Republished by the LSU Medical Reseach Law Project

DR. SHAPIRO: All right, if we can, then let's go back to the issue we were discussing, the set of issues we were discussing yesterday when we began to deal with issues of identifiability. And we had a rather long and very useful discussion regarding issues of community, community harms, and how that would relate to IRBs and the obligations it would have and so on and the obligations on investigators. We will, on the basis of that discussion, begin to rework some of these recommendations, expand them, fill them in, get them a little denser, so to speak.

But I'd like to move for this moment right now to the top of page 5 in the accompanying memo which deals with what is called here "Guidance 3." As I understand this issue, it raises the concern about whether we think anything further needs to be said in this context regarding the issue of minimal psychosocial risk. Is that something that we want to say anything about, should say anything about? I think it's here because not a lot is said about it in the current context, in the current regulations and so on.

The question is, in my mind, one, are we talking about risk to individuals or, again, are we talking about risk to groups? And in either case, whether we're talking about one or the other, is there something we think that we want to say about this? Now, the way it's written here is we're just saying someone else should start thinking about this, right, is what substantively that statement says. But that's just a convenient placeholder here.

The real issue for us is to think about psychosocial risk, is there something in the context of this problem that we want to raise? And this could be a recommendation, it could be an exhortation, it could be anything. But the question is, should we take on the issue?

David?

DR. COX: I feel very strongly that we need to take on the issue. And I'll start off by making a statement and then seeing—well, I doubt people will be shy about it. There are two extremes by which one can look at minimal risk with respect to the genetic information in tissue samples. One extreme would be to say that all such research is minimal risk. And the other extreme would be to say that no such research dealing with genetics is minimal risk. I would argue that neither of those is a tenable position.

And so, what it means is that genetic research or research involving genetic information is situational in its character and to use some examples of those situations where we would say in this situation a particular type of risk is deemed minimal, but the same type of research in a different situation is not minimal. Use specific examples. So that's my suggestion of what we could do as a Commission.

DR. SHAPIRO: Bernie?

DR. LO: I agree with that very strongly. I want to sort of flip it around and ask David if he can give us two examples; one, an example of genetic research on existing samples that is commonly thought to be minimal risk that he thinks is not really minimal risk, and the other way around, examples of commonly designed studies where the risk truly is minimal.

DR. COX: This is tough, Bernie, because I was thinking about this this morning. I'll say at the outset that I don't have the appropriate cases nailed down yet. But I can give some general trends in that regard.

So, some research that would be viewed as not minimal risk because that family member involved in a research study comes to you, is very interested in being involved with the research, and, in fact, volunteers his or her family members for this, but it turns out that for those other family members it basically destroys their lives to be involved with this because for psychosocial reasons they don't want to know what the answer is. And so in that particular family context, this isn't minimal risk. It has to do with the family dynamics and the family structure.

Now, that's straightforward. This is faced in genetic counselling all the time. The question is, how the hell is an IRB going to deal with that kind of an issue? So this is why I have no problem defining that contextual situation of not minimal risk, but is it relevant to IRBs. So I need a better example.

On the flip side, something that would be viewed that may be not minimal risk that could be minimal risk is this information of coded materials. Coded means identified. So does that mean that all research done with coded samples has the same risk? If, in fact, there is no intention and it's very difficult for people to go back and get information through the code, then that's a very different type of risk than if the code leaks like a sieve. So those are two situational examples. But I think we need much better ones than that that can be very precise. And I think they need to be directed specifically to how IRBs would adjudicate or consider the issue.

DR. SHAPIRO: Alta, then Arturo.

MS. CHARO: I find myself thinking about what Allen Buchanan said yesterday. And among many things, he reiterated a very commonly understood notion; that is, that risk is not simple harm, it's harm times the likelihood that the harm is going to come to pass. And I think that makes it possible—I'm watching your face, Diane, only because it's—you know, if you have a terrible thing that's only got a one in a billion chance of happening, it's not considered to be as risky as if it has got a one in two chance of happening, right?

The reason I think of this as being tremendously important is because this is where I think that the discussions about coding can be brought back in in a much more robust fashion now that we've lost some of their value in the discussion about identifiability. Because there are three kinds of harms that we worry about here primarily. There's kind of a "peeping Tom" invasion of privacy harm, there is a backflow of intermediate ambiguous information harm, and there's a kind of breach of confidentiality stigmatization/discrimination type harm.

And this is where if samples that are being used by researchers are coded, we have substantially reduced the likelihood that there is an actual invasion of privacy in the form of "peeping Tom" because the researchers at least do not have any way at the time they're doing the work of linking what they're seeing to any particular person.

And this is where David's long-ago-made suggestion about some kind of filter or filtering group that can regulate the backflow of information should some findings that might have clinical significance be developed becomes very important in trying to regulate the flow of information back to tissue sources should there be some inclination to do so. A researcher sees something that looks like it might be of clinical significance but it's ambiguous enough that it might just be aggravating, and rather than making the decision on his or her own to not go or to go and to wrestle with that alone, there's a filtering group, whether it's the IRB or another, that helps to decide if this is finally the appropriate moment to send information back down.

So that in the discussions about minimal risk, I think we'll have a great deal of difficulty in coming up with some nice hierarchy of which harms are significant and which harms are not. But what might be easier to do is to come up with a set of procedural devices that can reduce the likelihood of any of these harms coming to pass. So that in many, many cases it will be possible to conclude that overall a protocol is, in fact, a minimal risk because the likelihood of any harm, substantial or otherwise, has been minimized so substantially.

DR. SHAPIRO: Alex?

MR. CAPRON: I think that's a very sensible way of proceeding. I wonder whether the reluctance to address the magnitude side is because we don't think that there really is some kind of a metric; if not very precise, at least sort of these kinds of things are more risky than others. And I can understand the conclusion, well, it's very hard, it's so much a matter of opinion. But then I come to the problem of the way in which this is simply then put off onto the IRB.

And my current thinking about IRBs is they are a pretty weak read for doing a lot of that work, not just because they're overworked and so forth, but because it's an extremely difficult thing to have an issue like that thrown at you if you're an IRB that hasn't spent a lot of time. Now, maybe the IRB at Stanford, if David's doing a lot of research of this sort, sees these kinds of proposals all the time and becomes quite expert, the same way we know that some become expert on research with the mentally impaired. But for a lot of them it is going to be out of the blue.

