 I think I'm going to go ahead and start and just say good morning and welcome. Thank you. And I'd like to welcome you to the Social and Behavioral Research Branch's biannual seminar series. The goal of this seminar series is to highlight speakers, scientists who are doing innovative research at the intersection of genomics and social and behavioral sciences. Today, though, we have a little treat. It's a little special. So in addition to our stellar speaker, Dr. Kim Cappings, we're also celebrating SBRB's 15th anniversary. And so that's 15 years we're in our ongoing research has focused on translating genomics through a social and behavioral science lens, as well as translating the social and behavioral sciences through a genomics lens. So what I'd like to do is just take a few moments to reflect on where we've been. So a little bit of our history as a branch, as well as where we are going before I introduce Dr. Cappings. And I would also like to invite all of you to stay for a little bit after Dr. Cappings' presentation to celebrate and reminisce with us, have some cake, have some coffee. We'd love to have you stay for a few minutes after the presentation. So to give you a little bit of history, in December of 2003, at the 10th anniversary of the Intramural Research Program, Dr. Colleen McBride was introduced to our faculty in the Intramural Research Program, as well as the Extramural Research Program and our distinguished guests as the chief of the newly established social and behavioral research branch. And for those of you who have the pleasure of the 25th anniversary, you can imagine that we do these celebrations from time to time. And so this was our inaugural introduction of the branch. Now, at that time, there were really no well-established research programs out there that were looking at the translation of genomic information to improve health outcomes. While there was a lot of excitement about the new emerging technologies, there was some reticence in terms of the view that they were ready for translation. As well, if we think about that time in our scientific history, there were really few social and behavioral science enterprises related to genomic translation. The majority of the ongoing research at that time was really focused on identifying individuals that were at increased genetic risk of disease. Indeed, some, and myself, I agree, refer to the establishment of this branch as a visionary. If we think about NHGRI at that time, it was the home of the world-renowned ELSI Extramural Research Program, which funded biomedical ethics research at the time, and it's still here today as the Genomics and Society Extramural Research Program. As well, in the NHGRI's Intramural Research Program, there were leading researchers in both the biomedical ethics research that was Dr. Wilfand and Dr. Hall were looking at the ethics surrounding genomic translation, as well as Dr. B. Secker and Mr. Hadley were pursuing innovative research, looking at the psychosocial implications of genetic services largely in the context of inherited cancer susceptibility syndromes and rare disorders. So, it was with this sort of infrastructure that the leadership at NHGRI decided to invest in an intramural research home for the social and behavioral sciences, and so established the Social and Behavioral Research Branch. So, 15 years ago, the founding faculty within this branch included Dr. Colleen McBride, Dr. Wilfand, Dr. Hall, Mr. Hadley, and Dr. B. Secker. And shortly thereafter, we, Mr. Bonham joined in 2004, and then our distinguished speaker today, Dr. Keppings, myself and Dr. Persky, came on board in 2005. After a few years in 2011, we successfully recruited Dr. Shaw to join the branch as an Earl Statman tenure track investigator. And more recently, Dr. Larry Brody's group and Dr. Sharon Davis's group has joined the branch. Now, 15 years later, we have quite a diverse faculty coming from a variety of disciplinary perspectives that help shape the culture and the climate of our scientific enterprise. We have representation from sociology, psychology, social epidemiology, health services research, as well as systems biology. So, we have a breadth of expertise within the branch that is shaping our ongoing research moving forward. In 2017, the faculty met to think about a new strategic vision for the branch. And this is SBRB's vision, where our focus is really on the optimal integration of genomic discoveries, not just in the clinical context, but also within the community context. And we focus on both rare conditions, as well as common complex conditions that may have a major public health impact. And finally, we prioritize training. We are training the next generation of scientists and clinicians who will be pursuing transdisciplinary research. So, with this in mind, how do we do this? These are our foci, so to speak, in terms of our strategic goals moving forward. We advocate for the field of social and behavioral sciences in the genomics discipline or context. We are eagerly developing and fostering and nurturing partnerships with both intramural and extramural collaborators. And we prioritize training the next generation of transdisciplinary scientists, all with an eye towards establishing the social and behavioral research branch as the intellectual home of scientific excellence in genomics and social and behavioral science research. So, in the spirit of advocacy for the field, and I would like to introduce our distinguished speaker, Dr. Kimberly Campingst. And it's really my pleasure to welcome Kim back to the social and behavioral research branch to present some of the work that she's been doing over these last 15 years. So, Dr. Campingst is an internationally recognized communication scientist who's conducting cutting-edge research related to genomic literacy and cancer communication. She received her doctorate in health and social behavior from Harvard University School of Public Health with a background in molecular and cellular biology. Little fun fact there for everyone. Following a postdoctoral fellowship at Dana-Farber Cancer Institute, Dr. Campingst then joined the social and behavioral research branch and was my dear friend and support during that tenure-track process that I'm happy to reminisce about at the close of her, at the close of her talk. Currently, Dr. Campingst is a professor of communications at the University of Utah and an investigator in the cancer control and population sciences program at the Huntsman Cancer Institute. She has authored or co-authored over 100 peer-reviewed publications and high-impact journals that have focused on, for example, genetic and genomic literacy, returning results from genetic and genomic testing that meet information needs of both patients and providers, and integrating communication sciences in the return and translation of genetic and genomic information. So today, Dr. Campingst is going to share with us her perspective on informing the translation of genetic and genomic information with communication research. There