 Take us away. It's Jackie Ogis, just to clarify that. Good afternoon. I'll be presenting on the Institute's Inclusion of Women and Minorities in Clinical Studies for Fiscal Years 2013 and 2014. So I'll give a brief history of the NIH-wide mandate to report enrollment. I'll provide a few definitions that we used in writing our biennial report. I'll explain the exclusion of enrollment data from our largest funded study, 23andMe. And finally, I'll discuss the NHGRI's enrollment data and analyses for Fiscal Years 2013 and 2014. So as we all know, increasing enrollment of women and minorities in clinical research is scientifically important. And responding to a lack of population diversity in clinical research, Congress passed the NIH for Vitalization Act of 1993, which requires all institutes to report enrollment every two years. And this includes data from intramural and extramural research programs. And this ensures that all ICs are in compliance with the NIH Inclusion Guidelines and allows us to look back and observe trends in enrollment over time across the institutes. And all investigators are required to report enrollment unless they're exempted in cases such as small sample size or duplicate reporting. And program officers and review staff keep track of the target or proposed data that PIs submit in their progress reports and their applications. But for our biennial report, we looked at actual enrollment data. These numbers are cumulative, meaning that tracking begins at the start of the project's funding period. So the NIH Inclusion Guidelines define a minority group as a readily identifiable subset of the US population that's distinguishable by racial, ethnic, and or cultural heritage. And a significant difference is what we're looking for in increasing enrollment of women and minorities is a difference that is of clinical or public health importance. And for the purposes of our biennial report, we defined a protocol as a single grant or IRB approved project. And the protocol is what we use to group and simplify our enrollment numbers by project. And when PIs submit their enrollment data to the institute, they populate tables similar to this one here. They break down their enrollment into ethnic, racial, and gender categories. And these categories are set by the US Census. So for most studies funded by NHGRI, participants self-report their race, ethnicity, and gender. But for our largest funded study, 23andMe, in which close to 580,000 participants are enrolled in a large online study, these data are collected differently. In that first, the data are collected online. And about 30% of participants did not report their race. For participants who did not specifically identify Hispanic, investigators assumed that they were not Hispanic. And then gender was determined by genetic data, as opposed to self-reporting. So when we added in the 23andMe numbers, we saw that our proportions of unknown race were significantly skewed. And so we tended to exclude these numbers from our analyses and reported them separately. So from FY13 to 14, I show ERP here. Without 23andMe data, we had 50 protocols involving close to 240,000 participants. IRP had 94 protocols with 65,000 participants. And across the institute, there was a total of 144 protocols for comprising a total enrollment of about 300,000 participants. And sorting these protocols by sample size, about 60 protocols had a sample size below 100 participants. And another 60 had a sample size between 100 and 600. Fewer protocols had larger sample sizes. And IRP had a higher proportion of protocols with smaller sample sizes. And again, these exclude 23andMe. So breaking down our enrollment by racial categories, I show here IRP data, I mean ERP excluding 23andMe. A little bit more than half of our participants identify as white. And about a fifth of our participants identify as black or African-American. Fewer participants identify as members of other minority groups. And we see here about 9% of our participants report unknown race or unknown in terms of race. And when we investigated that further, we saw that there were three studies that seemed to skew these numbers and led to a higher proportion of unknown race. There was one study in which race and ethnicity were combined as one variable. And then there were two studies in which Hispanic participants did not report race. And investigators for these studies explained that participants who identify as Hispanic, particularly recent immigrants, they were less familiar with the US Census categories and therefore declined to report their race being declined to define themselves in those terms. And looking at IRP, we had a slightly higher proportion of participants being white. And in both ERP and IRP, we have about a fifth of our participants identifying as black or African-American. And a smaller proportion of participants were of unknown race. And here we have 16% identifying their ethnicity as Hispanic and 4% of unknown ethnicity. And in IRP, we had about 3% identifying Hispanic and 2% of unknown ethnicity. And in ERP, we have a majority of our participants identifying as female. And in IRP, about 51% identifying as male with 5% unknown gender. So as you remember, we treated 23andMe data separately. So I show it here. 58% of participants identify as white. And about a third of these participants identify as unknown race or did not report their race for reasons explained earlier. And as a reference, I show IRP excluding the 23andMe data. There's a slightly lower proportion of participants who are white and much greater proportion of participants who identify as