WEBVTT 00:00:08.000 --> 00:00:10.000 >> Recording in progress. 00:00:10.000 --> 00:00:24.000 This event will be live captioned. 00:00:24.000 --> 00:00:54.000 Okay. We will get started in just a moment. 00:01:07.000 --> 00:01:37.000 All right. So I am delighted to welcome you today to ELSI Friday Forum. 00:01:39.000 --> 00:02:01.000 I am Sandra Soo-Jin Lee, Professor of Medical Humanities and Ethics at Columbia University, and co-director with Mildred Cho of the Center for ELSI Resources and Analysis. For those of you who are joining us for the first time, ELSI Friday Forum is a monthly webinar that we organize through CERA, which also provides resources and other events through the website ELSIhub.org. 00:02:01.000 --> 00:02:22.000 CERA is funded by the National Human Genome Research Institute at NIH. Today's session is our 50th forum, in our ELSI Friday Forum series, and it is entitled: Rigor, Reproducibility, and Responsibility: ELSI Questions in Population Data Practices. 00:02:22.000 --> 00:02:47.000 Now, I'm going to give you some logistical information about closed captioning, the Q and A feature, and the chat, for this webinar, and that can be found actually in the chat, in terms of how to access and to submit your questions today. 00:02:47.000 --> 00:03:12.000 But before we get to today's discussion, I have a few quick CERA announcements. The first is that the American Journal of Bioethics now welcomes proposals for target articles related to ELSI Friday Forum topics. Including the topic today. And those explored in our Future Forums. 00:03:12.000 --> 00:03:31.000 In the chat, you will see a link to submission instructions, and the monthly deadlines for that opportunity. The ELSIhub website has many resources related to today's panel. A compiled list of today's resources will be published on ELSIhub, and a link to this page will be provided in the chat. 00:03:31.000 --> 00:03:52.000 And then finally, I want to encourage each of you to join the ELSI Scholar Directory, and you can also sign up for our newsletter and connect with us on LinkedIn and BlueSky. And there is information in the chat about that. 00:03:52.000 --> 00:04:09.000 I also want to recognize that this session is being held as a Zoom meeting, so we can all see one another and connect via the chat. When you have questions, however, please use the Q and A feature, and keep yourselves muted for the duration of the event. 00:04:09.000 --> 00:04:39.000 When we transition to the Meet the Speakers Afterparty at the top of the hour, we will cease the recording and invite you to turn on your camera and participate in the informal part of our session today. 00:04:41.000 --> 00:04:55.000 Now it is my distinct pleasure to introduce our moderator for today's session. Full bios for all our speakers will be put in the chat. So I'm just quickly going to introduce Dr. Aliya Saperstein. Dr. Saperstein is a professor of sociology and the Benjamin Scott Crocker Professor in human Biology at Stanford University. 00:04:55.000 --> 00:05:18.000 She served as a member of the National Academy's consensus study that is a focus of today's forum, and I am absolutely delighted to hand it over to her to introduce today's topic. Aliya? 00:05:18.000 --> 00:05:42.000 >> Thanks so much, Sandra. I'm very excited to be here for today's discussion. As Sandra said, this ELSI Friday Forum is inspired by the recent National Academy's consensus report, Rethinking Race and Ethnicity in Biomedical Research, which was first released to the public in late October, and officially published at the beginning of this year. 00:05:42.000 --> 00:05:55.000 The 2025 report was sponsored by the Doris Duke Foundation and the Burroughs Wellcome Fund. For background, the official statement of task for the report required both assessing the current use of race and ethnicity in biomedical research and providing recommendations to guide the scientific community in their future use. 00:05:55.000 --> 00:06:25.000 This included identifying circumstances in which it would be appropriate to use race and ethnicity in research, and circumstances in which race and ethnicity should not be used. 00:06:27.000 --> 00:06:56.000 As Sandra said, I had the pleasure of serving on the committee that produced this report, along with 15 other biomedical scientists, physician and nurse scientists, epidemiologists, and social scientists. Our team included experts in law, anthropology, ethics, and clinical informatics, and we assessed practices across a range of research contexts, including race correction, medical devices, 00:06:56.000 --> 00:07:14.000 secondary data analysis, and clinical decision making tools, such as the expanding use of artificial intelligence in clinical algorithms. Our work also built on the efforts, conclusions, and recommendations from the 2023 NASEM report on Using Population Descriptors in Genetics and Genomics Research. The 2025 report is both broader and deeper, as suggested by its title. 00:07:14.000 --> 00:07:33.000 It is broader in aiming to provide guidance for all biomedical researchers, not just geneticists, but deeper in focusing solely on the use of race and ethnicity, rather than on descent-based descriptors more generally. 00:07:33.000 --> 00:07:52.000 Despite these scope differences, there are also important similarities between the two reports. For example, both stress the need for not only more transparency in decision making, but also more accountability for implementing what are clear consensus best practices. 00:07:52.000 --> 00:08:19.000 The Rethinking Race and Ethnicity includes 9 recommendations aimed at a range of audiences, from biomedical researchers to journals and funders, and our panelists will be digging into some of the specific guidance during their presentation. 00:08:19.000 --> 00:08:38.000 A key theme across the recommendations is the need for deliberate and sustained consideration of the use of race and ethnicity at every stage throughout the research process. That is: The decisions are not simply semantic, terminological tweaks that researchers should make to satisfy journal guidelines at the time of publication. 00:08:38.000 --> 00:09:08.000 They also involve carefully scrutinizing the existing evidence base, having a principled scientific rationale for either use race/ethnicity or not, and carefully weighing the potential implications, limitations, benefits, or harms of the research design or chosen dataset. 00:09:10.000 --> 00:09:40.000 Importantly, both NASEM reports extend decades of concerns, critique, and scholarship on this topic. When I was in graduate school, 20 years ago, I devoured work by the likes of Audrey Smedley, Thomas LaVeist, Evelyn Higginbotham, Tukufu Zuberi, and many more, who challenged me to think critically about when, why, and how to use the concept of race in my research. 