you go. Thank you. Thanks so much for such a nice introduction, Laura. It's so nice to be back at SBIRB as part of the anniversary celebration. I was reminiscing, this is where I gave my very first job talk in 2004, so both good and bad flashbacks, I guess, being in this room. But what I wanted to focus on today is some of the work that I've done really since leaving SBIRB, so over the last eight or nine years in informing the translation of genetic and genomic technologies. This is our lovely cancer center in the foothills of the Wasatch Mountains, for those of you who have not been to Huntsman. So really briefly, I'm going to give some background on why I started focusing in genetics, why I think it's such a fascinating area to study health communication and genetic literacy. And then I'll focus on two particular projects. One is a scoping review that I just completed with colleagues at NCI, looking at the state of the genetic communication literature. And one is a project looking at factors affecting communication preferences, which has an emphasis on genetic literacy as well as numeracy. And then talking about some of my current research directions at Huntsman. So my area, genetic communication. I think of it as how you take all of this information that's being generated by genetic and genomic technologies and translate it and convey it to people in ways that are understandable and usable. And I will probably use genetic and genomic kind of interchangeably, not nearly as rigorously as I think bench people do. I always get in trouble for this. So I will say genetic and genomic, and we can debate the differences between those later. But why was this such an interesting area when I was a new health communication scholar and thinking about what to focus in? So as Laura mentioned, I have a long ago master's degree in molecular and cellular biology. So I knew some of the jargon. I was actually a drosophila geneticist in the very early part of my career. So I knew some of this. But I found genetics to be really fascinating as a communication challenge. So you're conveying a lot of information. It's often complex. It's unfamiliar. It's often uncertain. So you have to think about how to convey probabilities, risk information. And I study literacy. So the fascinating jargon in this area, how do you convey terms and concepts that are so unfamiliar to people? I probably don't need to tell anybody in this room, this background. But now my work is largely in cancer. I am in a cancer institute and an institute that is all genetics all the time. So I do cancer and genetics. And of course, there is a growing use of genetic and genomic testing in both cancer prevention, but also in the care of oncology patients. And so the communication challenges are continuing to grow in this area. And this is partly also because of the increasing affordability and accessibility of this type of testing. And so a lot of our literature base, which I will talk about, is from people who are highly educated, white, high-income, great insurance. Those are the people we've studied communication challenges with. But greater proportions of our population are going to be starting to get this type of testing. And so communication challenges are even growing. And both the complexity as well as the uncertainty of the information has a lot of implications for clinical care as well as public health. So there's certainly great potential for confusion among both patients and providers on how to interpret this information. The uncertainty drives a lot of difficulty with interpretation, making testing decisions, decisions about surveillance, and making sure that people have access. And of course, as we move to next generation technologies like multi-gene panels, so for instance, at my cancer center, almost everybody gets panels now, not single gene tests, but most of our literature still is from single gene tests. There is even greater communication challenges. So thinking about communication of variants of uncertain significance or VUS, the possibility of secondary or what used to be called incidental findings, so findings that are not related to the indication for testing, but you still might need to convey to patients. Ambiguous are possibly non-actionable results. What do people do with that? There is really good evidence that providers feel completely unprepared to talk about this type of testing, particularly as we move into sequencing or multi-gene panel testing. And so it's been my position and the position of colleagues that as you think about how to translate this information in ways that are useful for patient-centered care and improving patient outcomes that you have to think about genomic literacy. So I will, I know some of you know the definition of genomic literacy. Laurie and I are actually talking about this next week at ASHG, but just so we are all in the same definitional space. So in my work, I used this definition of genomic literacy, which is actually from a workshop that was conducted here, I was going to say a couple of years ago, but I think it was actually more like six years ago, to talk about what genomic literacy actually meant. So this was the definition that came out of that meeting, that genomic literacy is the degree to which individuals have the capacity to obtain process and understand basic genetic information and services needed to make appropriate health decisions. So you can see that there is a lot packed into that one definition. You've got a lot of skill domains going on there. So accessing, processing, understanding, decision making. And from an operational standpoint, I've actually found a framework based on health literacy to be more useful in operationalizing how do we actually measure genetic literacy or genomic literacy in research studies. So this framework breaks down genomic literacy into four major domains of skills and knowledge. So conceptual knowledge is what is considered to be the background concepts or knowledge that people need in order to understand a piece of health information. And one of the issues with genetics and genomics is that a lot of these concepts are really unfamiliar to the U.S. public. So for example, many people don't know that they have two copies of a gene. And so the distinction between having one copy and two copies doesn't make sense if you don't actually know you have two copies of genes. So even if you don't know the term heterozygous or homozygous, which I can guarantee almost nobody does, they also don't have the underlying concept. So even if you substitute in a simpler word, the conceptual knowledge is still an issue. Oral literacy is listening and speaking skills. So this is particularly important in studying provider patient communication about genetics. Genetic counseling is where a lot of this literature lies at the moment. So how do people understand verbally delivered genetic and genomic information? Print literacy is I think what it sounds like, reading and writing skills. So how do people interpret genetic test