1 00:00:12,112 --> 00:00:12,946 Thanks for having me. 2 00:00:12,946 --> 00:00:15,349 I'm excited to be here. 3 00:00:15,349 --> 00:00:17,184 So I'm going to talk about, 4 00:00:17,184 --> 00:00:20,053 the ethics of returning genomic research results. 5 00:00:20,053 --> 00:00:23,021 There are a lot of topics I could talk about in the area 6 00:00:23,021 --> 00:00:23,657 of genetics 7 00:00:23,657 --> 00:00:27,961 and, research ethics and human subjects, research. 8 00:00:27,961 --> 00:00:30,297 But but this is the one that really 9 00:00:30,297 --> 00:00:30,964 has been, 10 00:00:31,665 --> 00:00:34,101 taking the most attention, 11 00:00:34,101 --> 00:00:36,470 over the past decade and a half. 12 00:00:36,470 --> 00:00:39,180 And so I think, I think it has particular 13 00:00:39,180 --> 00:00:39,907 relevance. 14 00:00:40,874 --> 00:00:43,180 I do sometimes say controversial things, 15 00:00:43,180 --> 00:00:44,044 and when I do, 16 00:00:44,044 --> 00:00:45,779 they are my views and my views alone. 17 00:00:45,779 --> 00:00:48,717 I do not represent anyone, of authority 18 00:00:48,717 --> 00:00:50,751 in the federal government. 19 00:00:52,019 --> 00:00:54,221 So let me give you some history, some background, 20 00:00:54,221 --> 00:00:57,224 just so you can see where how we got here. 21 00:00:59,059 --> 00:01:03,096 It was about 2008 that we really had a revolution 22 00:01:03,096 --> 00:01:05,612 in how we did sequencing, the sequencing 23 00:01:05,612 --> 00:01:06,366 technology. 24 00:01:07,534 --> 00:01:09,408 Sometimes people call it next generation 25 00:01:09,408 --> 00:01:09,970 sequencing. 26 00:01:09,970 --> 00:01:12,473 It went from the Human Genome Project, 27 00:01:12,473 --> 00:01:15,108 which cost $3 billion and took a decade 28 00:01:15,342 --> 00:01:18,745 to, to the ability to sequence 29 00:01:18,745 --> 00:01:22,182 a whole human genome pretty quickly in a few days. 30 00:01:22,182 --> 00:01:25,085 For at that point it was about $10,000. 31 00:01:25,085 --> 00:01:27,120 Now it's it's like $1,000. 32 00:01:27,120 --> 00:01:30,322 It's very, very cheap, to, to do sequence 33 00:01:30,322 --> 00:01:30,791 data. 34 00:01:31,758 --> 00:01:34,903 And so this was sort of a paradigmatic 35 00:01:34,903 --> 00:01:38,131 shift in, in, in the way we do things. 36 00:01:38,131 --> 00:01:39,769 And I compared to it's compared here 37 00:01:39,769 --> 00:01:41,134 to Moore's Law, which is the, 38 00:01:41,601 --> 00:01:43,767 is the rule or the, the rule of thumb 39 00:01:43,767 --> 00:01:45,405 about computing technology, 40 00:01:45,405 --> 00:01:47,950 the supercomputers that we all carry around in our 41 00:01:47,950 --> 00:01:48,408 pockets. 42 00:01:48,642 --> 00:01:51,020 The idea there is those double in power 43 00:01:51,020 --> 00:01:53,580 for the same price every 18 months or so. 44 00:01:53,847 --> 00:01:55,662 And so thinking about how much computers 45 00:01:55,662 --> 00:01:57,250 have advanced over the past decade 46 00:01:57,250 --> 00:01:59,454 and a half, you can see that sequencing 47 00:01:59,454 --> 00:02:01,488 technology has been that much more, 48 00:02:02,622 --> 00:02:04,424 impressive. 49 00:02:04,424 --> 00:02:06,793 So it's a powerful research tool, 50 00:02:06,793 --> 00:02:09,058 generates huge amounts of data about an 51 00:02:09,058 --> 00:02:09,696 individual 52 00:02:09,696 --> 00:02:12,286 and really beyond that, which is necessary 53 00:02:12,286 --> 00:02:14,568 to answer most scientific questions. 54 00:02:15,602 --> 00:02:17,803 And when you're generating huge amounts 55 00:02:17,803 --> 00:02:20,173 of data like that, you can obviously find 56 00:02:20,173 --> 00:02:24,211 clinically relevant things in, in, someone's genome. 57 00:02:24,578 --> 00:02:26,967 And so the ethical question broadly stated 58 00:02:26,967 --> 00:02:29,016 is what sort of ethical obligations 59 00:02:29,483 --> 00:02:32,319 do we, as researchers have with regards 60 00:02:32,319 --> 00:02:34,801 to these potentially clinically relevant 61 00:02:34,801 --> 00:02:35,422 findings. 62 00:02:36,657 --> 00:02:39,192 Let me just give you a few terms that I'm going to throw around. 63 00:02:39,192 --> 00:02:42,195 So we're all on the same page. Really. 64 00:02:42,529 --> 00:02:44,332 So sometimes people talk about Gwas 65 00:02:44,332 --> 00:02:46,033 genome wide association studies. 66 00:02:46,299 --> 00:02:48,348 That's when you just take little snips, 67 00:02:48,348 --> 00:02:48,769 little, 68 00:02:48,769 --> 00:02:50,620 slivers of a genome across the entire 69 00:02:50,620 --> 00:02:50,971 thing. 70 00:02:50,971 --> 00:02:52,939 That's not what we're talking about today. 71 00:02:52,939 --> 00:02:56,543 We're talking about, spelling out the entire genome, 72 00:02:56,543 --> 00:02:58,949 whole exome or whole genome sequencing 73 00:02:58,949 --> 00:03:01,481 exons, just the protein coding regions, 74 00:03:02,149 --> 00:03:03,780 the whole genome, genomes, the whole thing 75 00:03:03,780 --> 00:03:04,518 or next generation 76 00:03:04,518 --> 00:03:07,521 sequencing is sort of the the umbrella term. 77 00:03:07,721 --> 00:03:10,350 I will talk about incidental or secondary 78 00:03:10,350 --> 00:03:10,991 findings. 79 00:03:11,758 --> 00:03:13,593 I'll probably use them interchangeably. 80 00:03:14,561 --> 00:03:15,395 Incidental 81 00:03:15,395 --> 00:03:18,726 findings or findings concerning an individual research 82 00:03:18,726 --> 00:03:19,466 participant 83 00:03:19,733 --> 00:03:22,963 with potential health or reproductive importance 84 00:03:22,963 --> 00:03:23,703 discovered 85 00:03:23,703 --> 00:03:25,628 in the course of doing the research, 86 00:03:25,628 --> 00:03:27,874 but not related to the aims of the study. 87 00:03:28,575 --> 00:03:31,978 So there's sort of the we stumble upon them kind of 88 00:03:31,978 --> 00:03:32,512 finding 89 00:03:32,512 --> 00:03:35,294 secondary findings are related, but, 90 00:03:35,294 --> 00:03:36,917 related in the sense 91 00:03:36,917 --> 00:03:40,025 that they're, they have these, these important health 92 00:03:40,025 --> 00:03:40,787 implications 93 00:03:40,787 --> 00:03:43,300 and not related to the aims of the study, 94 00:03:43,300 --> 00:03:43,790 but we, 95 00:03:44,491 --> 00:03:47,461 we intentionally look for them, 96 00:03:47,661 --> 00:03:50,272 actively analyze the genome, looking for 97 00:03:50,272 --> 00:03:50,664 them. 98 00:03:52,132 --> 00:03:53,944 I just explain this primary results, 99 00:03:53,944 --> 00:03:55,202 just just for reference, 100 00:03:55,202 --> 00:03:57,554 are things that are related to the, 101 00:03:57,554 --> 00:03:59,906 the condition under investigation. 102 00:04:00,807 --> 00:04:03,140 So I'm going to primarily be talking about secondary 103 00:04:03,140 --> 00:04:03,543 findings 104 00:04:03,543 --> 00:04:06,546 because we've sort of moved past incidental findings, 105 00:04:06,947 --> 00:04:09,683 things that are important that are, 106 00:04:10,851 --> 00:04:13,406 but unrelated to the research that we're 107 00:04:13,406 --> 00:04:13,854 doing. 108 00:04:14,221 --> 00:04:16,923 So let me give you a warm up. 109 00:04:16,923 --> 00:04:18,859 Okay. Excuse me. 110 00:04:18,859 --> 00:04:21,962 This is actually a real case 111 00:04:21,962 --> 00:04:22,462 that I had. 112 00:04:22,462 --> 00:04:24,905 That was one of the first, ethics consults 113 00:04:24,905 --> 00:04:26,533 that I ran as an attending. 114 00:04:27,200 --> 00:04:30,203 So imagine a clinical researcher studying, 115 00:04:30,937 --> 00:04:34,040 the genetic etiology of breast cancer, 116 00:04:34,040 --> 00:04:36,388 in patients, patients that are presenting for 117 00:04:36,388 --> 00:04:36,910 treatment 118 00:04:36,910 --> 00:04:39,208 at an academic medical center, they get 119 00:04:39,208 --> 00:04:41,448 they've given their clinical consent. 120 00:04:41,448 --> 00:04:44,584 They also get research specific informed consent. 121 00:04:44,584 --> 00:04:48,822 And the team is going to then sequence, their, their, 122 00:04:49,256 --> 00:04:52,583 their genetic material looking from surplus tumor 123 00:04:52,583 --> 00:04:53,126 tissue, 124 00:04:53,393 --> 00:04:56,396 looking for novel BRCA, 125 00:04:56,897 --> 00:04:59,833 variants that might predict disease. 126 00:04:59,833 --> 00:05:01,592 I should say this is an understudied 127 00:05:01,592 --> 00:05:03,303 population of minority population. 128 00:05:03,570 --> 00:05:06,573 So this is, there's there are lots of BRCA variants 129 00:05:06,573 --> 00:05:09,078 that have been identified, but not as many 130 00:05:09,078 --> 00:05:11,344 in this group, at least at that time. 131 00:05:12,512 --> 00:05:15,015 They proposed to de-identify the sequence 132 00:05:15,015 --> 00:05:15,382 data. 133 00:05:15,382 --> 00:05:18,137 They don't plan to return any results, 134 00:05:18,137 --> 00:05:20,821 even though they know they're there. 135 00:05:21,121 --> 00:05:23,872 They're looking for for novel, disease 136 00:05:23,872 --> 00:05:25,392 associated variants. 137 00:05:25,725 --> 00:05:27,781 They know that they're in some percentage of the 138 00:05:27,781 --> 00:05:28,295 population. 139 00:05:28,295 --> 00:05:30,050 They're going to discover already 140 00:05:30,050 --> 00:05:31,698 established BRCA variants that 141 00:05:31,698 --> 00:05:33,809 could be clinically relevant, particularly 142 00:05:33,809 --> 00:05:35,569 for near-term treatment decisions. 