 Good afternoon. My name is Sarah Hull. I am the director of a unit that's known as the NHGRI Bioethics Corps, and I am also the vice chair of the NHGRI intramural IRB. And I share a joint appointment, a faculty appointment with the clinical centers department of bioethics. And what I hope to accomplish today, in the next hour, is giving you a very practical perspective on the ethical and IRB issues associated with doing whole exome sequencing research. I want to note ahead of time that I'm possibly going to use the terms next generation sequencing, whole exome sequencing, and whole genome sequencing a little bit interchangeably, realizing that there's huge technical differences between these, but the ethical considerations maybe not quite so much. So if my use of terminology seems a little bit loose, I apologize for that, but we've been seeing protocols that use all of these technologies, and that's been influencing our thinking. Let's see. So I'm not sure how to advance. Oh, the other keyboard. There we go. So the views and the opinions that I'll be sharing with you are my own and not those of any official perspective of the agencies that employ me. That said, it doesn't actually feel quite right to say that these views and opinions are my own. They're very much a product of many interactions that I'm quite privileged to be able to have being here in the intramural program at the NIH. So I want to acknowledge, I want to do my acknowledgements upfront because I think that will help you to understand all the experiences that my comments are grounded in and based upon. First, I want to acknowledge my colleague, Ben Berkman, who with me runs the NHGRI bioethics core, and that involves providing research ethics consultation to our intramural scientists as well as helping to run and provide the infrastructure for our IRB. And so over the last few years, we've had the opportunity to do dozens of ethics consults related to questions that are arising about next generation sequencing research. From the earliest design stages all the way through to dissemination and questions that even come up at the time of publication or attempted publication. Too many of these consultations to count, but those are very much influencing my thinking about all of this and the evolution of my thinking over time. My role with the NHGRI IRB has also been very important and we've been, in a way, we're sort of a specialized IRB. We've been addressing issues related to genetics and genomics for the almost 13 years that we've now been up and running. And as early as 2006, we were starting to think in an anticipatory way about some of the questions related to next generation sequencing thanks to the ClinSeq protocol, which was one of the earliest ones that was starting to try out some of these kinds of techniques. And we've actually spent the last two annual retraining retreats that we run. We run half-day retreats and we've spent the last two dedicated to exploring in more depth issues that had been coming up around next-gen protocols in our IRB. I also want to call attention to the Clinical Center's Ethics Consultation Service, which is one of the roles that I'm involved with as part of my faculty appointment with the Department of Bioethics. Anybody who's here, who's part of the NIH, can call up our paging operator and request to get in touch with the Bioethics Fellow who's on call to ask for a consult, an ethics consult, to help with questions that might come up at any stage in the conductive research. This is a clinical service that's available to you. A number of consultations that I've been involved with through this referral mechanism have also dealt with questions around the ethics of whole exome and whole genome sequencing. I also just want to call out one additional person, Bashali Meskeri, who I know was here earlier. She's a part of the NISC program, and she illustrates the spirit of cooperation and collaboration between those who are working on the ethical and oversight side of things, like myself and the scientists who are at the computer and seeing the subject. She's become an example of a very good gatekeeper when somebody comes and talks to her and wants to set up sequencing at NISC. She knows when to refer them to talk to us about the ethical oversight and review. She makes sure everything is in place before they get too far down the pipelines that you've been hearing about today, and that's been just a really important aspect of our work. And there's many interactions with the other IRBs here at the NIH, as well as collaborators. A lot of my own research program is inspired by the kinds of questions that come out of these consultations, and I've listed all of the various collaborators on papers that are starting to come out about these questions, and I'm very grateful for the opportunity to be able to work on. And so let me get to the point. I think many of you here are here because you have a very basic question. You want to know, how do I get my whole exome sequencing protocol through the IRB? And to draw on David's presentation earlier, I think research in general, protocols in general, are all on a chariot race of sorts. Everybody wants to use this technology, get through the process, do the science, get the paper out, and we really don't want the IRB process to hold things up for sure. That said, I'm going to very politely suggest that this is not the right question to start with. First of all, there's actually more than one ethically appropriate way to do whole exome sequencing research. So on the next few slides, I'm including quotations from a qualitative project that I'm working on with Ben Berkman and others. We just started the analysis, but I have a few quotes from a question that we posed. This was interviews with IRB chairs and staff at three different sites, the NIH and academic medical centers in the Boston area and in the Seattle area. And we asked them about policies that their IRB has, either explicit or de facto policies about the ethics of whole exome sequencing research. And in this case, one administrator said, it's much more case by case. What are the protocols? Who are the people? What's the relationship between the investigators and the people whom they're studying? So here we get the