results that they might get in print? And then numeracy, which I'll talk about a few times as being a really understudied area actually in genetics and genomics. Numeracy is essentially number skills, so various domains of number of skills depending on how you define it. There are surprisingly few communication studies that I've actually looked at numeracy to date. So then this framework has really defined for me a number of major communication challenges that we grapple with in the work of my group. I am fascinated by this idea of how do you deal with the unfamiliar concepts that come with genetics and genomics? How do you decide, in fact, which of the unfamiliar concepts to convey, which relates to the second challenge of information overload? Because as you start thinking about everything, you can tell people about genetics and genomics. If you think about a multi-gene panel that might have 32 genes, you can easily overwhelm people just the sheer amount of information that you convey. So how do you decide which pieces of information to convey and how to do it? I have been for a long time and still remain fascinated with jargon. How do you deal with the jargon in this area and convey it clearly? And how do you deal with the probabilistic information? But the research base, as I mentioned, is still a little limited in terms of population, but it's also quite diffused. So if you've ever tried to look for studies on genetic communication, it has spread across a lot of different literatures. There's no one place to find it. And so moving on to the first study that I wanted to talk about, one of the studies that I've done recently was actually a collaboration with the National Cancer Institute, a group led by Sylvia Cho and Bill Klein, to do a scoping review. So this type of review, it's not like a systematic review where you really narrow in on a particular question. This is meant to be a landscape or a mapping of a large area of literature. So we were interested in doing a scoping review of what we thought would be a relatively contained question of the literature on communication of genetic information in the cancer context. And the goal was to synthesize the recently published literature from diverse disciplines on communication of cancer-related genetic and genomic testing to patients or the public. And I will also mention one of our secondary aims, which was of particular interest to me, which was to examine whether and how these communication studies assessed or described genetic or genomic literacy, as well as the related domain of health literacy, numeracy, and genetic or genomic knowledge. I won't go into the review methods in great detail, but just to give you a sense of the scope of this, what turned into a monster project. So this was a comprehensive search of six databases. A lot of the literature was in Medline, but not all of it. And so if you're only searching Medline, you will not get all of this literature. Some of it resides in Psychinfo or the nursing literature. We looked at articles published between January 2010 and 2017. And the rationale for that is that there were a lot of major reviews that came out in 2008 and 2009. So we wanted to look past those reviews, but we also wanted to capture the timeframe when newer technologies were being introduced into the clinic and into research settings to see if that changed the communication literature, where people starting to study multi-gene panels or sequencing or were we still sticking with the same single-gene tests. All of the articles we looked at were English language. They had to have at least some cancer focus. So there could be other indications for testing, but cancer had to be at least one of them. We looked only at communication of genetic or genomic information to patients or the public. So we excluded papers that focused only on family communication or communication within families or provider-focused papers. So provider knowledge, provider attitudes and so forth. All of the articles had to have some empirical data. So we excluded bioethical or legal normative analysis from the scoping review. But because it was a scoping review, we were able to include all study designs and outcomes. So we weren't limited to only, for instance, quantitative studies because of a meta-analysis. So we identified more than 14,000 possible papers. Even after removing duplicates, 9,243 papers were identified. And what rapidly became clear is there are no one set of key terms, key words, mesh headings to find this literature. You just got to slog your way through everything that says communication and genetics, essentially, in a much fancier version, and look and see what you got. So we had a large team of research assistants who actually screened the titles and abstracts of 9,243 papers, which still left us with almost 1,300 papers for a full text review. I was not popular among the research assistants at this point, to say the least. But we made it through this entire process, and in the end, included 513 studies and did extraction on all of those. So what I'd like to present is just some of our high-level findings that I think are of particular interest, perhaps, to this audience. So this is still largely an observational literature. Overwhelmingly, the most common type of design was a quantitative observational design. So either a cross-sectional survey or increasingly actually longitudinal studies are coming into play. So over half of the studies, I think about 60%, were quantitative observational. The next most common was qualitative. So a lot of interview and focus group data in this literature. There were about 14% of the papers that were randomized controlled trials or quasi-experiments. So largely any sort of experimental or quasi-experimental design, we lumped them all together and that was 14%. And then about 8% were mixed methods, which was largely observational, a quantitative and observational qualitative together. So as a social and behavioral scientist, I was also interested in how people were using theory. And the short answer is they are not, actually, in this literature. So we said what I think was a very low bar. Did the paper say the name of a theory or framework? Yes or no? And only 26% said the name of a theory or framework. 