And if we can't at least develop some discussion—I'm not talking about regulatory language where you would say if it's this it's automatically hot—but some illustration of well, if it's research on a disease that is treatable and where it's familiar to people and so forth, it's this kind of thing; if it's research on something which is lethal and untreatable and where the fact that you have this could be devastating psychologically and in terms of ability to get insurance and so forth, that's different; if it's research on sexual orientation or alcoholism or something that's very socially fraught.

These are the kinds of things where I think we need to ask somebody to talk to geneticists about examples and have real-world examples and have some discussion of it so that we provide some help to the people at DHHS, if this is maybe going to be illustrations in the official IRB guidebook or something, assuming that IRBs may not all read this report. But somewhere along the way, I'm with Trish, we want to be helpful with these documents. And if we don't think that there's any way of saying much about it, what are we thinking the IRB is going to do about it?

MS. CHARO: Just by way of clarification, I was never suggesting we don't try to look at the other side. But I just want to emphasize that the Mayo Clinic approach, described last time, in which every form of genetic research is non-minimal risk is one that has the advantage of simplicity and it certainly gets high levels of compliance because everybody knows that they have to go to the IRB and they can't waive consent and it's done. But I don't think it accurately reflects the real level of risk out there that could be achieved.

MR. CAPRON: Well, I wasn't thinking I was disagreeing with you, Alta. I was in part responding to the combination of your comments following David's, the argument that the probability issue is something which we can highlight and discuss and show how lower probabilities reduce the sense of "risk," and I agree with that.

But I didn't want to leave David's sense, well, he couldn't think of anything off-hand and his answers were all on the probability side, too, to say that we can't provide some guidance. Once the IRB at Mayo or any other place gets these things and they are weighing potential harm against potential benefit, one of the issues is, well, what is the harm? And just giving people some sense so they'll be familiar with it. That's all I was suggesting.

DR. SHAPIRO: Eric?

DR. MESLIN: This issue was raised at the last meeting in Cleveland. And in the revised staff draft, we took some time to identify cases in the material. It appears beginning on page 230 and goes on for another three or four pages. These were cases that were constructed, with David's input and others, that attempt to identify those sorts of examples. One possibility for Commissioners to consider is to review those cases and give us staff some feedback as to whether they would appear to be the kinds of cases that form the boundaries that you're describing, David. They can certainly be supplemented by others.

DR. COX: They do sort of. But I'm actually very much in favor of combining, and I think this is what Alex is saying. The problem with those cases right now is that they're just sort of there floating and it doesn't have the principles of which the cases illustrate. So I hear Alta saying let's have some principles and then here's the application of them to real cases.

The fact that we may not have those cases perfect right now doesn't mean we can't get them. I'm quite convinced we can. But what I haven't heard us do is really articulate the principles yet. Alta started. We don't need a lot of those. We say this is a hard problem, here are some guiding principles and then here are some cases that they're applied to. Because if you just have the cases, it will be like Al Jonsen spoke yesterday, this wooden interpretation, it's going and trying to fit everything into these cases. We need the principles.

DR. MESLIN: That was one of the reasons why the revised chapter 3 was placed where it was. Pedagogically, we might want to move it and provide that backdrop for the case analyses.

DR. COX: Yes. I actually think that these principles are woven through a lot of our discussion and a lot of the text now. But what we need to do is extract them. And so we have this thing "Principles for Idiots," and we say here it is, three lines, here they are, these are the principles. That way, for busy people at IRBs or doctors or researchers, they can carry around a little card in their wallet to remember the principles.

MR. CAPRON: Maybe the idea without having the title printed on it.

[Laughter.]

DR. SHAPIRO: "Dummies" is the more common title these days.

Let me ask a question. What's wrong with the idea, which I take it doesn't have any support here, of saying that there is no minimal risk, everything is more than minimal risk? Now, I understand that makes a lot more review cases. I understand that technical issue. But is that a serious objection? It does have the benefit of being very simple; it's probably therefore wrong and not very helpful. But I'd like to hear some comment about that.

DR. BRITO: I think what Alex was saying, that it puts the onus then on the IRB to make that decision because then everything has to be reviewed by the IRB.

DR. SHAPIRO: Alta?

MS. CHARO: Actually, everything would have to be reviewed by the IRB anyway. The key here is whether or not you could waive consent, because to waive consent you have to show both that it's minimal risk and that you couldn't practically get the consent—and we'll deal with practical later, which I think is even more of a morass than this.

MR. CAPRON: And expedite review.

MS. CHARO: Yes, although that's really in some ways maybe it was a poor choice to highlight that as much as we did because it's really just a matter of whether it takes a whole committee or one committee member to get the thing through.

The simplicity is actually a huge advantage because it takes away so much uncertainty and it simplifies life tremendously for the IRBs. All they've got to say is fine, get consent, come back when you have. But the reality I think is that many collections are going to have samples taken from people who have long ago moved along, and the United States is a tremendously transient population and some people will have died, others are just unlocatable. And it not only reduces the universe of samples that you can use in an absolute sense, it changes the statistical validity of the samples that are left if one suspects that there are some in fact selection factors in who has died or who has moved on which might make the remaining samples that are useful in the sense that you could get consent still useless from the point of view of scientific validity.

So an insistence on a consent requirement without exception has the potential for being a tremendous obstacle in the research endeavor and it's not an obstacle to be placed there lightly.

DR. SHAPIRO: Just to push it one tiny step further and before dropping it. I think those points are extremely well-taken, especially if there's no exception to the consent. But we have examples of other cases where there is exception to the consent, they're not practical or whatever issues like that.

MS. CHARO: But to get the exception to consent you have to show that it's minimal risk. So if you say nothing is minimal risk, then no exception can be made.

DR. SHAPIRO: That's the current—

MS. CHARO: Except for totally unidentifiable anonymous samples, absent minimal risk, there's no exception to the consent requirement.

DR. SHAPIRO: Right. But that's the current regulation.

MS. CHARO: That's true.

