they're at risk for. Nobody's going to get zero, nobody's going to get 15. So they're going to get a full range of results. And this is the way the results look. We had to go through the process of developing risk communications so that individuals could decide if they even wanted this test. And I'll show you the study design in a minute. But this is what it looks like and this is why it could be, one could have some skepticism about the impact of this, both positive or negative. So in this case, this is the KCMJ11 gene that's associated with type 2 diabetes. And there are three different what we call versions of risk that an individual could have. They could have zero risk versions and those 35 people out of 100 would get type 2 diabetes. If they had one risk version, that goes up to 37 so that's only two more people more. And if they have two risk versions, it goes up to 43. And this is a standard, this is an evidence-based method for conveying this kind of risk information. Very, very small risks. And then we also gave them the frequency of falling into one of these categories. So 65% of people have zero versions, 29% have one and so forth. And this was to give them, and we did this for each of the 15 genes. And we also gave them information about each of the diseases and then asked them if they wanted to undergo testing. So just to give you a little perspective on this compared to behavioral risk factors, this is taken from a current study that's ongoing where they're using FTO gene risk feedback that's related to obesity with enrollees in the Corielle healthcare program. And what they show here is this is the genetic risk if you have a variant of FTO, a risk variant of FTO, which is 1.3, a 30% increased risk. But if you sit on the couch more than, I think it's 40 hours a week watching TV, you have a 1.9% risk. So again, these are, these are thinking about how populations might respond to this. It's important to see how these risks compare to other, other things that they're doing every day. Maybe not as scary as we all think. So here's the study design. This was at NCI. We worked with the NCI funded cancer research network which I think has 15 different sites around the country. We worked with Henry Ford healthcare system because it had a demographically diverse patient base and they were a very willing partner. And the group health cooperative which is another one of those was the survey site so they conducted all the surveys. These were healthy adults who were enrolled at Henry Ford healthcare system in Detroit. They were ages 25 to 40. And they couldn't have any of the health, 15 of the eight health conditions that were on the test. This is the feedback that we had to develop among those who opted to be tested. They received in the mail. So this is again, in keeping with the public health approach, the, they received their test results in the mail that were personalized that said this is what your results are. And, and then a lot of caveats about this doesn't mean you're going to get the disease and so forth. And then we also had, this was all in a folder, we also made sure that we addressed that to try to avert any genetic determinism sort of conclusions, all the different things that influence their risk that weren't genetic. So here's the study flow. We identified from the patient enrollees so that we could have a known denominator. All of those who potential participants that met our age criteria and through the through the medical records we could see that they didn't have a diagnosis of some of these diseases. We conducted a baseline survey with that group. I think it was about 6,000 individuals that we surveyed. We offered to about I think it's 1,959 individuals access to the website to come online to consider if they wanted this free genetic testing. They came online if they, if they went through the three modules and they decided, yes, we want to be tested then they had to schedule a clinic visit to have a blood draw and then they completed some more informed consent process to finally decide that they wanted to be tested. And then about three months later they got their test results. This was when it was much moving much slower again by mail and within one week we scheduled a telephone call with a research educator who talked to them about their test results basically just trying to make sure that they understood what the results were telling them. But more importantly understood what the results weren't telling them. And then we did a three month follow-up survey. So when we think about just this, this was the baseline data in terms of the concerns that we have about genetic determinism that if people don't understand it the public doesn't understand it and so we need to be concerned about this and this is a baseline before they've had any education about genetic testing and what you can see here is they were asked two questions. They were asked how much do you think what you do, your behavior contributes to each of these eight health conditions and a separate question asking how much do you think genetics contributes to each of these health conditions and what you can see is that across the board behavior trumps genetics and across the board we'll see that genetics is considered to be less influential but this isn't just a sort of there's a nuance sort of understanding of this if you look at say lung cancer for example there's a much bigger difference between endorsement of behavioral contributors to lung cancer because of our major public health campaigns on tobacco use and much lower endorsement of genetics. So the idea that the public doesn't the public is susceptible to genetic determinism doesn't look like this data would suggest that's as big a concern as what the public discourse would suggest. Also again bringing sort of data to these concerns about there's going to be the stampede for this testing it's expensive to do everybody's going to want it and it's going to have that's going to have a negative effect and what you can see here is of the 1959 people who we offered the genetic testing to only 612 actually went to the website to consider it fewer wanted and actually had their blood drawn to have the test done and that constitutes if you look at this as the denominator this is about 14% of those that we touched got tested. So