74% did not. We also looked at what the most commonly mentioned theories or frameworks were. So health belief model was actually number one. Self-regulation theory or the common sense model. Combining those, that was actually the most common Theresa Martau's version of the common sense model popped up. The theory of plan behavior, a theory of reasoned action, and the Ottawa decision support framework were the four most common. But as a communication person, what was striking to me is there is very few communication frameworks or theories that are being used in the communication literature. So things like uncertainty management theory, which is kind of tailor made for this context, is just not being used. And I think there's a great opportunity there to really both actually increase our use of theory but draw from other disciplines that have studied uncertainty, risk, conveying probabilities and so forth. I have a doctoral student who's particularly interested in this question. So she's actually going through now and mapping, if people said the name of a theory, how they actually used it. So do their measures map onto the theory or did they literally just say the name of a theory and then that was it, which I think at least in some cases is how they were used. We also looked at the types of genetic or genomic information that were being studied. So we first looked at the Tier 1 evidence tests, so BRCA and Lynch syndrome. I think this was actually really surprising to me, even knowing this literature as I do. Overwhelmingly, I think which no one would be surprised about, this literature focuses on BRCA testing. So almost half of the studies focused on single gene testing for BRCA 1 and 2. But surprisingly few actually focused on Lynch syndrome or HNPCC. So only 7% of the sample as opposed to 47% for BRCA. And of course, both of these have very strong evidence. Lynch syndrome in particular is a big emphasis in clinical care at the moment and we're kind of lagging behind in communication. And then looking at were we studying newer technologies, the answer is largely no. We're still back in communication about single gene testing. So only 12 papers or 2% looked at communication using multi-gene panels and 5% or 27 papers looked at sequencing. Some of those are mine and I can tell you with confidence that some of those are looking at hypothetical sequencing information. So this was actually before sequencing was introduced. And so a lot of this is more anticipating rather than actually conveying sequencing. But very few papers actually looking at newer technologies. And then we looked at outcomes across the different types of studies. This is what I am showing you for outcomes in intervention studies. But the pattern looked identical no matter which of the studies we looked at. So observational, quantitative, qualitative and intervention studies, overwhelmingly the most common outcomes are still psychosocial and distress and anxiety are number one and number two. Actually anxiety is first and then distress. And so I think we have the need to move beyond the psychosocial outcomes and focus more on behavioral outcomes, communication outcomes as well as decision making. And then the findings that were closest to my heart, I was kind of surprised about this and kind of not. So only five papers out of 513 assessed or described genetic or genomic literacy. 2% or 12 papers assessed or described health literacy. And this was actually the most surprising to me. Only 21 papers assessed numeracy, which again seems like such an obvious thing to look at in this literature. And then more, 18% assessed or described genetic or genomic knowledge, but I will describe some of the issues with those papers and I think two slides. Looking just at the five papers that focused on genomic literacy, three of these were actually qualitative papers. And so genomic literacy was a theme, but it was actually closer to genetic knowledge. So what they were looking at was a theme that reflected understanding or knowledge of genetics or genomics among patients. There were two quantitative papers. So one used the rapid estimate of adult literacy in genetics, which I think is familiar to some of you at least in this audience, that looked at how genetic literacy was related to learning from videotaped genetic counseling sessions. And then the other paper I think is also familiar to this audience, the Abrams paper from 2016 that used a genetic risk communication measure. Looking at how genetic literacy was related to forming opinions about surgical decisions, but this is really a wide open area of the literature at the moment. It's not something people are really focused on. So drawing these findings together, really few papers are assessing literacy or numeracy. And there are even fewer that have assessed associations with any particular outcome. So it really wouldn't be possible at this point to do something like a meta-analysis. There just isn't the evidence base to look at how numeracy relates to risk perceptions or something like that. There are lots of different outcomes and very few papers. More studies have assessed knowledge. But what we are finding, it is almost impossible to tell what people are actually measuring because people do not provide their items. So they will say something like we adapted Lerman's measure. Oh, what does that mean? Is it the same adaptation as when other people adapt Lerman's measure? We don't know. So we often know how many items they had, how they were scored. But we don't know exactly what people are asking. And we don't know if they're using the same measure as other people. So I have doctoral students trying to track down these measures to even get a count of them, how many different measures are actually in this literature. But it's incredibly difficult to really compare findings across studies. You actually have no idea what people are measuring. And I think there's a real need for greater clarity, both conceptually and actually conveying what you're doing in selection of measures and items. And you will know if it is me as a reviewer of one of your papers, because I will ask you to give the knowledge items. Now this is my thing whenever I review other people's papers. Like I don't care if it's a supplement or what, but I want to see what those knowledge items are. So certainly I think there is need for further work, both theoretical and empirical, on how genetic literacy or genetic related knowledge and skills actually affect information processing. So in that vein, I wanted to talk about a second of my projects. This was an RL1 funded through the LC mechanism, which was a mixed-max study to look at communication preferences for genome sequencing results. And I decided to talk about this study in particular because numeracy turned out to be interesting in this particular analysis. So the first aim of the study was qualitative. We did a series of semi-structured interviews to investigate communication preferences for what at the time was a whole genome sequencing results, although now we've dropped the whole, so it's just genome sequencing results, among patients diagnosed with breast cancer at a young age, which we defined as 40 or younger. And we looked at what people wanted to know, when they wanted to get the results back, who they wanted to hear the results from, and how. So what we called content and delivery preferences. And then the second aim was a quantitative aim to examine factors that influence these preferences and compare communication preferences among subgroups. So I created, when I wrote the grant, a whole series of very elegant hypotheses about how communication preferences were going to be influenced by people's family history of breast cancer, their mutation status, did they carry a known mutation