00:09:45.000 --> 00:10:07.000 Over the years since, as I and countless others have worked to change research practice within and across our disciplines, progress often seemed frustratingly slow, and sometimes entirely elusive. Suddenly, in 2020, addressing the use of race and ethnicity in research gained renewed urgency in the midst of the pandemic and in the wake of worldwide protests against police brutality that galvanized greater public and scholarly attention to structural racism. 00:10:07.000 --> 00:10:34.000 Now, just as suddenly, the political pendulum has swung dramatically in the opposite direction, with federal research funding being slashed on banned topics such as DEI and health disparities, and the launching of lists of "trigger" words, including race, ethnicity, and racism. 00:10:34.000 --> 00:11:04.000 It is amidst this upheaval and uncertainty that we meet today to discuss both the importance and the implications of these two NASEM reports, to consider what constitutes good science in this space. Responsible, rigorous, transparent biomedical research. And how we can continue to make progress in our population data practices. 00:11:04.000 --> 00:11:07.000 Here to join me today in this conversation is Gen Wojcik, an associate professor at the Johns Hopkins Bloomberg School of Public Health in Baltimore, Maryland, and Annette Flanagin, the executive managing editor and Vice President of editorial operations for the Journal of the American Medical Association, the JAMA Network, and executive editor of JAMAevidence. 00:11:07.000 --> 00:11:33.000 I'll now hand it over to Gen for her comments. 00:11:33.000 --> 00:12:03.000 STUDENT: Thank you, Aliya. Let me share my screen real quick. Where is it? Here we go. Okay. I was also on the committee with Aliya, and I'll be giving a brief set of remarks about the report, as it relates to the rigor, reproducibility, and responsibility of our research, as well as some thoughts of my own at the end. 00:12:05.000 --> 00:12:21.000 Okay. So there are two complementary resources, as Aliya said before. The report that was released in 2023, as well as one released last year in 2024. The 2023 report really is grounded in genetics. And so the exposure of genetics and making decisions about descent-based descriptors -- if you are doing genetic research. While the newer report comes at it from a different angle. 00:12:21.000 --> 00:12:51.000 Where you're not necessarily looking at genetics, but rather any biomedical research. Whether race or ethnicity or these descent-based descriptors are looking at -- as an exposure, as sort of a confounder, much broader ways of looking at things. 00:12:51.000 --> 00:13:07.000 So they're both available. We have the QR codes here. As well as in the notes somewhere. So there were four major considerations that the committee had. And on the right hand side here, I'm showing you the process, the biomedical research process. And the points along that process in which different things need to be considered, when relating to race and ethnicity in biomedical research. 00:13:07.000 --> 00:13:33.000 And this goes from the conception all the way through to study design, data collection, and then to reevaluation. And the point of this cycle is that it should always be iterative. It should always come back around, and past work should inform future work. 00:13:33.000 --> 00:13:56.000 With, at all stages, this consideration of community engagement and partnership. As well as study methodology, inclusion and equity, and the overall consideration of use of race/ethnicity. I'm gonna focus on two parts of this, which is assessing the use of race and ethnicity, as well as evaluating datasets and study methodology, just to focus on something with these short remarks, as well as -- again, we're focusing on rigor. 00:13:56.000 --> 00:14:14.000 And reproducibility with this. So the first part is the assessing of use of race/ethnicity. This is the full details in the report. This is recommendation number 4. There is a table, table 6-1, which is listing the race, ethnicity, and the associated concepts, which really helps people dial down into what they're actually asking. 00:14:14.000 --> 00:14:43.000 And there is some background in chapter 5. And the key questions for this -- part of it is saying: What is the purpose of race/ethnicity? How do these actually relate to health? Not necessarily as a proxy, but what are you trying to proxy? 00:14:43.000 --> 00:14:57.000 Getting to the root causes or the root variables you're thinking about. Now, this can be things such as: You see a few things listed here as examples, where there's structural racism, social determinants of health, genetic markers, ethnicity, immigration status, different biomarkers, indigeneity, and so these are all very different concepts that you might have, that differently react with the research question of interest. 00:14:57.000 --> 00:15:23.000 And so the point of this is to analyze the data, to tease apart these underlying mechanisms. Right? So again, not to rely on race or ethnicity categories as a proxy, but rather to dial down into what you're actually trying to ask in your study question. 00:15:23.000 --> 00:15:47.000 Now... The other component here is evaluating datasets and study methodology. And this really goes to the reproducibility of your work. And so how do you explicitly define race/ethnicity? How it's collected. And these key questions. Right? Which is a source of the data. How is it collected? What are the limitations of this? If it's secondary datasets or legacy data, what does that mean? 00:15:47.000 --> 00:16:01.000 What are the biases? If it will be used in the analysis, how and in what context? And so... This is really important for authors to think about, when putting it into their work. And this is also -- even if race is not an input variable in the analysis, that if there are subgroup analyses that are conducted, to report them to assess fairness and bias in performance. 00:16:01.000 --> 00:16:17.000 And the full details for this are in recommendation 2 and 3, as well as box 6-1, which is questions for the researchers to consider. And then there's an entire -- well, two chapters, really, to focus on this, if folks want more details on this. 00:16:17.000 --> 00:16:33.000 So these were the recommendations from the committee, and they basically boil down to: What are you trying to capture with race/ethnicity? And if you are trying to capture something specific, that you should measure that and put it into your research. 00:16:33.000 --> 00:16:52.000 Now, we know this is not really easy all the time, given legacy data. But the idea and the intent is to understand how race or ethnicity may or may not be appropriate variables with the research question to do better. 00:16:52.000 --> 00:17:13.000 Now, one of the things that I think is important to know for this is that this is not just a thought experiment. As Aliya said, this is not sort of semantic, about what the right words are to use, although that is also very important, but also, there are real world consequences for our work. We don't sort of work in a vacuum, in which everybody takes everything at face value. 