again I would say I mean I have said in audiences like this well that doesn't seem like a lot to me that doesn't seem like a stampede but again if you think about this at the public health level 14% of the healthy population wanting genetic testing would quickly overwhelm our capacity to do that so again you have to ramp this up to the public health level. We also looked at factors about how what we gave them this intervention this set of materials this website how much how that influenced whether they got tested or how easy it was to make the decision so you assume if the decision is easier to make and this is regardless of whether they decided to be tested or not one of the things that we're trying to do is help people make a decision and the easier it is for them to make it maybe that's a success and what you see is that the more pages that they read online the more likely they were to decide the test but even if they didn't decide to test they found the decision to be much easier and then you see things like you would see in our mediating mechanisms model genetic self-efficacy that's confidence to understand genetics was also associated with desire to get tested and with how easy they found the decision and thinking it's really important to learn more about genetics also was associated with their decisions so again back to this this rhetoric about what's going to happen one of the again the stampede of health care delivery we asked people if they intended to talk to their doctor now I want to back up for a minute and say this this testing was done completely separate from the health care delivery system so the doctors were completely looped out of this so that we would have no no problems around privacy of this information finding its way into the medical record because we didn't we don't know that it has any medical benefit at all so this was really up to the individual did they want to have who would they talk to about it and as you can see most of them would talk to their immediate family and actually more wanted to talk to their friends about it or plan to talk to their friends about it then to their health care provider so again not seeing this sort of sense that they're going to be running to their health care provider and asking for other tests or asking for inappropriate services the benefit of working in a health care delivery system is that we actually had automated records of use of health health care and so what you see here is the is the time before the genetic testing was offered and the quarters after the genetic testing was offered and we have the four three groups here the multi the folks the solid line of the folks that got tested who underwent the multiplex testing the the dotted line is those who went on the web and considered testing but didn't didn't get tested and then this broken line here is just those folks who completed the baseline only so what you can see is that the individuals who got tested were higher utilizers before they got tested and they stayed about the same so there was no impact of this testing on their use patterns and so again no data to support that somehow this was going to prompt some sort of over utilization of healthcare services so I'm going to turn now to thinking about motivation and using behavioral risk or genomic risk information to motivate behavior change which has been again something that you see a lot in the literature is that if people understand that they're at heightened risk for these diseases they're certainly going to change their health behaviors and that's really based on on this idea that has some conceptual support which is that if we can increase individuals feelings of susceptibility to risk and by personalizing whatever risk communication we're giving we're also increasing the relevance of that risk to that individual that that's going to heighten that's going to be more appealing and catch their attention in a different way at the same time though we have to balance this with and this is the other set of concerns you see in the literature is that if we start talking about genetics then people are going to think it's not controllable and that's going to diminish their motivation or their confidence that they can change and that there's a whole bunch of factors that we have to consider again these are the background factors that I showed you in that model of things that people bring to this sets of assumptions that they bring to whatever risk communication is being given to them that have to be considered but it all focalized focalizes and then there's context obviously is this being done in a healthcare system is this being done when somebody's been diagnosed with cancer is this being done just at a local church that you're getting your information and all of those will make a difference in terms of how that information is received but the centerpiece of all of this is motivation and that the assumption underlying this is that people are going to present not motivated and so we have room to move them and I'm going to make the case that that hasn't been what we've been seeing so this work started actually 1997 Karen Lairman and colleagues at Georgetown here started with smoking cessation and giving cigarette smokers information about their susceptibility to lung cancer as a means to get them to engage in smoking cessation efforts and the first trial here showed absolutely no effect this was using the SIP 2D6 so that basically everybody fell if you were in in the genetic testing group virtually everybody had the high risk category which is somewhat different and as we move along and you can see here that there was no effect of these different different kinds of the genetic feedback is the blue no effect of that on smoking cessation rates and then a few years later a colleagues and I at Duke did a similar trial where we used a different marker the GSTM which is involved in metabolizing the carcinogens and tobacco smoke and here you have about 50% of the population who are good metabolizers and 50% who aren't good metabolizers and what you can see here is when you compare the individuals who got the genetic feedback based on whether they got the high risk result which is the blue the dark blue and the or the low risk result again no effect on smoking cessation outcomes and this