in BRCA, and their prior genetic testing. And to give away the punchline, none of that mattered. The clinical factors did not affect communication preferences at all. What did matter was some psychosocial variables that I talk about, as well as numeracy. And we looked at both self-reported numeracy ability, as well as people's preferences. And what the measure of preferences means is essentially, do you like to get information through words or through numbers? So I'll talk about why that mattered. But first, why young breast cancer patients? Young patients are of course more likely to carry mutations in cancer susceptibility genes. But at the time I started the project, I was at Washington University, and they were gearing up to do a very large sequencing study of something like 2,000 women, and nobody was thinking about return of results. So were they giving anything back? If so, what were they giving? And I thought, huh, somebody should ask the patients what they want and what they're expecting. And clinical sequencing at the time was also starting to become important. And it seemed again as though somebody should be asking the patients. The grant had a huge advantage, which was being able to use the Young Women's Breast Cancer Program cohort, which was originated out of Washington University. And at the time we started was 2,200 women who had been identified and agreed to be part of the cohort who had been diagnosed with breast cancer at age 40 or younger. And this was an incredibly well characterized cohort. So we knew if they'd had genetic testing, we knew what their results were. So we knew, for example, if they had BRCA, did they have a mutation? Did they have a VUS? Did they have no mutation? We also, for everybody, although the diagram started to complicated so I dropped out some of the boxes, we had the strength of their family history of breast cancer. And unlike most of my studies where you're stuck with a survey item of, do you have a family history of breast cancer? Yes, no. This was actually characterized from a three generation pedigree and scored by a genetic counselor. So it was a really good measure of family history. The cohort was largely Caucasian. Their mean age of diagnosis was 35 years. And this box is actually to mainly remind me to say the mean time since diagnosis was seven years. So this was largely not a newly diagnosed cohort. And we did that purposefully so that people could look back over their illness trajectory and think what would have been helpful at the time of diagnosis and what would have been helpful later of possible return of results. So as part of the qualitative aim, we completed 60 semi-structured interviews that turned out to be on average 90 minutes each. So for those of you who know qualitative data, that is a huge amount of qualitative data. We stratified recruitment by the variables that we thought at the time would be important, family history of breast cancer, whether they had had prior genetic testing or not, as well as BRCA mutation status. And this was a semi-structured methods. So we looked at their beliefs and attitudes toward genome sequencing. But what I'll focus on is their interest in different types of possible results they could get back, as well as how they would want those results provided. The qualitative analysis was really time consuming, but used the software in Vivo 10. So I'll just highlight a few things from the interviews, and then mainly focus on the more quantitative data. But we looked at why people were interested in six possible types of results that you could get back from genome sequencing. So we looked at or asked them about variants that related to risk of preventable or treatable disease, risk of unpreventable or untreatable disease. We never use the words pharmacogenomic, but essentially effects on treatment response or medication response, uncertain clinical significance, which is how we describe VUS, carrier status variance, as well as variants that did not have a health meaning. And the examples we gave them were ancestry and physical traits. And what I think was interesting here is that the reasons for interest differed, depending on the type of result, which is one of those things that seems obvious. And yet the different types of results at that time tended to be lumped together in the literature, as though people were only interested in actionable results because they wanted to do something with them. So that was true for variants that related to risk of preventable disease and pharmacogenetic variants. So people were interested in those types of results because they could do something with them. But they were interested in other types of results as well, both because of future meaning, because they wanted to use them as planning part of their life. These are largely young women. Not all of them had completed childbearing, and so they wanted to use some of the information as part of reproductive decisions. The major outlier was the non-health meaning, so the ancestry and physical traits, where they just were like, oh, I don't know, that seems kind of fun. But the rest of it was really either for present use or future use in terms of life planning and prevention. We also asked them why they might not be interested in the different types of results. And this was largely not applicable. Everybody wanted the variants related to risk of preventable or treatable disease as well as variants that might affect how they responded treatment. But overwhelmingly, they were not interested in other types of results. It was because they felt like it would worry them. It would stress them out. They'd be concerned. So they didn't want, for example, VUS or risk of preventable or treatable disease because they thought I can't do anything with it and it's just going to worry me. Carrier status, it was a bit of an outlier just because people who didn't have kids thought that that was less relevant for them. But overwhelmingly, it was, I can't use that information or it's going to worry me. So I won't go through all of the six in detail, but I wanted to just show one example of what this data looks like. So these are reasons for interest in VUS findings in particular for people who were interested in this type of finding. Largely, they totally understood that that was not relevant information at the moment. So sometimes when you read the literature, you get the impression that people speaking about return of results don't think that patients understand what VUS are. And that was not what we found. They really did understand this was not information that was useful at the moment. But they also understood if they're part of a research study, this might be the only time they ever get that result. That they're not going to be able to go back in 20 years, find the researchers, and get their variants. And so they essentially want to just squirrel the information away in case somebody figured out what that meant in the future. Or there was also a strong altruistic vein. So they wanted to know what