DR. SHAPIRO: One could use that—well, I'm not sure, I haven't thought this through to be honest, I'm just sort of raising the question. I will think about it more carefully.

DR. COX: Harold, this is my concern, and I can certainly see how people may minimize what I'm going to say. I don't think it's on really strong grounds but I feel it strongly. And that is that by doing that, by basically having nothing "minimal risk," what you basically do is take any responsibility off the researcher because then the researcher says, oh, all of this stuff somebody else has to figure out what's risky and what's not.

That's the situation that we're in right now. And that anything that we do that makes it so the researchers become less invested in this issue of figuring out what is risky and what isn't, I'm against.

Now, I realize that from a regulatory point of view this is a very appealing fix. But I think part of the problem with these complicated problems is that unless people get engaged in them, what we are going to have are regulatory fixes and we will never solve the problem. So that's my concern, but maybe it's too ethereal.

MR. CAPRON: I would like the Chair to follow up on what I found his very helpful intervention in this discussion just now. When are we going to see language that would suggest that on the flow chart if the answer is initially yes, this may involve more than minimal risk, you have a chance to revisit it on the same kind of practicality grounds, because the practicality grounds are not unconnected with the risk.

I have Dr. Hook's slides from the last time, and looking at the risks, if you're thinking about a situation of the type that Alta describes where the problem is you're dealing with samples of people who have died of the disease that you're eventually studying and so it's impractical, it would certainly make sense to say that the risk to them of insurance and employment discrimination, of discovery of unwanted or uncertain information of psychosocial or emotional harm and their right not to know aren't risks to them because they're deceased.

And so some kind of a process that says, yes, in principle, we're looking at this, we see that there are risks, they are more than minimal risks, but as to this group that you cannot get consent from but whose tissues you want to use, those risks fall out and it moves to a lower category. So it's a somewhat more refined examination. As to other people in the sample group, yes, you're going to have to try to trace them down and get their consent. In other words, a process that is a little different where we're going to have to be recommending that there be potentially an alteration in the process and the IRB ask different questions.

Are we anticipating that out of a discussion like this, if that idea that the Chairman threw out, which is, well, that's only the current regulations, there could be other ones, are we going to have that in front of us as language?

DR. SHAPIRO: We could certainly put it in front of us if it seems like a helpful idea.

MR. CAPRON: It seemed helpful to me.

DR. SHAPIRO: Alta?

MS. CHARO: I want to make sure I understand the suggestion. And just by way of clarification, if the people whose tissues are being used have died, it doesn't matter what the level of risk is because they're no longer considered human subjects. The problem is that you don't usually know that they've died. You are faced with a collection of samples, you are told you have to get in touch with people, you send out a series of letters, a certain percentage comes back as undeliverable and you don't know why it's undeliverable, so you don't know which of those people are no longer human subjects and you can use their samples freely and which of them have simply moved away. But more than six months ago and their mail is not being forwarded, having suffered that problem myself, a six and a half month sabbatical is very inconvenient.

But I want to make sure I understand more generally the example. The idea is we might be discussing a possibility in which researchers are encouraged to describe subpopulations within the samples—those for whom this is clearly minimal risk, and those for whom it is not yet clearly minimal risk. There's probably no reason why that can't be done at all under—

MR. CAPRON: It may even be a matter of making clear the way it could be done under current regulations. I just wanted us, to the extent that what I took away from the Mayo situation was the advantage of presumptively saying if the probability and the magnitude are such that this is "more than minimal risk," and let's start off assuming that most genetic research is, we get the IRB and the investigator looking at it, not just the investigator and not just to check off minimal risk, I don't have to do anything.

MS. CHARO: And, of course, in an ideal IRB, not the ones that you're worried about, it doesn't happen that way. But there already is a presumption of non-minimal risk in the sense that an investigator essentially has to petition for a minimal risk declaration by an IRB and has to prove that the protocol is eligible for it.

And if what's really being suggested is that that be somehow highlighted, emphasized, and that in the area of genetic research—and I have a feeling now Steve Holtzman and Carol Greider are going to flip when they read the transcript—in the area of genetic research there's such a public concern about it that one wants to strengthen that presumption and force an even stronger level of proof that it's really a benign kind of genetic research, that might be a very helpful way of flushing out this term.

MR. CAPRON: Right. I guess I would go back to the point that I mentioned a moment ago that you kind of dismissed, which is the difference not on the consent issue but on the expedited review. I think there is a huge difference between expedited review, which involves either the chair or the administrator checking off, yes, expedited review granted—I'm always glad when I get expedited review, because it's easier—but when you're dealing with these things, if our discussion, and this is not a regulatory change, but the discussion was the presumption should be at that point this should go before the whole IRB and get the normal process, and then coming out of that the recognition. But the conclusion can be as to certain parts of the population, or maybe the entire population you're looking at depending on who they are, this drops back down and you are allowed to have a process that does not involve individualized consent because it turned out that the risk is more minimal. That's I guess where I would put our discussion. And, again, it may not be regulatory change, it may be change in the way the regulations are understood.

DR. COX: I like this discussion because I see it being a process discussion in that it doesn't lay the burden on any one individual or group of individuals to deal with these frontier issues. So the concept of the researcher having to petition because he or she believes something is minimal risk, goes to the IRB with an actual brief on it, the person has thought about it, the researcher has thought about it, then the IRB, even the chair of the IRB has that brief as the working point to say they either agree or disagree or need to get more information. So it's actually a substantive process that increases information about the frontier.

Now, you can tell this is like a scientist's way of thinking about things. There are other ways to do things in life. But as it gets more and more information, I think we will actually get smarter and smarter about what the right answers are. But if we don't get more and more information, if we just sort of lay it off on the one person or the other and then that person doesn't have any information to make the judgment either, I don't think we get any smarter. So the process I think could be a very good one and we lay that out.

DR. SHAPIRO: I think the notion that came up yesterday in another forum that really a lot of these procedures can be strengthened by having a higher quality process in which a lot of people are involved is interesting. It came up yesterday, at least I interpreted this coming up in that, in Alta's suggestion, or not suggestion but pointing out that there's a lot of initiative left to the investigator early on. That investigator has a lot of responsibility at the beginning in making various choices. It's something I think we should emphasize that's a little bit like the responsibility for the design of the whole thing, which is also a major responsibility of the investigator. And I think if somehow we can formulate the language as we go through these kinds of things to emphasize those aspects, it would be very healthy and will clarify a number of things that we try to work on language.