143 00:05:35,569 --> 00:05:37,325 Remember, these are women with breast 144 00:05:37,325 --> 00:05:39,272 cancer who are presenting for treatment. 145 00:05:39,272 --> 00:05:42,764 And so knowledge about the their their genetic variants could be 146 00:05:42,764 --> 00:05:43,310 relevant. 147 00:05:43,710 --> 00:05:46,313 And so the question is should these researchers be allowed 148 00:05:46,313 --> 00:05:47,747 to design the study in this way 149 00:05:47,747 --> 00:05:49,439 where they're going to find these things 150 00:05:49,439 --> 00:05:50,750 that are clinically important, 151 00:05:50,984 --> 00:05:54,346 but not give them back to the patients, 152 00:05:54,346 --> 00:05:56,156 to the participants? 153 00:05:56,389 --> 00:05:58,825 The puzzle here, 154 00:05:58,825 --> 00:06:01,528 is there something beneficial 155 00:06:01,528 --> 00:06:04,531 that the team could provide to these participants 156 00:06:04,531 --> 00:06:05,874 that they might not otherwise know 157 00:06:05,874 --> 00:06:07,534 that could be important for their health? 158 00:06:07,968 --> 00:06:10,423 So there's this beneficence argument that 159 00:06:10,423 --> 00:06:12,939 you could make the investigator, however, 160 00:06:14,040 --> 00:06:16,710 can and in fact did say, 161 00:06:16,710 --> 00:06:19,713 you know, we don't have any genetic counselors on staff. 162 00:06:19,713 --> 00:06:21,181 It would cost us a bunch of money 163 00:06:21,181 --> 00:06:22,782 to have to validate these findings, 164 00:06:22,782 --> 00:06:25,886 which we would have to do by law before we can return them. 165 00:06:26,086 --> 00:06:27,420 It would slow us down. 166 00:06:27,420 --> 00:06:30,153 It would actually delay the time to us 167 00:06:30,153 --> 00:06:31,591 finding these novel 168 00:06:31,591 --> 00:06:34,590 BRCA variants that could help thousands of 169 00:06:34,590 --> 00:06:35,161 people. 170 00:06:35,862 --> 00:06:38,598 And so there's this burden on the research enterprise 171 00:06:38,598 --> 00:06:41,635 kind of argument that, is often made. 172 00:06:42,302 --> 00:06:44,671 Also, he made the argument, you know, 173 00:06:44,671 --> 00:06:46,773 or could have made the argument, this isn't my job. 174 00:06:46,773 --> 00:06:48,995 I'm a researcher. I'm not a I'm not providing clinical 175 00:06:48,995 --> 00:06:49,242 care. 176 00:06:49,242 --> 00:06:52,112 If they want to get testing, they can go get testing. 177 00:06:52,112 --> 00:06:54,814 Why should I have to provide this service? 178 00:06:54,814 --> 00:06:57,817 But I think we've moved to a point, 179 00:06:58,418 --> 00:07:00,200 where where those kinds of arguments 180 00:07:00,200 --> 00:07:01,388 that the researcher was 181 00:07:01,388 --> 00:07:04,724 making are no longer, being well received. 182 00:07:05,458 --> 00:07:07,394 So let me give you some early views. 183 00:07:07,394 --> 00:07:08,899 I talked about the stumble strategy 184 00:07:08,899 --> 00:07:09,329 early on. 185 00:07:09,329 --> 00:07:11,440 We're just we weren't actively looking for 186 00:07:11,440 --> 00:07:13,400 things are just stumbling across them. 187 00:07:14,200 --> 00:07:16,136 And doing a case by case analysis. 188 00:07:16,136 --> 00:07:17,370 Would we return results? 189 00:07:17,370 --> 00:07:18,789 Well, we've got to look at all of the 190 00:07:18,789 --> 00:07:19,172 different 191 00:07:19,172 --> 00:07:21,608 contextual features of the situation. 192 00:07:22,575 --> 00:07:25,108 IRBs were very reluctant to return 193 00:07:25,108 --> 00:07:25,779 results, 194 00:07:26,613 --> 00:07:29,582 because they were worried about psychosocial risks. 195 00:07:29,916 --> 00:07:31,885 But today 196 00:07:31,885 --> 00:07:34,222 and this is sort of the top line, thing 197 00:07:34,222 --> 00:07:34,821 I want to 198 00:07:34,821 --> 00:07:37,978 I want to stress genomics is is cheap, it's 199 00:07:37,978 --> 00:07:38,858 ubiquitous. 200 00:07:38,858 --> 00:07:41,861 It's all throughout our research enterprise. 201 00:07:42,195 --> 00:07:45,332 There's a ton of proliferation of expertise and guidance. 202 00:07:46,866 --> 00:07:49,603 We've moved past this idea that it's really 203 00:07:49,603 --> 00:07:50,303 dangerous, 204 00:07:51,371 --> 00:07:54,140 that towards the idea that genomic information 205 00:07:54,140 --> 00:07:56,041 is just another kind of medical 206 00:07:56,041 --> 00:07:56,776 information 207 00:07:56,776 --> 00:08:00,480 and towards the idea that, these psychosocial risks 208 00:08:00,480 --> 00:08:02,530 that we understandably were concerned 209 00:08:02,530 --> 00:08:03,416 about early on. 210 00:08:03,416 --> 00:08:04,317 Really actually, 211 00:08:04,317 --> 00:08:06,069 we've done tons of research and the 212 00:08:06,069 --> 00:08:07,921 the risks seem to be pretty minimal. 213 00:08:08,488 --> 00:08:11,191 Except in some, some isolated, 214 00:08:11,191 --> 00:08:14,194 and very distinct discrete cases. 215 00:08:15,161 --> 00:08:16,763 And so we, we now broadly 216 00:08:16,763 --> 00:08:19,275 have this view that there's some obligation to 217 00:08:19,275 --> 00:08:19,766 look for 218 00:08:20,100 --> 00:08:23,383 and return a defined set of secondary 219 00:08:23,383 --> 00:08:24,270 findings. 220 00:08:25,872 --> 00:08:27,273 But that's just 221 00:08:27,273 --> 00:08:30,543 sort of the top line, the top line, argument. 222 00:08:30,543 --> 00:08:33,780 Like we need to flesh out the contours 223 00:08:33,780 --> 00:08:36,783 of what that obligation might look like. 224 00:08:37,083 --> 00:08:38,845 And one of the ways we need to do 225 00:08:38,845 --> 00:08:40,286 that is to talk about why, 226 00:08:41,154 --> 00:08:44,257 there's a duty to look for genetic research results. 227 00:08:44,257 --> 00:08:47,193 So I've already explained that it's largely a beneficence 228 00:08:47,193 --> 00:08:49,512 based argument, to give people back 229 00:08:49,512 --> 00:08:51,631 important clinical information. 230 00:08:52,766 --> 00:08:54,634 But research isn't clinical care. 231 00:08:54,634 --> 00:08:55,735 I think that's probably something 232 00:08:55,735 --> 00:08:56,302 you guys learned 233 00:08:56,302 --> 00:08:59,339 in the first session of this course. 234 00:08:59,339 --> 00:09:00,802 There are important distinctions, 235 00:09:00,802 --> 00:09:02,575 ethically relevant distinctions between 236 00:09:02,809 --> 00:09:04,077 research and clinical care. 237 00:09:04,077 --> 00:09:06,449 And so researchers can't be on the hook 238 00:09:06,449 --> 00:09:07,847 to give the same kinds 239 00:09:07,847 --> 00:09:11,384 of services, as a doctor would. 240 00:09:11,384 --> 00:09:14,166 A doctor has to do what's best for the patient in front of 241 00:09:14,166 --> 00:09:14,454 them. 242 00:09:14,921 --> 00:09:16,719 A researcher has other obligations 243 00:09:16,719 --> 00:09:18,992 to the research enterprise, to their data, 244 00:09:18,992 --> 00:09:21,177 to whatever generalizable knowledge 245 00:09:21,177 --> 00:09:22,862 they're trying to produce. 246 00:09:25,065 --> 00:09:28,001 And so from my perspective, the best model 247 00:09:28,001 --> 00:09:30,567 for thinking about this and for balancing those 248 00:09:30,567 --> 00:09:31,604 considerations is, 249 00:09:32,272 --> 00:09:35,075 there's a duty to rescue or ancillary care 250 00:09:35,075 --> 00:09:36,075 kind of model, 251 00:09:37,377 --> 00:09:40,936 ancillary care for those of you, if you don't know, is is a is a 252 00:09:40,936 --> 00:09:41,214 role 253 00:09:41,214 --> 00:09:44,384 specific obligation derived from the duty to rescue, 254 00:09:44,384 --> 00:09:47,387 which is a general obligation. 255 00:09:47,620 --> 00:09:50,644 Ancillary care is the idea that researchers sometimes will 256 00:09:50,644 --> 00:09:50,957 have, 257 00:09:51,558 --> 00:09:53,597 participants, groups of participants 258 00:09:53,597 --> 00:09:54,561 in front of them 259 00:09:54,761 --> 00:09:56,785 with known medical needs unrelated 260 00:09:56,785 --> 00:09:58,631 to the research they're doing, 261 00:09:59,199 --> 00:10:02,114 and the researcher could help them meet 262 00:10:02,114 --> 00:10:04,804 those needs at relatively low cost. 263 00:10:05,071 --> 00:10:07,379 A classic cases is if you were going to do 264 00:10:07,379 --> 00:10:07,874 research 265 00:10:07,874 --> 00:10:10,877 in a region where malaria is endemic, 266 00:10:11,945 --> 00:10:13,947 malaria has nothing to do with the research you're doing, 267 00:10:13,947 --> 00:10:16,002 but you could bring along a bunch 268 00:10:16,002 --> 00:10:18,618 of anti-malarial medication at fairly low 269 00:10:18,618 --> 00:10:20,862 cost and give it out to, the people, 270 00:10:20,862 --> 00:10:23,356 and you provide a huge service to them. 271 00:10:24,524 --> 00:10:26,993 But it's unrelated to your, your research. 272 00:10:26,993 --> 00:10:28,261 So that's ancillary care. 273 00:10:28,261 --> 00:10:31,135 I think that's the best model for thinking about returning 274 00:10:31,135 --> 00:10:31,531 genetic 275 00:10:31,531 --> 00:10:34,501 research results. 