VUS they carried so that they could go out and find clinical studies or research studies that were trying to figure out what the variants meant. Lifestyle changes were really a distant third. Largely, people were thinking about future use or helping research. But this was really a typical response. So just to tuck it away, just to kind of know, okay, this one's here, 20 years down the line, we might know more. And I think this is also reflective of their age. These are people who are diagnosed with cancer in their 30s. Most of them had an excellent prognosis. And so they're thinking about their entire lifespan and how they might use genetic information over their life. And then again, for reasons if they were not interested in VUS, why not. They thought it was either uninteresting or useless or that would stress them out or worry. And I think what was striking to me and hadn't thought of before the interviews was people who were interested in this type of information tended to focus on future use. People who were not interested tended to focus on present use. And so it was how they were thinking the information was useful. So this is one really typical quote. There's nothing that they can do, nothing I can do. So it's not really a value of or interest to me at this point unless they find out there's something they can do. And we also saw a lot of ambivalence actually about everything except the risk of preventable or treatable disease. And this is I almost showed the whole quote and then it just looked really too crowded on this slide. So if you know what transcript data look like, this was essentially a page and a half of single-spaced transcript data where a woman is sort of weighing all of the different considerations on why you might want VUS or why you don't. And this is kind of toward the end of that very long segment. Do I want to know it when you've just diagnosed me with a cancer that I'm likely to die from in six months? No, clearly. It's not a worry that needs to be on my radar screen. There's a very long pause and then she says maybe I should know so that I can have my family tested so they could follow up, but probably not. And so by the end of this very lengthy sort of going back and forth she's decided she would not want VUS results. So we took the interview data and used it to create the survey that was done and aimed to where we looked at a quantitative way at what type of factors that affected what people wanted or what we called content preferences. So we did the survey of an entire the entire Young Women's Breast Cancer Program. We got a 60% response rate which is actually really pretty good for this cohort. We gave them a choice of doing the survey online, by phone or by mail. Almost everybody did it online using call trucks. And we were interested in identifying factors that might be important to tailoring return of results or test feedback strategies. We used the interview data to select the survey domains that we thought might be important. So we created communication preferences items. We also looked at genome sequencing knowledge using our measure from ClinSeq. Their preferences for decision making. The worry actually was something that was really changed when I looked at the interview data. So I had thinking about worry in the context of cancer, cancer recurrence. But what came out of the interviews were people were also worried about genetic risks that were not cancer related. So we looked at both domains. We looked at health information seeking. This issue of present versus future orientation that came through really clearly in the interviews. How important they thought health information was. And then Angie Fagerland's measure of self-reported numeracy or subjective numeracy that separates both self-reported ability from self-reported preferences for numeric information. A little bit about our respondents. 11% carried a known mutation in BRCA one or two. And in fact for two unlucky people they had mutations in both of the genes. 83% had had prior genetic testing. So they were familiar with genetic counseling and genetic testing largely. 28% had a strong family history of breast cancer. 15% had had a second primary cancer. Most of them 69% had biological children. Most were married or living with a partner. And this was largely a pretty highly educated population. So only 21% had not completed college. Their current age at the time of the survey was 46. And they were about 10 years out from diagnosis on average. So for content preferences what we did was ask them about the six types of variants that we asked about in the interviews. But we asked about it on a seven-point scale where one met not at all interested in getting that type of information back. And seven met very interested in getting the type of information back. And what I'm showing you here are just the proportions that said very interested. So the other people were not necessarily not interested. They often fell somewhere in the middle. But people were most interested in what you think of as the actionable results. So variants that affected risk of preventable or treatable disease as well as medication response. Carrier status was a close third. And I think again that reflects the age of the women who were doing the survey. They were much generally less interested in variants that affected risk of unpreventable or untreatable disease, the ancestry and physical traits, and VUS. Although again what they actually did was they measured somewhere between three and five normally. So really someone interested rather than very interested. It was actually pretty rare that someone said not at all interested in any of the six types. We asked them about who they wanted to get results back from. And there were actually 10 options but these were the top four. So genetic counselor and clinical geneticist which we actually defined, I'm not sure we used that term, I think a genetic specialist doctor is what we call them. I'm an oncologist and a primary care provider. So if you combine the genetic counselor and clinical geneticist columns there, overwhelmingly people wanted these results back from a genetic specialist. And when we looked at the open-ended data what came clear was that they wanted somebody who was able to interpret the results but then who was able to explain that to them. And so there were a lot of anecdotes about how, I don't think my oncologist or my cancer doctor knows how to interpret these results. So I want somebody who knows what they're talking about. And a lot of these people had seen a genetic counselor so they actually knew what genetic counselors do and so they understood why. The exception was people wanted oncologists to provide the pharmacogenetic variance back and I think of course what they were thinking was chemotherapy. So variants that might affect response to chemotherapy they thought an oncologist would be good for that. But what was also striking is people really do not want their primary care providers giving this information back and that really