sense that there's not a one-size-fits-all approach for handling this. Okay, another issue is that IRBs, and I'm perfectly willing to admit this, we're all still trying to figure this out. We're really struggling with the ethics of reviewing whole exome and whole genome sequencing protocols. So this quote is from an IRB chair. We certainly don't have a policy, and I don't know that we really have come to a firm conclusion. I mean, it gets discussed every time and there's disagreement every time. So this suggests that within the IRB, protocol to protocol, there are some disagreements. And I'll also mention that from IRB to IRB, even here within the NIH, there are some differences of opinion and different outcomes that we see. So I can feel the sense of frustration welling up in some of you, at least, who are here in the room. Another IRB chair said, we do not have an institutional policy. I think we've gone through several different discussions in our IRB for each specific protocol, but I think we're still at the stage where we hear from investigators what their approach is, and then we decide at the meeting if this sounds reasonable. I included this quote because I hope it gives you the sense that the process is a little bit more constructive, that IRBs are actually quite open to hearing from investigators what a reasonable plan is for their set of circumstances and to review it on a case-by-case basis to come work together to come out with the right set of answers and the right protocols and solutions to the ethical questions. Now, we actually do have a policy here at the NIH in the Intramural Research Program that pertains to the ethical review of whole exome and whole genome sequencing protocols. It's not a very prescriptive policy. It talks about the level of review that's required. So the policy states, since there are ethically relevant distinctions between traditional genetic research and whole exome and whole genome sequencing research, the use of these sequencing technologies must be reviewed explicitly by an IRB and or OHSRP. Now, OHSRP stands for the Office of Human Subjects Research Protections. It's an oversight office within the Intramural Program here at the NIH. So existing protocols will need to be amended to include whole exome sequencing and whole genome sequencing. And this was, this is a policy that was endorsed by a group called HISRAC, which is a collection of all the Intramural IRB chairs as well as higher leadership back in July of 2010. So you might be thinking to yourself, well, what if I already have consent to do genetic research on the specimens that I've collected from subjects in my protocols? This would still require ethical review oversight. It might be as simple as an amendment to add the new analyses to your protocol, but those would have to be reviewed by an IRB. Well, what if I'm only using coded specimens that were given to me from an outside, from an extramural collaborator? Or two would need some kind of institutional review. What if I'm not planning to return any incidental research findings? That would solve the problem, right? Well, that might be an acceptable approach depending on the circumstances, but it's an approach that would still need to be reviewed either by an IRB or OHSRP. So now you might be getting the sense that we're actually proceeding very cautiously deliberately when it comes to the ethical oversight of whole exome sequencing research. And you'd be correct about that, and that leads to another question, which is what's the big deal? What is so different about whole exome sequencing that everybody's taking it quite so seriously? Glad you asked. Here's the now somewhat familiar curves that map out the extraordinary speed in which sequencing technologies are able to generate kilobases and the extraordinary decline in costs that we've been talking about. And what this means is that it's much more efficient, much more accessible to use next-generation sequencing technologies. Investigators are scrambling to incorporate these into their protocols, and it's going to result in much, much more, exponentially more genetic and genomic information being generated about individual research subjects. And it's that magnitude that presses on some of the ethical questions. So in general, my colleagues and I have argued that these technologies don't actually alter or raise new ethical concerns. The concerns that we've been talking about in relation to genetic research, issues around adequate informed consent, how to protect confidentiality, how to minimize risks and maximize benefits and issues of inclusiveness and fairness, these are all still stable and intact and still apply to whole exome sequencing research. The issue is that because of the magnitude of information that's going to be generated, this will in turn magnify and make more concrete many of the concerns that were relatively hypothetical or theoretical until this point, and it's this that has implications for how we conduct ethical review of whole exome sequencing research, especially in the context of the regulatory framework that we're working within. So I think that these, the ethical issues that emerge from this context can be sorted out into three general categories. There's issues related to the management of all of this information and specifically what we do with individual research results. Both those that are considered to be primary or research related results as well as those that we've been calling secondary or often people refer to these as incidental, in other words incidental to the core research questions being asked. There's also a lot of interesting and challenging questions around the ethics of broad data sharing requirements and how that will play out in the context of large scale genomic sequencing. And then important questions to ask about the content of informed consent and whether it's adequate to cover all of these new kinds of plans and issues. And in some cases perhaps whether re-consent, re-contacting people and giving them new information might be required. So this is going to, I'm going to use this framework to organize the remainder of my comments this afternoon and I'm going to focus in mostly on questions around incidental research findings, incidental genomic