276 00:10:36,970 --> 00:10:39,698 And really, the the reason I think it's the best 277 00:10:39,698 --> 00:10:40,039 model 278 00:10:40,406 --> 00:10:42,455 is because it specifies conditions 279 00:10:42,455 --> 00:10:44,444 when results should be returned, 280 00:10:45,812 --> 00:10:48,348 i.e. high benefit conditions 281 00:10:48,348 --> 00:10:51,234 where there's low burden to the research, 282 00:10:51,234 --> 00:10:53,486 research team and where there's 283 00:10:53,486 --> 00:10:55,465 the researchers in a sort of unique 284 00:10:55,465 --> 00:10:57,557 opportunity to provide that benefit. 285 00:10:57,557 --> 00:10:58,625 If someone else would be able 286 00:10:58,625 --> 00:11:01,446 to provide the malarial drugs like the researcher doesn't have 287 00:11:01,446 --> 00:11:01,628 to. 288 00:11:01,861 --> 00:11:04,779 But if the researcher is the only one who could, that 289 00:11:04,779 --> 00:11:05,164 helps. 290 00:11:05,732 --> 00:11:08,701 Helps helps limit the cases in which, 291 00:11:09,068 --> 00:11:12,071 ancillary care obligations might be produced. 292 00:11:12,639 --> 00:11:16,209 So it sort of balances the benefit to the participant 293 00:11:16,209 --> 00:11:18,250 and the burden to the research enterprise 294 00:11:18,250 --> 00:11:19,345 in a way that I find, 295 00:11:19,345 --> 00:11:21,381 very attractive in this kind of case. 296 00:11:22,382 --> 00:11:23,983 But it's not a 297 00:11:23,983 --> 00:11:26,400 perfect model, for a number of reasons 298 00:11:26,400 --> 00:11:27,353 I can go into. 299 00:11:27,353 --> 00:11:31,079 But the big one is that it makes return of results 300 00:11:31,079 --> 00:11:31,824 dependent 301 00:11:31,824 --> 00:11:35,862 on researcher expertise and protocol specific resources. 302 00:11:35,862 --> 00:11:37,964 If you're talking about, 303 00:11:37,964 --> 00:11:40,303 burden to the research team as one of the primary 304 00:11:40,303 --> 00:11:41,067 considerations, 305 00:11:41,067 --> 00:11:42,879 if the research team doesn't have 306 00:11:42,879 --> 00:11:44,470 the expertise, doesn't have, 307 00:11:45,371 --> 00:11:48,245 the staff to do this, then they'd say the burden is 308 00:11:48,245 --> 00:11:48,808 too high. 309 00:11:49,108 --> 00:11:50,710 I have no obligation. 310 00:11:50,710 --> 00:11:53,549 And obviously, if we're making return results 311 00:11:53,549 --> 00:11:54,180 dependent 312 00:11:54,180 --> 00:11:56,128 on those kinds of considerations, 313 00:11:56,128 --> 00:11:58,017 that can lead to inefficiencies 314 00:11:58,017 --> 00:12:01,754 and justice concerns, because similarly situated 315 00:12:02,221 --> 00:12:05,029 participants would get different, access 316 00:12:05,029 --> 00:12:07,627 to genetic information based on what 317 00:12:07,627 --> 00:12:10,097 sort of resources or expertise their team 318 00:12:10,097 --> 00:12:11,965 had, which doesn't seem ideal. 319 00:12:13,700 --> 00:12:15,301 And here's a paper we wrote about this. 320 00:12:15,301 --> 00:12:17,737 If people are are interested. 321 00:12:20,873 --> 00:12:23,276 So, I and others have 322 00:12:23,276 --> 00:12:25,310 argued that we should start thinking 323 00:12:25,310 --> 00:12:27,513 about an institutional duty of rescue, 324 00:12:27,513 --> 00:12:31,255 that this obligation shouldn't fall to the individual research 325 00:12:31,255 --> 00:12:31,618 team. 326 00:12:31,618 --> 00:12:33,699 It should fall to the institution 327 00:12:33,699 --> 00:12:35,655 supporting that research team. 328 00:12:37,156 --> 00:12:39,394 And, and there are a number of reasons 329 00:12:39,394 --> 00:12:40,159 why this is, 330 00:12:40,660 --> 00:12:42,328 this is attractive. 331 00:12:42,328 --> 00:12:45,199 The, the main ones are that since 332 00:12:45,199 --> 00:12:47,200 individual researchers 333 00:12:47,200 --> 00:12:49,000 are going to lack the right expertise 334 00:12:49,000 --> 00:12:50,703 if if the institution is providing 335 00:12:50,703 --> 00:12:52,267 those that can sort of centralize 336 00:12:52,267 --> 00:12:53,973 that expertise and those resources, 337 00:12:55,842 --> 00:12:59,345 and do it much more efficiently and effective 338 00:12:59,646 --> 00:13:02,415 and can take this off of the plate of the researchers 339 00:13:02,415 --> 00:13:05,718 and let them focus on their primary research aims. 340 00:13:05,718 --> 00:13:09,288 So it's efficient, it solves justice concerns and, 341 00:13:10,056 --> 00:13:12,933 it allows the researchers to actually do what they're 342 00:13:12,933 --> 00:13:13,693 aiming to do. 343 00:13:17,897 --> 00:13:21,534 So, let me transition here. 344 00:13:22,902 --> 00:13:24,570 I mainly do research, 345 00:13:24,570 --> 00:13:28,441 but sometimes I help, do some policy things too. 346 00:13:28,441 --> 00:13:31,811 And so the NIH intramural IRB, 347 00:13:34,781 --> 00:13:37,316 asked me and my colleague Sarah Hall 348 00:13:37,316 --> 00:13:41,454 to, to chair a working group to come up with, guidance 349 00:13:41,921 --> 00:13:45,476 for how our intramural program should deal with research, the 350 00:13:45,476 --> 00:13:46,059 research, 351 00:13:46,059 --> 00:13:49,128 the genetic research results problem. 352 00:13:49,128 --> 00:13:52,092 And so it's going to build on some of the arguments I just 353 00:13:52,092 --> 00:13:52,398 made. 354 00:13:53,132 --> 00:13:54,892 So this is this is our sort of I can't 355 00:13:54,892 --> 00:13:56,836 call it a policy because that's not fair. 356 00:13:56,836 --> 00:14:00,039 But this is sort of our, our IRBs recommendations or 357 00:14:00,039 --> 00:14:00,840 guidance to, 358 00:14:01,741 --> 00:14:04,744 expectations for for researchers. 359 00:14:05,044 --> 00:14:08,881 So the first argument is it's time for specificity. 360 00:14:08,881 --> 00:14:11,066 We've we've existed in this sort of limbo 361 00:14:11,066 --> 00:14:12,719 for 15 years where we've said, 362 00:14:12,719 --> 00:14:14,220 oh, this is a hard problem. 363 00:14:14,220 --> 00:14:16,688 Here's some interesting ideas, but no one's been willing to 364 00:14:16,688 --> 00:14:17,023 sort of 365 00:14:17,023 --> 00:14:22,254 put pen to paper and say, we need a firm, well-defined 366 00:14:22,254 --> 00:14:23,029 policy. 367 00:14:23,029 --> 00:14:24,597 And I think that's wrong. 368 00:14:24,597 --> 00:14:26,921 I think we're far enough into this debate 369 00:14:26,921 --> 00:14:29,302 that the sort of Wild West, like everyone 370 00:14:29,302 --> 00:14:32,283 doing whatever they want and case by case 371 00:14:32,283 --> 00:14:34,974 analysis, we need to move past that. 372 00:14:35,942 --> 00:14:36,476 I think 373 00:14:36,476 --> 00:14:39,238 it's great to defer to IRBs generally, 374 00:14:39,238 --> 00:14:40,546 but in this case, 375 00:14:41,514 --> 00:14:43,783 we used to have 12 IRBs. 376 00:14:43,783 --> 00:14:45,451 Intramural in the NIH, 377 00:14:45,451 --> 00:14:48,554 and they all had different ways of dealing with this problem. 378 00:14:49,355 --> 00:14:51,324 Now we have a single IRB, so that's better. 379 00:14:51,324 --> 00:14:54,827 But but it would be ideal for institutions 380 00:14:54,827 --> 00:14:57,830 or even the whole research enterprise, 381 00:14:58,131 --> 00:15:01,134 nationally or internationally, to to try to come up to a, 382 00:15:01,434 --> 00:15:04,437 consistent way of dealing with this. 383 00:15:05,404 --> 00:15:08,014 Like I said, existing guidance is very high level and 384 00:15:08,014 --> 00:15:08,407 really, 385 00:15:09,075 --> 00:15:12,618 sometimes avoids making specific or controversial claims, which I 386 00:15:12,618 --> 00:15:12,945 think 387 00:15:12,945 --> 00:15:15,269 we're at this point where we need to start 388 00:15:15,269 --> 00:15:17,316 actually making those, those claims. 389 00:15:17,316 --> 00:15:21,487 So, the IRB asked us to create this policy. 390 00:15:23,055 --> 00:15:25,863 And we argued that there is a broad 391 00:15:25,863 --> 00:15:26,826 but shallow 392 00:15:26,826 --> 00:15:29,513 obligation to return genetic results 393 00:15:29,513 --> 00:15:31,230 generated in research. 394 00:15:31,798 --> 00:15:33,627 Broad in the sense that it applies 395 00:15:33,627 --> 00:15:35,134 to most research protocols, 396 00:15:35,601 --> 00:15:37,369 but shallow in the sense that there's a 397 00:15:37,369 --> 00:15:39,272 very high bar for the kind of information 398 00:15:39,672 --> 00:15:40,907 you have to return. 399 00:15:42,008 --> 00:15:43,242 The American College of 400 00:15:43,242 --> 00:15:46,245 Medical Genetics and Genomics, Acmg, 401 00:15:47,580 --> 00:15:50,333 curates a list that it updates every year or 402 00:15:50,333 --> 00:15:50,583 so. 403 00:15:51,484 --> 00:15:53,719 And right now there are 404 00:15:53,719 --> 00:15:56,088 how many truly how many things on the list? 405 00:15:56,088 --> 00:15:58,658 83 things on the list right now. 406 00:15:58,658 --> 00:16:01,828 So it's a, it's a, it's a very, 407 00:16:01,828 --> 00:16:05,131 very high bar for what, what counts as the kind of thing 408 00:16:05,131 --> 00:16:07,700 that needs to be returned. 409 00:16:07,700 --> 00:16:08,968 Which protocols? 410 00:16:08,968 --> 00:16:11,971 We argued that, 411 00:16:13,206 --> 00:16:16,042 That the deeper the clinical relationship, 412 00:16:16,042 --> 00:16:18,557 the stronger the presumption in favor of 413 00:16:18,557 --> 00:16:19,312 disclosure. 