seemed to relate to this issue of expertise. So they didn't feel like their primary care providers could really do that. And I know this slide is a little busy but I wanted to make one point here. We also asked with this issue of information overload what they wanted to know about each variant. So we asked about 14 different things. What would you want to know if you get one of these types of results back? Overwhelmingly what they wanted to know and I think this will not be a surprise to any of you. They wanted to know implications for prevention or treatment and they wanted to know the effective gene variant on disease risk or the risk for their family. So what do I do about it and how important is this is essentially what they wanted. What they wanted to know less is often what we talk about. So how sequencing is done, the prevalence of the variant in the United States, how the gene works in the body which is always so tempting to talk about because it's interesting. That was much less interesting to them than the health implications for themselves and their family. So then I've combined actually six multivariable models onto this one slide. What you're looking at in each column is a multivariable logistic regression model where the outcome variable is being very interested in each of the six different types of results. And I decided to smush all these multivariable models onto one slide to show that there was a really consistent set of patterns. So the variables that were predictive are in hopefully showing up sort of bold blue. There were really three variables that mattered in predicting content preferences. So being very interested in all of the six different types of results. That was people who had higher knowledge about the benefits of sequencing, higher worry about their genetic risks. And actually the strongest predictor was people who had a strong orientation toward health information. And when we looked at the open ended data what became clear is they were really thinking of this as just another type of health information. So people who are generally interested in health information wanted their genome sequencing results back because they're storing up information about their health. Numeracy was the other variable that was predictive in more than one model. But interestingly it was numeracy preferences, not ability. And my hypothesis had been that it would be ability. So people who prefer to get numeric information as opposed to words wanted three of the different types of results back to a greater extent. So again, none of my lovely hypotheses about the importance of being a mutation carrier, having a stronger family history, or having had prior genetic testing, those were largely totally unpredictable when looking at content preferences. So just to summarize this area, patients I think, as I think anyone would expect, have varying information needs. We're not going to be able to come up with a one-sides fits all approach to return on results. What was important was their genetics-related knowledge and worry and their general importance of health information. Are these people who have strong value in health information or not? Numeracy preferences but not self-reported ability was predictive, which I'm still thinking through. And if anyone has ideas I would love to hear about why that's important. And my clinical factors were not significantly associated with content preferences. Nothing was happening there at all. So the argument that swayed Elsie on why carrying a known mutation would matter did not hold up in our data. All right, so I know that I stand between you and cake. So I will wrap up in a few slides of some of my current directions at Huntsman. I wanted to mention that we have a new UO1 that was funded under the Cancer Moonshot mechanism for which I serve as a PI. It was actually funded on my birthday. So thank you to NCI for that lovely birthday gift. But what we are doing is actually building on a larger infrastructure grant that we have at the University of Utah that is developing a clinical decision support tool that mines the electronic health record using both the structured elements as well as natural language processing to identify people who might be at high familial risk for breast or colorectal cancer. And that is the work of my colleagues Ken Kowamoto and Guillemir Adil-Fiole who are biomedical informaticists. We built our UO1 on top of it. So taking advantage of the clinical decision support tool that is finding these people in the University of Utah system, we are testing two different ways of possibly delivering genetic services to them. So a traditional pre-test, post-test, provider-driven model and then a patient-driven model where we do some online pre-test education as well as having patients drive more of the genetic services delivery. And the reason I wanted to mention this particular project is both because this was a lot of work and I gave up my Christmas to write this grant. So, you know, I'm excited that it was funded. But also because a lot of my thinking about the patient driven modality actually came out of my work on the multiplex initiative which is familiar to many of you in this room. And so a lot of my thinking from a decade ago is actually coming into this grant again. And I will mention so a lot of my work now is on communication and delivery of genetic services for rural and frontier populations. So I did not know the term frontier until I moved to Utah. Frontier is a very sparsely populated area of usually a county or a census block. So frontier means there are less than seven people per square mile. And there is a lot of Utah that is frontier. Actually, 70% of our landmass is frontier. So very sparsely populated. What is also true if you live in a state like Utah is that there is actually surprisingly little research on communication with rural and frontier populations. And there are a lot of different ways to look at this. This was the sites of the inaugural all of us locations. What do you notice if you live in the western half of the country is that the vast majority of the locations are on the eastern half of the country and congregated on the Atlantic seaboard. Then there is the area of the country that people think of as flyover country. And then there's California. I've labeled Utah because I've learned over time no one knows what Utah is. It's like one of the boxy states in the west. So what you can see is there's a vast majority of the country that was entirely left out of the all of us initiatives. Even in the newer funded sites, so Washington state and Baylor, we're still not really considering how do you do something like a large genetic cohort with rural and frontier patients. This was an analysis that we did really recently looking at ovarian cancer cases in Utah. So this is again the state of Utah for those of you who don't know what it looks like. And why they may or may not get tested. And one of the analysis we did was looking at distance to genetic services, which I think will