Alta?

MS. CHARO: I must say at that point just for the sake, in case he's reading the transcript, that Eric Cassell will be very happy to hear that, because I think that's his constant point about the research community's own education.

I would like to take up the other half of the equation that Alex was emphasizing earlier, if it's appropriate now, which is the nature of the harms and how one might view them. The usual way in which minimal risk is described in IRB review is by comparison to the risks of everyday life. And in the context of physically invasive research, that has been complicated enough, but people have basically done it by having a sense of the likelihood that I'll have a puncture wound today versus getting an IV.

In theory, it should not be any more difficult to make the same kind of assessment. It will be just as muddy, no more muddy, for psychosocial, and yet I find myself finding it much more difficult. I think maybe it's because the subjective experience is even more varied in the area of psychosocial than it is in the area of physical invasiveness, even given the range of reactions to pain and such.

I'm not yet sure that this particular way of thinking about minimal risk is, in fact, useful in this context but am open to persuasion that it can be used as the kind of principle of measure. I was wondering if anybody here had thought this through better than I had?

DR. SHAPIRO: Alta, I'm not sure I understood the last part of what you said. Could I ask you—

MS. CHARO: Sure. I haven't been able to think of a way that I could use the experience of everyday life as my benchmark against which to think about the psychosocial harms associated with genetic research that may be on the association of dark hair with curls or could be on the association of some particular marker with a high probability of breast cancer. I was wondering if anybody here had been able to figure out a way to use this kind of everyday life experience as a benchmark?

DR. SHAPIRO: Alex?

MR. CAPRON: I agree with Alta, it's complicated. And I want to add an additional complication. I don't think that when we think of the physical risks of everyday life we think that most people are going to be subject to malpractice and get injured by their doctors in ways. But here, certainly one of the things that is most often cited is the breach of confidentiality.

I think we might want to say something quite directly that it cannot count towards one of the risks of everyday life that your physician or nurse or hospital is going to be very loose-lipped or very casual about your medical records and breach your confidentiality as a way of saying this happens in research, it happens all the time anyway. Because that enterprise is too closely connected to the research enterprise and you don't want to have people degrading the standards.

And this is not hypothetical. Certainly, in California one of the reasons a special statute was passed very explicitly about consent and about confidentiality of the HIV test results when the first antibody tests were developed was because everybody knew that ordinary medical records were simply not being adequately treated on the level of confidentiality. And this was sensitive enough that they established requirements, which ought to apply to everything but it was sort of like that's a lost cause, that on this issue we're serious about it, you must tell people you're conducting a test, you can't just take their blood and then do a test on it, and you must keep the results confidential and only reveal them when they explicitly say yes, it's all right to tell so and so.

So I think we need to talk about that because certainly on this psychosocial level that's the thing that gets mentioned most often, that information. And you mentioned it under two headings; the kind of voyeurism or peeping Tom, just the nurse that lives down the block from you and she tells somebody so and so was in and got a test the other day, or the actual breach of confidentiality because the report is published in a way where your pedigree is obvious, that's obviously your family and now you're shown all these people with X's through them to indicate they have breast cancer or whatever and you don't want that out there. I think we need to address that.

I agree with you though that in many ways the other kinds of things are very hard. What is the risk that you'll be stigmatized? It doesn't seem as though that prejudicial reaction that people have in society ought to be carte blanche to say, well, so here's another stigma to get stuck on people. People have to deal with stigma, that's just life. So I agree with you, I think it's actually quite difficult.

DR. SHAPIRO: David?

DR. COX: So this sort of falls under the category we were talking about yesterday where should the IRB make judgments about something if it really can cause tremendous social mischief but is a high quality scientific question?

MR. CAPRON: I don't think this is the same thing, David.

DR. COX: But it's an analogy. It's a different issue. And the analogy as I see it is the following. It is that these are highly intertwined issues that can't have one institution or one body dealing with them all. And so the idea that you have subtle social prejudices and harms coming out of stigmatization, one of the ways of sort of dealing with that is with the privacy laws. And I'm saying I totally believe we should have those. But anybody that believes that we can solve the problem of discrimination and harms by writing laws, it's wrong. It doesn't mean that we shouldn't have the laws, but there have to be many layers and levels by which this gets dealt with.

Some of the them are going to be the IRB, some are other social institutions. So that you don't throw up your hands and say it happens so we can't write any laws. You don't write the laws and say that that's going to solve the problem. You acknowledge it's complex, we can deal with the practical things of privacy, but at the same time we point out that that's not enough and other social institutions have to pay attention to discrimination, whether it's on a person's skin color or whether it's on genetic information.

So I think it's by trying to make these things simple that we get into trouble, because they're not simple. They are going to be multifaceted checks and balances and we're not going to solve them all.

DR. SHAPIRO: Well, with respect to the question I thought you asked Alta, that is, whether one could use something like everyday, that it would seem somehow acceptable in another arena, my own judgment, and it's no more than a judgment, is that you can't. Not only for the reasons that Alex mentioned. But because even putting those issues aside, which are very important, it's really that you gave yourself the key explanation; namely, the variance is at least thought to be so much higher between individuals in that area that there's really no common way to think about it. Now, we may be fooling ourselves in the other area, maybe because of different reactions to pain and so on, as you said in your comments. So I don't think we can use that framework.

MS. CHARO: Well, if that's the case, and certainly where I was leaning, although, like I said, I was open to being persuaded it could be used. The alternative is to try to develop some kind of list, recognizing that it's highly imperfect, of red flags. That research that involves these kinds of things should be understood to probably be research that raises the possibility of a significant harm rather than an insignificant harm, and at this point, unless you can show that you've reduced the likelihood quite substantially, you really are not going to be allowed to call it minimal risk.

