 Welcome. I'm Eric Green, Director of the National Human Genome Research Institute, and I'd like to welcome you to the 13th installment of the Genomics and Health Disparities Lecture Series, which is cosponsored by the National Human Genome Research Institute, the National Heart, Lung, and Blood Institute, the National Institute on Minority Health and Health Disparities, the National Institute of Diabetes and Digestive and Kidney Diseases, and the Office of Minority Health and Health Equity at the Food and Drug Administration. As many of you know, the series aims to highlight the opportunities for genomics research to address health disparities. Speakers have been specifically chosen to present their research and their ideas about how genomics can improve health of all populations, and also discuss the challenges in making sure these improvements are accessible and applicable to all groups. The speakers in the series approached this problem from different areas of research and different areas of expertise, including basic science, population genomics, and translational and clinical research, among other areas. Today's distinguished speaker is Dr. Katrina Armstrong. Dr. Armstrong is currently the Jackson Professor of Clinical Medicine at Harvard Medical School, Chair of the Department of Medicine and Physician in Chief at Massachusetts General Hospital. She's an internationally recognized investigator in medical decision making, quality of care, and cancer prevention, and outcomes. She's also an award-winning teacher and a practicing primary care physician. She has many honors associated with her distinguished career, including being elected to the American Society of Clinical Investigation and the Institute of Medicine. And she's also served on multiple advisory panels for academic and federal organizations, and I can tell you from NHGRI's perspective, we have been a beneficiary of that, as she was a key advisor and actually former chair of the advisory committee of our clinical sequence and exploratory research program. And so, without any more delay, we're delighted to have Dr. Armstrong with us today to give a talk entitled, Can Equity in Genomic Medicine Become a Quality Gold? So, thank you so much for having me. It's an incredible blast to be here. I was actually sharing with many of the people I saw today that I, after I finished college, I actually came and worked at NIH. That was in about 1920. And some things have changed, but it brings back a lot of fantastic memories. Actually, I bought my first car here on a parking lot for $500 from a postdoc who was going back to China, I think, at the time or maybe it was Japan. And I loved that car. Let me just tell you, it was like the greatest car. So, I am here to talk about equity in genomics, and I really am grateful to be here. I'm really, unbelievably grateful for the leadership that NHGRI has shown in this area over the last decades. It's really changed my career and I think has changed the opportunities for our country dramatically. I have no disclosures. What I'm going to do today is actually share some thoughts I have about this field, starting with what we're experiencing now in terms of widening disparities in the country and life expectancy, and how that may have implications for thinking about equity in genomic medicine. I'm going to talk a little bit about how I see some of the primary drivers of disparities in genomic medicine, including issues around tool development where I spend less of my time, and then particularly issues around access where I spent most of my career. And then I want to finish by how I'm now currently thinking about the next steps in equity in genomics, particularly in how we may be able to engage the quality community to help us advance this agenda. So, I think we all know that disparities across our country are wide. This is a slide that I came to very soon after I finished training in the late 1990s. We're really showing the disparity in life expectancy between blacks and whites. I grew up in Tuscaloosa, Alabama, and I think if you grow up in small town, Alabama, this was a reality that I didn't need a slide to see. I think that one of the things that I shared today at lunch that's striking, though, is that if you go back and look at that number, I don't know if I go, I can't go back. It's a beautiful thing that, you know, we're talking about life expectancy differences in there, five to six years. And often we seem to think of that as being something that is not that big of a deal when we talk about a life expectancy in the 80s for most individuals. But what I often want to show people is that actually if you look at it, that's a huge gap in life expectancy. And that the estimates are that if we were to cure cancer in our country, that we would save a total of three years of life expectancy total. And these are some of the estimates for other diseases. So huge gaps in life expectancy by race in our country. And of course, it turns out that that's not the only characteristic where there's a disparity. We know that the same thing exists for economics here showing US life expectancy at age 40 for women on the left and men on the right, both by race, but then by economics, whether you're in the poorest 1% or richest 1% at birth. Huge gaps in life expectancy. The reality, however, is that one of the most striking things to me about those data are that they're actually getting worse. And so there have been many studies now that show that this is the gap, let's say, if individuals were born in 1930. And we can see this is for men if you were in the bottom quintile of socioeconomic status born in 1930. The expectancy was that at age 50, you would have 26 years to live. And if you're in the top quintile, you would have 31 years to live. But that number has changed dramatically. So that if you were born in 1960, I was born in 1966, that the gap in a life expectancy at the age of 50 had really now increased to over 12 years, an enormous gap. So why is that? And there's a lot of hypotheses, but there's one that I think is particularly relevant to genomics that I want to talk to you about because I don't think we pay enough attention to it. And I'm going to come back to it as we go through the talk. So when you look at life expectancy over the history of humankind, let's say, go back to 1700 and map out what has happened to life expectancy. People often talk about there being four phases of what has happened. So back in phase one, and it obviously goes on before that, is when we actually first started to have nutrition and housing and life expectancy increased as we didn't die of starvation and cold, we then had a phase here, phase two, where life expectancy actually went down during the Industrial Revolution. But I want to focus for a second on phase three and four, because these are the ones that we kind of own in the biomedical world. So in phase three, life expectancy increased as it did in phase four. But people look at the causes of that