not surprise you to know that it was highly predictive. How far people lived from Huntsman Cancer Institute, which is largely the only game in town when it comes to genetic counseling in our state. So Salt Lake City is that little area with the three stars up until about 2014. That was the only place you could get genetic counseling or testing in our state. So the little box down here, this is St. George for any of you who might have come out to our big national parks. They're down there in the southern part of the country. It is about a five and a half hour drive from the southern part of the state up to Salt Lake City. And so that turned out to be highly predictive in whether people had gotten genetic testing or not. So we're doing a lot of work on telemedicine, telehealth, and other ways of delivering genetic information remotely. And then I think in the interest of time, I will just skip my last slide, but show the acknowledgments. This was the work of a huge amount of people that I talked about here. So my wonderful team at Huntsman, my fantastic collaborators at NCI led by Sylvia Cho and Emily Peterson, who was a post-doctoral fellow. And then my collaborators in the communication preferences study are funding from NCI as well as some support from NHGRI. And then I wanted to leave you, this is the view from my back deck on Sunday. So you can see we have snow on our mountains now for those of you who are skiers. But this is what I look out at every day. So when people ask me why I live in Salt Lake City, this is why. Thank you very much. It's been a pleasure. It's a good question. I think what we are seeing is more numbers. So it was bad enough when we're talking about one single gene at a time where you're giving one number or a couple numbers. Now you're potentially returning results from multiple different things. And so you're dealing with multiple numbers, but also greater level of uncertainty of those numbers. And so trying to grapple with both the greater volume as well as the greater uncertainty. Not that I have any answers, but those are definitely the communication challenges. I don't work. I do know her. Yeah. Yeah. Yeah. No, it's such a good question. So I had expected actually to see an uptick in studies around Lynch syndrome because of changes in, for instance, at Huntsman Cancer Institute, the directive is to test everybody who's diagnosed with colorectal cancer. And that's fairly standard in a lot of places. But there has not been a lot of communication research. And I think what is interesting to me is as we expand the definition of the syndrome from the sort of really traditional HMPCC families, how people understand not only where they just diagnosed with colorectal cancer, but now they find out that they have an inherited cancer syndrome that they don't really have many indications for. I think it's such an interesting communication question. But apparently I am alone in that because we don't seem to have many communication studies that have looked at it. Now, I think it's a great question. So this is a good question. And partly the answer is we are currently negotiating with our providers how much of this can actually be patient driven. But my goal for it is that rather than reaching out to patients who are identified, patients will first have the opportunity to reach out to us to see if there is uptake of genetic counseling without the immense outlay of staff time that are going to be required to actually reach out to people and identify them. The goal is probably we'll give a certain amount of time for patients to reach out to us once they do pretest education and then we'll reach out to them because of safety issues and the IRB is really unhappy with my patient driven model at the moment. But I think I'm interested in if you provide people with information, you do some pretest education online, is that enough for them to reach out or do we still need to reach out to them? Because the underlying focus of this particular U01 mechanism is large scale implementation of genomic medicine. How do we do this as we start to integrate into large healthcare systems like the University of Utah where we have the luxury of having eight genetic counselors at Huntsman but that's still not a lot when you're talking about thousands of patients who might need outreach. And so can we motivate patients to call us and take more of the lead or not? And that is to me an open question and with a negotiation about the IRB who is convinced that this will be very damaging to patients despite all of my efforts to show them data that it will not. But it reminds me actually of multiplex and some of the IRB issues from so long ago. I wish I could say the IRBs have advanced. Yep. Yep. They do not. Yes. No, it's a good point. Yep. So we have actually been testing models of telehealth or telegenetic counseling which has worked actually pretty well but our original ideas about video were actually had to change because a lot of people even if they have internet access, it is not broadband. It's slow. So we have to kind of go back to old school and how you do things on the web in the days when people had dial-up connections. But actually the outcomes have been pretty similar. So there are a couple major randomized controlled trials that have looked at telephone counseling or video counseling compared in person. And uptake is a little less with telephone counseling. But a lot of the psychosocial outcomes and decision-making outcomes are very similar or at least not statistically different. So we actually, and I don't know why I'm saying we because I have actually nothing to do with it, the genetic counselors at Huntsman have established six telemedicine sites or remote sites where people can go to clinics in their local community and then we have better broadband access and better ways to do the video counseling. But they don't have to drive seven hours in a lot of cases. So there are no genetic counselors in the entire state of Wyoming. There is one in Idaho and so we've established telemedicine sites up there. But it's hard unless you live in an area like this to understand how far people have to go to do things. So our nearest cancer center neighbors are seven hours away in Denver on one side and 10 hours away in San Francisco on the other side. And so the idea of traveling an hour to a remote site isn't bad for most people. They just don't want to drive the entire day to get up to Huntsman. And so the combination of telehealth but also remote clinics seems to be working really well. Yeah, they tend to be right. So that's a good question. Two of them are Huntsman clinics and four are contract sites. So they're actually community hospitals in where they're essentially renting out part of their space to do the clinical sites. Yeah, it's a really interesting services challenge. I wouldn't have thought of myself as whole services research, but here I am, you know, doing some of that. So I will thank you. Thank you. I just realized I forgot to get my little flash drive out.