414 00:16:19,545 --> 00:16:22,236 So if you're just doing secondary research 415 00:16:22,236 --> 00:16:24,350 with samples that were collected 416 00:16:24,350 --> 00:16:26,549 elsewhere where you have no relationship 417 00:16:26,549 --> 00:16:28,254 with the participants, there's 418 00:16:28,254 --> 00:16:30,627 that's not a case where you need to return 419 00:16:30,627 --> 00:16:31,757 secondary findings. 420 00:16:32,892 --> 00:16:35,995 If you have a much more clinical like relationship 421 00:16:35,995 --> 00:16:40,029 where you're seeing someone extensively over time, repeat 422 00:16:40,029 --> 00:16:40,666 workups, 423 00:16:40,666 --> 00:16:42,716 maybe you're providing standard of care 424 00:16:42,716 --> 00:16:44,503 as part of the research protocol. 425 00:16:44,871 --> 00:16:48,416 That's the kind of thing would probably would involve returning 426 00:16:48,416 --> 00:16:49,542 secondary findings. 427 00:16:50,810 --> 00:16:53,351 It's a little trickier when you start talking 428 00:16:53,351 --> 00:16:53,746 about, 429 00:16:53,746 --> 00:16:56,749 minimal interactions, like one time interactions. 430 00:16:57,917 --> 00:16:59,218 It's just a one time blood draw. 431 00:16:59,218 --> 00:17:02,488 Maybe not, but if it's a one time or two 432 00:17:02,488 --> 00:17:05,491 time, extensive workup, maybe, 433 00:17:05,791 --> 00:17:08,027 we argued that in this gray area, 434 00:17:08,027 --> 00:17:11,349 at least to start, there's no need to return secondary 435 00:17:11,349 --> 00:17:11,964 findings. 436 00:17:11,964 --> 00:17:14,033 You're welcome to, it'd be. 437 00:17:14,033 --> 00:17:16,636 It'd be lovely, but you don't have to, 438 00:17:16,636 --> 00:17:19,154 but as it becomes easier and easier 439 00:17:19,154 --> 00:17:21,240 as more centralized services 440 00:17:21,240 --> 00:17:25,111 are brought online, the presumption in this category 441 00:17:25,111 --> 00:17:28,080 could evolve, and expand. 442 00:17:29,615 --> 00:17:32,251 A few more things about the IRB expectations. 443 00:17:32,251 --> 00:17:34,654 This is only prospective, older studies, 444 00:17:34,654 --> 00:17:36,756 studies that had already generated 445 00:17:36,756 --> 00:17:38,748 sequence data didn't need to go back 446 00:17:38,748 --> 00:17:41,127 and analyze and return secondary findings. 447 00:17:41,127 --> 00:17:43,726 Only new or new studies or ones that were 448 00:17:43,726 --> 00:17:45,565 being substantially revised. 449 00:17:47,733 --> 00:17:48,534 This is important. 450 00:17:48,534 --> 00:17:51,182 There's no need to generate genomic data 451 00:17:51,182 --> 00:17:52,638 beyond that necessary 452 00:17:52,638 --> 00:17:54,607 to answer the research question. 453 00:17:54,607 --> 00:17:57,539 So if you're only interested in a particular region of the 454 00:17:57,539 --> 00:17:57,944 genome, 455 00:17:57,944 --> 00:18:00,145 and that's all you need to generate 456 00:18:00,145 --> 00:18:01,781 to answer your questions, 457 00:18:01,781 --> 00:18:07,053 there is no obligation to, to do secondary analysis. 458 00:18:07,053 --> 00:18:08,321 You don't have to get 459 00:18:08,321 --> 00:18:11,435 a whole genome sequence if you don't need a whole genome 460 00:18:11,435 --> 00:18:11,991 sequence. 461 00:18:13,459 --> 00:18:15,024 And I think a lot of investigators were 462 00:18:15,024 --> 00:18:16,629 worried that the IRB was going to start 463 00:18:17,363 --> 00:18:18,297 requiring them 464 00:18:18,297 --> 00:18:20,509 to generate all sorts of data that wasn't 465 00:18:20,509 --> 00:18:22,234 even close to what they needed. 466 00:18:22,234 --> 00:18:25,237 But that's, that's that's decidedly not the case. 467 00:18:26,939 --> 00:18:28,993 And as I said, we're distinguishing between 468 00:18:28,993 --> 00:18:29,375 studies 469 00:18:29,375 --> 00:18:32,378 based on the depth of the clinical relationship. 470 00:18:34,413 --> 00:18:37,316 So that's sort of the sketch of the, 471 00:18:37,316 --> 00:18:39,618 the, the, the policy. 472 00:18:39,618 --> 00:18:40,720 What I want to do now, 473 00:18:41,854 --> 00:18:44,957 is to, to run through relatively quickly. 474 00:18:45,791 --> 00:18:49,034 Some related ethical and policy issues 475 00:18:49,034 --> 00:18:50,229 that come up. 476 00:18:50,863 --> 00:18:52,795 If that's the 10,000ft level, these are, 477 00:18:52,795 --> 00:18:54,100 these are some of the more 478 00:18:54,467 --> 00:18:56,715 interesting substantive problems, 479 00:18:56,715 --> 00:18:57,737 some of which. 480 00:18:57,737 --> 00:18:59,276 Well, I have answers for all of them, 481 00:18:59,276 --> 00:19:00,773 some of which people agree with me, 482 00:19:00,773 --> 00:19:03,776 some of which people don't. 483 00:19:03,976 --> 00:19:06,379 But these are the kinds of sort of sticky problems 484 00:19:06,379 --> 00:19:08,494 that we are still working through, 485 00:19:08,494 --> 00:19:10,049 or need to work through. 486 00:19:11,350 --> 00:19:15,588 So, the first one has to do with sequencing data quality. 487 00:19:15,588 --> 00:19:18,057 Quality. This is a more technical one. 488 00:19:18,057 --> 00:19:20,782 But, as I said, researchers are only required to 489 00:19:20,782 --> 00:19:21,293 generate 490 00:19:21,661 --> 00:19:24,664 the sequence data necessary to answer their question. 491 00:19:26,198 --> 00:19:29,246 But more than that, if the kind of quality they 492 00:19:29,246 --> 00:19:29,635 need, 493 00:19:30,469 --> 00:19:32,511 the sort of resolution of the data 494 00:19:32,511 --> 00:19:33,472 that they need, 495 00:19:33,973 --> 00:19:36,365 if the, the research resolution is lower 496 00:19:36,365 --> 00:19:37,143 than perhaps 497 00:19:37,143 --> 00:19:40,846 the clinical resolution, if you can't use the data to, 498 00:19:41,881 --> 00:19:43,249 to if it's not 499 00:19:43,249 --> 00:19:45,897 amenable to, to a secondary analysis, 500 00:19:45,897 --> 00:19:47,186 then that's fine. 501 00:19:47,186 --> 00:19:50,222 You don't have to produce, data 502 00:19:50,523 --> 00:19:53,492 of higher quality or of different type, 503 00:19:53,793 --> 00:19:56,762 to, to do sequence, to do secondary analysis. 504 00:19:57,229 --> 00:20:00,433 The second issue is equity within a protocol. 505 00:20:00,433 --> 00:20:03,091 This is something we talked about a lot in our working 506 00:20:03,091 --> 00:20:03,436 group. 507 00:20:04,703 --> 00:20:06,809 So there's a baseline expectation 508 00:20:06,809 --> 00:20:08,340 that similarly situated 509 00:20:08,641 --> 00:20:11,101 participants would be treated the same, 510 00:20:11,101 --> 00:20:12,678 but it can be acceptable 511 00:20:12,678 --> 00:20:15,748 to treat different groups within a protocol differently. 512 00:20:16,082 --> 00:20:20,198 So, so let's imagine you have a protocol where you've 513 00:20:20,198 --> 00:20:20,586 got, 514 00:20:21,687 --> 00:20:24,457 one group of people who have like an active disease 515 00:20:24,457 --> 00:20:26,448 and they're getting lots of clinical care 516 00:20:26,448 --> 00:20:27,760 and being seen repeatedly. 517 00:20:28,060 --> 00:20:30,721 And you've got another group of people 518 00:20:30,721 --> 00:20:33,032 who are, maybe relatives of that 519 00:20:33,466 --> 00:20:35,772 of of those people who don't have disease, 520 00:20:35,772 --> 00:20:37,803 but maybe have some genetic markers, 521 00:20:38,404 --> 00:20:41,373 and are just being seen once or twice, for a sort of, 522 00:20:42,308 --> 00:20:44,910 control or kind of, 523 00:20:44,910 --> 00:20:46,011 purposes. 524 00:20:46,011 --> 00:20:48,005 It's okay to treat those two groups 525 00:20:48,005 --> 00:20:50,282 differently if we're, if we're focusing 526 00:20:50,282 --> 00:20:53,385 on, clinical relationship, if different groups 527 00:20:53,385 --> 00:20:55,045 within a protocol have, different 528 00:20:55,045 --> 00:20:56,956 clinical relationships with the team, 529 00:20:56,956 --> 00:20:59,465 then then it's okay to give some of them 530 00:20:59,465 --> 00:21:00,092 secondary 531 00:21:00,092 --> 00:21:03,095 findings and others not, 532 00:21:04,763 --> 00:21:07,533 and, and particularly remember, we've got an insular care model 533 00:21:07,533 --> 00:21:09,339 where that one of the, one of the, 534 00:21:09,339 --> 00:21:11,570 considerations is whether the researchers 535 00:21:11,570 --> 00:21:15,875 are in a unique position to give secondary findings. 536 00:21:16,275 --> 00:21:18,380 If participants have already had clinical 537 00:21:18,380 --> 00:21:18,944 sequencing 538 00:21:18,944 --> 00:21:22,114 and already have access to this, this information, 539 00:21:22,114 --> 00:21:24,534 then the researchers shouldn't be the ones to provide 540 00:21:24,534 --> 00:21:24,717 it. 541 00:21:24,717 --> 00:21:26,346 They already have gotten it through 542 00:21:26,346 --> 00:21:28,254 the normal course of their medical care. 543 00:21:30,556 --> 00:21:33,559 And I'll talk about this again later. 544 00:21:33,559 --> 00:21:36,025 There's some end of life situations where 545 00:21:36,025 --> 00:21:38,430 the the beneficence reasons to provide, 546 00:21:40,332 --> 00:21:43,335 this information to people become less salient. 