life expectancy. And if we go back to it, as you all remember, the difference that what happened with disparities during those periods, as really coming from different sources. So in phase three, we gained life expectancy from interventions like vaccinations and antibiotics. And in phase four, it turns out that we changed and our life expectancy gains since 1950 have come much more from individual interventions like bypass surgery, TPA, different types of thrombolytics and chemotherapy. And it turns out as we look at the categorizations of how we're affecting health, we went from a phase where healthcare was a relatively simple social good that people could access and that we were able to deliver across populations to a formulation of healthcare that has become increasingly complex and private and dependent upon your ability to pay for that healthcare. And so I want you to remember as we now talk about genomic medicine that we're facing a reality, we're in our era of biomedicine, the era that in many ways NIH created with the unbelievable understanding of biology, that we've led to a period where disparities are getting much, much wider. And that there's a growing belief that those disparities may be widening because healthcare has become increasingly complicated and increasingly dependent upon an individual ability, individuals ability to pay for that healthcare. So there's data somewhat to support this, I'm just going to show you some of those data. This is a slide that I often show about cancer disparities. And I think the paradox and this has been published in many places is that interestingly, the cancer disparities are greatest for the diseases where we have effective treatment. It also turns out if you characterize diseases by whether we they're preventable or not preventable, the disparities are widest when we have a prevention for that disease. And if we look at one particular example that I grew up with, I trained in the era of the HIV epidemic. When the HIV epidemic came out, we really had no effective treatment, no highly active antiretroviral therapy. And the reality is that although those medicines have transformed that disease, they are complicated. And access and getting those medicines is tricky. And if you look at it that actually since the advent of heart, that the disparities between blacks and whites with HIV and their mortality have grown substantially. And so I want us to keep that in the back of our mind as we think about where we're going in the era of the genome. So my area of focus in genomics and in most of my career has been on the role of genomics in addressing cancer disparities. And I've particularly been interested in issues related to breast cancer. I'm going to ask my buddies up in AV to see if they can help me. Whoa, there we go. And so one of the things about the same time when I ended up finishing training in the mid 1990s was there was a growing recognition that along with other disparities that we were experiencing a substantial disparity from breast cancer mortality. And in fact, this had become my passion in part because of a patient who I'd taken care of on the Hopkins Medical Service, a woman who came to see me in primary care with very advanced cancer, and a woman who'd actually spent her career as a janitor at Hopkins and ended her life, I think in a really challenging way, experiencing a tremendous amounts of financial toxicity from her own cancer treatment. And it was that story in many ways that has motivated me to try to understand how we can do better in addressing this particular disparity. So one of the areas my K award, which actually came from the NCI a very long time ago, was actually about the idea that we were going to help fix that disparity by the application of genomics, kind of where Dr. Green started, that by understanding the human genome, and in this setting, BRCA one and two, that we would be able to test women and be able to identify women at increased risk, particularly African American women, and then we'd be able to intervene to do something to make that curve better. And so I set up a cohort study, and I was actually in Philadelphia at the time, to take all the women who were coming in to be able to test them for mutations to look at their sequence and to follow forward to be able to understand the genetic epidemiology and black and white women of BRCA one and two mutations. But the reality was that despite that aim, and I think that was aim one, that when we went to look at the data, because we were here in Philadelphia city that is 50% African American, we had essentially no black women in our data set that the women who were coming in for testing for BRCA one and two were almost all white. And so we changed our study and looked forward to try to understand what was it that was driving that disparity in use of genetics? Why was it that we were seeing really only a highly selected group, even in a health system population that was all insured, and was over 40% at the University of Pennsylvania African American? And so what we found then now some 12 years eight years or so years ago was that white women in our system with the same family history, socioeconomic status, risk perception, all the same risk factors that we could find were about four times more likely to be pursuing BRCA one and two testing them were black women. So I've spent a lot of my time trying to understand why that is. And what is it if that's true for BRCA one and two? What is it as a community that we need to do to make sure that that story changes as we create more and more ways to make our to do better in terms of risk prediction disease treatment from genomics? And so I want to talk about the two main areas that I think are critical for this, primarily again, focusing on access as that's where my research has been. But also, I think as we know, and as many of you sit here that the tools that are created from the human genome project from the NHGRI, the role of those tools is critical. So as we all know, there were some initial challenges and how those tools were going to be used or how effective they would be for minority and other populations. This is a slide that has shown a lot that just talks about the difference in the representation in GWAS populations. But from the GWAS over on your right to the US population on your left, I think those are data that you all know, I'm sure very well. I think it also meant as you all know that that meant as we were trying to use these tests, and I remember doing this in clinic in Philadelphia, that the rate of variance of unknown significance when we would send a test in a minority or in a group, an ancestral group that was not well represented in the reference data was much, much higher. And so the challenges of using those tests back in 2002, if you can see these are data from myriad about the rates of variance of unknown significance here by different ancestries. And you can see the top line is for African populations. And at that