00:17:13.000 --> 00:17:38.000 And we all know sort of what people are talking about when they say things. And this is sort of... This confluence of both eugenics and scientific racism that is on the rise, currently. This can be... These scientifically erroneous and immoral theories of "racial improvement", with this idea that you can eliminate social ills through genetics and heredity. 00:17:38.000 --> 00:17:48.000 And then at the same time, you have the scientific racism that argues for superiority of white Europeans and the inferiority of non-white people, and when you get them together, right now what you see is that these forces have combined, and a lot of groups now support the sort of contemporary xenophobic ideas, antisemitism, sexism, colonialism, and imperialism, in the United States. 00:17:48.000 --> 00:18:18.000 And so we're seeing this right now. So what does this look like in practice? I think we can all think about, in theory, what could happen. But it's actually happening, and I think it's important for us to highlight that. 00:18:18.000 --> 00:18:39.000 So one thing that has come up in recent months is the idea of race as biology. This is an example from a confirmation hearing of RFK junior, with an exchange with Angela Alsobrooks, in which RFK junior had previously stated that Black people should have a different vaccine schedule than white people, because their immune system, quote-unquote "is better than ours". 00:18:39.000 --> 00:19:04.000 So there's a debate about what vaccine schedule should Angela Alsobrooks receive? And he cited this article, the Poland Article. So we can look up that article, we can see what that article is, to see where he's getting those conclusions. And here we can see this is an article from 2016. It's fairly recent. 00:19:04.000 --> 00:19:21.000 From a group, from Gregory Poland's group at Mayo. And what you can see here is -- looking at just the title and sort of a perfect storm for misinterpretation of what they did -- as well as some decisions that they made in the processes, which didn't help. One is this idea of genetically defined race here, where you have these groups that are racial categories. 00:19:21.000 --> 00:19:34.000 That are apparently genetically defined, which is not possible. As race is not genetically defined or biological. But then you see that it's associated with responses to vaccination, in a case... This was a very controlled experiment. Right? 00:19:34.000 --> 00:19:49.000 So you have both this intersection of these improper usage of genetics and race in the article, which can lead to what we are now seeing as scientific racism in policy. 00:19:49.000 --> 00:20:11.000 And as we increasingly move towards this world of AI and looking at data driven approaches, I think it's important to take a step back and say: Okay. But driven where? Just because it's data driven doesn't mean it's driving us to a good place. 00:20:11.000 --> 00:20:32.000 And that these groups cannot be a better solution if the system and the data themself are actually faulty to begin with. And this is true, because many times we rely on social constructs for data collection and groupings. And this bleeds into the AI algorithms. Right? These algorithms work with what they have, and that can result in issues. 00:20:32.000 --> 00:20:49.000 We can see this in real work. This is real data. This is a real published data on the top. It's a GWAS. And what they did is they did genetic clustering to find these groups. But if somebody had more than one group, they were likely to be more than one group, they had maybe mixed ancestry, they had a hierarchy in which they classified people. 00:20:49.000 --> 00:21:09.000 In which they arbitrarily assigned African over European, Hispanic over Latin American, et cetera. And what this does is it gives a veneer of this datification, this idea... Oh, it's algorithmically derived. Therefore it can't be racist. Which is just not true. 00:21:09.000 --> 00:21:27.000 Because that hierarchy had to be decided by somebody. And what it does is actually... We can look at sociological theory and find these two axes of subordination here, in which this arbitrary assignment directly follows these axes, as proximity or being distal to whiteness here. Right? 00:21:27.000 --> 00:21:28.000 And so you're seeing that you're recapitulating these rules of hypodescent. So that's all I have today. I'm a little over. I apologize. The PDF is able to read down at the link, and there's a lot of cool interactive materials there that would be great to look into. 00:21:28.000 --> 00:21:35.000 Okay. Thank you. 00:21:35.000 --> 00:22:05.000 >> Thank you, Gen. And next up is Annette. 00:22:17.000 --> 00:22:36.000 >> Thank you, Aliya. Gen, that was fantastic. Thank you. I'm going to share my screen. Okay. How is that? Good? Thank you. Okay. Great. Thank you, Sandra, Aliya, very much, for inviting me to do this, and to be taking part in this forum with Gen. 00:22:36.000 --> 00:22:59.000 I'm already learning so much. I am going to be representing someone who is a recipient of the guidance of the two NASEM reports. And to talk to you a bit about how they are aligned with some guidance that we've put together from the JAMA Network journals and the AMA Manual of Style. 00:22:59.000 --> 00:23:15.000 These are my disclosures. I'll just have this up for a minute, so you can quickly see that. So... I will start with the 2025 report, and how it is aligned with recommendations that we have put together. 00:23:15.000 --> 00:23:27.000 I'm gonna focus specifically on recommendation number 8, which you can see here is directed to funders, sponsors, publishers, and editors. And I'm going to be talking specifically as a representative of a publisher and an editor. 00:23:27.000 --> 00:23:39.000 There are three parts to this. That we, as publishers and editors, should provide consistent guidance to assist researchers in developing and examining their work and publishing their recommendations. 00:23:39.000 --> 00:23:54.000 That we should require researchers to provide a scientific rationale for their use of race and ethnicity, describe the data provenance, and acknowledge limitations of their use. 00:23:54.000 --> 00:24:11.000 And we should also provide reviewers with specific guidance for reporting race and ethnicity on the papers that they are examining. And then lastly, that we should periodically evaluate our policies to see how we can improve them. 00:24:11.000 --> 00:24:37.000 As things change, as norms change, as understanding and awareness changes. So in 2021, we released updated guidance on reporting race and ethnicity in medical and science journals. 