The list would, like I said, be highly imperfect but not at all non-obvious. It would be research that involves conditions that are life-threatening, disabling, or commonly understood to be socially embarrassing. It would include research that involves results that could generate a clinically ambiguous kind of information that puts people in an impossible position of having to figure out how to pursue their own care in light of truly intermediate insufficient findings. Conditions that regardless of the reality of them generate reactions of discrimination in the employment or insurance context or social stigmatization. And that this be a kind of set of guidelines. And that instead of trying to look at the overall risk by looking at the harm and the likelihood and comparing it to everyday life, in this particular area it may be necessary to break it down into its components a little bit more in a more reductionist fashion to come to some conclusion about eligibility for minimal risk.

DR. SHAPIRO: Yes. I think that sounds very helpful. I think that is the way we're going to have to go.

I think we've had a number of very interesting and helpful comments which will enable us to write some interesting proposals here for dealing with this particular aspect of it.

Why don't we go on then to the next issue, which is an issue I must say has me a little bit puzzled, but I hope I'll get some clarification here. That is the question of living relatives, what status they have. If I understand the situation, it is that the current regulations seems to indicate, at least some people think they indicate, that in fact we might ned their consent because we're going to find out some things about them even if we're using a sample from a deceased relative. The question is, how do you deal with this issue of living relatives? Are they subjects or aren't they subjects in the research?

Now, I, very frankly, haven't thought of a good way to deal with this issue. It's just here to try to see if Commissioners feel that it is an important issue that we need to deal with, and, if so, how we should go about it. I think, and maybe Alta you could help me with this, I think currently people just treat them as not subjects. That is what's happening out there. And treating them as subjects seems mind-boggling to me. But yet I think I've learned that, in fact, the way the regulations are written that at least there's an argument to be made that they are, under existing regulations, subjects. The question is what do we want to say about this.

MS. CHARO: Your understanding of the regulations is right. The dead person is clearly not a human subject.

DR. SHAPIRO: Correct.

MS. CHARO: And the regulations are written in a generic way so that information about any identifiable living individual constitutes information about a human subject. But rarely does anybody in the research community take seriously the notion that these kin are, in fact, being transformed into human subjects. It's counter-intuitive to them because these kin are not the ones who are the primary subjects of study. They are incidental. And I don't have any idea how we could limit this universe, since the study of any dead person is information about many, many, many other people in a quite indirect fashion.

MR. CAPRON: Maybe the probability arguments we were having a moment ago bear on here. It's certainly a lot more information about their first degree relatives than it is about a distant relative. So the probability of being harmed by that information, that you'll actually be able to say anything about that distant relative, is perhaps smaller. We may be able to employ some of the discussion we were just having. That's just a thought.

The notion that if my parents were deceased, their tissues were going to be looked at, both of them, to find genetic markers that would have direct implications for me at a 50 percent level or a 25 percent level, depending on what you're talking about, seems something in which I would have a rather great interest. But the notion that Great Aunt Matilda, the implication is much, much more remote.

And, of course, if my parents are willing to undergo the study, the fact that it is going to have implications for me doesn't allow me to stop the research. So we're dealing with a situation where we don't have that balance of consent. Just to add to the complication. I'm as much as subject, as it were, in the case where my parents consent as in the case where they can't consent because they're deceased.

DR. SHAPIRO: Bette?

MS. KRAMER: I was wondering, would this be embraced within the consideration of what's practicable?

MS. CHARO: It might come up there as well, but this actually cuts to an earlier decision point. The earlier decision point is are you doing research on a human subject. That's the second decision point. The first is are you doing research, then it's are you doing research on a human subject. If the answer is yes, then you head down to your nearest IRB and you start filing a lot of paperwork.

If we want to understand research on cadavers or samples from cadavers as research on a human subject because of its information yield on first degree relatives, we have in some ways a special case of the group effects from discussion yesterday. It's just one special case about the way in which research on one person can actually be of interest and significance to a larger number of people.

We also have the discussion we were having yesterday about the role of those secondary people in vetoing or giving consent. In your example, Alex, of research on the tissue of one or both of your parents, it would not only be you, it would be all of your siblings, and if you're saying first degree relatives, it's all their siblings, and if they're still living, their parents who are interested in this.

MR. CAPRON: Exactly. Right.

MS. CHARO: And one possible conclusion is that since all of them are now being made into subjects of research by virtue of the research on just your parents' tissue, each of them individually in all three of those generations has a right to veto this research, because any one person denying consent shuts the research down. Which is a pretty significant finding. It may still be justifiable but it is the natural conclusion.

DR. SHAPIRO: I have to say, we have to think about this and talk about it more, it doesn't feel right to me. I understand the issue of probabilities in establishing risks, that risk itself is a product of two elements, one of which is probability of harm. But as I start thinking of treating first degree relatives as subjects, it sounds to me like an unmanageable problem. That's how it feels.

DR. COX: It's even more complicated because anybody that is involved with genetic counselling knows that for some people you tell them they have a 50 percent risk and they say, "Whew, I thought it was 100 percent." And then other people you tell them they have like a 1 in 1,000 risk and they break down in tears.

MS. CHARO: You been talking to my mother again?

[Laughter.]

DR. COX: No. So as a scientist, Alex, I agree with you that this makes sense. But from the point of view of how people perceive risks quantitatively in terms of whether they would want to keep something from happening, it becomes a very complicated issue.

MR. CAPRON: I was actually raising it from the other point of view, thinking of the more distant relative being someone for whom the claim is tenuous. In other words, because the probability falls off. I'm saying that if we start putting probability in here, we ought to—

DR. COX: No, I understand. But that's these people that are the 1 in 1,000 and they say, well, that's still a gigantic risk to me and I don't want to deal with it. So I don't see how we can—these discussions don't pass the red face test. But I'm trying to see if there is some way that we could incorporate this. And I don't see right now any mechanism. I'm struggling with a mechanism. Because if we could address it, it would be good. But I don't see how.

DR. SHAPIRO: Alta?

MS. CHARO: All right. Now to say this as somebody who is not yet a complete fan of the mechanism, but we chose it yesterday and I'm looking forward to seeing it. In the context of group harms yesterday, a mechanism that was chosen is as follows. Even when something is not required to go to the IRB because, in the example that Alex gave, you're using totally unidentifiable samples but it has the possibility of implications for an entire identifiable community, we were going to write something that urged PIs to go to their IRB anyway. Now, I've always been skeptical because it's hard enough to get them to go when they have to go, let alone when they don't have to go.