these two trials have been used and several other studies that have shown not all targeting tobacco use some some with weight management that are finding similar things that this genetic risk communications are just not influencing behavior change and there's been two reviews most recent being a Cochrane review that basically says this ain't gonna work folks roll up the sidewalk and let's go home and I think that's probably premature probably not something because these studies and most of the studies in this field are flawed they're flawed in a number of ways one they're usually people who are highly motivated to make the behavior change who show up for these studies to most of them rely on a single marker that just ain't the way it is anymore and they're also tend to be self-selected populations so problems in terms of generalizability so lots of weaknesses in these study designs and most of the study designs actually weren't ours being one of them really weren't designed to answer the question does this kind of respect feedback actually change behaviors and so we launched a study when I first got here to the genome institute looking more at context and thinking about could we draw in smokers who weren't particularly motivated to quit and in fact not even ask them to quit smoking just say we want to give you this information and we want to see what you do with it and so we targeted the context of a family member having a diagnosis of lung cancer and that's often where genetic testing is talked about so these were all individuals at Moffitt Cancer Center in Tampa who's had a blood relative who had who was diagnosed with a late stage lung cancer most lung stains lung cancers are late stage and we offered them genetic testing if they wanted it and what we see again here in this context and these are younger smokers who are at least our goal was to reach younger smokers regionally distributed throughout the country because they didn't have to live near the cancer patient and what you see here is 118 showed up and said they would consider genetic testing at this time point and actually I think virtually everybody who logged on opted to get tested we didn't get that nice 50 a nice effect of having some people consider and then leave but what you can see here is that at baseline on a seven point scale these folks were highly motivated to quit smoking so again the context was drawing in individuals who were already motivated to quit smoking and what is risk communication supposed to do if it's supposed to motivate and you're already motivated and somebody says hey I'm really motivated to quit smoking and then you say yeah you're a really high risk that's not particularly helpful it's like help me quit smoking so probably the wrong target audiences have been in basically in most of these studies so how do we and when you look here we offered them so we didn't this wasn't a smoking cessation study we told them that up front because we wanted to draw on these people who weren't particularly motivated we didn't succeed in doing that but we did at the end say hey you know if you'd like some free assistance at quitting here's some different options and we gave them nicotine replacement therapy for free we gave them self-help books we offered them six telephone counseling calls these are evidence-based shown to be effective ways to help people quit smoking and what you can see here is by test result no difference in uptake of these lower sort of effort intervention approaches slight different but not significant uptake those who were in the higher-risk category taking up the genetic the telephone counseling which really requires a lot more commitment so maybe getting that high-risk message made them take a little bit more effort towards quitting but not it's not clear so when we think about okay let's back up let's reevaluate this obviously this sort of putting out our store front and saying genetic testing you know this could be helpful to motivate you to make a behavior change isn't working so what is the optimal place to think about and so I'm going to take you back to this public health model again which is that you know we have this idea of intervention dose and what we're trying to do honestly is to get as far down here as we can because that's where it's going to be cheaper and we're also looking at motivation and people over here probably don't need as much perhaps as people who are down here so what's the sweet spot in terms of finding where to target genomic information what's the population that we should be thinking about when we think about doing this and we turn to college smokers so college smoking smokers being their transition to that first freshman in college at first year are at high risk for adopting risk behaviors to at high risk for any experimentation they've done with tobacco becoming entrenched as a habit and this is amazingly though college students are in most social science studies they're not targeted for healthy behavior change and so this this is a group that actually is in need and some of our previous work in historically black colleges was showing that these individuals were very interested in genomic testing in genetic testing but they were only interested in it if they thought they would be found to be at low risk so there's a nice possibility for bias that we probably would want to overcome in in our interventions and there's lots of conceptual models mapped to why this is a good target group to to explore so we we started a study that was funded by the National Cancer Institute and Isaac Lipkus is the PI at Duke and our baseline survey we asked smoker we asked these young adults freshman in college and it was about four colleges half historically black half UNC type system schools and what they're sort of senses were about their desire to quit smoking which as you see here much lower than those sixes and five point eights that we saw in our in our lung cancer study they have by they have misperceptions about how easy it is going to be for them to quit they think that they can probably quit any time and they don't they think that the harms that they're going to experience from tobacco are far afield so this looks like this could be a really good group to target and in some of the qualitative we did some focus groups with these folks