00:24:37.000 --> 00:24:58.000 This guidance took 18 months of updating review, posting for public comment, receiving the public comment, revision again, and then finally releasing this again in August of 2021. This guidance is freely available on the JAMA website and in the AMA Manual of Style, which is a stylebook -- a very large book -- that is used by many medical and health journals. 00:24:58.000 --> 00:25:17.000 These are the key principles of the guidance. First, an acknowledgment that race and ethnicity are social constructs, which I'm sure you all well understand. However, the terms may be useful as a lens through which to study and view health equity disparities and racism in healthcare. 00:25:17.000 --> 00:25:41.000 Second, we acknowledge the historical and current sensitivities and controversies related to language used to describe race and ethnicity. And we recommend understanding how important it is that researchers and authors are clear, precise, fair, and consistent in the terms used. 00:25:41.000 --> 00:26:07.000 Third, we acknowledge that race and racism do not exist in isolation, and therefore racial and ethnic descriptions should not be reported alone. Other sociodemographic factors and contexts should be included as much as feasible. And that these variables are key to the study of situations of disparities and inequities in health and healthcare. 00:26:07.000 --> 00:26:26.000 Fourth, in research articles, we acknowledge that it is very important for determining who classified individuals by their race and ethnicity, what categories are available, and how these determinations were made with a reference for self-determination. And finally, our guidance, like many guidance, is not final. We continue to collect feedback, and will update the guidance as needed. 00:26:26.000 --> 00:26:53.000 We are deep into, right now, updating our guidance on the reporting of gender, sex, and sexual orientation. And that will be forthcoming. These are the components of our guidance. It includes definitions of commonly used terms associated with race and ethnicity. 00:26:53.000 --> 00:27:08.000 With acknowledgment that many of these terms have changed over time. That some are out of date. And that the nomenclature is going to continue to evolve. The guidance acknowledges concerns and controversies in healthcare and research, including the intersectionality of ancestry and heritage, social determinants of health, and other socioeconomic, structural, institutional, cultural, and demographic factors that need to be attended to. 00:27:08.000 --> 00:27:24.000 We have guidance on the use of collective or umbrella terms for racial and ethnic categories, with preference and terms to avoid. We have guidance on capitalization, abbreviation use, and alphabetical order. 00:27:24.000 --> 00:27:44.000 As well as recommendations on usage for geographic and regionalization considerations. And examples are included to help guide authors, and we continue to collect feedback and provide more examples. 00:27:44.000 --> 00:28:04.000 This is a summary table that we have developed, that I'll briefly go over now. Not all of it. Just to give you sort of a sampling of some of our guidance. Some of it is really quite simple. But with important rationale for it. This guidance is freely available on the JAMA website. 00:28:04.000 --> 00:28:30.000 Our guidance, as originally published, recommended capitalizing all race and ethnicity terms that were based on geographic terms, like Asian or African American. But like other style manuals at the time, initial capitalization was not applied to the terms "black" and "white". 00:28:30.000 --> 00:29:00.000 Due to many concerns about racism and biases and things we were seeing specifically in this country, at that time, we revisited this guidance and determined that the most fair, sensitive, and consistent approach would be to capitalize all race and ethnicity terms, including the capital B in Black and the capital W in White, except of course when such capitalization could be inflammatory or inappropriate. 00:29:01.000 --> 00:29:23.000 We recommend avoiding use of the forward slash, otherwise known as the virgule. This is the second row here, in talking about race and ethnicity. Many categories fit under that umbrella, and because studies in race and ethnicity are inclusive, that is, they don't separate ethnicities from race but include it as a single demographic variable, say in table 1, and because that virgule can be confusing, we now recommend the term or phrase 00:29:23.000 --> 00:29:44.000 race and ethnicity always. Our policy also recommends listing race and ethnicity categories in alphabetical order instead of order by prevalence. Multiracial would be included alphabetically, but vague terms like other and unknown would go last. I'll talk about other in a second. 00:29:44.000 --> 00:30:00.000 The guidance here is modified on our patient first language. In that we never address people as some characteristic or quality they possess. For instance, we don't say addicts or asthmatics or diabetics. In that same sense, we don't call people by their race or ethnicity and we say do not present these terms as nouns, but use them as modifiers. 00:30:00.000 --> 00:30:30.000 We also have some guidance to avoid using the term "mixed race", if possible. Unless of course that is the term that is used for data collection. Preferred terms include terms such as multiracial or multiethnic. 00:30:35.000 --> 00:30:57.000 The guidance recommends avoiding the use of abbreviations for race and ethnicity unless due to space constraints, say, for instance, in a table or a figure. Some abbreviations such as AA can obviously have multiple meanings, and others, such as AA and HPI may not be understood well by international readers. So we always recommend expanding them if possible, at least at first mention. 00:30:57.000 --> 00:31:24.000 A lot of topics where we received a lot of feedback when asking for public comment was use of the term "minority" and "minorities". They didn't all agree. So we now recommend not using "minority" as a noun, because it could be inaccurate or stigmatizing. But instead to partner it with a term or a descriptor such as racial and ethnic minority. 00:31:24.000 --> 00:31:46.000 However, some may not prefer to use that collective term either. And you can see here terms that can be used, such as underserved population, when referring to health disparities among groups, or underrepresented populations. And there's also good cause for use of the terms "minoritized" or "marginalized", as described here. 00:31:46.000 --> 00:32:06.000 Back to the term "other". We recommend avoidance of the term other, due to the connotations of otherness, and because it is unclear who is included. Thus categories included in this phrase, other, should be defined and reported, and authors are advised to be specific as possible when reporting on racial and ethnic categories. Even if these categories contain small numbers. 