time, there's tremendous rates of VUS in that group. And luckily, as we've used these data that has come down, but significant barriers to the effectiveness of the tool when we don't have the representation in the research. This is just one study comes from folks up in Boston that emphasizes that point. There's lots more of these studies. And basically suggesting that when we have tools here, we're doing panel testing for cardiomyopathy, that we still have challenges with the rate that the effectiveness of that tool, the accuracy in different groups. Here you can see for white underrepresented minorities and Asians, the difference in the rate of the positive inconclusive negative or no interpretation test. I will just point out that this problem also exists for trials where we're using this for genetically guided trials. So these are data that a postdoc pulled some time ago for what I call genome guided trials. So these are trials registered on clinicaltrials.gov between 2009 and 2014. Very hard. Most of them don't give you the race of the participants. All of these trials were using some kind of genetic tests or direct therapy. And interestingly, most of the studies that reported race were actually sponsored by industry. And of the African Americans included in the study, only a small in those studies, only a small minority were not included in a study that was specifically designed for African American populations. And so the representation in trials also remains an issue. I guess I will say that despite what I'm talking about in terms of challenges and creating tools, that I am unbelievably optimistic about this part of addressing equity and genomic medicine. And I'm optimistic for several reasons. One is I just think that there has been transformative leadership by NHGRI and I'll call Vence out here. And when he asked me to come do this, I said I'll come because I think you have done more for this field than anybody I could ever have imagined. And so I just want to say that without his leadership, I do not think this part would ever have gotten to where it is. And I will tell you that now as I read the literature, when I first started in this field, I would go to meetings with geneticists and I would talk about inclusion. And they I mean, it was like as soon as I said that they were on their computers. They didn't have any idea what that word meant in terms of diversity in this way. All we cared about was population stratification at that moment. And so Vence and many others have changed that dialogue that when we now look there in these studies, there is a focus both on doing this for equity and social justice reasons, but also because we now know that by looking into diverse populations, we're going to make better tools. We're actually going to discover variation that we wouldn't find by looking under the lamp post again and again and again. And so it's been a transformative time for this. And as you guys read papers from fundamental scientists now, I just read a bunch in cell, they're arguing for the inclusion of diverse populations. And I think that is the biggest sign of Vence's success that I could ever imagine. I think that has shown this is a recent review from cell, which actually talks about the diversity of patients who are coming into biobanks. I'm not going to go through it. But just the fact that we're now emphasizing this and moving this forward gives me a tremendous amount of optimism. I also am optimistic because I believe, and I'm going to talk you through the theory behind this for a second, that actually having these good tools is in itself going to reduce disparities. And I'll show you a couple of pieces of data to support that. So one of the seminal works in understanding disparities in healthcare comes from the Institute of Medicine. I'm sure many of you know unequal treatment. And one of the places that they articulate as a source of disparities is down the bottom here, which is the uncertainty about what to do in any clinical setting. And when that uncertainty exists, we can end up with our biases and stereotypes driving decisions. It turns out that this has been shown now, that there are many studies have shown here one recent meta analysis, that if you look at studies about bias and stereotypes at 14 out of these 15 studies, found negative bias and stereotypes related to African American patients among physicians. And that bias was associated with differences in treatment and disparities in treatment for multiple conditions, including HIV and diabetes. So we can help make better clinical decisions. Maybe we can reduce the disparity by reducing the effect of bias stereotypes and uncertainty on those clinical decisions. And so I just want to give you one example that a fellow did in my group some a couple of years ago. And so what she looked at was the effect of using Oncotype. So using a gene expression profiling test. And if women were getting Oncotype, would they potentially have a smaller disparity in the use of chemotherapy? Oncotype is used in breast cancer to help decide about the use of chemotherapy. And so you could imagine if we often use bias and stereotypes to make that decision, maybe having genomic data to help guide us would be good. So there are about 900 and so women in the study who had incident breast cancer and were eligible for gene expression testing. About a third had undergone chemotherapy and about a third had undergone gene expression testing. This is an observational study. People weren't randomized to these different things. But what she found was that if women underwent Oncotype or gene expression testing, there was no racial difference in chemotherapy. It looked like the decision to use chemotherapy was driven by their clinical factors. But if they hadn't undergone testing, there was a significant difference in chemotherapy. Here in Philadelphia and Pennsylvania, actually this was Pennsylvania and Florida, actually the black women were more likely to get chemotherapy, which has been shown in some settings. But a disparity that existed when you didn't use the test compared to when you did. We've also, this is the work by the same young investigator, has also shown that we know that in many of our predictive models that we do a poor job in certain populations. This was a study that she did just looking at women who were undergoing breast cancer screening and trying to see if we could improve breast cancer risk prediction in African American women. We know the current risk prediction models were developed in white women and they work less well in African American women, meaning that fewer African American women are identified as high risk. Here again she took about 800 women and I just want to suggest the only point here is really that if we look amongst the African American women that actually more of the women when we included the SNPs were identified at high risk. And I'm not showing you the data here that shows that they ended up doing a better job at predicting cancer