showing them the genetic information and asking them sort of what they how influential they thought this would be what we found was some really interesting ways in which our public health messages have backfired and this may be another avenue for how genomics could be beneficial for helping people understand disease risk but what you can see here is that most of these freshmen or a good group of these freshmen thought you know this is tobacco nicotine is really helpful for studying it's really helpful socially to fit in and you know all the public health messages say you know you can quit and recover from the harms of smoking if you quit young enough so I'm going to smoke through college and then I'm going to quit and you hear this about among young women I'm going to quit I'm going to smoke until I get pregnant and then I'm going to quit and and these folks underestimate the powers of addiction and so what you see is that they and that they also think that that they don't have to worry too much about it because there was also that they're that they also didn't understand that less exposure mean when you are susceptible less exposure is going to have a big impact so this is also reflecting that misunderstanding that genetic susceptibility might make it worse for you to keep smoking younger even though for some people they can recover so these two kinds of this G by E sensibility not there and there's lots of researchers who have shown that genetics and behavior are really on separate tracks how we think about how the public thinks about how that influences health are on separate tracks and we see this played out here and that they also underestimate the potential for addiction and it may be rather than talking about risk communication we should be talking about our you know disease risk communication we should be talking about addiction propensity communication that that's really where we're going to get them and so we are in the midst of that trial I will have results probably the next time I speak in this course but in the meantime we've been thinking about next steps of being sort of do we do we use sort of new technologies and and the internet blogging and so forth to have young adults as a sort of environment for young adults to learn about G by E and to think about the genetics of addiction as a next step so now we're moving even younger so if we think about let's say we can influence we we're influencing healthy young adults who aren't particularly motivated to change their behavior what if we tried to influence I mean public health is all about primary prevention that's where the cost saving is that's where the health health outcome bang for the buck is why shouldn't we be targeting kids and you know kids and genetics very very sensitive topic the idea that we would do genetic testing on kids and give feedback is something that again there's been a huge amount of public or not public scientific concern by and large this isn't a good idea and part of that is that we think that if we give information obviously if we test a kid we're going to give the information to their parents and parents have a major effect on how kids live their lives and their behaviors and health habits they have how is that going to affect parenting practices and I think there's lots of concerns particularly in the area of eating the parents will become overly restrictive of the diet of a kid's diet if they think the child is at high risk genetically that misunderstandings about you know that being a deterministic kind of finding will influence their parenting practices and yet at the same time and the reality is that this risk is really quite big family history of obesity is a much bigger predictor than our genetic snips and genetic variants are what you can see here is that if you have if the child has one parent who's overweight there's about a doubling of risk and if both of the parents are at risk it's almost a quadrupling of risk so the idea then it does sort of make you think gee these are going to be pretty strong messages that a parent's going to get and is that going to then perhaps not result in the outcome we want if parents over restrict their kid's eating habits we know that backfires and that those kids may become more food focused and it may make the problem worse so here's a big challenge when you think about translation research not probably advisable to go and test kids maybe not even advisable to actually do a clinical clinically implemented family history because unless you've got a huge infrastructure to support that parent once they leave so how could we do this well we we've employed a technology of immersive virtual environment testing so that we can start to ask some of these questions before these things are actually clinically available and and so this helps us be nimble we can change with the changing nature of genomic technology we can anticipate future situations and and envision them and then try to operationalize them for an experiment and then some of these concepts are very very complicated and would be difficult to control in real world settings and really sort out what the effect of the genetic information was above and beyond what else was happening in the clinical setting and so we've been using this immersive virtual environment testing and I'll show you we've developed one in our social behavioral research branch and you'd be more than welcome to come and do a tour if you'd like to but it it what is typically being used in the field right now are hypothetical vignettes asking people to imagine if you got this information what would you do we know there's huge gaps between what people think they would do in a situation and what they actually do we can get rigorous behavioral outcomes so most of what the research we do we have to rely on people telling us so a self-report there's a lot of error in self-report here we can observe what a person's doing in that setting and we can measure it and and then when we're talking about eating patterns you know preparing food and having it stay not stay fresh and and not poison our participants is also a major challenge so here is the immersive virtual environment testing area the subject is wearing a head mounted is looking into a head mounted display in which this woman is seeing the world we've