00:32:06.000 --> 00:32:36.000 And if they could, if the numbers are small, so as to potentially identify study participants, we ask that that be made clear. We're still seeing a lot of the use of the term "other", versus "additional categories", but we're hoping that more researchers and authors will be using different terms. 00:32:36.000 --> 00:32:55.000 Ideally, any statistical analysis of race and ethnicity should not be a dichotomy of White versus non-White. If race and ethnicity are valid comparisons to make, all categories possible should be included. However, if there is justification for comparison of one racial or ethnic category in a specific type of research, we just ask for that rationale to be clearly and carefully explained in the methods section. 00:32:55.000 --> 00:33:18.000 I think you are all familiar with the problems with biological explanations for healthcare disparities or inequities. And we have specific guidance on this. We recommend avoiding biological explanations for healthcare disparities or inequities between racial and ethnic groups. 00:33:18.000 --> 00:33:44.000 Given the fact that race is not a biological construct. And that associations between race and ethnicity and health outcomes may intersect with ancestry, heritage, social determinants of health, as well as other socioeconomic, structural, institutional, cultural, and demographic factors. 00:33:44.000 --> 00:34:00.000 There are concerns with some of the age old use of clinical algorithms, such as inclusion of race in estimations to estimate glomerular filtration rate for kidney functioning, that can perpetuate biases, discrimination, disparities, or risk scores such as the Framingham risk score and genetic studies based on databases that primarily include populations of only one particular region. Most commonly European descent. 00:34:00.000 --> 00:34:10.000 In such cases, we urge authors to use caution in interpreting and generalizing such findings. Much of what I've just given you is summarized in our instructions for authors. Which is fairly detailed. I'm giving you a summary here. 00:34:10.000 --> 00:34:32.000 There's a link here, and I'll be sure that my slides are available to participants after the meeting. So you can see this is just a summary of what I've just gone over. 00:34:32.000 --> 00:34:53.000 So now to the 2023 NASEM report, which, as has been mentioned, was focused on using population descriptors in genetics and genomics research -- following the release of that report, Greg Fiero, who was associate editor at JAMA, pulled together editors at JAMA and six genetics journals. 00:34:53.000 --> 00:35:16.000 And we came together to issue consensus 10-point recommendations in the form of guidance for authors and reviewers of manuscripts, that include use of population descriptors as proxies for genetic ancestry groups and some things to think about in that area. 00:35:16.000 --> 00:35:46.000 These 10 points, these 10 recommendations, many of them are extensions of the basic guidance for reporting race and ethnicity, which I've just gone over. However, several are specifically directed to genetic research. And you can see point number 2 here, that race and ethnicity should not be used as proxies for genetic ancestry groups. 00:35:46.000 --> 00:36:05.000 Or to represent the genetic diversity of study participants, and are not recommended for use as analytic variables. And I'm sure many of you are familiar with this. I will actually summarize the last two points that I think, in our guidance for authors, are important. And number 9: When using legacy datasets, it may not be possible to derive and use more accurate measures of genetic ancestry groups. 00:36:05.000 --> 00:36:32.000 And in these instances, authors need to address and explain the limitations of their datasets. And point number 10, which might seem a little prescriptive, but we think this is important: That generally titles and conclusion sections should avoid the inclusion of race and ethnicity as population descriptors for genetic ancestry groups. 00:36:32.000 --> 00:36:59.000 Because they can clearly be misleading. As you saw with the article on the so-called genetically defined race that Gen Wojcik just mentioned in a few of her other examples. Of course, such usage may be entirely appropriate for studies of historically isolated populations or those about healthcare disparities. Where the social constructs of race and ethnicity are important contributors to the outcomes. 00:36:59.000 --> 00:37:10.000 So how can journals assess the extent to which policies such as the NASEM statements are being followed? Well, we can publish research that assesses the quality and completeness of reporting. Here you're seeing a graphic summary of the findings of a cross sectional study of all randomized clinical trials that were published in three medical journals. 00:37:10.000 --> 00:37:26.000 JAMA, the Lancet, and the New England Journal of Medicine in 2015 and 2019, a comparison determining the changes in reporting policies for race, socioeconomic status, and sex. 00:37:26.000 --> 00:37:45.000 This included 688 studies. Approximately half of the studies reported race over all. 48% in 2015 and 51% in 2019. But less than 15% reported on SES. On socioeconomic status. 00:37:45.000 --> 00:37:58.000 While more than 98% reported on sex. And if we looked at age, we would see something very similar here. Of the studies that reported on race, demographic characteristics, representation of White participants was the highest, as you can see in this graph. 00:37:58.000 --> 00:38:19.000 These authors concluded that limited progress has been made in reporting and representation of race and socioeconomic status within medical research during this time frame. This is up to 2019. 00:38:19.000 --> 00:38:47.000 And we have quite a ways to go. So what else can we do? Well, we as editors can actually conduct our own research. We published a study that compared race and ethnicity reporting in research articles published in three of our journals before and after implementation of our updated guidance in 2021. 00:38:47.000 --> 00:39:10.000 And we found that reporting of race and ethnicity increased from 57% in 2019 to 67% in 2022. So still some improvement. But a ways to go. We also found that the proportion of articles that defined categories included in that so-called "other" category increased from 27% in 2019 to 70% in 2021. 00:39:10.000 --> 00:39:27.000 To 85% in 2022. So we're quite pleased with that improvement. We also saw significant differences after the reporting guidance was implemented for articles listing racial and ethnic categories in alphabetical order, versus order of prevalence, or order of majority. And for articles indicating how race and ethnicity was determined. 