amongst the African American women, suggesting that as we develop these tools and as they become inclusive and effective for other populations, they may have an opportunity to not only lead to new treatments, to new ways of understanding disease, but to making us help us make better decisions across groups and to make better risk prediction. So I want to spend the rest of my time really focusing on if we have a tool. Let's imagine we do. Let's imagine we have a tool that works. How are we going to deal with the fact that right now we have tremendous inequities in access and what do we need to understand about access and about delivery that can overcome that. So one of the frameworks that I like for access is a framework that talks about it having multiple dimensions and I want to focus on three here today. Dimensions of affordability, acceptability, and availability. I'm going to spend most of my time on availability because I'm guessing probably people have talked a lot about acceptability in this times before. And I think there's an important role as we develop these tools to be able to understand how these different dimensions of access are driving disparities and what we can do about them. So I won't spend a lot of time on this because it's hard sometimes to have a research agenda around this, but it's very clear that the affordability of the tools is a critical part of us addressing disparities in access. This is a great slide that comes from Joe Lenners who's actually a Mass General who just looks at how complicated, if we want to get back to what I was saying about complexity earlier, how complicated it is to try to get insurance coverage. And if you think about that hypothesis about complexity leading to disparities in access and an uptake, you can see some of the trouble that we're facing. Of course, that's for insured populations and we all know that the reality is that there still remains tremendous pockets of uninsurance across our country. I live in a state now that has very little uninsurance, but even in our state, it's crazy hard to get Medicaid to cover these types of tests. And if you look at the other states, particularly where Medicaid has not been expanded in orange, we have substantial rates of uninsurance. And so without the ability for us to understand how we're going to pay for these tests for patients who do not have an individual ability to pay, clearly access will remain limited. So we think about acceptability and it's something that the human genome research has supported a lot over the last years is really understanding the acceptability of genomics and genomic testing to different groups. We just talked about it a little bit at lunch. And I think this bi-directional dialogue and understanding how the communities can inform the tools that are developed in the implementation is critical. I spent a good bit of time on this early in my career trying to understand the effects of trust and distrust on the acceptance of technologies by disadvantaged communities. And one of the areas that comes out and is still an area that we should continue to explore is the reality that most groups believe it turns out in the positive aspects of health care. So in national and regional samples, if we think about genetics, over 85% of both blacks and whites believe that genetic testing is going to be used to help manage their health care. The challenge that we have is then the beliefs about what else could happen. What are the bad things that could be done by the information? And so we know that whites in these samples is just one study or less likely to believe that health insurance coverage is going to be used to limit health insurance coverage to label groups as inferior. And so thinking about how we're going to manage the acceptability of the tools that are being developed here and how do we continue that bidirectional dialogue to sort out when we're making choices about the tools? What can we do to address these questions of acceptability? I think are critical as the Human Genome Research Institute moves forward. But I want to spend most of my time talking about an issue of availability, which I think really comes down to the crux of how these tools will be delivered. One of the realities growing up in the South and now living in Boston is that if we look at how resources are delivered in healthcare across our country, there is tremendous regional variation. And right now we do not have a good understanding of what that means for our genomics infrastructure. What does it mean for our ability to affect individuals to bring these advances to communities across our country? And what types of infrastructure currently exists? And how does that infrastructure need to be adapted and supported? Part of the reason that matters so much is that it turns out our country is highly segregated. So that different racial groups live in different parts of our country. And so to the degree that there's inequity in infrastructure is going to lead to inequity in all sorts of other different ways. We did a study some time ago, and these are relatively minor differences. Looking at primary care providers in all these four regions and asking them whether they knew where to send a woman to get BRCA counseling, so to get counseling. And these are relatively small differences and I think the numbers would be higher now. But you can see that in the areas in the Midwest and the Northeast where most of the population is of Caucasian background and white, it turns out that there's relatively higher rates of understanding and infrastructure of what to do in that setting than in the Southwest. And those lead unintentionally to disparities in access. It also turns out that they're even within a region, so if we think about D.C., and sometimes I ask you all to think even now, some of you are only here a little while, about how healthcare is delivered where you are and where people go for their care. It turns out that even within a region in Boston in particular, in Philadelphia, that healthcare is remarkably segregated. That we know that that segregation also means that the diffusion of technologies varies across those institutions, varies across region as we show, but across city, across hospital and provider. And there's been an interest for some time in the hypothesis that minority patients we know that minority patients in the U.S. are more likely to not have a regular source of care, so to get a different type of care, to see a primary care provider who has more difficulty and to be treated at hospitals with who are that are slow in adopting new technology. And so maybe as we think about the availability of genomics, part of the story here is going to be that we've got to be paying attention to the diffusion of genomics in these different facilities, to the infrastructure and the segregation between the genomics infrastructure and the race of the patients. We've shown some of this for prostate cancer and endometrial cancer and other