constructed her movements are being tracked through that antenna by these cameras that are in the room so you see nothing is in the room but what she's seeing is a buffet a food buffet in which she has food options to choose from to fill a plate for her five-year-old child and what we've done is we've randomly our question was if we gave mothers who were overweight information about their child this family history information either one parent is overweight or two parents are overweight would that result in her restricting the foods that she selected for her five to six-year-old child and we would know that because we'd be able to weigh and measure everything that she put on the plate and then we would also be able to see how the family history affected that so here is now an experimental design study quantitative study where we we did some advertisements and identified potential people they came online they decided we decided if they were eligible or not they came to the iVita lab and did a screening they practiced the buffet because we didn't want learning effects to influence how they performed in the buffet and then they were randomized to one of three conditions a food safety condition just learning about how to not have food spoil or get food poisoning for children behavioral risk information talking about all the behaviors that influence overweight and obesity in childhood and then this third group got this behavioral risk information plus a family history assessment of their child now all of the mothers were selected to be overweight or obese so all of the kids would get at least a one parent feedback that they were at increased risk and then some had two biological parents who were overweight and they would get the two message and then we did post surveys after they got that information but before they went into the buffet to fill their plate for the child and then we did the post buffet survey so here again is we had sugared beverages which is a major contributor to overweight and children and you can see that this looks very realistic and the participants rated as very realistic they feel very immersed in this world so what did we find? so what we found here is the outcome is total calories on the plate what we have is the what the main finding was that in the behavioral risk arm we see the mothers filled the plate with 35 fewer calories than the food safety group but that was not significantly different from the food safety group what was significantly different from the food safety group was the family history feedback those mothers filled their plate the child's plate with 45 fewer calories now 45 calories per meal adds up and that can be a major influence on we had a dietitian involved in this on calories per day that could result in weight loss so it did look like mothers were restricting if they got the family history information and restricting whether that would be harmful or beneficial it really is a judgment call when we looked just amongst those though this is really really fascinating when we looked just among those who were in the family history group what we see is that the mothers who got the two overweight parent feedback actually filled their child's plate with 71 more calories so restriction was happening only in the mothers that got the one risk message and it was the mothers weight that was influencing the child's obesity risk so what you see here is that is that these groups are all the same so what was happening is in the two and these groups are the same so it's just the one parent in the behavioral risk and family history group mom is the one mom is the cause of this child's obesity risk and she restricted the calories we're exploring whether or not that's a guilt effect or responsibility effect and how would you harness that guilt or responsibility effect to actually have a positive benefit to the child so we immediately go to the oh my god that's harmful we can't do it rather than thinking from a social behavioral science standpoint how do we use that information to motivate the desired behavior and the idea that mothers will do things for their children that they won't do for themselves could be a way of influencing more than just the child oh and we saw the same effect I'm sorry we saw the same effect for sweet and beverages this is total calories and saw the same effect they were restricting sweet and beverages as well oh sorry so now I want to say I want to turn this to a global health applications because honestly the chronic disease problem is worldwide and if we're going to have traction in public health we're going to have to bring these things to bear in the lower socioeconomic countries and part of that also is where disparities are really quite great so when we think about bringing could we bring genomics to benefit disparities perhaps global health is one arena to do that so we've been working in Ethiopia and rural Ethiopia on a condition called podoconiosis which is a form of elephantiasis that is caused by an inflammatory lymphatic response to exposure to the soil the soil is in Highlands Ethiopia has a lot of mica in it it shines this mineral base crosses the skin barrier gets into the lymphatic system and causes this pretty horrific swelling this is a young child so it's not quite as severe but these feet look are like sort of Frankenstein type feet and this is preventable entirely if the foot is protected from soil exposure but of course there's a genetic aspect to this that some there's a heightened it there's a heightened risk among a select group of folks who have I think recently the gene has been identified as in this inflammatory pathway that they are going to benefit much more from wearing footwear than others although there's many other ways in which footwear can be beneficial one of the challenges in these kinds of resource poor settings is that in this case 50 percent of the population are under the age of 15 the infrastructure public health infrastructure in Ethiopia is virtually absent and so the idea that even if there are shoes available and companies like Tom's shoes and others have offered to provide shoes to population I think you're forgetting your jacket I think you're forgetting your jacket anyway we'll auction it off sorry about that but we digress so the population so distributing