00:39:27.000 --> 00:39:50.000 So we were pleased with that. A higher proportion of articles still reporting a participant's age and sex or gender overall, and a much lower proportion reporting on socioeconomic status. So we have some ways to go in that area. 00:39:50.000 --> 00:40:08.000 This is what the New England Journal of Medicine is recommending, and is actually using. You'll see this is an example of a table. Authors are asked to report on the demographic variables of study participants. Sex, gender, age, race and ethnic origin, and geography, with a comment about the representativeness and generalizability of the study findings. 00:40:08.000 --> 00:40:38.000 To help improve the diversity of research studies. This table is included as a supplement in a published article, and it has to be referenced in the main article. And there are a lot of examples of this. And I think this is actually a nice solution. We're beginning to think about how we might do this in the JAMA journals. 00:40:43.000 --> 00:41:05.000 Maybe not have it in a supplement, but to include it in the main article. So back to the NASEM report: Recommendation 8 for providing guidance to peer reviewers, we added this required question in 2023. And since then about 5% of reviewers have indicated that there are EDI concerns. And regarding these concerns and the relevant comments that peer reviewers are providing both the authors and the editors, the most common comments are considering issues... 00:41:05.000 --> 00:41:21.000 Categories of responses include things like lack of clarity of race and ethnicity categories, statistical or methodologic concern, particularly with regard to why particular groups were included in an analysis, or not included, and ambiguity in language around demographic identifiers. 00:41:21.000 --> 00:41:39.000 We're studying this now to see how we can further improve this guidance for peer reviewers, and then also for authors. And with that, I will thank you all. And turn this back over to Aliya. 00:41:39.000 --> 00:41:56.000 >> Thank you so much for that, Annette and Gen. So we have about 20 minutes left for discussion. And I do also want to remind you to drop questions into the Q and A box, if you have them. 00:41:56.000 --> 00:42:26.000 We will do our best to get to as many of your questions as we can. And if for some reason you're not comfortable putting your questions in the Q and A box, please do stay for the afterparty, where we will have turned off the recording, and you can ask your questions, perhaps, more anonymously. 00:42:28.000 --> 00:42:33.000 But as we wait for questions to roll in, I have a couple that I wanted to ask both Gen and Annette, building off some points that they brought up in their presentations. Gen, you highlighted how current research practices can reinforce hypodescent norms like the One Drop Rule. The new NASEM report notes a related concern: How studies often ignore, elide, or exclude participants who identify as multiracial, multiethnic, or have mixed ancestries. 00:42:33.000 --> 00:42:56.000 Can you share with us what tools or recommendations the report offers around this issue? 00:42:56.000 --> 00:43:23.000 >> Of course. Thank you, Aliya. As you know, this is a topic that's near and dear to my heart. But I think a really important part of the report that sets it apart from different previous iterations of this topic is that there is an entire section on multiracial individuals and strategies that can account for what you're actually trying to ask. 00:43:23.000 --> 00:43:52.000 And so there's a table. That's table 5-1. In which it provides different categorization usage. And one thing to note is that there's no one right way. It depends on the research question you have and what part of the multiracial or multiple ancestries, identity, that you're actually interested in for your research question. And so one example would be to use an entire category of multiracial. 00:43:52.000 --> 00:44:07.000 But the problem with that is that you lose any sort of substructure or nuance for what that actually means in the contexts of both populations and their health. You can also do sort of this additive approach, of different identities. Or if the hypodescent aspect of it, with the least advantaged group, is the most relevant to your research question, maybe you're looking for different elements of structural racism or perceived race. 00:44:07.000 --> 00:44:23.000 Then these are aspects you can also look into. And so the important part, again, is not to just disregard these individuals completely. But to consider what they can contribute and how that factors into the research question at hand. 00:44:23.000 --> 00:44:40.000 This is very similar in some ways to the previous report of genetics, in which the push to not categorize individuals into these sort of genetic ancestry groups would lead to better inclusion of individuals who fall between these sort of really tight clusters. 00:44:40.000 --> 00:45:00.000 And this part, you're saying -- you know, it can be complicated. But just because something is complicated and hard doesn't mean we don't have to do it and don't have to have the same care and consideration for those participants as others. Thank you. 00:45:00.000 --> 00:45:15.000 >> Thanks, Gen. Annette, I wanted to build on you. There was some hope in your slides. At least, I picked up on them. You seemed to find some success with the latest JAMA guidelines that were put forth in 2021. 00:45:15.000 --> 00:45:45.000 As we all know, over the last probably two decades, there have been numerous best practice guidelines for reporting race and ethnicity that have been put out there, with relatively little improvement in kind of the specificity and transparency and rigor of published research. 00:45:45.000 --> 00:45:49.000 Could you tell us a little bit more about what you think the journal -- what did the journal do that seemed to make a difference? You showed us the kind of radio button question that was posed to reviewers, but that wasn't released until 2023, so probably doesn't account for some of the changes that you saw earlier. Maybe that will make it better. But can you share with us some more takeaways from the JAMA experience about assessment and accountability in particular, 00:45:49.000 --> 00:46:10.000 that other journal editors or actors in this ecosystem can take advantage of? 00:46:10.000 --> 00:46:25.000 >> Sure. First just to state the obvious. We were doing this in what I'll call the near before-times. When there was so much awareness going on about the importance of -- namely race and racism. And so that gave us sort of a cultural foundation to step on. 00:46:25.000 --> 00:46:43.000 When we released the guidance, we released it in two parts. First, we published it, and we asked for public comment and feedback. And we got a lot of attention for that. And I think that helped us improve the guidance and improve the acceptance. 00:46:43.000 --> 00:46:59.000 We published it in August of 2021. We made it freely available. It's not behind any kind of paywall. We did that table. And then we went on the road. The AMA Manual of Style is a committee of nine individuals, who are affiliated with the JAMA Network. 