diseases. I'm not going to go through the data here, but just to show you at the top line that black patients tend to congregate within certain hospitals. So if you look at hospitals that have a proportion of 30 percent black, there are very few white people in those hospitals. So we did a study some a couple of years ago, I guess it's probably almost eight years ago now, but then we looked at this question of is it the segregation that is driving the disparity in BRCA testing? Because if the problem is how we're diffusing technology, that's a different set of research questions the next step than if the problem is how we're treating people from within our institution. So we enrolled 3,000 women with newly diagnosed breast cancer in Pennsylvania and Florida as well as all their cancer providers. And we found yet again that there were substantial disparities in BRCA testing amongst eligible women in this population. I think you can see high risk, medium risk and low risk disparities with let's just pick the high risk line to have black women 44 percent undergoing testing compared to 68 percent of white women. We also found that the care of black of women with breast cancer in these states not to our surprise is relatively highly segregated. So black and white women go to different hospitals for their care in those states. And the index of dissimilarity tells us that we'd have to move about 68 percent of those women to a different place if we were going to fully address that disparity. So what we found is that the providers when we survey both the providers about their probability of rating of recommending it and then matched it to the to the proportion of black women in their practice. And then when we asked individual women about what their experience was that both surgeons and medical oncologists were much less likely to have recommended testing to black women than they were to white women, whether we looked at all women or women less than 50 or essentially any risk category that we picked up. And that when we looked at the clinical disparity altogether and this is just a regression analysis that the majority of that difference, the odds ratio between black and white of point five here on the left, once we adjust just for risk of imitation, that the biggest effects that were explaining that difference were the clinical and sociodemographic factors, but then were the physician recommendation, that it wasn't because people were seeing women were seeing different doctors, it was because the doctors were telling black women different things than they were telling white women. And so the difference that we found was coming from this difference in treatment within a practice. And I think this gets back to the initial point around the tools and the uncertainty and the complexity. And so what we find as we went in to follow up on this is that in many places, as we saw, there have been lower utilization of genomic tests among certain groups, which led to less information about inner individual variation in those groups, which led to more uncertainty and lower utilization. And so our cancer providers had gotten into a vicious cycle, where they were recommending fewer tests in the setting, believing, which is not true, that the mutation rate would be lower and the information would still be less. And so I would argue for us to achieve equity here, we really need to do two things. We need to develop the tools, but once we have an effective tool, a disparity in the use of that effective tool becomes an organizational quality favor. So at Mass General, if I know that I have a test, a BRCA test that I should be delivering to all women who meet a certain criteria, then the lack and lack of equity, a disparity in that is one of the actual Institute of Medicine quality dimensions here. And so often in quality, I think we focus not enough on equity. We think about safety, we think about effectiveness, timeliness or increasingly about cost, but the reality is that for genomic medicine to succeed, we're not only going to need to think about quality in these other dimensions, but it is absolutely imperative that equity and genomic medicine quality become part of the future of our research and of our clinical programs. So what are the ways we can do that? Well, it turns out people have been working on quality research for a long time. It's a big part of how we think about cancer care now, a big part of how we think about HIV, a big part of how we've made such an impact in so many diseases. And that can include the need to develop quality metrics to understand what does it mean to see a high quality genomics provider? What is the right test? What are the outcomes we're trying to mean? To create data tools in this area to be able to measure quality, to develop decision support, quality improvement approaches, pay for performance strategies and to understand how they'll work in this era, as well as the potential role of public reporting. But I think for disparities, when we think about quality, we think about that toolbox, but we add on a whole additional set of strategies that need to be evaluated and move forward in genomics. Long traditions in genomics of understanding community engagement and how engagement of those communities may make a better service, more acceptable, better access in a higher quality service. Translation and interpreter services have become a major focus in my world. We have at the MGH dramatic shifts in the languages seen on our services from year to year. And so our ability to keep up on actually creating interpreter services that can actually reach those patients is constantly put to the test. Issues of patient navigation, I talked to people at lunch about navigators, community health workers, other programs that have been developed to improve quality that can be adapted and tested and studied in genomic medicine. One of the things that brought me to Mass General was actually their willingness to embrace equity as part of quality. And I'll just show you that once we've developed the tools to know what we're trying to measure and achieve in genomic medicine to address disparities, once we've done the research, we can come together then and measure whether we're doing it and hold ourselves to making that better. So this is something from something called the Disparities Solution Center run by Joe Bettencourt and as we attend McGrory at Mass General. And over the last years, they've put together Disparities Report from our Center for Quality and Safety. And so our annual report on equity, you guys can find this online. Public reporting is critical, I think, to show it. And this goes through, as you can see, this is just one image I picked, but there's lots and lots there. And looks to see are we able to achieve equity in the metrics we've picked for high quality across different groups. These are some of our ambulatory quality practice measures here. The difference between the MD and the practice is just whether or not the patient