547 00:21:43,903 --> 00:21:46,334 If someone's going to die imminently, 548 00:21:46,334 --> 00:21:48,240 the fact that they have risk 549 00:21:48,240 --> 00:21:49,308 of some sort of cancer 550 00:21:49,308 --> 00:21:51,108 that's going to develop 15 years from 551 00:21:51,108 --> 00:21:52,811 now isn't really relevant to them. 552 00:21:52,811 --> 00:21:55,156 And so it's okay, again, particularly 553 00:21:55,156 --> 00:21:57,817 in the, in the cancer genetics realm, to, 554 00:21:58,350 --> 00:22:01,487 to, to treat people differently for that reason. 555 00:22:02,655 --> 00:22:05,336 But whatever you're going to do, you have to clearly set 556 00:22:05,336 --> 00:22:05,958 expectations 557 00:22:05,958 --> 00:22:07,970 and explain it to people in the protocol 558 00:22:07,970 --> 00:22:09,228 and in the consent form. 559 00:22:11,564 --> 00:22:14,733 Re-analysis, this question comes up a lot. 560 00:22:14,733 --> 00:22:17,703 You've you've got the genome sequenced. 561 00:22:18,070 --> 00:22:20,072 You do the analysis secondary analysis. 562 00:22:20,072 --> 00:22:22,041 You find out that they do or don't have something. 563 00:22:22,041 --> 00:22:25,411 But our knowledge about genetics is constantly evolving. 564 00:22:25,411 --> 00:22:27,833 And the question is whether researchers 565 00:22:27,833 --> 00:22:29,882 have an obligation to go back to 566 00:22:30,149 --> 00:22:33,390 the sequence data they've already obtained and re 567 00:22:33,390 --> 00:22:33,919 analyze 568 00:22:33,919 --> 00:22:36,963 it periodically as the state of genomic 569 00:22:36,963 --> 00:22:38,524 knowledge advances. 570 00:22:38,524 --> 00:22:40,092 In our argument was no, 571 00:22:40,092 --> 00:22:43,596 you can in the research context, research is in clinical care. 572 00:22:43,596 --> 00:22:45,798 So one time secondary analysis is 573 00:22:45,798 --> 00:22:46,599 sufficient. 574 00:22:46,832 --> 00:22:48,993 You're welcome to do more if you have 575 00:22:48,993 --> 00:22:51,036 the resources and the inclination. 576 00:22:51,036 --> 00:22:53,872 But but one time is is enough. 577 00:22:53,872 --> 00:22:57,309 Again clearly explain this in the protocol and the consent. 578 00:22:59,044 --> 00:23:01,547 And and let people know that 579 00:23:01,547 --> 00:23:04,483 sorry that if they want 580 00:23:04,483 --> 00:23:08,153 more clinical genomic testing that's available to them. 581 00:23:08,153 --> 00:23:11,156 But outside of the research context, 582 00:23:11,890 --> 00:23:12,891 negative findings. 583 00:23:12,891 --> 00:23:15,854 So the Acmg list is the one that people use most 584 00:23:15,854 --> 00:23:16,595 often. And, 585 00:23:18,264 --> 00:23:19,598 I don't know the exact number right now, 586 00:23:19,598 --> 00:23:24,169 but estimates are about like 3 or 4% of people will have, 587 00:23:25,104 --> 00:23:27,906 one of those findings when you analyze their genome. 588 00:23:27,906 --> 00:23:30,843 So it's not a huge number, but it's substantial. 589 00:23:30,843 --> 00:23:33,989 But that means that 97, 98, 90, 96, 590 00:23:33,989 --> 00:23:35,247 97% of people 591 00:23:35,681 --> 00:23:39,666 aren't going to have a positive finding for that high bar acmg 592 00:23:39,666 --> 00:23:40,052 list. 593 00:23:40,252 --> 00:23:41,787 So the question is, how do you convey 594 00:23:41,787 --> 00:23:43,322 that negative information to people? 595 00:23:43,756 --> 00:23:46,892 And we argued, in our report that you don't 596 00:23:46,892 --> 00:23:50,162 need to generate and convey a negative findings report. 597 00:23:50,162 --> 00:23:52,164 This is something you do in clinical care. 598 00:23:52,164 --> 00:23:54,596 If you've got clinical genetic testing 599 00:23:54,596 --> 00:23:56,068 and nothing came back, 600 00:23:56,068 --> 00:23:59,071 you'd get a report saying nothing came back. 601 00:23:59,672 --> 00:24:01,430 But we said, that's too much work 602 00:24:01,430 --> 00:24:02,975 for the research enterprise. 603 00:24:02,975 --> 00:24:03,776 Like we don't. 604 00:24:03,776 --> 00:24:06,052 We need to clearly explain to people 605 00:24:06,052 --> 00:24:08,580 that they're only going to be contacted 606 00:24:08,580 --> 00:24:09,915 if they have a positive finding, 607 00:24:09,915 --> 00:24:12,619 and not to assume that they've got a clean bill of 608 00:24:12,619 --> 00:24:13,052 health. 609 00:24:13,052 --> 00:24:15,603 But the, the, the going through the extra 610 00:24:15,603 --> 00:24:17,656 step of generating and conveying 611 00:24:17,656 --> 00:24:19,972 a negative report was, was going to be too 612 00:24:19,972 --> 00:24:21,627 much in the way of resources. 613 00:24:24,063 --> 00:24:24,363 Okay. 614 00:24:24,363 --> 00:24:26,501 Now moving on to some more interesting 615 00:24:26,501 --> 00:24:28,133 sort of normative questions. 616 00:24:29,335 --> 00:24:32,071 What do you do when you've got a, 617 00:24:32,071 --> 00:24:35,502 pro band who's deceased, but the information you've 618 00:24:35,502 --> 00:24:36,175 generated 619 00:24:36,175 --> 00:24:39,311 about their genome could be relevant to, 620 00:24:39,311 --> 00:24:42,314 to their relatives. 621 00:24:42,314 --> 00:24:45,451 And so I have, 622 00:24:45,451 --> 00:24:46,852 I have the paper here. 623 00:24:46,852 --> 00:24:48,787 Yeah. So we wrote a paper. 624 00:24:48,787 --> 00:24:50,322 Side note, this was a paper I wrote. 625 00:24:50,322 --> 00:24:52,760 The former fellow who, if any of you watch 626 00:24:52,760 --> 00:24:54,560 jeopardy bench here, and he's, 627 00:24:55,294 --> 00:24:57,318 he he had a run there for a while 628 00:24:57,318 --> 00:24:59,465 that, that got a lot of attention. 629 00:24:59,465 --> 00:25:02,468 He's the first author on this paper. 630 00:25:02,968 --> 00:25:04,069 So we argued that. 631 00:25:04,069 --> 00:25:07,072 Well, well, obligations to relatives. 632 00:25:07,773 --> 00:25:09,517 Have to be less than the obligations 633 00:25:09,517 --> 00:25:10,776 you have to the problems. 634 00:25:11,043 --> 00:25:12,814 There can be cases where the clinical 635 00:25:12,814 --> 00:25:14,346 information is important enough 636 00:25:14,613 --> 00:25:17,316 that you'd want to try to get it to them. 637 00:25:17,316 --> 00:25:19,243 And so, but we argue that that's a very 638 00:25:19,243 --> 00:25:21,220 high bar, and you should only really be 639 00:25:21,754 --> 00:25:24,597 returning findings when the information has 640 00:25:24,597 --> 00:25:25,324 potential. 641 00:25:25,324 --> 00:25:27,777 Direct and really substantial 642 00:25:27,777 --> 00:25:28,961 implications, 643 00:25:29,661 --> 00:25:32,989 for their health, and the depth of the relationship with the 644 00:25:32,989 --> 00:25:33,599 probe and, 645 00:25:33,599 --> 00:25:36,735 and their family is relevant here as our protocol resources. 646 00:25:36,735 --> 00:25:39,089 So this this is not an expansive 647 00:25:39,089 --> 00:25:39,972 obligation. 648 00:25:40,406 --> 00:25:43,242 And really a reasonable effort standard is sufficient. 649 00:25:43,242 --> 00:25:44,943 But you should, you know, document 650 00:25:44,943 --> 00:25:45,644 what you did. 651 00:25:45,644 --> 00:25:46,378 There's no 652 00:25:46,378 --> 00:25:48,380 no expectation here that you've got to hire 653 00:25:48,380 --> 00:25:50,883 a private investigator and track down a family member, 654 00:25:51,984 --> 00:25:53,585 to, to return information to them. 655 00:25:53,585 --> 00:25:55,663 But if you if you can do so relatively 656 00:25:55,663 --> 00:25:57,523 easily and it's really important, 657 00:25:58,323 --> 00:26:01,326 it's a good thing to do. 658 00:26:01,727 --> 00:26:02,761 Another big question 659 00:26:02,761 --> 00:26:05,764 that we've been struggling with is, is 660 00:26:05,764 --> 00:26:08,634 how to think about settings that have different 661 00:26:08,634 --> 00:26:11,031 standards of care or different, access 662 00:26:11,031 --> 00:26:12,671 to health care resources. 663 00:26:13,038 --> 00:26:16,189 So the question is, is there a duty to return 664 00:26:16,189 --> 00:26:17,309 findings in low 665 00:26:17,309 --> 00:26:19,601 resource settings where benefit will be 666 00:26:19,601 --> 00:26:21,246 substantially less certain. 667 00:26:21,246 --> 00:26:24,383 So, to return to breast cancer case, 668 00:26:25,217 --> 00:26:28,720 if you found out that someone had a BRCA variant, 669 00:26:29,888 --> 00:26:33,826 but they live in a in a country where there are no surgical 670 00:26:33,826 --> 00:26:34,293 suites 671 00:26:34,293 --> 00:26:37,262 that could could treat breast cancer, 672 00:26:37,563 --> 00:26:40,065 should you be returning this information 673 00:26:40,065 --> 00:26:43,068 in that kind of setting? 674 00:26:43,435 --> 00:26:45,938 We call it the Actionability problem. So. 675 00:26:45,938 --> 00:26:47,661 So you only want to return things 676 00:26:47,661 --> 00:26:49,541 one of the, one of the core reasons 677 00:26:49,541 --> 00:26:51,576 we're returning things where we have the 678 00:26:51,576 --> 00:26:52,644 we feel this there's 679 00:26:52,644 --> 00:26:54,355 this idea, there's this obligation 680 00:26:54,355 --> 00:26:55,814 to return secondary findings 681 00:26:55,814 --> 00:26:57,455 so that people can do something about it 682 00:26:57,455 --> 00:26:58,851 so they can improve their health. 