00:46:59.000 --> 00:47:14.000 And we went giving talks at professional societies for the Council of Science Editors, for example. The American Medical Writers Association. We gave a talk for the clinical editors of the Nature journals. So we were going international. 00:47:14.000 --> 00:47:29.000 We got invited to give talks for researchers and at professional societies. But I think that helped quite a bit. We did a lot of education and training of our manuscript editors. All of our manuscript editing is done in-house. 00:47:29.000 --> 00:47:41.000 And we trained them and we had listening and learning sessions for them to come back and tell us how authors were receiving this guidance. So there's always a negotiation. Any of you who are authors, your words are your children. And you don't want anyone to change them. 00:47:41.000 --> 00:47:54.000 So I think that helped quite a bit as well. And we're continuing that. This forum is an example, I expect. 00:47:54.000 --> 00:48:15.000 >> Thanks, Annette. This is a question -- so either of you could take this question. That building off what we were just talking about, about the importance of thinking about what you're saying at the time of publication... 00:48:15.000 --> 00:48:30.000 The new report is very clear that that's not the only time -- and in fact, that's often too late to be thinking about issues of race and ethnicity and research. So could you share with us why it's so important to be aware of concerns about the use of race and ethnicity? And population descriptors more generally? From the very beginning of a research project? 00:48:30.000 --> 00:48:50.000 Or put another way, perhaps more scientifically, why is it necessary, but not sufficient to pay attention to terminology and definitions when writing up results for publication? 00:48:50.000 --> 00:49:15.000 >> I can start. I mean, I think one of the sort of biggest points that I think about is that... You can't hack your way out of a lack of representation. Right? There's no statistical methods that you can use that will make up the fact that some people are just not included in biomedical research. 00:49:15.000 --> 00:49:31.000 And so... When we think about... You know... How you consider race, ethnicity, from the beginning... A good example of this is the All Of Us study, in which there was recruitment very early on with the sort of equity approach. As to who was underrepresented previously. And therefore they overrepresented those individuals, and those groups, in their studies. 00:49:31.000 --> 00:49:48.000 Now, you can say that... You know... That also is a necessary but not sufficient part of it. Which... What you saw with the All of Us paper that was published last year, and sort of the uproar, as to how these groups were defined, and labeled... 00:49:48.000 --> 00:50:05.000 But that is sort of why, from the beginning, and also... The point about it is that there is a lot of work that is done just because it's routine. Just because it's always been done to stratify or it's always been done to have these groups. 00:50:05.000 --> 00:50:27.000 And we've come to a point where both the data and the methods might not necessitate it anymore. The situations and the understanding that we had of these groups is past that sort of... Rote reaction to our analytical plans. 00:50:27.000 --> 00:50:32.000 And so thinking about it from the beginning can help you do better science. Right? To reach better answers for the questions that you have. And so it's not just about the sort of lip service and having the check boxes, but rather about doing better, more rigorous science, that's also more reproducible, because you've thought about these variables from the beginning. 00:50:32.000 --> 00:50:58.000 And then can go all the way through. 00:50:58.000 --> 00:51:23.000 >> Yeah. I would echo that. Speaking as an editor, and being on the receiving end of research, when the manuscript comes in the door, and does not completely or adequately address all the populations that were included... Does not provide granular descriptions. But lumps people into one or two categories. 00:51:23.000 --> 00:51:34.000 For which the findings are not necessarily generalizable to others. And so the likelihood of success of a research project, when it comes to JAMA, is greatly improved if all of the population descriptors, not only race and ethnicity, as I mentioned, are adequately -- at least reported. Now, not all of them can be analyzed. 00:51:34.000 --> 00:51:43.000 That's gonna depend obviously on the protocol. The study protocol. But at least are reported completely. And fairly. And consistently within that manuscript. 00:51:43.000 --> 00:52:01.000 >> Thank you both. I'm gonna synthesize some questions that we received in advance, when people registered, with a question that popped up here in the Q and A. 00:52:01.000 --> 00:52:26.000 But... So this is about methods. So it's probably more for Gen. But Annette, you may be seeing manuscripts coming in that you're also excited about, that might fit this question too. Are there new methods needed to help address these issues? 00:52:26.000 --> 00:52:32.000 And avoid some of the mistakes of the past? What are some exciting new developments or approaches that seem especially promising? And perhaps on the flip side, to what degree do you think a lack of workforce knowledge of statistical methodology is suitable for analyzing heterogeneity, such as hierarchical linear modeling and others? 00:52:32.000 --> 00:53:01.000 Or access to statistical expertise could be barriers to adoption of some of these recommendations? 00:53:01.000 --> 00:53:20.000 >> Coming from a methods development sort of space, and working with these questions, you know, there's always room for better statistical methods to handle structured data. Especially as the data becomes wider, as well as longer, in terms of the number of variables that are included and the idea of integrating multiple levels of a hierarchy in terms of scale. 00:53:20.000 --> 00:53:43.000 On the other hand, I think there's a more fundamental knowledge gap when it comes to how our environment interacts with people and people's understanding of race and ethnicity. It's often... People are very quick to say, you know, that race is not biological. 00:53:43.000 --> 00:53:57.000 But when it comes to their work and what that means, there is a sort of gap, in terms of how they can functionalize that belief and that fact. Right? And so that is sort of the bigger issue, I would say, because you can... There has always been this emphasis on... We're academics. We like to know everything. We like to be experts in everything. 00:53:57.000 --> 00:54:15.000 But as you consider these questions, no one person is sufficient to do this well. Right? It takes multidisciplinary teams to be able to understand all of the variables, and to model them appropriately. 