research findings and whether or not they should be disclosed to participants because this is really where some of the most challenging questions have been coming up for us and where we've been spending a lot of time and effort in our work with investigators. So just a couple of conceptual clarifications in the world that we've seen shift in the last 13 years from a candidate gene targeted approach to genetic research to now next generation sequencing. A couple of our assumptions have changed. So an existing assumption is around traditional genetic research is that it will produce relatively few clinically significant incidental findings. It would be relatively rare to see these in the context of research and our revised assumption now that it's not so much a question of whether we're going to see clinically relevant results emerging from research and how many and of what character they will be in individual participants. A second assumption that I want to put out there that we've had is that it's actually really easy to distinguish between so-called incidental or secondary findings and those that are explicitly related to the original study hypothesis or disease focus and I think as research evolves over time and as we get focused on broader models of genomic medicine I think it's at least arguable that research that uses next generation sequencing is going to blur the distinction between what counts as incidental and what counts as related to the research. It's all going to be there and related to the research in some way and this distinction might become less meaningful over time. A third assumption is one that I've, and my colleagues informally refer to as don't look, don't tell. I understand that phrase is now available and what it refers to is that researchers generally, the idea that researchers don't have an obligation to be clinicians and to affirmatively search actively for incidental findings in their data. This was put out for example by Susan Wolfe and her colleagues in some of the very important papers they wrote early on about incidental findings in genetic research. The first assumption is that the nature of this technology makes it such that the act of looking for all possible results becomes much more practical. It's all there and very sifter. You just enter in the variance right into a computer program and it's a more fundamental component of the analytic approach. So it's going to be harder to tell people that you can just avoid, avert your glance and not see the incidental findings that are right there in the computer before you. This leads to some really interesting questions. The fundamental question here is whether all of this new technology changes the scope of the obligation that investigators have to disclose incidental genetic research findings. And I want to give a very cautious view on the answer to this question that I think that there's an emerging consensus that yes, there is some obligation to now be actively and deliberately disclosing incidental findings of genomic research to subjects, at least sometimes and in certain contexts. And that seems to be a trend that I'm seeing here across the NIH and that bioethics colleagues elsewhere are also talking about. So this represents a paradigm shift in how we think about incidental findings. And so given these new technologies, what does this mean for IRBs and how we review investigators' protocols and their approach to managing incidental findings? What I'd like to do now is present you with a case to chew on. This is a composite of a few investigators who have come before our IRB, but it's very realistic and grounded in their actual experience. So I'm going to present you with how they approach their protocol and I'm going to present you with the specific concerns that the IRB raised in response to the proposal. And hopefully that will illustrate for you how this is playing out in practice. So our investigator, and I call her well-intended, but I want to emphasize that everybody that I've been talking to is very well-intended here and wants to do the right thing, even while they're really enthusiastic about incorporating these technologies into their research. So I don't mean to single out that feature, but just to highlight it and be grateful for it. So a bench scientist, a PhD scientist who studies a common complex disorder and it could be cardiovascular disease, it could be diabetes. That's not quite so important, but to understand that it's more common might be helpful. This investigator wants to collect samples prospectively and to use whole exome sequencing analysis to help pinpoint some of the variations that they're interested in. The protocol that she puts together involves a one-time blood draw here at the clinical center. And the intention is that there's not going to be an ongoing clinical relationship or even research relationship between the subjects and the researcher team. And although the researcher's done a good job of partnering with some clinical investigators and certainly is working within the clinical center, there's nobody directly on her team who is a clinician and very relevantly, there's no genetic counselor that she has access to or counseling resources immediately at hand as part of this protocol, nor any funding to be able to bring one on board. And she very much wants to do the right thing, but is conflicted and a little bit ambivalent about what this actually means in practice. So she does her best to put together a protocol that tries to straddle this balance between disclosing results and doing it in a way that is practically feasible. So this is submitted to the IRB, and the IRB returned with a couple of comments in the category of whether or not the investigator should be engaged in disclosing incidental findings. The IRB felt, picked up on the fact that the protocol didn't clearly and consistently address the study procedures and most importantly the plan for providing genetic findings to subjects. And what they were picking up on here is the consent form and the protocol weren't exactly consistent with each other and even in the consent form it tried to both say that this probably wasn't going to happen but might happen and they were concerned that it just wasn't a coherent and