683 00:26:59,151 --> 00:27:00,219 But if they can't actually 684 00:27:00,219 --> 00:27:02,286 improve their health because they can't 685 00:27:02,286 --> 00:27:04,089 access resources, what do you do? 686 00:27:05,157 --> 00:27:07,693 Some people have argued that 687 00:27:07,693 --> 00:27:11,465 you should only return things when they're actionable in the 688 00:27:11,465 --> 00:27:12,030 setting, 689 00:27:12,030 --> 00:27:15,033 in the in the community where the person lives. 690 00:27:15,801 --> 00:27:18,036 I disagree. 691 00:27:18,036 --> 00:27:21,039 We wrote another paper about this, but, 692 00:27:22,007 --> 00:27:23,375 I think the argument is 693 00:27:23,375 --> 00:27:26,378 we should default to returning 694 00:27:26,912 --> 00:27:28,480 research results. 695 00:27:28,480 --> 00:27:31,483 Genetic research results in these settings. 696 00:27:31,817 --> 00:27:32,618 For a number of reasons. 697 00:27:32,618 --> 00:27:33,519 I'm happy to get into 698 00:27:33,519 --> 00:27:35,929 if people are interested, but there are a couple of 699 00:27:35,929 --> 00:27:36,355 caveats, 700 00:27:36,355 --> 00:27:38,891 a couple reasons that this case is different. 701 00:27:38,891 --> 00:27:41,936 First, because communities cultures 702 00:27:41,936 --> 00:27:43,328 have different, 703 00:27:44,029 --> 00:27:46,498 relationships with genetic information, 704 00:27:46,498 --> 00:27:49,902 it's it's appropriate to consult with those those 705 00:27:49,902 --> 00:27:50,736 communities 706 00:27:50,736 --> 00:27:52,872 before you start returning the findings 707 00:27:52,872 --> 00:27:55,173 to see whether these are actually wanted. 708 00:27:55,908 --> 00:27:57,840 And two, I'll talk about the right 709 00:27:57,840 --> 00:27:59,545 not to know in a minute, but, 710 00:28:00,612 --> 00:28:01,713 I'll tip my hand. 711 00:28:01,713 --> 00:28:05,751 I'm a right not enough skeptic, but this is one case where, 712 00:28:06,118 --> 00:28:08,813 I think it's important to ask people, individuals, 713 00:28:08,813 --> 00:28:09,621 whether or not 714 00:28:09,621 --> 00:28:11,089 they want to know this information, 715 00:28:11,089 --> 00:28:14,385 because it really if you can't do anything about 716 00:28:14,385 --> 00:28:14,660 it, 717 00:28:15,160 --> 00:28:16,848 that's a case where someone might 718 00:28:16,848 --> 00:28:18,230 or might not want to know. 719 00:28:22,668 --> 00:28:24,903 All right, a few more, 720 00:28:24,903 --> 00:28:27,205 legacy samples. 721 00:28:27,205 --> 00:28:29,007 This is what I call the freezer problem. 722 00:28:29,007 --> 00:28:32,802 So a lot of a lot of protocols have these samples that were 723 00:28:32,802 --> 00:28:33,445 collected 724 00:28:33,445 --> 00:28:36,230 decades ago with sort of minimal informed 725 00:28:36,230 --> 00:28:37,249 consent before 726 00:28:37,249 --> 00:28:41,186 even genomic sequencing was, was, possibility. 727 00:28:41,587 --> 00:28:43,589 And then they want to go back to these samples. 728 00:28:43,589 --> 00:28:45,728 Sometimes they're very valuable 729 00:28:45,728 --> 00:28:46,625 collections, 730 00:28:46,625 --> 00:28:49,194 and sequence them now that the technology exists. 731 00:28:49,194 --> 00:28:51,171 And the question is what do you do about 732 00:28:51,171 --> 00:28:53,198 genomic findings when you sequence them? 733 00:28:54,633 --> 00:28:56,807 Oftentimes they have very general language 734 00:28:56,807 --> 00:28:57,636 in the consent, 735 00:28:57,636 --> 00:28:59,807 like we want to do genetic research 736 00:28:59,807 --> 00:29:02,040 that doesn't mention anything about 737 00:29:02,507 --> 00:29:05,510 sequencing or secondary findings or anything. 738 00:29:06,311 --> 00:29:08,086 Is it ethical to allow researchers to 739 00:29:08,086 --> 00:29:08,614 sequences? 740 00:29:08,614 --> 00:29:11,002 I've got a lot of thoughts on this, 741 00:29:11,002 --> 00:29:11,617 but but, 742 00:29:12,417 --> 00:29:14,786 generally my view is, 743 00:29:16,688 --> 00:29:18,883 that really we should only be thinking 744 00:29:18,883 --> 00:29:19,691 prospectively 745 00:29:20,025 --> 00:29:23,161 that it's fine for people to use these samples, 746 00:29:24,296 --> 00:29:25,831 and sequence them. 747 00:29:25,831 --> 00:29:27,322 But that we shouldn't be returning 748 00:29:27,322 --> 00:29:28,900 incidental findings in these cases. 749 00:29:29,267 --> 00:29:31,117 A lot of times these people are deceased 750 00:29:31,117 --> 00:29:33,105 or they're going to be impossible to find. 751 00:29:34,373 --> 00:29:36,108 And the, 752 00:29:36,108 --> 00:29:37,965 and or they'd be shocked to learn 753 00:29:37,965 --> 00:29:39,878 that these things are being done. 754 00:29:39,878 --> 00:29:43,348 And so, so I think it's okay if you've got consent 755 00:29:43,348 --> 00:29:46,351 to use these samples, it's fine to sequence them. 756 00:29:46,351 --> 00:29:49,289 But that that we don't need to be returning secondary 757 00:29:49,289 --> 00:29:49,788 findings 758 00:29:49,788 --> 00:29:52,791 in these kinds of cases. 759 00:29:52,791 --> 00:29:55,711 There might be some exceptions to that rule where, where it's 760 00:29:55,711 --> 00:29:56,094 sort of 761 00:29:57,129 --> 00:29:59,364 very defined small set of samples. 762 00:29:59,364 --> 00:30:01,500 And we could, we could recontact them. 763 00:30:01,500 --> 00:30:02,834 But generally unless 764 00:30:02,834 --> 00:30:05,347 unless that's the case, I don't think we have to return 765 00:30:05,347 --> 00:30:05,804 findings. 766 00:30:07,706 --> 00:30:08,674 This is another 767 00:30:08,674 --> 00:30:11,510 technical one that I only include 768 00:30:11,510 --> 00:30:13,732 because I'd get a question about it 769 00:30:13,732 --> 00:30:16,081 if I didn't, clear the clinical lab. 770 00:30:16,281 --> 00:30:17,849 Laboratory Improvement act. 771 00:30:17,849 --> 00:30:20,752 This governs how, 772 00:30:20,752 --> 00:30:23,755 sort of quality control in labs. 773 00:30:24,122 --> 00:30:26,124 And there's this rule under clear 774 00:30:26,124 --> 00:30:27,459 that you can't return 775 00:30:28,260 --> 00:30:30,458 information that will be used for clinical 776 00:30:30,458 --> 00:30:30,929 decision 777 00:30:30,929 --> 00:30:33,464 making without having validated it 778 00:30:33,464 --> 00:30:35,701 in a clear, compliant manner. 779 00:30:36,635 --> 00:30:39,805 And so researchers who don't live in the clinical 780 00:30:40,072 --> 00:30:43,275 testing space often say, well, do I have to do that? 781 00:30:43,742 --> 00:30:45,811 And the answer is yes. 782 00:30:45,811 --> 00:30:48,447 I've talked to various lawyers in, 783 00:30:48,447 --> 00:30:51,450 our institution in, in Egypt. 784 00:30:51,683 --> 00:30:55,654 So that's actually CMS, that that governs Clea. 785 00:30:56,121 --> 00:30:58,824 And the answer is yes. 786 00:30:58,824 --> 00:31:00,614 The law doesn't have any wiggle room 787 00:31:00,614 --> 00:31:01,460 for researchers. 788 00:31:01,460 --> 00:31:04,027 If you're going to be giving something back that they could 789 00:31:04,027 --> 00:31:04,463 plausibly 790 00:31:04,463 --> 00:31:07,499 use to actually make decisions, you have to validate it. 791 00:31:08,266 --> 00:31:10,102 But this is getting easier and easier. 792 00:31:10,102 --> 00:31:12,337 And so this really shouldn't be a barrier. 793 00:31:12,337 --> 00:31:15,273 It's not that expensive. And if you plan ahead, 794 00:31:16,575 --> 00:31:19,022 the big thing is you need a second sample to 795 00:31:19,022 --> 00:31:19,578 validate. 796 00:31:19,811 --> 00:31:22,215 So if you have a second sample you can 797 00:31:22,215 --> 00:31:24,049 you can have it in your back 798 00:31:24,049 --> 00:31:27,486 pocket to go, validate before you return results 799 00:31:27,753 --> 00:31:30,489 to participants. 800 00:31:30,489 --> 00:31:31,723 Pediatric findings. 801 00:31:31,723 --> 00:31:34,760 There's a whole I've written a lot about this, 802 00:31:36,261 --> 00:31:38,630 I don't want to get into too much detail here, but 803 00:31:38,630 --> 00:31:40,341 but there's some really interesting 804 00:31:40,341 --> 00:31:40,732 points. 805 00:31:40,732 --> 00:31:42,673 There's a debate about the right to an open 806 00:31:42,673 --> 00:31:43,034 future. 807 00:31:43,034 --> 00:31:46,104 If you're doing pediatric genetic research. 808 00:31:47,506 --> 00:31:49,674 The idea is 809 00:31:49,674 --> 00:31:52,234 you have the right to an open future is an argument that we 810 00:31:52,234 --> 00:31:52,711 should try 811 00:31:52,711 --> 00:31:54,780 to preserve, 812 00:31:54,780 --> 00:31:57,151 a child's ability to decide for themselves 813 00:31:57,151 --> 00:31:58,450 when they're an adult. 814 00:31:58,450 --> 00:32:01,820 As much as possible, and shouldn't be giving them 815 00:32:01,820 --> 00:32:04,321 genetic information that when they're 25, 816 00:32:04,321 --> 00:32:06,091 they might not want to know. 817 00:32:07,559 --> 00:32:10,629 I think there's sort of evolving views on, 818 00:32:10,629 --> 00:32:14,466 this, people are for, for the first, 819 00:32:14,466 --> 00:32:18,537 I don't know, 15 years of the, of the genetic research era. 820 00:32:19,304 --> 00:32:22,407 This was sort of dogma, but now, 821 00:32:22,407 --> 00:32:25,410 the second article is a, 822 00:32:28,113 --> 00:32:29,781 a sort of professional guidance, 823 00:32:29,781 --> 00:32:32,048 a professional society guidance document 824 00:32:32,048 --> 00:32:34,486 I worked on, where we started walking away 825 00:32:34,486 --> 00:32:37,155 from the idea that this right to an open 826 00:32:37,155 --> 00:32:37,622 future 827 00:32:37,956 --> 00:32:40,959 is of, such, 828 00:32:41,493 --> 00:32:43,862 paramount importance and that there can be conditions 829 00:32:43,862 --> 00:32:46,865 under which it's okay to give families access to, 830 00:32:48,867 --> 00:32:50,268 information about their kid 831 00:32:50,268 --> 00:32:53,271 that wouldn't be relevant, until their adult years. 