00:54:15.000 --> 00:54:42.000 And this has not been incentivized, currently, in our climate. And this is partially -- when we think about methods, I think we always default to statistical methods, but I think there needs to be better methods of collaboration. Better methods and frameworks that are incentivized to bring those together. 00:54:42.000 --> 00:55:09.000 And speaking as somebody who comes from the genetics field, it's very... For a while, we've been very, very fortunate in that a SNP is a SNP is a SNP. It's a SNP no matter how you measure it. Now, when we try to go into environmental variables, high air pollution in New York City is very different in terms of what it means for socioeconomic status and environmental factors than in LA. Right? 00:55:09.000 --> 00:55:29.000 And so as geneticists, the idea that a variable that's measured in the exact same way, in the exact same sort of metrics, can mean wildly different things in different places, is sort of a foreign concept. And this includes race/ethnicity. Right? It's a variable. But you can't treat it as sort of your standard SNP. Because what it means is different, depending on the age of the individual. Right? 00:55:29.000 --> 00:55:52.000 It's a sociopolitical construct. Which means that it depends on the society in which it exists. And having that expertise I see as being a more urgent need than newer statistical methods. I think our method development is sometimes ahead of our understanding of how that's important for these concepts. 00:55:52.000 --> 00:56:22.000 >> Yeah. I agree completely. Just as someone would be ill advised to begin a research study without a protocol, and then to begin a research project without a statistician, in these areas, you probably need to include a social scientist or two, to get to that. 00:56:23.000 --> 00:56:33.000 I would also say that, again, on the receiving end of manuscripts reporting things like exposure to particulate matter in the air, and trying to connect that to everything, dementia, Alzheimer's disease, we urge researchers to be cautious about hyping a pretty small statistical significance, versus understanding the a priori minimal clinically important difference that's gonna have meaning to clinicians and patients. 00:56:33.000 --> 00:56:53.000 And the public. And that's just another thing we try to guide authors on, when they first submit their paper to us. Which is why we have peer review and why we have revisions. 00:56:53.000 --> 00:57:15.000 >> Thank you both. I'm gonna pick up on something -- Gen, you used the word "incentivize". So I think it may be a good time for us to bring into this conversation funders. And there is a question/comment in the Q and A that there's an elephant in the room these days, when it comes to discussions of research funding. 00:57:15.000 --> 00:57:38.000 Have these standards been applied to, required for funding and grant requests, either... I think they mean biomedical research. Or indeed in other areas of funding? Have they been required for IRB approval? I think I can answer that question, and the answer was mostly no. There was some guidance that has since been rescinded. 00:57:38.000 --> 00:57:52.000 And so that leads me to perhaps a provocative question that will lead us into the afterparty: What can be done, even in the current political environment, to implement these recommendations? How can journalists support researchers in the current political environment? How can researchers in our roles as reviewers help provide accountability? 00:57:52.000 --> 00:58:18.000 What role do private funders, professional societies, and research institutions still have to play, even now? So if you could... Each of you pick up on some part of that, and then we'll wrap up here shortly. 00:58:18.000 --> 00:58:40.000 >> I'll start with journals. So first of all... Yes. There was a tremendous amount of movement. Especially in funders. That has been rescinded. In this country. In February, we published an editorial in JAMA, reaffirming our commitment to... Basically to researchers and to public health. 00:58:40.000 --> 00:58:56.000 We acknowledge what we were seeing, because of researchers' fear for their funding, fear for their livelihood, and so we wanted to be public that we're here. And we are committed to doing everything we can to support researchers, so that science is not silenced. 00:58:56.000 --> 00:59:09.000 But that includes being flexible, if we need to. So if a researcher comes to us as an example, and has to change the word "gender" to "sex" in table one, we will work with them to do that. 00:59:09.000 --> 00:59:29.000 We will not work with them to remove participants who might have identified or been identified by a specific gender. That would require a reanalysis and a rereview of a paper. 00:59:29.000 --> 00:59:49.000 But we are still publishing quite a lot in these areas. You know, the NSF has a list of 50 banned words. One of them happens to be "women". I'm looking at four of us right now. And... You know... That's just... Well... That's just not gonna happen. 00:59:49.000 --> 00:59:58.000 So... I have seen more recently... And it's probably because there are other distractors out there... Less requests for change. But we're in a long tail of what this is gonna do to research in this country. Biomedical research in this country. 00:59:58.000 --> 01:00:00.000 >> Gen, do you want to take a quick stab? Or do you want to hang on for the afterparty? 01:00:00.000 --> 01:00:06.000 >> Hang on to the afterparty. 01:00:06.000 --> 01:00:18.000 >> With that, then, I will turn it over to Sandra to wrap us up. Thank you so much, both of you. 01:00:18.000 --> 01:00:25.000 >> Thank you, Aliya, and so much gratitude to you and Gen and Annette for today's presentations and the discussion. We're gonna continue this discussion. 01:00:25.000 --> 01:00:39.000 So if you can stick around for a little bit, you can transition with all of us to the informal afterparty. 01:00:39.000 --> 01:01:08.000 I did want to mention that our next ELSI Friday Forum will happen on July 11th. And we are planning a discussion on biosecurity in the age of AI and genomics. 01:01:08.000 --> 01:01:28.000 The link to register for that session should be in the chat. Also, there will be a postevent survey that will pop up, once you leave the Zoom room. And I encourage you to complete this. Our organizing committee takes your comments very seriously, and it really does help us, as we think about new topics and speakers to bring to you all. 01:01:28.000 --> 01:01:39.000 So now I think we can transition to the discussion portion of the forum. And we invite you to turn on your cameras if you can. And to raise your hands to ask questions. There were some questions that we'll port over from the recorded session. 01:01:39.000 --> 01:01:40.000 And for those of you who can't join us, thank you for attending, and I hope you have a great weekend. But let's go ahead and stop the recording, if we haven't already.