internally consistent plan. The IRB went on to say that it's the decision of the investigator to determine which genetic findings will be returned to subjects. So this was in effect saying that the investigator has latitude to figure out what it is that they're going to be focusing in on and needs to come back to the IRB with a reasonable plan. So we didn't want to, there is no set list of variants that we could provide the investigator and it was up to her to come up with a plan to figure out what that was going to look like and come back to the IRB. Some other issues that came out around informed consent. The investigator has a duty to inform subjects clearly during the consent process about the possibility that using these techniques the study itself may identify genetic variants that could be of relevance to subjects but only some of which would be of known clinical significance. So it need to put out there and inform people that stuff is going to be learned basically through these techniques and then the consent form would also have to address whether or not the subjects would be provided with any of these results or perhaps if they would be given a choice about whether or not they wanted to be recontacted and to receive some of these results. And then some other issues as well. If a decision is made to return results they would have to be confirmed in a CLIA certified lab and that they would have to be disclosed with appropriate counseling available to the subjects. And the IRB suggested to the investigators they look into whether their institute, their home institution and or the clinical center might be able to provide some support in the area of genetic counseling consultation. So this came to the investigator in the form of a series of recommendations. The original protocol was tabled and this does happen and the ideas that the investigator can go back incorporate those suggestions and come back to the IRB and receive a full review. So the investigator did that, consulted with the bioethics core about the plan and then resubmitted a revised, a dramatically revised protocol to the IRB which was quite consistent about a plan for disclosing incidental findings and provided some criteria about what clinically relevant might look like. And the protocol also made clear that it was expected that the number of variants that would realistically be identified and disclosed were quite small and this was reflected in the consent form in a consistent way as well. So this time the protocol was approved but with a series of stipulations that still were attached to it. The IRB still wanted more specificity about the process that would be used to review incidental findings to determine which ones were clinically significant. And based on the IRB's review of other whole exome sequencing protocols it made a recommendation that the investigator might want to consider putting together a multidisciplinary committee and this committee would be used to review the variants that emerged in the analysis and help to determine which ones were actionable and would rise to the level of something that ought to be disclosed to a subject or at least that they'd be given the option of receiving. The IRB suggested that this should... The committee should include a broad range of expertise including medical genetics, genetic counseling, ethics and the IRB perspective and it also pointed out that this review should be occurring before any decisions are made about which variants should be confirmed prior to disclosure. And this was in part because we had a sense that the list was going to be a lot longer than the investigator was assuming. So the IRB also felt that it would be beneficial for the investigator to talk to some others who have been involved with this kind of research. People that we've been hearing from today, people who have been dealing with this and know a little something about how to interrogate the sequence data that they're generating for clinically significant variants. And the IRB also commented that their experience... The experience of these other protocols is such that probably a larger number of incidental variants would be identified as I just mentioned. It's really interesting for me to take a step back now and look at that last bullet point because I think maybe as an IRB, our own IRB has shifted somewhat in its thinking about this issue over time. It's clear to us now that the number of incidental or secondary variants that's going to be generated really depends on how hard you're looking. And I think Jennifer's presentation just before gave us a sense of how possible it is to find incidental variants that might be of clinical significance, but also really how complicated it is at the same time. And I think the IRB has developed a more nuanced perspective on this question since these comments were made that suggests that we're not quite sure in different studies we think different amounts of looking might be happening and might be indicated. And so we've seen some more variation in how investigators are now proposing to handle the how hard should we look question. So I hope that gives you a little bit of a sense of what themes are coming up at the IRB. What are some of the principles and what are some of the questions that we're asking? I highlighted the example of a PhD scientist to show that not everybody doing this is trained in the field of medical genetics and maybe we'll have to draw on the expertise of others. But I think some of the same questions and issues that I raised that the IRB raised will apply to other kinds of studies as well. Now what I'd like to do is take a step back and just look a little more deeply into what is behind some of these IRB concerns? What ethical principles are at stake and why is the IRB so concerned about some of these questions or why this is also complicated? And there's good sound ethical arguments in favor of disclosing incidental findings and arguing against such disclosure. And that's the nature of bioethics, the nature of a true dilemma. My colleague, Frank Miller, who's a faculty member in the Department of Bioethics and one of our past fellows, David Shallowitz, argued that an obligation to return findings is grounded in the principle of respect for