But given that we've chosen that mechanism, that sounds like one that would work here, if we could make it work at all. Which is, you're dealing with cadaver tissue, you're not clearly falling under the regulations in terms of the first degree or any degree kin as far as them being turned into subjects, but would you please go to the IRB anyway so that they can work with you on trying to minimize any kinds of harms that might flow to this kin group.

Indeed, there is a lot of stuff in the IRB guidebook already about the special problems of pedigree studies, the sort of obligations to go and ask for consent even where it's not required that you get consent just in order to make sure that this is done in the best possible manner. It goes back to Bernie's request for kind of a best standards kind of approach. And it may be that that's going to be the compromise. You can't make it a requirement but you can make it an aspirational standard.

DR. COX: I hadn't appreciated that. So if you think of it in the context of community, and it is a type of community, it certainly falls under that framework.

DR. SHAPIRO: So what you're saying, if I understand, or what you're suggesting here is they not be considered subjects because that would automatically require consent and so on in the case we're talking about, but that we develop some kind of another red flag or another issue that we want to point out for people to be conscious about and try to design their work to minimize any harms that might flow to this group even though we don't give them the status of subjects from the point of view of the regulations. I'd almost be satisfied with anything that didn't make them subjects. And that sounds like something to think about anyway.

MS. CHARO: At the same time, I'd also like to add that, and he's not in the audience today, I'd like to get the Director of OPRR to actually engage in a conversation, verbal or written, with the Commission on this. Because one of the problems here has been the perceived absence of clarity in the interpretation of the definition of human subject. And if there is a good case to be made that these relatives should be understood to be human subjects and that there's a workable way to do that, I would love to know how that office has been in fact implementing that particular provision and take that into account as we move towards conclusions.

Right now, I feel like we're operating slightly in a vacuum. So if we can add that.

DR. SHAPIRO: We can certainly get that information. My preference right now, and I have to say I haven't thought enough about this, but my preference right now is that we should say something saying that they are not human subjects. And if it requires clarification of the regulations, then it requires clarification of the regulations. Then go on to say what we hope IRBs, investigators, and others might think about this because this is something which is a nontrivial issue.

MS. CHARO: That's right. In the end, we are not the authoritative interpreters of these regulations.

DR. SHAPIRO: Right. Correct.

MS. CHARO: There are many interpreters and the one that tends to be most authoritative is the agency that administers the regulations or that wrote the regulations.

MR. CAPRON: Well, on this subject, what our recommendation is is asking them to address it.

MS. CHARO: Right.

MR. CAPRON: And so we have to tell them in the process what we ideally would like to see them be doing.

MS. CHARO: Right. That's right. And I would love before we come to a firm conclusion to at least hear the other side, if there is another side to be made.

DR. SHAPIRO: We'll certainly do that.

David, then Arturo.

DR. COX: Why I'm so keen on this is it again comes back to taking the onus off the IRB and putting the onus on the investigator. So I think of this as an investigator and I say, all right, so I'm working with autopsy material and I have to think about is if this going to have any impact on the community of relatives. Why am I going to think about it? Because I want to make the argument that I'm doing my research in a way that it won't have that impact. So I think about it because my motivation is to get that through so that I can do my research but that I'm not going to be able to do it unless I think it through clearly so that it just whizzes through the IRB.

Now what that means then is that there will be a clear brief for the IRB to either agree with it or not agree with it. So it's a process that puts the onus where it should be, which is on the researcher, but then it has the oversight where it should be, which is on the IRB.

DR. SHAPIRO: Okay. Any other comments?

Yes, Arturo, I'm sorry.

DR. BRITO: At the risk of going backwards a little bit, on Guidance 2 we talked about the implication of anyone other than the individual, the index subject. How does that differ from when the human research subject is deceased, a cadaver or what have you? I guess what we're worried about here is the implication of what this may have on a living relative.

DR. SHAPIRO: Right.

DR. BRITO: How do Guidance 2 and Guidance 4 truly differ? And is there a way to really just combine what we're talking about in Guidance 4 within Guidance 2, assuming we're not going to consider living relatives human research subjects?

DR. SHAPIRO: I think there may be, as we work through this, opportunities to combine some of these things. A number of the items here need to be combined and recombined to make coherent sense. And I was very conscious that that's a suggestion you made yesterday, I think. We certainly want to think about that carefully and see how when we actually articulate this. I think there are some differences but some commonalities, so they're not exactly the same. But I take your suggestion and we'll certainly think about that.

MR. CAPRON: I hope that as part of this process—I think Alta has underlined a number of times the initial difficulty here that we're relying on investigators, against their own self-interest, coming forward and saying I'm willing to take the time before I do this research to sit down and discuss it with you, as David says, prepare a brief, and explain why I think I've done the right things. I could simply look at the existing regulations and say I don't have to do any of this.

One way that Arturo's point here I think ties in about combining them is, and Alta didn't say this, but one way of understanding this is that without saying they're subjects, we say they are enough like subjects that you need the IRB review to make sure that you've done all the things that David describes. And so, in a way, that's much clearer, Arturo, when you're thinking about family members than it is when you're thinking about this more amorphous group. But the argument there by analogy is that in a way they're sort of subjects, too, because you're really planning to reach conclusions, generalizations about this group, whether it's a Portuguese family living in San Francisco with this interesting pattern where you're looking at the deceased relatives but making these implications about all the first degree relatives about them, or it's X, Y, Z, other community of this or that group.

So what we're saying here though has implications, in effect, for sort of a regulatory change. Which is, that they're a special category in the genetics area of saying that, because of this analogy, the subjects, even if you answer the question no, there are no subjects in the traditional sense because this is dead people's tissue, but will implications or conclusions be drawn with direct relevance to identifiable people, the answer is yes, then you have to go through a process which doesn't turn them into subjects but has the same effect of getting some IRB review. And that's awkward but I think it's worth thinking about because at least it says it's not just at the discretion of the investigator to be big-hearted about it and say, sure, I'll take the time and do this.

DR. SHAPIRO: David?