832 00:32:54,472 --> 00:32:56,467 Re consent at the age of majority is 833 00:32:56,467 --> 00:32:57,409 another big one. 834 00:32:57,409 --> 00:32:59,731 You've sequenced someone's genome 835 00:32:59,731 --> 00:33:01,279 when they were a kid. 836 00:33:01,279 --> 00:33:03,748 They turn 18. Now, what do you do? 837 00:33:03,748 --> 00:33:06,117 Do you have to go back to them to get re consent, 838 00:33:07,185 --> 00:33:08,253 to continue using their 839 00:33:08,253 --> 00:33:11,323 samples to return results? 840 00:33:11,323 --> 00:33:13,084 Some colleagues and I have argued 841 00:33:13,084 --> 00:33:14,793 that you don't have to do that. 842 00:33:14,793 --> 00:33:17,796 This is one that where we are very much in the minority. 843 00:33:18,730 --> 00:33:20,644 Generally the policy is that you go to IRB 844 00:33:20,644 --> 00:33:21,099 and they, 845 00:33:21,099 --> 00:33:24,769 they ask for, you ask them for an exception. 846 00:33:24,769 --> 00:33:28,512 So functionally that the cases that you don't have to get re 847 00:33:28,512 --> 00:33:29,074 consent. 848 00:33:29,074 --> 00:33:31,543 But, but we don't for various reasons. 849 00:33:31,543 --> 00:33:33,527 We don't think you actually even need 850 00:33:33,527 --> 00:33:34,546 to take that step. 851 00:33:35,180 --> 00:33:36,481 And misattributed parentage. 852 00:33:36,481 --> 00:33:39,451 This comes up surprisingly frequently. 853 00:33:40,018 --> 00:33:43,054 This is another early case that I had, 854 00:33:43,054 --> 00:33:46,161 in my ethics consultation career, but, 855 00:33:46,161 --> 00:33:47,959 you're doing testing. 856 00:33:47,959 --> 00:33:49,942 You're comparing the genomes of the, 857 00:33:49,942 --> 00:33:51,429 the kid and their parents, 858 00:33:51,830 --> 00:33:53,619 and it's oftentimes you find that 859 00:33:53,619 --> 00:33:55,300 the parents aren't genetically 860 00:33:55,300 --> 00:33:57,845 related to the kid, or the father isn't related to 861 00:33:57,845 --> 00:33:58,303 the kid. 862 00:33:59,070 --> 00:34:00,872 And and what do you do? 863 00:34:00,872 --> 00:34:03,174 Do you disclose that information to people? 864 00:34:04,576 --> 00:34:05,243 I've argued. 865 00:34:05,243 --> 00:34:06,811 No, you you you don't. 866 00:34:06,811 --> 00:34:09,993 But, this is one where I think reasonable people could 867 00:34:09,993 --> 00:34:10,582 disagree. 868 00:34:12,350 --> 00:34:16,187 I've got about ten minutes, so I'm going to move on 869 00:34:16,521 --> 00:34:19,530 to the last big point that I'm going to set up and 870 00:34:19,530 --> 00:34:19,891 then, 871 00:34:20,191 --> 00:34:23,695 leave time for questions if people want to discuss it. 872 00:34:23,695 --> 00:34:27,232 This is something I've looked at quite a bit. 873 00:34:27,666 --> 00:34:30,669 And really a topic that, 874 00:34:31,369 --> 00:34:33,138 that was, 875 00:34:33,138 --> 00:34:37,208 was very controversial and and sort of, 876 00:34:39,077 --> 00:34:40,652 surprised that the vehemence with 877 00:34:40,652 --> 00:34:42,514 which people thought about this topic, 878 00:34:42,781 --> 00:34:46,450 which was surprising to me, which is why I've, I've sort of 879 00:34:46,450 --> 00:34:46,885 chosen 880 00:34:46,885 --> 00:34:49,461 to dive into it and really explore it in a lot 881 00:34:49,461 --> 00:34:50,021 of depth. 882 00:34:50,422 --> 00:34:52,958 Let me give you a case to set this up. 883 00:34:52,958 --> 00:34:56,194 So imagine P is having her genome sequenced. 884 00:34:58,163 --> 00:34:59,764 Could be research, could be clinical care. 885 00:34:59,764 --> 00:35:02,367 Let's say this is research ethics class. 886 00:35:02,367 --> 00:35:06,425 So for research purposes and during the informed consent 887 00:35:06,425 --> 00:35:07,005 process 888 00:35:07,238 --> 00:35:09,485 there's a checkbox that says we might find 889 00:35:09,485 --> 00:35:11,142 important medical information. 890 00:35:11,142 --> 00:35:13,945 One way circuits sequencing your genome. 891 00:35:13,945 --> 00:35:16,214 Would you want to know it yes or no? 892 00:35:16,214 --> 00:35:18,650 She says no. 893 00:35:18,650 --> 00:35:20,552 But during their analysis, 894 00:35:20,552 --> 00:35:24,526 the researchers find evidence of, high genetic 895 00:35:24,526 --> 00:35:25,390 risk for, 896 00:35:26,157 --> 00:35:30,095 non polyp forming colon cancer, which is the kind of thing 897 00:35:30,095 --> 00:35:33,531 where if you don't touch it, it's pretty uniformly fatal. 898 00:35:34,532 --> 00:35:36,835 But if you do know you have genetic risk, 899 00:35:36,835 --> 00:35:39,938 there are, enhanced screening techniques 900 00:35:39,938 --> 00:35:41,759 that you can use and you can catch it 901 00:35:41,759 --> 00:35:42,941 early and it's curable. 902 00:35:43,208 --> 00:35:45,073 And so, like knowing that you have 903 00:35:45,073 --> 00:35:47,212 this genetic risk is really important. 904 00:35:47,512 --> 00:35:49,358 It's it's literally a matter of life and 905 00:35:49,358 --> 00:35:49,681 death. 906 00:35:50,915 --> 00:35:52,543 And so the question is she checked 907 00:35:52,543 --> 00:35:53,118 the no box. 908 00:35:53,118 --> 00:35:56,287 She said she didn't want this kind of secondary information. 909 00:35:56,287 --> 00:35:58,656 Should they disclose the finding? 910 00:35:58,656 --> 00:36:01,985 Should the researchers disclose the finding even though she said 911 00:36:01,985 --> 00:36:02,193 no? 912 00:36:02,193 --> 00:36:05,597 This is sort of the classic case in the right not to know. 913 00:36:06,931 --> 00:36:08,500 Let me just give you a little bit of set up, and then 914 00:36:08,500 --> 00:36:08,767 I'd love 915 00:36:08,767 --> 00:36:12,003 to hear your thoughts on that kind of situation. 916 00:36:13,571 --> 00:36:15,306 Early on in the genomic 917 00:36:15,306 --> 00:36:18,309 findings result, research. 918 00:36:19,277 --> 00:36:20,178 Debate. 919 00:36:20,178 --> 00:36:22,494 This was actually an area where we thought we had 920 00:36:22,494 --> 00:36:23,014 consensus. 921 00:36:23,014 --> 00:36:24,082 We should only be. 922 00:36:24,082 --> 00:36:24,749 The idea was 923 00:36:24,749 --> 00:36:26,465 we should only be returning findings 924 00:36:26,465 --> 00:36:28,420 when they're desired by the participant. 925 00:36:28,853 --> 00:36:31,644 There's an obligation such to accept there's an 926 00:36:31,644 --> 00:36:32,357 obligation. 927 00:36:32,357 --> 00:36:34,871 It's only to offer individual findings 928 00:36:34,871 --> 00:36:36,261 to research subjects 929 00:36:36,261 --> 00:36:38,629 and let them know that they could say 930 00:36:38,629 --> 00:36:39,397 no, solicit 931 00:36:39,397 --> 00:36:41,278 their preference, and only give it to them 932 00:36:41,278 --> 00:36:42,801 when there's some sort of opt in. 933 00:36:42,801 --> 00:36:45,061 When someone has positively indicated 934 00:36:45,061 --> 00:36:47,505 that they want these kinds of findings. 935 00:36:49,274 --> 00:36:51,142 Then the acmg 936 00:36:51,142 --> 00:36:54,145 produced these recommendations or very controversial, 937 00:36:55,046 --> 00:36:57,449 where they, produced this list. 938 00:36:57,449 --> 00:37:00,452 And I should stress the list that I referenced earlier, 939 00:37:00,452 --> 00:37:02,712 these are all things that are associated 940 00:37:02,712 --> 00:37:03,955 with serious disease, 941 00:37:05,423 --> 00:37:08,159 where there's something you can do about it 942 00:37:08,159 --> 00:37:11,830 and where the evidence base is, is relatively strong. 943 00:37:11,830 --> 00:37:15,867 So not Huntington's disease, not Alzheimer's, 944 00:37:15,867 --> 00:37:18,870 where there's nothing you can do about it. 945 00:37:19,871 --> 00:37:22,407 But in these recommendations, they controversially argued 946 00:37:22,407 --> 00:37:24,743 that they should just this should just be the default, 947 00:37:24,743 --> 00:37:26,458 that we should just be returning these things to 948 00:37:26,458 --> 00:37:26,745 people, 949 00:37:26,745 --> 00:37:29,748 and we should not be asking about their preferences. 950 00:37:31,015 --> 00:37:33,618 The world exploded. 951 00:37:33,618 --> 00:37:36,824 It was like every every meeting I went to, every conference I 952 00:37:36,824 --> 00:37:37,455 went to for 953 00:37:37,455 --> 00:37:41,907 for three years, people devolved into shouting matches and their 954 00:37:41,907 --> 00:37:42,393 tears. 955 00:37:42,393 --> 00:37:43,628 It was it was a very dramatic, 956 00:37:45,663 --> 00:37:47,298 eventually you walked back 957 00:37:47,298 --> 00:37:49,738 their recommendations, sort of walked back their 958 00:37:49,738 --> 00:37:50,602 recommendations. 959 00:37:51,836 --> 00:37:53,973 And so there's this debate about the 960 00:37:53,973 --> 00:37:56,407 right now to know proponents argue that, 961 00:37:56,407 --> 00:37:58,773 you know, we we if you return information 962 00:37:58,773 --> 00:37:59,811 without people's, 963 00:38:02,547 --> 00:38:03,715 without people's being, 964 00:38:03,715 --> 00:38:06,225 without their permission, it's a violation of their 965 00:38:06,225 --> 00:38:06,718 autonomy. 