has an individual MD, whether or not we've been able to assign that patient and they're coming to see a regular MD. And you can see, as you can see, the yellow bars mean that that group is not doing as well as our reference or as our mean in our groups. And then you can see people at different levels. We can also do that for measures that are self reported quality measures back to the acceptability. What is the experience of individuals undergoing genetic counseling? How do we decide what we're going to hold ourselves to in terms of that experience of care? These are measures just from our CG caps, one of our ambulatory quality surveys. And as you can see, we've got a long way to go. And so when I see these measures and I'm being told that the difference in different groups between how our patients, white patients rate providers at 80 per six percent say that they would come back and see that provider compared to 81 percent or 72 percent, those are the data I need to be able to do the studies to move the effort forward to actually address this. And it's the data we need for genomics. So I'm just going to finish by going back to where I started. So we're at an incredible time in medicine in so many ways. The investment that has come out of the NIH and so many other areas into understanding human biology, I think has the ability to transform what we're doing in medicine over the next decades. I often joke when I'm recruiting residents to Mass General that when I was a medical student in whatever, I still can't keep these years straight, but let's say 1987, most of what I learned, most of what you guys are what medical students have now, we hadn't even heard of. It was completely different curriculum. But yet if I go to my medical services now, half of what we're doing on the medical services, 90 percent of what we're doing on the medical services is still the same. I often joke we're mostly taking fluid, giving people fluid and taking it away for the doctors in the audience. But this is going to change. We are transforming in genomics and immunology in every place our understanding of biology is changing what we're doing for patients at a pace that I never thought I would see. The challenge we have is it's coming at a time where the inequities have never been greater. And what I want to suggest to you is as you move forward in genomics research, then not only are we facing a need to understand how those tools are going to be effective and how we're going to get access to them. But I think it's a social imperative that we understand how we make the best possible tools that can be reached by different groups. And that the health care that we create out of this biomedical revolution isn't so complex and so difficult to use that it will never actually get back to reducing that disparity. I think when I think about genomics, right, I grew up in BRCA 1 and 2 and we had a hard enough time getting that out to creating a quality metric to understand who should get that test, what type of follow-up should they do, getting that report into the EMR. And now this is what my patients are getting. So they don't get BRCA 1 and 2 testing, they're getting panels of genes. I couldn't, you know, most of my primary care providers couldn't tell you what any of these genes are. It's an incredible thing to have this power, but we've made it really complex. I often think about this. So this was a huge thing in my world. So the development of a PARP inhibitor as a treatment for BRCA related cancer was the pipe dream when I first started in this area. But if you think about what has to happen now for some woman to get on a PARP inhibitor who presents with advanced breast cancer, they have to be identified as eligible, be offered it, be able to afford it, have informant, you can read it. So think of the number of steps that have happened and have to happen and understand why we see that growing gap in disparities. I guess I want to suggest that it often feels like that may be the way it has to be, that it's just going to get more and more complicated. But there's some incredible examples where it's gone the other way. And I would push us all to try to think that way as we create our tools. So when I was training, this is how we used to do cervical cancer prevention. So this is a PARP smear algorithm. I'm sure you guys can't read it, right? I couldn't read it. I still couldn't remember it. My husband's a G1 oncologist. And so my solution to an abnormal PARP smear was to call my husband. Because it turns out, like, I can't figure this out. But if you think of what has happened, right, is that we've actually moved science to the point where not only did we start to understand the cause of cervical cancer, but we're developing interventions that come more from that phase three of medicine, the phase of simple and socially acceptable, socially available interventions. Most say it's acceptable here, but simple and actually able to be delivered. And so as we move forward in genomics and addressing equity and genomics, I guess I wanted to leave you by saying that I think we have made unbelievable advances in developing genomic applications that are effective across groups. Those advances are only going to continue with the leadership of people here in this room. I was joking. I was telling somebody, I can't remember, that I spent a good bit of time talking to folks at my institution yesterday to Mark Daley and some of the other geneticists there about upcoming plans. And the biggest discussion at that time was one of the major areas of focus was about inclusion. How are we going to be changing some of the tools that were being developed into think about inclusion? And so that is a huge win in where the tools are going to go. I also think that if we have those tools, we're going to be able to use them to actually reduce disparities in care by making better clinical decisions and taking bias and stereotypes out of the exam room. And that genomic equity means that we have to develop access to these. We have to understand and then address the barriers in affordability, acceptability and availability. That we need to begin a research agenda that is actually taking genomics now, which has become part of medical care and actually defining and implementing what it means to do genomic quality of care. What does that mean just like we did for oncology care some decades ago? And then I personally believe that as we move forward, it's going to be critical to keep our eye on this complexity issue. If we make medicine so complicated and so difficult, then I think we are never ever going to go back to being able to bring those disparities together. And my hope is that this room is part of who will change that. So I want to thank everybody for having me. Thank everybody who's funded me over the years and I'd be delighted to take questions. We have microphones in each aisle if people will come up to a microphone to ask their questions. And let me start actually at a couple