DR. COX: Yes. If we don't make it clear that investigators have to spend their time dealing with these issues, then some of them will but a lot of them won't, that's the situation that we have right now, even though the law says they have to deal with it right now. I actually am concerned that unless people have to come forward, and this is the point you were making, Harold, about having everything less than minimal risk, because what it does is it says that people have to address the issue. My argument against that was that you want people to be engaged. But I don't think that those have to be dichotomous, that's where I'm coming to now, because I don't like the idea of sticks by themselves, you have to really have people wanting to be engaged, but I'm becoming worried based on these comments that if you don't have any stick at all then people aren't going to even play. So there has to be some happy medium here.

DR. SHAPIRO: Right. All right. That's been a very helpful conversation the last fifteen minutes or so. I think we will be able, again, to articulate something quite interesting out of this for us to consider as we get through the next draft of this.

Let's just proceed along here. The way this memo is drawn up we're now at sort of Roman II, so to speak, and there are a bunch of things called recommendations that follow that. I think some of these are repetitive, things we already discussed and/or are made obsolete by things we've already discussed.

Recommendation one, which deals with encryption schemes, I think is, by and large, moot at this point. Something is identifiable or not, and that's what we decided on yesterday. Personally, I don't think those encryption things are possible, but I'm not the expert here. But I just think it's moot; we don't need to spend any time discussing that today. We'll just put it aside for the moment. I could add a whole long list of these.

[Laughter.]

DR. SHAPIRO: But then recommendation two here really is what we've just been discussing. It is not a separate issue, at least as I understand it, Eric.

MS. CHARO: The answer to that one is no?

DR. SHAPIRO: Yes, subject to all the things that we've been talking about here and the processes around which we want to surround this to try to achieve the aims that we've just talked about in the last fifteen or twenty minutes. But that's just not a separate issue. It's something we've already dealt with.

Now what about recommendation three? Do you want to say anything about recommendation three here?

DR. MESLIN: Three and four, in fact, three, four, and five are variations on a theme and they were put there to try and distinguish different nuances of the same issue. All I think we want to mention in three is a concern about publication, which raises a particular type of risk. And in recommendation four, we are speaking about inadvertent identification and whether or not that should be drawn to the IRB's attention.

My view is that three and four can probably be folded together if the group feels that a guidance regarding publication of pedigrees would be of use. This is a moving target. There has been research very recently published on this by Jeff Botkin and others. So it may very well be that we can turn that into a guidance of sorts rather than a recommendation.

Please don't take the words recommendation versus guidance too literally in this memo. They are quite fungible.

MS. CHARO: The discussion yesterday, obviously, is richer in detail than the language that's here. So need we at this point in the transcript try to reiterate what was said yesterday, or simply say refer back?

DR. SHAPIRO: I think refer back if fine.

DR. MESLIN: I suggest that we fold it together. I think that will make it much easier.

DR. SHAPIRO: I think whether these come out as recommendations or guidances, whatever they come out as, these are not easy recommendations to write I think. I think there's a lot of problems in writing these and saying something we don't intend to say. It's a sensitive area.

DR. MESLIN: The only thing I would add is we discussed yesterday the possibility of writing a one-sentence finding or general conclusion that would precede the guidance statement. I know staff would be grateful if Commissioners might think about what that finding, that one sentence statement would be that would allow us to then say so it is our guidance that certain types of information should be provided.

I would be happy to work with any Commissioner who would like to do that. But that would be very helpful.

MS. CHARO: Yes. I know even in light of the complications of FACA and requests, et cetera, I would love to urge a return to the use of e-mail to allow for more interaction between meetings to get the text kind of refined.

DR. SHAPIRO: Yes, that's fine. FOIA is no problem in this area. No reason why.

Let me ask questions then about three and four. When I read it, my first reaction was one of being somewhat troubled by them. I understand the need to protect privacy of innocent people who are not participating in that. And if that's what's meant here, then I don't have any problem with it. If that's what's meant one way or another here, then I don't have any problem with it. And the question is, was anything else meant by this?

DR. MESLIN: No.

DR. SHAPIRO: In that case, whatever we say about this ought to be able to be said in a really compact form because I think it's pretty straightforward. And in what way would that differ from anything that's extant right now?

MS. CHARO: I'm sorry, we were on—I was dealing with another—

DR. SHAPIRO: Three and four in some combination. It would seem to me similar. It would need to be combined in some way.

DR. MESLIN: The only distinction is whether we feel it is important to make clear what is either ambiguous or silent in existing regulatory language regarding the ability of an IRB or others to focus explicitly on potential harms to others. And we've listed the types of others and have put language that says there should be concern and sensitivity to this. But up until we say that the regulations should be changed that explicitly empowers IRBs to do this, then, no, there's nothing more involved than that.

DR. SHAPIRO: Putting IRBs or particular institutional arrangements aside here, who is it that is going to, if there is some responsibility here, you're trying to articulate a responsibility, who is it that has the responsibility and what's the nature of the enforcement action here that we're talking about? That's the part that's completely unclear to me.

MS. CHARO: Right. This is what was going on yesterday. As I understood the conversation yesterday, we were placing responsibility first on investigators to voluntarily present themselves to the IRB in situations where the regulations don't require it to ask for discussion, guidance, changes, et cetera. We were asking IRBs to take this request seriously and to develop expertise in how to minimize harms or how to look at the social implications of research, and if they are seemingly dangerous, how to review the scientific validity of the study even more closely. Because none of it was required. It was all voluntary.

Now that said, on a variety of these things it should be mentioned perhaps that IRBs are never forbidden to go beyond the federal regulations. They can always go beyond. They can do it on an ad hoc basis, or they can formalize it in the form of their Multiple Project Assurance with OPRR in which they promise to do it on a regular basis. If they do that, there is an enforcement mechanism because they are bound by that MPA and if they fail to comply with it they are subject to investigation and sanctions from OPRR, which may be a disincentive for them to formalize these supra-regulatory tasks. But those are ways in which this can be incorporated a little more formally.