black. And there is a 9% unknown race as opposed to 30%. And here, 11% in 23andMe identify Hispanic in the city. And that's compared to 16% in ERP. And a greater proportion of male participants in 23andMe as opposed to a majority of female participants in ERP. And tracking enrollment and putting together these reports every two years allows us to look at trends in enrollment over time. And we see here, I've grouped 2010 data in the blue column, 2011 to 2012 data in the red column, and then 2013 and 2014 in these two columns. This is excluding 23andMe data. And this is including 23andMe data. And I compare all of these to the 2010 census data. We see a slight decrease in the proportion of white participants. And this is below the US census. And we see a slight increase in the proportion of black participants. And this is higher than the US census. We see no major trends in the proportion of participants identifying as Asian. And we have been trying to reduce the number of unknowns across all of these categories. But we see an increase in the number of unknown race in this chart here. This is something that we recognize. And this is something that we'll continue to address going forward. And also, these categories below are the groups in which we saw major trends in enrollment. And here we see an increase in the proportion of Hispanic participants in our studies. And a decrease in the number of unknowns in terms of ethnicity. And an increase in the number of female participants. And if we add in 23andMe, we see a slight increase in the proportion of males. So in summary, NHGRI has increased the enrollment of minorities and women in fiscal years 2013 and 2014. And in comparison to the 2010 US census, it was able to increase the population diversity of its cohorts. Because race, ethnicity, and gender were reported differently by 23andMe, we tended to exclude these numbers from our analyses and reported them separately. And the high proportion of unknown race for ERP is due to the collecting of race and ethnicity in a single item or Hispanic participants declining to report race. Thank you. Council reviewers are Drs. Bohrwinkel and Hughes Halper. Dr. Bohrwinkel. First, thank you for making the report on our behalf. I didn't realize that little entity. And also, congratulations on the trends, both to you and the community. I think it's a real win because the community realizes that diversity is important not just for social reasons but also scientific reasons. And I think that's great. One technical comment, I would advise that we don't report data and have sentences that focus on 23andMe. I would think we would focus on NHGRI with and without 23andMe and not have anything about just 23andMe. I don't think it's helpful. And I'm not sure it's appropriate. So that's my opinion. So for example, Table 2B, I would eliminate. So was 23andMe funded by NHGRI? Yes. And what was the purpose of the project? But I think if we go down that route, we need an appendix then about every project. I'm just I'm blocking it showing the data from a single project. Eric, the reason we did that was because it was so large. It's three times larger than any others. But your point is well taken. And we'll just do with and without them. So I agree with everything that Eric has said. One observation that I thought of as you were presenting and just thinking about some of the comments made earlier today about what our experience has been with limited racial and ethnic diversity and samples is significant that about 20% of the individuals who've been enrolled are from African-Americans. And I just wondered if that's due to a select number of studies that have a specific focus on racial and ethnic minorities, or if that 20% is distributed across all of the projects. So one study, for instance, I'm just thinking it. So you could think about is it that, I don't know, one study is able to have like a racially diverse sample, or is it that you have one study that enrolled 2,000 African-Americans, and that's included in your percentages? That's a good question. The 20% of African-Americans are not restricted to the one or two studies. That's across all the studies. Correct. Not every study, but it's almost across all the studies. We don't have a specific cohort or studies that exclusively recruit African-Americans. So I guess I would just push back a bit. In one of the slides, I had a study that was funded maybe in 2011, 2012, that only recruited African-Americans. And so we didn't have any whites, any Asian-Americans. And so I mean, I think in that we included 1,000 people, which is substantial. I'm not saying that our study did all the heavy lifting here. But I do think it's worthwhile to maybe, I would just be curious to see the distribution across studies and if that 20% holds up across all groups. That's a very good point. That's a good example for the next year. Next time we'll report the data. Actually, for the intramural, we have one study. That's Dr. Gary Gabbens' study. His study is to exclusively recruit African-Americans. Yeah, we will include information for that. Just speaking for the intramural research program, that's one important study that actually the numbers will start to increase. But we do have a few studies intramurally. I think it looks a little different than extramural, where recruitment was either happening in West Africa and persons of African descent in the United States. A few studies of things like sickle cell disease where the disease is occurring because of founder mutations are the reasons in certain populations. So the numbers are smaller. But that's more likely