of things. I just want to clarify one thing. You're using the phraseology tools, but I think you mean more than just the technologies, the experimental methods. It's sort of the resources. Application. The knowledge bases. The whole kit and caboodle, right? Yeah, yeah, absolutely. Although I mean, I guess in the end, I would say that if I'm seeing a patient, I have a slew of things I can offer to try to make that patient better or keep them healthy. And I think there's an enormous amount that comes out of the knowledge that's created from these large investments and all the rest of it. But in the end, what I want to be able to do is to have tools or have applications, have things I can use that are going to be, they're going to work for people irrespective of their background. And that includes therapeutics. Absolutely. Yeah, I just want to clarify, because sometimes in the genomics world, when people say tools, they think just about the experimental, the sequencing methods, the checks, but you mean a much broader conceptual definition. Absolutely. No, I thought so. Yeah, Tom? Hi, thanks, that was a great talk. A little louder? Sorry, hi, great, thanks. That was a great talk. I started my medical career knowing a guy named Bob Guthrie, who actually at the beginning didn't trust physicians to implement this. And his thoughts was to go back to phase three and make it a public health issue by just testing all newborns. Yeah. Well, genomics and the tools of genomics has that opportunity, especially, because I noticed the focus of your talk are the common, if you will, epidemiologically approachable diseases. And if rather than in complexity, if you'll humor the heritability issue of rare disease, large numbers of rare diseases versus common polyfactorial disease controversy, which is still going on. Yes. If we get to a large number of rare diseases, this basic idea is gonna start to break down again because having practiced a rural America myself, primary care physicians do have tough time with rare diseases. So could you just mention what your thoughts are about taking this back to phase three and doing sort of prenatal or neonatal genome-wide testing? So, I mean, for some of the people who know me in the audience, I think one of the things that I just will never understand and probably never be able to do anything about is the division between our public health and primary care systems, right? And so the reality is that I do think you're exactly right is that genomics in particular, we can make it crazy complicated. And that doesn't mean the rare disease, we see patients, we have a lot of undiagnosed rare disease use of a sequence and clearly there's a role that you could call that complicated. But if you're talking about this other using it for common diseases, as you're talking about, there's a tremendous amount of beauty almost and imagining that genomics for all this complexity could be the place that actually brings healthcare back to public health. And merges these two pieces. The reality is that our public health system is underfunded. And so the reason we don't do that is that hospitals, like I'm probably going on record, so I'll be a little careful here. Let's just be a little careful here. So the money is held places that keep that money for lots of good reasons because those places believe they're doing really important things with that money. And so I think for us to be able to do that is going to take some major change and understanding how we can move it back. But I think that's an, I think it's a, absolutely to move it back in that model would make a ton of sense in addressing the disparities issue. The other thing I wanted to ask you, it strikes me the other level of complexity, which you gave examples of certain population groups, but the disparities and the equity issues are going to vary among different subgroups. And how do you, how do we dissect that out and sort of approach that knowing that what might be a disparity for African-Americans might be very different when you go to rural population might be very different than when you go to Hispanic groups and other groups. Yeah, no, and I think there's been tremendous work in trying to recognize in some ways that all healthcare is local, right? And so to some degree the role of the NHGRI in thinking about developing these applications, developing therapeutic, wherever we're gonna go can't be owning all of the ecosystem at the individual or at the community level because there has to be a way that that can be adapted. I guess I would say, so the theory that underlies some of this focus on complexity is actually something that's been written about called the Fundamental Cause Theory. And actually it would argue actually differently, Eric, which is that actually the one thing that we might be able to change that could actually address some of the problems in all of these groups is actually how we think about making the interventions more public and more available and simpler. And so it's an interesting, it's not clear that would work, I wanna be careful, but I think as we think about it, if we think about what's going on in cervical cancer, I'll just take that for a second because my husband spends a lot of time working on cervical cancer in African countries. And the reality is that if you go to a lot of those countries with very little infrastructure, that deciding you're gonna start wide complicated screening programs, whether it's rural or socioeconomic or racial barriers is really hard. But the concept of being able to think about delivering vaccination is a different ballgame. And so I think as we think about the tools that we develop is just trying to hold ourselves to those types of things, recognizing they've still gotta be put down into the reality of the individual community. Let me press you on that, but is genomics different? So I mean, if you look at other tools and other interventions where you might, there might be some commonalities so that you could reduce disparities across the board, is there something fundamentally different about genomics? And some people, the way they either fear or adverse to genomic interventions. Well, so I, you know, I spend a lot of time studying that. I don't know, people in the audience will have much stronger feelings about this than I do. So I wanna be a little bit careful about that one. You know, I will say the probably, I'll tell you a story. Maybe this is the best way that I can share this is I once did a study that was supposed to be about the acceptability of genomics in Western, in West Philadelphia. So we actually had this whole plan. So we were working in beauty shops in West Philadelphia. And so we did a bunch of focus groups and beauty shops. It was all about how we were gonna roll up genomics in African-American women in West Philadelphia. And we would show up with our, you know, our focused group stuff and all the rest of this kind of stuff. And to a person, all of the women would say something