persons. So they said it would be disrespectful to treat research volunteers as conduits for generating scientific data without giving due consideration to their interest in receiving information about themselves derived from their participation in research. So some of what's behind this is the idea that there's some genetic information that's actually quite important, quite relevant. And if somebody had it in their possession, it might enhance their decision-making. It would enhance their ability to make autonomous decisions about their health and other life considerations. And also, I think what's embedded in this argument is an idea that returning results also recognizes and sort of repays a participant's contribution to research. There are a number of other arguments that are made in support of an obligation to return genetic research results. The principle of beneficence, the general idea of doing good, of doing something that's beneficial or sort of raising the prospect of benefit of participating in a study. The notion of reciprocity, the idea that is actually related to the point on the last slide about giving participants something back, recognizing their contribution and realizing it's fair to give something back if you can. The principle of justice, it's a little more of a nuanced point, but it has to do with the idea that some of our research participants are going to be in categories that are considered more vulnerable. They may have less access to health care resources. And if they're in a study where it may be their only access to certain kinds of services and certain kinds of information, that might heighten the obligation for us to disclose what we know because we can't reasonably expect them to be able to find out this information somewhere else. So a principle of fairness would say we might treat certain populations in a particular way and we might give them certain kinds of information as a result. The principle or issues around investigator integrity and professional responsibility, I think of these as having to do with what some medical geneticists have described to me as the feeling of wanting to be able to sleep at night. The idea that if they're engaged in this research and they know enough to know that what's sitting there on their computer is actually relevant to somebody's health care, they think it's a matter of professional responsibility to make this information available. These are often going to be investigators who are playing a dual role. There's a dual aspect to their profession and a research aspect, but the clinical aspects tug at them and would keep them up at night, and this is what drives their sense of obligation to return some findings. Okay. So on the flip side, some arguments against an obligation to return incidental findings. So there are some who dispute that these principles that I've just described, beneficence, respect for persons, atrocity and even justice, they dispute that these are violated by a failure to disclose or a lack of disclosure. They will emphasize that the core purpose of research is to generate generalizable knowledge. Research is different than clinical care and we shouldn't be in the business of recreating a clinical setting. And in fact, there's risks of doing so that conflating research and clinical care to do things like lead to therapeutic misconceptions where people start to expect more from the research environment than perhaps they should. We perhaps also don't want to overestimate the value of some of these results to individuals or we at least want to be really sure that they're of the sort that are clinically relevant, actionable and meet certain criteria and that many emerging results will not. A whole different set of arguments is made around practical considerations such as resource limitations. It is very resource intensive to do the sorts of analyses and disclosure that you've heard described here. It requires lots of technical expertise, lots of access to bioinformatics tools, access to counseling resources, the ability to pay for clear validation and we all know that we're operating in an environment without overflowingly abundant resources. We're living under resource constraints and so there might be some practical considerations that weigh against our obligation to disclose incidental findings that we have to take into account. At the very least, acknowledge that more resources are needed to be able to do this well across the board to be able to extrapolate from the ideal models that we're hearing described. So many have attempted to develop guidance and to sort of assimilate all of these competing values into a set of coherent principles that will help us figure out what do we actually do with all of these different commitments and obligations and as you can see there's actually quite a few guidelines out there. There's the NHLBI guidelines which are very commonly cited and were referred to earlier but there's a lot of other approaches out there as well. And when you look across all of these approaches they do have some common features. So they do come up with a generally consistent set of factors that should be applied to figure out which kind of results are the sort that generate an obligation for disclosure. So the characteristic of analytic validity, we want to make sure that this is an accurate, a real finding. Clinical relevance, it should be related to something that's actually important and relevant to one's health and healthcare. Actionability, the idea that there's something that can be done to alter the course of a negative outcome whether it's a therapeutic intervention, screening procedures, perhaps a different reproductive decision depending on what kind of a result you're talking about. And a fourth feature of almost all of these guidelines is that a result should be desired that individuals should be asked whether they want this information and that their decision about whether or not to receive it should be respected. But there's disagreements across the different guidelines as well. One area where there's some considerable disagreement is the level of clinical relevance that we should... What is the threshold? Where is the cutoff between clinically relevant and not? And there's different ways