966 00:38:07,519 --> 00:38:10,588 And they would say to that first case that I started with 967 00:38:10,588 --> 00:38:13,542 that autonomy concern should take priority over 968 00:38:13,542 --> 00:38:14,359 beneficence. 969 00:38:14,592 --> 00:38:16,861 That person said, no. 970 00:38:16,861 --> 00:38:19,297 If her colon cancer risk, 971 00:38:19,297 --> 00:38:22,300 we shouldn't return that to her. 972 00:38:23,101 --> 00:38:24,202 I'm a skeptic. 973 00:38:24,202 --> 00:38:27,205 I think I said that already. 974 00:38:27,672 --> 00:38:30,074 I can read my whole argument if you want, 975 00:38:30,074 --> 00:38:32,464 but I think it's philosophically shaky 976 00:38:32,464 --> 00:38:33,344 as a concept. 977 00:38:33,344 --> 00:38:34,712 It's actually only the right notes. 978 00:38:34,712 --> 00:38:37,482 And it was only about 25 years old. 979 00:38:37,482 --> 00:38:41,052 And the early literature was what's much less, 980 00:38:41,386 --> 00:38:45,089 strident about the strength of someone's right not to know. 981 00:38:46,124 --> 00:38:46,691 People make 982 00:38:46,691 --> 00:38:50,041 arguments about the right to have the constitutionally 983 00:38:50,041 --> 00:38:50,662 protected 984 00:38:50,662 --> 00:38:52,440 right to refuse medical treatment, 985 00:38:52,440 --> 00:38:54,532 implying a right to refuse information. 986 00:38:55,300 --> 00:38:57,096 I think that's wrong for a variety 987 00:38:57,096 --> 00:38:59,103 of both legal and conceptual reasons. 988 00:39:00,238 --> 00:39:04,175 I've got some empirical data that show how easily, how easy 989 00:39:04,175 --> 00:39:04,576 it is 990 00:39:04,576 --> 00:39:06,837 to shift people's preferences about this, 991 00:39:06,837 --> 00:39:08,713 depending on how you word things. 992 00:39:10,582 --> 00:39:12,183 I think they're concepts like moral 993 00:39:12,183 --> 00:39:15,220 distress and genetic exceptionalism, 994 00:39:15,220 --> 00:39:17,282 that weigh in favor of being skeptical 995 00:39:17,282 --> 00:39:18,857 about the right not to know. 996 00:39:19,424 --> 00:39:20,925 And for a variety of reasons. 997 00:39:20,925 --> 00:39:23,171 I think if we really enshrined as a policy 998 00:39:23,171 --> 00:39:25,096 this strong conception of the right 999 00:39:25,096 --> 00:39:27,595 not to know that would do a lot more harm 1000 00:39:27,595 --> 00:39:28,266 than good. 1001 00:39:28,533 --> 00:39:32,370 The main argument I would make is from some, 1002 00:39:33,471 --> 00:39:35,940 an empirical project I did a few years ago 1003 00:39:35,940 --> 00:39:39,995 where we asked people who said no, who checked the no 1004 00:39:39,995 --> 00:39:40,378 box. 1005 00:39:41,813 --> 00:39:43,056 And we went back to them, give them just a 1006 00:39:43,056 --> 00:39:43,381 little bit 1007 00:39:43,381 --> 00:39:45,915 more information and said, if you please, make another 1008 00:39:45,915 --> 00:39:46,384 decision. 1009 00:39:46,651 --> 00:39:50,655 And, about half of them said no a second time 1010 00:39:50,989 --> 00:39:53,191 and we call them strong refusers, 1011 00:39:53,191 --> 00:39:55,393 but about half of them said yes. 1012 00:39:56,394 --> 00:39:57,729 And we call them weak refusals. 1013 00:39:57,729 --> 00:40:00,832 So they said no initially and then said yes later. 1014 00:40:02,433 --> 00:40:05,403 And the the big thing is 1015 00:40:05,403 --> 00:40:08,757 that those strong refusers remembered what they said 1016 00:40:08,757 --> 00:40:09,274 before. 1017 00:40:10,141 --> 00:40:11,568 They remembered that they said no, 1018 00:40:11,568 --> 00:40:12,910 and they said no a second time, 1019 00:40:12,910 --> 00:40:15,913 but the weak refuse, or 75% of them 1020 00:40:16,347 --> 00:40:17,849 misremembered what they had said. 1021 00:40:17,849 --> 00:40:21,563 No, they clearly made a mistake the first time, for most of them 1022 00:40:21,563 --> 00:40:21,853 did. 1023 00:40:22,654 --> 00:40:26,024 And they thought they had agreed to receive secondary findings. 1024 00:40:26,024 --> 00:40:30,434 And so if you enshrine this idea that we have to ask everyone 1025 00:40:30,434 --> 00:40:30,795 for, 1026 00:40:31,629 --> 00:40:33,113 their view, for their preferences 1027 00:40:33,113 --> 00:40:34,866 about whether or not they want to know 1028 00:40:35,099 --> 00:40:37,773 secondary findings, some of them are going to make a 1029 00:40:37,773 --> 00:40:38,236 mistake. 1030 00:40:38,536 --> 00:40:40,317 And this raises, for me at least, 1031 00:40:40,317 --> 00:40:42,206 a normative question about whether 1032 00:40:42,206 --> 00:40:45,137 we should be constructing right not to know policies to 1033 00:40:45,137 --> 00:40:45,777 accommodate 1034 00:40:46,344 --> 00:40:48,162 this tiny group of people who are strong 1035 00:40:48,162 --> 00:40:49,981 Refusers are really don't want to know, 1036 00:40:52,150 --> 00:40:54,285 given the significant harms 1037 00:40:54,285 --> 00:40:57,865 to patients or participants that might make a mistake and 1038 00:40:57,865 --> 00:40:58,556 mis report 1039 00:40:58,556 --> 00:41:00,158 their preferences on a consent form, 1040 00:41:00,158 --> 00:41:01,359 or to put a different way, 1041 00:41:01,359 --> 00:41:02,826 whose interests are more important 1042 00:41:02,826 --> 00:41:03,861 the weak or the strong? 1043 00:41:03,861 --> 00:41:06,324 Refusers the people who really don't want 1044 00:41:06,324 --> 00:41:06,864 to know, 1045 00:41:07,098 --> 00:41:08,923 like imposing information on them 1046 00:41:08,923 --> 00:41:11,302 that they might not want, or weak refusers 1047 00:41:11,302 --> 00:41:14,116 who are making a mistake and who really do want to know 1048 00:41:14,116 --> 00:41:14,372 this 1049 00:41:14,372 --> 00:41:15,873 potentially life saving information. 1050 00:41:15,873 --> 00:41:18,676 But just to have checked the wrong box. 1051 00:41:18,676 --> 00:41:20,956 I'd argue that we should protect the weak 1052 00:41:20,956 --> 00:41:21,512 refusers, 1053 00:41:21,512 --> 00:41:24,482 but again, reasonable people can disagree. 1054 00:41:25,216 --> 00:41:28,219 And this is, Oops, sorry. 1055 00:41:29,754 --> 00:41:31,422 I've been playing around with this idea. 1056 00:41:31,422 --> 00:41:33,301 We wrote a couple fellows, and I wrote 1057 00:41:33,301 --> 00:41:35,426 a law review article a couple of years ago 1058 00:41:35,626 --> 00:41:39,230 about the idea of choice masking nudges, the idea that 1059 00:41:39,230 --> 00:41:39,697 people 1060 00:41:39,697 --> 00:41:42,447 can still have a choice about whether or not to know 1061 00:41:42,447 --> 00:41:43,134 information. 1062 00:41:44,602 --> 00:41:46,963 But we can sort of hide that choice 1063 00:41:46,963 --> 00:41:47,705 from them. 1064 00:41:47,705 --> 00:41:49,107 So they'd have to raise their hand. 1065 00:41:49,107 --> 00:41:51,673 They'd have to those strong refusers would have 1066 00:41:51,673 --> 00:41:52,110 to say, 1067 00:41:52,110 --> 00:41:53,934 you know, I don't want to know this 1068 00:41:53,934 --> 00:41:54,612 information. 1069 00:41:54,612 --> 00:41:55,880 There'd be no question. 1070 00:41:55,880 --> 00:41:58,883 There'd be no information in the consent form. 1071 00:41:59,183 --> 00:42:01,714 Except for information about how to contact people with 1072 00:42:01,714 --> 00:42:02,220 questions. 1073 00:42:02,220 --> 00:42:04,522 There'd be no prompting. 1074 00:42:04,522 --> 00:42:06,090 I'm attracted to this idea. 1075 00:42:06,090 --> 00:42:08,292 I acknowledge that, 1076 00:42:08,292 --> 00:42:10,294 in the research ethics and bioethics 1077 00:42:10,294 --> 00:42:11,963 community, the idea of choice 1078 00:42:11,963 --> 00:42:14,538 masking nudges could be controversial, 1079 00:42:14,538 --> 00:42:16,300 depending on the context. 1080 00:42:16,300 --> 00:42:19,303 But, I think it's an interesting, 1081 00:42:19,670 --> 00:42:22,340 an interesting concept to play with. 1082 00:42:22,340 --> 00:42:22,673 All right. 1083 00:42:22,673 --> 00:42:25,308 I am nearly out of time. Let me just give you my. Oops, 1084 00:42:25,308 --> 00:42:25,643 sorry. 1085 00:42:25,643 --> 00:42:28,679 Bottom line, I think that, 1086 00:42:29,547 --> 00:42:32,250 we shouldn't explicitly solicit preferences 1087 00:42:32,250 --> 00:42:34,360 during the consent process about the right 1088 00:42:34,360 --> 00:42:36,320 not to know if people raise their hand 1089 00:42:36,320 --> 00:42:37,588 and say, I don't want to know. I think 1090 00:42:37,588 --> 00:42:38,790 that's fine, and we can honor that. 1091 00:42:38,790 --> 00:42:42,727 But that's the extent of, what I think we need to solicit, 1092 00:42:44,128 --> 00:42:46,641 with the exception that when there are people for 1093 00:42:46,641 --> 00:42:46,898 whom 1094 00:42:46,898 --> 00:42:49,257 the genetic findings might not be clinically 1095 00:42:49,257 --> 00:42:49,901 actionable, 1096 00:42:50,802 --> 00:42:52,845 it's okay in those cases to solicit 1097 00:42:52,845 --> 00:42:53,604 preferences, 1098 00:42:53,604 --> 00:42:56,523 but that's an exception, I think, to that to my general 1099 00:42:56,523 --> 00:42:56,841 rule.