shoes to all of these kids would be extremely difficult and so the idea of targeting high risk again which could be a public health strategy to those who are genetically at risk could have benefit for at least reducing the burden of this particular condition on the system so we've been exploring can you target these shoes to genetically high-risk families without causing other social problems or exacerbating other social problems and so we started with some preliminary about I guess four or five years ago qualitative focus group study with four different sites and we selected the sites so that they would reflect different numbers of individuals in that community that had podoconiosis and there's an NGO that's been distributing shoes the Mossy Foot Foundation that we wanted to have the communities vary on the length of duration of that relationship and then the distance also because many of these communities are quite remote and so we wanted to make sure that we had fair representation of that and so we really did overkill in terms of focus groups most focus group studies you see will have very few participants maybe 20 people and we had 28 focus groups 38 individual interviews and we talked to 307 participants and these were adults and children that we asked about their attitudes and beliefs about what causes podoconiosis because the mental models that they bring the explanations that they have for how that causes what causes that disease are likely going to influence what they think the solutions are to those diseases do shoes make sense should I be wearing shoes and so what you see here is what we found in this and this is just a quick summary what we found in this qualitative work is that a sizable group did believe and acknowledge that heredity played a role in the development of podoconiosis and those individuals perceived it to be less important to do preventive behaviors like wearing shoes because the sense is it couldn't be prevented so why bother and those individuals also tended to endorse more interpersonal stigmatizing behavior they didn't want to for example marry into the family and they didn't want to work closely with or eat with or socialize with individuals who had the condition which is all indicators of stigma and we don't want to we don't want genomic information to exacerbate stigma but there was also a sizable group of people who perceived that it was not heredity and in those those individuals were much more likely to endorse wearing shoes that made sense because it could be prevented they were more empathic to people who had the condition but they thought it was contagious and so they were fearful of being again they wanted more social distance they wanted they wasn't so much partner selection but they just didn't want to be around these individuals and individuals so this is this is individuals who are unaffected by the condition but it also includes individuals who were affected so both of these folks had these beliefs and for those who were affected by the condition some of these beliefs affected how they felt about themselves and whether they could fit in socially so we use that to develop a questionnaire that we implemented with six communities and what we did was and I'll show you this trial and this is the step first step in our community-based intervention trial is we had about 600 individuals who we knew had podoconiosis in their family and then we took two neighbors who lived close to these individuals so we got about 1200 unaffected neighbors and we asked them these questions about the extent to which they thought podoconiosis was inherited it was contagious they could prevent it they couldn't prevent it they could they couldn't prevent it they could prevent it and it was spread by wearing other people's shoes because there's a lot of shoe exchange because this is a resource poor so when you grow out of a shoe you would give it to someone else what you can see here is the yellow bars are the affected individuals they had podoconiosis in their family they're much less likely to endorse these kinds of things more likely to believe it can be prevented they've been involved with this Mossy Foot Foundation for some time so they've been getting lots of education about prevention and less likely to have sort of misconceptions about how the disease is conveyed look at the neighbors the neighbors are all endorsing really a lot of misinformation so any intervention that we went into where we focused just on the high-risk individuals and gave them this genetic information would likely flop and maybe only make these folks feel worse about their situation not better and so when we also look at the experiences of stigma or the willingness to stigma stigmatize these individuals what you see here is that these are the affected families so these are the folks saying I feel ashamed I feel inferior people distance me don't want to be around me is equal to what you see in terms of the unaffected families saying that they're willing to a hypothetical family they don't want a family living next door who has POTO they don't want to share meals with a family they don't want to help them treat their feet because they think it could be contagious and they certainly don't want to marry into the family so again an intervention that targets the neighbors is obviously going to be really important and so here is the trial that is currently underway in six communities this is the baseline assessment I just showed you some of the data from that those individuals then again this is a quasi-experimental study so now we're not randomizing we can't really randomize communities doesn't work that way we now assign communities to a condition we often stratify them based on characteristics so we have characteristics that are important like distance from the NGO and size and so forth we stratify on those variables and then we assign them to one of three groups so we have two communities in each of these groups our comparison group doesn't get anything other than what the Massey Foot Foundation is currently doing which is distributing free shoes to young children the affected household neighbors get nothing the standardized health education looks very much like a standard public health intervention I'll show you in a minute some examples posters, campaigns