that this gets defined and debated in all of these guidelines. There's even disagreement about this question of desirability, whether we really want individual subjects to actively consent to receive all kinds of results. There may be, some people argue, there may be a lot of results that are just so important that it wouldn't be reasonable for somebody to say they wouldn't want them. It's a life-altering result, an easy intervention that would change the course of their health outcomes. And we might override their right not to know or override their interest in making a decision and say, no, you have a duty to disclose that. And there's a lot of debate about that very question at the moment in the bioethics field. There's a lot of questions about analytic validity and what the standards are there. There are some people who question whether CLIA certification or clinical validation in a CLIA-certified lab is required. Now, it would be irresponsible of me to suggest that that's actually a valid question to ask here at the NIH. It is our policy, it is a policy out there that we follow here at the NIH that, yes, if you are planning to disclose to a research subject any kind of clinical result for that matter, it does have to be confirmed in a CLIA-validated lab first. But this is an area of debate that's out there and whether this is an appropriate or even good enough standard in some circles. Okay, so up until now, I've been talking about the ethical arguments and guidelines about the kinds of information, the kinds of genomic information that might generate an obligation for disclosure. But there are some additional layers that I want to throw out there and encourage you to think about that have to do with contextual factors, either about the study characteristics or population characteristics that might alter our obligations to disclose incidental findings. So some of these factors have to do with the nature of the study. Is a clinical trial a treatment intervention where somebody is coming with a hope and expectation for some kind of benefit? Is it a natural history study? Is it a basic science protocol? And do these purposes have some influence on expectations? Study resources. What resources are available? I've already hinted at this. Are there adequate resources for appropriate counseling and confirmation? What's investigator expertise to be able to do the right kinds of analyses? What are the aims of a protocol and what counts as incidental to those aims? And other practical considerations like how long has it been since we've been in touch with these subjects? And can we recontact them? Do we have current information if we want to follow up with them and give them results? And then some factors related to subject characteristics. And again, I've also mentioned many of these. Do they have alternative access? Are they on this particular study for care and for information? What degree of vulnerability does a particular population have? And another interesting question. What's the depth of the relationship between the subject and the investigator? Is this a long-term study where there's repeat follow-ups where they're going to have multiple opportunities to talk about some of these issues or is it a one-time interaction? Is this somebody known to them or somebody who's a stranger to them? Or is it a secondary and tertiary user of the data or samples as they've been drawn out of repositories? And how does that bear on the ability and obligations to disclose? So I want to go back to a question that I raised earlier, the question of what, if I'm using coded specimens that were given to me by an extramural collaborator? Okay. And I have a different case study that I developed around this. I identified a source of clinical samples from patients with a rare condition who are being seen at University X. And the samples were collected under University X's IRB. They were collected with informed consent that that IRB approved. They're planning to code the samples and effectively they'll be de-identified to you. You won't have access to patient information or any other patient names or other identifiers. And you want to proceed with whole exome sequencing and so you set up your planning meeting with NISC and Bashali over at NISC asks you, do you have appropriate institutional approvals from an NIH IRB or from OHSRP? Good question. So if you are using coded de-identified samples, it's true that you might not need IRB approval from an NIH IRB, from your own institution. And this is because under current regulatory guidance, current interpretations of the common rule, which is a set of rules that governs the conduct of human subjects in this country, if the use of data that's considered to be de-identified is not considered to be human subjects' research under certain circumstances. And because of that, it's not subject to IRB review and other oversight requirements. However, you don't get to make this determination as an investigator. This determination has to be made by the NIH Office of Human Subjects Research Protections. Some of you might be familiar with an acronym OHSR. In the same office, they recently added the P for protections to their names. So when you see these acronyms, they're interchangeable and they mean the same thing. And you submit a request for review to them through a form that they have available on their website, which I've made available here. And this connects to a second part of the NIH Intramural Policy that I mentioned earlier. It states that investigators requesting OHSRP review of whole exome sequencing research activities will be asked to answer a set of supplemental questions to allow for enhanced OHSRP review. So the idea is they were concerned that it would be inappropriate to quickly approve under current regulatory guidance the use of whole exome sequencing technology in protocols here without looking closely at some of the ethical considerations that I've been talking about, because there's still a whole lot of information that theoretically by somebody could be connected back to an original research subject, and so they wanted to be able to look at this a little more closely. So here's a sampling of some of the questions that OHSRP is interested in. They want to know whether