to be a trend in intramural studies. My comment isn't intended to diminish the significance of what you've shown. I'm just pointing out that I think it's important to see across the lay of the land where some investigators are really meeting their accrual goals in terms of enrolling a diverse sample. You can make the same criticism of work that I've done that it wasn't diverse. It only included African-Americans. So again, my comments were more about just being more precise, I think, for this committee in terms of understanding the distribution. So I also want to add my voice to the chorus on seeing the trends going in the right direction. I think it would be great if you could make that data available for other people who are interested in looking at it. In particular, I had exactly the same kind of questions that I wanted to ask about. And I'd love to see study, total number of individuals, broken up by race, ethnicity, be good to know whether it's an emerged study or if it's a study that's got genotype data, exome data, full genome data, all those sorts of details I think people are going to be really, really interested in. And I was curious in particular about PAGE and how much PAGE contributed to that. Because PAGE is 50,000 people specifically geared towards non-European populations. And so I was curious about how much of this was driven by a couple of large studies that were doing multi-ethnic. And then the rest sort of being more of the where we've invested in the past. I think we can show you the breakdown of those numbers we have that collected. Yeah, I mean, it would be great to make it available. I mean, I think it's one of those things that people are really interested in. And when people like me harp on and say, oh, 95% of the studies are in populations of European descent. And Terry's called me on it a couple of times. Oh, it's gotten better. And look at this. And here are the numbers. And so if you guys really make those numbers available, I think it would just be really, really useful for the community so that we could see where change is happening. One of my biggest worries, for example, is that we're getting better in genotype data. But the full genome sequence data the trend may be getting even worse. And so those are the kinds of things that I think would be really useful to see, particularly because you have to report on it on a certain interval of time. But people want to look at it much more regularly, I think. Yeah, I think at your point in time, we're taken. Yeah, we have that data available in spreadsheet. I think for the report, if people self-identify as being male or female, then you can call that gender. But if you use genomic information, it's sex. Right. I'll make that distinction. So I was just, and maybe you went through this, but I was confused by the fact that Hispanics weren't a designated group in a lot of these. I know there's the issue of the unknown being because of a larger percentage. But Hispanics are the largest minority group, right? And I was just confused as to why I didn't see them broken out in the slide. Is it, yeah, is it within the why it's supposed to be? It has to do with the OMB classification. So you can be a black Hispanic, you can be a white Hispanic. So those are counted separately. Well, so they are broken out, Jim. So it's just that we do race, and then we do ethnicity. And they are two separate things, except for one very large study which collected historically as a single combined variable. So I'm just curious, and the instruments that collect these data, can individuals identify in multiple categories simultaneously? I think there is one category to record multiple race, yeah. So one category for multi-race? Yeah, multi-race for that category. Okay, I didn't see that number reported on here. Is that, was it there? If you look in the pie charts, more than one race is the option. Caring or orange, okay. And the other thing I'm interested in, it's not required, but the age trends, is that, is the age of the participants collected in, and are the age trends in these studies also matching the US census in some regard? You don't collect it. Not through this mechanism. Okay. I just wonder, because in the document background, policy applies to research subjects of all ages and NIH from the clinical research studies. So I was interested that age isn't tracked at all. And is that gonna be something that's important in the future to track? Yeah, for the future track, that we have as an aging person. Based on peer review, it's just child or adult. So it's are you 21 years of age or older, you're an adult. If you're below 21, you're a child. It's not binned. The data that are collected, come from that enrollment table, so it's fixed. And it does not include age, even though you may have a pediatric study, it doesn't show up any different from if there are adults in the study. Is it from the target enrollment table or actual enrollment table? From actual enrollment table, inclusion table. Are there questions or comments? Okay. Can I have a motion to accept the report? Second. All in favor. Any opposed. Any abstentions? Thank you. Thank you all. Yes, Betty. Microphone. So I'd like to just clarify that what you're voting on is the report that goes to Congress and separately you're asking at another time for a progress report on how well we're doing. Because I don't want to conflate the regular report to Council with a lot of things, Congress with a lot of things that you want to see. I assume this vote is with taking into account the comments that Council gave. We heard those and we will make those changes.