raising public awareness getting people who are who are leaders in the in the community to make public statements about the importance of wearing shoes and both affected and unaffected households are exposed to that campaign and the kids in the affected families get free shoes this last condition is asking this question does genomic information gene by environment interaction education add value to what is a standard public health intervention so here we have the same thing they get the public health campaign but both the affected and unaffected households get education about G by E now this is challenging because education about G by E in a largely illiterate population means no words only pictures only metaphors and so I'm going to talk to you about that in a minute so here's sort of the public health campaign the communities helped us develop posters that showed kids and all the different activities that they do and why they should be wearing shoes we had kids create a song that is played we drive around in these vans and play the song and bring people to the market and talk about the importance of wearing shoes and have important role models wearing shoes and so forth standard public health intervention now we have to think about the G by E education and here we turn to a science of metaphors believe it or not there is a lot of thinking about how do we take what the public already thinks and integrate new information new knowledge into that in ways that they make it more likely for them to understand and importantly less likely for them to misunderstand what it is we're trying to tell them so here the metaphor we use this is the Highland Ethiopia no trees lots of sun beating down and what you see here is that this is depicting that individuals and we have a lay health advisor who's teaching about this so in their name in their language what will attenya they're talking about this relationship of some people stand in the bare sun doesn't bother them at all some people bothers them a little bit but doesn't really that some people it bothers a lot and they have such a reaction to it they want to and you see this walking down the streets because everybody's walking to market there's very little car use having headbands wrapped around them or umbrellas showing a sensitivity to the sun there's no stigma to that there's no negative feelings about that it's just a biological difference so likewise that's the same thing with soil is that some people have everybody's getting exposed to the soil but some people because of this family heredity genetics react to it differently and for those people it's particularly important to wear shoes all right does this work and as we do this in the households we do an initial session and a booster session three months later and we and we will tell you later we're right now in the process of collecting the last bit of data on that and so again two years from now I'll give you the answers to that sorry so now lastly I'm going to turn to talking about this where we might be going in the future and I'm going to go through this a little bit more quickly that I did the others so that we stay on time and I give you some time to ask some questions but one of the big problems is that we our public health interventions have had have had pretty much the same message of you know eat five fruits and vegetables a day maintain a healthy weight don't smoke those sorts of things and the public is actually quite familiar with that but yet what we see in terms of what they're actually doing we don't see adherence being very high to most of these behavioral recommendations and that's been identified as a problem and in fact in some of these behaviors we even see sort of a trend that is very concerning they're not going in the direction we'd like them to go we also look at our behavior change interventions this is a meta analysis of behavior change interventions that have been targeted to low income populations which is where the prevalence of these health conditions these behaviors is highest and what you see is that these interventions are as likely to be ineffective as they are to be effective so we've got a sort of glass half full glass half empty kind of situation going on but if we focus on the glass half empty we should probably be thinking about some novel ways that we might engage the public into into these interventions we also have a pretty significant problem amongst those that we get to change behaviors I know this is sounding so negative but the people that we get to change behaviors actually tend to relapse pretty soon after the intervention is over this is in smoking pretty common curve we see it a lot in weight we see the drop as part of the intervention and then the return to the weight and sometimes they even gain more weight than they had before so we've got some problems here when we think about the randomized controlled trials I think we can get some information about that basically from the people who drop out so what you have here is a standard randomized controlled trial comparing basically it's a horse race design you've got these different kinds of diets pitted against each other and who does best and you start out with recruiting people and you randomize them to condition and you lose some people along the way you lose people because they you lose about 20 about a third by the first six months and half by the end of the trial so at the end of the trial you have the most motivated of the most motivated that have stayed in this study and even in those we see a very very modest weight loss usually less than 10 pounds about 10 pounds so why do people drop out? well they drop out because you know they're dissatisfied with randomization they had an idea what was going to work and they didn't get it but a lot of them drop out because they say they can't tolerate the diet their the pace of their weight loss is not going at the speed that they'd like and they just can't do it it's just too hard and we've pretty much ignored that and said you know it's mind over matter you know it's it's energy in energy out stupid just you know just restrict your calories and exercise more but what that's ignoring and we're learning more and more about this in the genetic field is that this is ignoring the widespread well documented increasingly