an IRB or an ethics committee somewhere has looked at this proposal and has approved the conduct of whole exome sequencing research, for example, with a set of samples. So if another institution has done this and is okay and has signed off on it, they will be reassured that somebody is paying attention to these questions. They also want to know whether it's possible that results from the proposed analyses will be returned back to subjects at that institution, and what this boils down to is they want to make sure that you and your collaborators have a clear plan in place for managing incidental findings. They also want to hear about any plans to deposit sequencing data into broad database sharing for data sharing purposes such as in DBGAP, the database for genotypes and phenotypes. They're also interested in knowing what kind of consent, if any, was ever given. And based on your answers to the questions on their checklist, they might ask you to go back to an institution and get some more information. They might approve it based on what you've submitted and say you don't need any further review here, or they might actually decide to send you on to an NIH IRB to have this looked at more closely than they're able to do. That provides a nice segue out of questions around incidental findings and into another topic that I'd like to cover, which is whether NIH data sharing policies apply to exome sequence data. The answer is possibly. And if I had more time to work on my slides, I think I would have revised this and said probably. So some of you might be familiar with the NIH GWAS, Genome Wide Association Study Data Sharing Policy. This is a data sharing requirement for all NIH-funded GWAS data that says there's an expectation that these data will be deposited in a repository like DBGAP or that investigators have to come up with some kind of alternative data sharing plan, either a different kind of a database or some other procedure for being able to collaborate with other investigators and share these publicly funded resources to help promote research. And we've heard, I think, in a lot of the talks earlier in the day, the value of this, how much of the information and comparison and reference data comes out of things that are available in databases like DBGAP. And so this is a very important principle. It promotes good ethical stewardship of resources and efforts to expand this policy to apply to broader genomic data types are underway. There's a committee looking at how this policy ought to be applied to broad data sharing. It hasn't yet been implemented. So there's not yet an across-the-board requirement that NIH-funded sequencing data has to be deposited into something like DBGAP. It is already happening and many programs have this as a requirement of their individual studies. Many investigators are starting to think through what a requirement like this would look like and how it would play out in their protocols. And I think this is a very good thing. I think we should be anticipating the probability that there will be some expanded data sharing requirements that will apply to whole exome and whole genome sequencing data. The way that this works at the IRB, when an investigator realizes that he or she is interested or required to deposit their GWAS data or exome and genome sequencing data into DBGAP, the IRB is asked to certify that the protocol and the consent form has taken several issues into account. They want the IRB to certify that the consent process covered the idea of broad data sharing and that research subjects are aware that this is happening and have given their approval. They also want the IRB and the institution to certify that appropriate confidentiality measures are being taken to de-identify the data that are being deposited, as well as that consideration has been given to minimizing risks in general to individuals, to families, and to communities as a whole. And I want to throw out there that there's another interesting and complicated question about how we think about the management of incidental findings that might be generated through secondary users of the data and whether there's an obligation to communicate back through the primary researcher these results so that they can sometimes, maybe in some very important cases, be disclosed to subjects. So I guess my advice or my encouragement is that, what I would say is that this process of certifying data for depositioning to DBGAP works best when investigators and IRBs have already anticipated that this is a possibility, when investigators have language already in their consent forms that addresses data sharing so that they don't have to re-contact subjects and get their re-consent around that issue and that the protocol has anticipated it that an appropriate level of engagement has taken place many steps beforehand to certain populations, certain communities will have strong opinions one way or another in favor of or against the idea of broad data sharing and figuring this out ahead of time is very useful and so I want to encourage investigators to start thinking about these questions in relation to whole exome sequencing to move on from the context of genome-wide association studies and it may be the case that there are some novel issues that come up. It might be harder, for example, to de-identify whole exome sequencing data. It might not, but I think it's very important that we're having this conversation at this point in time. Just a few comments about approaches to informed consent we're about to have a talk that I'm very much looking forward to from Julie Sapp, who is the expert, the in-the-trenches, has done this, knows how to do this expert on talking to subjects about these issues but I want to give a couple of framing comments about how IRBs at an earlier point in time have been thinking about informed consent. So there's a few elements of informed consent that I think can be highlighted as being specific to whole exome sequencing. So in general, just the broad scope of the analysis and the kind of results that are going to emerge and also the potential, the kinds of findings that might emerge, this is something that hasn't traditionally been covered in consent forms that may be relevant to an individual's decision