along the lines, look, I got like way bigger problems than that. Like, I can't pay my rent. Like, Comcast just took my cable. That was usually like number three. And so I think there is concern about genetics and there is differences about genetics. But I think if you spend a lot of time in communities, I think a lot of the reality, as you go back to that slide I showed, most people think it can be good. We can make it for the force of good. And if I was just talking about spending time in the Rosebud Reservation in South Dakota, it is not, there are many other problems that are bigger in communities than this. And so I think thinking about that makes sense. I do think we can create tools in genomics that are simpler. And I think we should push ourselves to do that. And I think that that is something that has to happen. Hi. I guess I'm focusing on your optimism and what I'm hearing, though, is that the disparities are growing larger. Nationally, we're seeing greater rather than lower segregation across racial lines. And so the access question is becoming bigger and bigger and people simply aren't getting access. In fact, more and more people are being recently deprived of access. In ivory tower situations on the NIH campus at MGH, where you can deliver reasonable services, that one thing is going on. But in the rest of the country, I'm wondering about the optimism. So I came here and spent some time at lunch with the trainees because I usually say that the one thing that I find that keeps me optimistic is if you spend time with young people who are gonna fix it, right? Whether it's my kids, the people in this room, my residents, the medical students, and the rest of it. I don't wanna disagree that we're at a time in our country where I think we could either become complacent about the disparities that we're experiencing or we could decide to prioritize them as a way of who we wanna be and what we do. I don't disagree with that. Am I optimistic that we are going to prioritize them? Yes, I am absolutely optimistic about it. But I think the reality is that there's huge inequities across place and I think trying to understand how you break down those inequities and that's partly why I focus most on the availability data. I think often we sit in our world and we see that we're okay where we are and we forget about the rest of the country. And so, I mean, I'll be concrete. We just talked to one of the students or one of the post-bacs who's here. I think it's our responsibility to do something about that. And so part of the reason that I'm optimistic is that I do think people are starting to stand up and try to do that, whether locally or elsewhere in those ways. I also think that we are making progress in healthcare, and so we have more tools to offer and the changes are happening very quickly. So I don't wanna be polyanist about where we are today in terms of the inequities, but I personally believe that there is a commitment and we are gonna face that commitment to trying that we will come together to take that on. Last question. Hi, I have a question. So I really admire how optimistic you are about genomic medicine and I think in our society today regarding this topic, we've grown curious and started companies like 23andMe where people will invest $100, $200. There are some disparities there though where some people can't afford to spend $100 on that curiosity. And I think this has kind of initiated this curiosity of genomic medicine. So I guess you can't really predict what the potential of simplicity of the medicine will be as far as like a kit, but how do you think you would go about convincing people or having people volunteer in this incentivized system of if you volunteer your data, we will secure it as far as like property rights of your own data, or how is it going to contribute to this quality goal? Yeah, so one of the complexities that we're sitting in genomic medicine and I don't know, Eric, where this is gonna play out in your strategic plan but I think you're hitting is this interface between the private sector and where the testing is being done now. And I would say the benefits of those data for discovery, for understanding, for research, and the reality that the private sector is not going to be driven to address disparities, to really think about these groups. And so usually when you say, so that means in some ways what you're articulating is that you're going to have what we would call a market failure, right? So if we move genomic testing to the private market, where it is run by private companies, then by definition it will not be accessible to people who can't pay for it. It might, if you can get competition down to 99 cents, I think on Amazon Prime Day, did you guys see that? There was like ancestry for $49 on Amazon Prime Day. But so the reality, so you are exactly right. And so the reason when you have that, and maybe I'll push Eric on it, but the reason is if the private sector takes over what needs to be a public good, then there has to be some other form of either governmental or other intervention to ensure that that public good is actually made, is leveraged for vulnerable and other groups. Does that make sense? Because that is not the job of the private sector. They're not going to do that. And so that's why we have governmental organizations and other groups. I bet, one of the challenges we have in disparities, and I don't know if I say what you would say about this, is that there aren't very many advocacy groups. You know, if you do breast cancer research, there's like a million advocacy groups, right? You do whatever, there's advocacy groups. And so part of the challenge is that the public sector responds to advocacy. And so for us to come together, I think we have to get to congressmen to, am I getting a little political here? But you guys need to get advocacy for what you're talking about, because it's not going to get solved by the market. Does that make sense? It just won't. But that doesn't mean it shouldn't be solved, and that's got to come from the public sector. Does that make sense? But all the trends say that the bulk of genomic data will be generated by healthcare, which probably means by private companies or hospital systems or medical systems by 2022. And there's no way that trend is going to be reversed. So that's just the reality we're going to live with. Yes. Yeah, I think, I don't disagree. We could argue about whether healthcare is in the private or public sector. But I think the reality is it's not going to come from research testing. It is absolutely going to come from this other type of sector. But that's why you need intervention to ensure that it actually reaches, one of the examples in HIV was the Ryan White Act that basically allowed people to get access to HIV treatment who had no other way to do that. So trying to understand how to advocate for that is critical. OK, at the end of the time, thank you very much for what we knew was going to be spectacular talk. Thank you.