well documented individual differences and that could explain why people are saying this individual differences in response to calorie restriction mood pace of weight loss ability to control one's appetite eating when they're not hungry likewise in exercise mood response to physical activity not everybody gets the endorphin rush thermoregulation overheating muscle adaptation not even getting muscle mass that is increasing metabolism in the way that you would want so there's quite a bit of differences and we're also learning from epigenetic finding that the environment that the individual lives in is probably also influencing their ability to comply with some of these things so why are we not paying any attention to that and I think that's really where we should be going in the future and what you see here is just again I'm going to hammer this conceptual model idea I'm sorry but now we need to start to think about bio behavioral conceptual models whereas before most of our models have focused a lot on what's going on in people's heads or what's going on interpersonally or in an organization now we're thinking about what's going on biologically that might be influencing so these differences in acute response to exercise also influence how somebody perceives what they're doing and I'll show you data on that in a minute and that in turn influences whether they're motivated or not to exercise which influences the likelihood that they will exercise and we learn more and more that exercise behavior also influences epigenetic changes that can be beneficial to health we're not currently measuring and that exercise also increases physical fitness which also influences so it's an iterative kind of thing that we're largely ignoring and I'm just going to show you some data here these are just gene these are just data and this is one study but there I could show you five or six for each of these usually small studies where they're looking at is there any evidence to support genetic variation accounting for response to interventions this is a Japanese study of middle aged Japanese women based on the beta-3 and renergic receptor gene has involvement in metabolism and fat cell development and what you see here is with the common variant you see this nice correlation between if you restrict your calories you're going to lose more weight so more weight fortunately this is kind of inversely but this is the weight loss higher weight loss and more calorie restriction and likewise you hear more steps and more weight loss so you see this nice nice correlation flat for those who about a third of the population fall into this risk variant group flat no association so these were tightly controlled experimental settings where these folks were exercising you don't see them getting the benefit likewise the model I showed you was developed by Angela Bryan and her colleagues and she's done some work again under controlled circumstances where someone's on an exercise bike so you can measure a lot of the parameters of this and what she's showing again this is the brain-derived neurotrophic factor which shows sort of how has an involvement in how people respond to exercise in terms of mood and what you see here is that individuals a positive affect in the GG in the AGAA is looking much more straight line-ish so these folks are getting more positive affect with more minutes of exercise whereas the GG's it's kind of leveling off and if you put that together with what they're perceiving in terms of their exertion that GG group is working a lot harder for less benefit so there is some plausibility to these assertions so I think what's really exciting about this and again I don't use these studies as sort of examples of this is the best science out there I use these as really hypothesis generators what you see here is a neutral genomics neutral genomics incredibly controversial and I'm not saying that we're necessarily ready to do that yet but in this case this is a Greek study where the individuals who were involved in the trial were those who had consistently failed at weight loss efforts and what they did is based on 19 genes they customized dietary recommendations to those 19 genes and it was really the customization that they did you know is somewhat questionable but what you see is really interesting is that this customization of diet compared to a control group that didn't get this had no effect in the short run so you could argue that the motivation that someone's bringing in to any initial weight loss or behavior change trumps anything it trumps risk communication it trumps everything but what happens is as the diet as you move out further where you saw those terrible relapse curves that's where you see the benefit of having this tailored diet so maybe it isn't going to help us get people to change behaviors in the first place it's going to help us keep people changed after they've changed and so that is another area I think a promise that we should consider so last but not least thank you for hanging in there um translation research I hope I've convinced you that this is important and that we should be doing it now and that we can be doing it now and we should be doing more of it I think that you should have seen lots of opportunities and avenues and these are just a sub-sample of what I find especially intriguing there's lots of other areas where we could be doing this work but it's going to be important that we go in with hypotheses that we go in with models what are we trying to do here rather than just taking the latest technology and throwing it at something and I think that unfortunately we've done way too much of that and we need to stop it I think also you've seen that we have a lot of different intervention or a lot of different methodologies that we can use quantitative, qualitative and bring that to these questions and in one study we may need to use a lot of different methods in order to get there and lastly this work is inherently interdisciplinary and I hope that those of you who happen to be geneticists or basic biologists maybe this will warm you up a little too when the next social and behavioral scientist approaches you and says hey let's collaborate that you'll think yeah you know maybe this could be really fun and with that I will close and thank you very much