making and might be something they would want to know as part of the consent process. More specifically, they are almost certainly interested in knowing whether results will be disclosed to them, whether they're related to the research question being studied or incidental to that question to the extent that that can be distinguished and whether or not they're going to be presented with choices and how to navigate those choices. This is a new element of informed consent that whole exome sequencing process to think about. And then the issue I was just talking about, plans for any data sharing of this sequence data and what the implications of that are to subjects. And there's some interesting questions about previously collected samples. So if somebody is enrolled here on a genetics protocol, they've been given some amount of information about genetic research and the attended risks and disclosure and so on. Is there a time and when is that time that they're going to need more information and give updated informed consent to anticipate in the expanded analysis? And a couple of things I'll say about that. I've been seeing a lot of reconcent going on. IRBs are asking investigators to get reconcent. And many investigators want to do this. They want an opportunity to talk to subjects again when their prior informed consent didn't anticipate these kind of issues. And I'll say that I think this is especially true when there are now more robust plans for the disclosure of incidental findings. But I would also say that this comes up even in protocols where there's a deliberate plan not to disclose findings. There's investigators and IRBs who still want subjects to know that the information is going to be there and maybe there are good reasons why they're not going to receive them. And then, again, choices about broad data sharing is another relatively new aspect of consent that many people are now anticipating is important in getting reconcent for. And what I've been seeing, I think that there's really three emerging models, three different kinds of protocols that are being approved here at the NIH. I can't speak as well to protocols that are outside of the intramural context. But there are some studies that have been approved where a very clear plan is articulated not to disclose incidental genomic findings to individual subjects. And the consent or a reconcent process actually covers this explicitly. Where I'm tending to see a cluster, the most studies that are being approved right now is something a little bit in the middle, a middle ground approach, where there's some limited commitment to providing incidental findings to subjects under very carefully specified, but relatively rare, certain stances. And then we're also familiar with approving proposals with very robust plans for the disclosure of findings. I apologize if some of that wasn't audible. The third kind of protocol that we're seeing, that we're comfortable with and that we've seen approved, features a very robust plan for the disclosure of findings and very detailed and often multiple time points associated with the informed consent process and a description of how subject preferences will be solicited over time. So as I was saying, what I'm seeing more of is a migration towards something in the middle, an emerging but limited obligation to, or a sense of obligation and plans surrounding this for the disclosure of incidental findings. And so I'm going to leave you with two main take home messages. If you've just tuned in now, the two most important things for you to know. In answer to the question, can I go ahead with whole exome sequencing based on protocols and consent forms that I already have in place? And the answer is no. The answer is not until you receive some level of institutional approval from either an IRB or OHSRP. But in addition to institutional approval, you should also be familiar with the many resources for consultation that you're able to draw on beyond this course, ethics consultation, the bioethics core, your colleagues who you've heard from who are all extremely generous about sharing what they know. It's useful to draw on the resources that we have around us as you put these plans together and submit them for approval. Okay, second take home message. What are the key things that an IRB or an entity like OHSRP is going to be looking for? And the answers are threefold, I think, that you and your collaborators have given thought to a plan for managing incidental findings that's coherent, internally consistent, takes into account things like your expertise that you have available to you, resources and subject expectations and contextual factors related to your study. And keep in mind that the IRB is open to hearing an account of what works in your particular circumstances and in working with you to figure out what the most ethically sound approach to your study is. A second thing that they're looking for is that you're starting to think about data sharing plans and what implications that has for your study. And third, we want to make sure that the informed consent process that either was in place or that needs to be in place reflects these various plans that you've developed in the context of a whole exome sequencing protocol. And the bottom line is this mostly has to do with expectation management. We want nobody to be surprised in any of this. We don't want an investigator who ends up with a result that they weren't expecting and is losing sleep over what on earth to do with it without having had the benefit of thinking this through ahead of time. We don't want subjects getting cold calls from researchers they've never heard of who are accessing their so-called de-identified data hearing about results that they weren't anticipating. We want everybody to go forward with as much advanced preparation and thought. And I guarantee you that there will be questions that come up along the way. That's why we have an ethics consultation service and we'll continue working with you on how to manage these questions. But we hope that we can help anticipate some of these problems and come up with ethically sound frameworks for doing whole exome sequencing research that starts to anticipate some of these challenges ahead of time. Thank you.