The other thing that can be done is, as David and I were talking kind of on the side here, is to aim some of these things at the world of scientific journals and at NIH, NSF, and other granting agencies. Journals and granting agencies are also perfectly free to say that we will give preference to those papers or those research projects that meet an even higher standard from our point of view of the protection of human subjects. They can add that. So in the study section, they can not only ask whether the appropriate IRB review was done, but whether or not there are any residual problems. Study sections that I've been on have certainly gone beyond the IRB review on occasion and said this protocol is problematic.

DR. COX: And professional societies. Because this is then the carrot approach and not the stick approach. But if professional societies say these are very important issues to get more data on so we know what to do, and so we want the researchers to basically pay attention to this, it's not being shoved down people's throats. On the other hand, I don't think it can be completely voluntary. But have different ways of enforcing this.

MS. CHARO: But it is true that in the end it is largely voluntary and it is circling around what Alex has called the "weak read" of IRBs. So that this is a bit of a house of cards.

DR. SHAPIRO: Okay. Well, in any case, we'll do our best there. We're going to combine these in some way to get it a little more effective. I still have some troubles, but I'll put those aside for now.

MR. CAPRON: One linguistic comment. The things under Roman I we were calling guidance, and here we're calling them recommendations. I realize that this is in some ways just a temporary memo and language was adopted for that reason. But things like this have a way of developing a life of their own. I see no reason to distinguish between recommendations that we make to the federal agencies and calling those recommendations one to whatever, and recommendations that we make for actual changes in the language of something, or IRBs, or whatever.

DR. SHAPIRO: Okay. Do you want to say anything further about recommendation five at all, Eric?

DR. MESLIN: That was a placeholder recommendation which is really folded under some others. Maybe Bernie wanted to say something.

DR. SHAPIRO: Bernie, I'm sorry.

DR. LO: I actually would like to say a little bit about five because I think it does get to this notion of group harm that we've been talking around. It seems to me there's a difference between the actual design of a study, going to a group that's been identified on ethnic or racial terms, often because they had volunteered for a previous databank, and using them mainly because they're convenient but the more you know the more interesting genes you find. That to me is different than you do a study where ethnicity or race is not a variable at the onset but when you finally do the analysis you find linkages or whatever.

It seems to me we ought to ask investigators to pay more attention to the former situation where it is part of the integral planning of the project. Again, I'm particularly concerned, I'm thinking about the news stories of concerns of Ashkenazi Jewish families in Baltimore who get used over and over again because there are lots of interesting markers and the samples are already collected and easy to get to. And yet they, at some point, maybe begin to have real concerns about the benefits being traded off against the sense that they have one serious marker gene after another. And although all the studies when they're published say this does not mean that this group is particularly susceptible or the other groups may not have as high a prevalence, the point is we just keep going back to them primarily because it's convenient because the samples are there.

I think the problem there is I don't know how we can get beyond the considered implications and act prudently. But I think maybe some guidance there as to what do you do in that situation. You're going to do a study on that group most likely, how do you mitigate it? It seems to me there is a real role for having a real dialogue with the community and at least explaining to them why you're going back to that same group, providing a lot of community education as part of the price you pay for having a really nice database that someone collected for you twenty or twenty-five years ago.

All the questions we raised yesterday where the synagogue has turned over, they've moved, the initial congregation is dispersed. I think that's all true, but that you shouldn't use that to say now I don't have to do anything. I think you do the best you can to find people who could seriously feel that they're implicated by this design of the study. And, again, it's more in the nature of an exhortation we're talking about and all the comments people that we're saying about using the carrot approach. True. But if we could be more specific about what constitutes best practices in that situation, it might be helpful.

DR. MESLIN: Bernie, one of the reasons why I made the comment that I did was in talking about Guidance 2 yesterday, which was that something should be developed that directs IRBs to address the issue of group or community harm, that's what I was referring to by this recommendation. What we discussed yesterday was sort of concentric circles, individual, family, social group, non-descript group. If you feel that it would be helpful for us to actually stipulate now that that guidance should include a kind of consultation or engagement which we've been discussing, that can easily be added in as a specific suggestion for what those—

DR. LO: I was just saying that in this context I would like the implication here that some things you weren't planning at the beginning but in the course of doing your research you began to identify certain groups or families. And there are other studies where right from the onset you know the results are going to be attached to a certain readily identifiable group because that's the way you designed it. That seems to me that latter situation ought to have more responsibility because it was much more foreseeable and sort of intentional.

MR. CAPRON: In many ways, I agree with you. But it would seem to me that if race or ethnicity is one of the factors which you are gathering in your database, we wouldn't want to have a situation set up in which we have a high hurdle for the first kind of study, the Baltimore study, and then people figure, well, I'm not going to do that but I'm going to gather the data, and I know I can always run the analyses and I suspect that I'm going to see some racial/ethnic, and I'll end up saying Jewish women not only have babies with Tay-Sachs, but they have breast cancer a lot, too, even though I didn't "design" it that way. That's not the right incentive to give.

DR. SHAPIRO: Alta?

MS. CHARO: Bernie, I think that it's possible that this set of concerns might be reflected again in the study that has been proposed on looking at the IRB system and human subjects protections generally. Because part of what has created this problem for us is the fact that the IRBs are not only under-staffed, overworked and under-rewarded within their institutions, it's that they are reflections of the professional expertise within the institutions, with very few exceptions, and that therefore they are in fact not a very diverse group of people and you don't have the kind of range of life experiences and sensitivities that you would have if you took a random group of twenty people off the street.

And so perhaps in that larger study one can address more widely some of the social costs that we are experiencing from having an IRB system that has been enfeebled by small size, overwork, et cetera, and the value of diversity in kind of preempting some of the problems. Most of the institutions, they don't have a lot of women, they don't have a lot of blacks, they don't have a lot of people who are disabled working there as professionals, relatively speaking, and so they don't show up on the IRBs either and their sensitivities don't show up on the IRBs in any particular way. So, it's all a part of a larger problem, and we can definitely make sure that it's highlighted.

DR. SHAPIRO: Okay. That's fine. Well, we'll certainly do so.

Let's go on and look at an issue which comes up as perhaps a little bit of a placeholder also, which is on the recommendation number six on page seven, which talks about waivers.

Alta?

MS. CHARO: This one actually got red pen on it from me, because I think it must be a typo. It recommends that when considering a waiver the investigator has to provide evidence it's not practicable. But that's already a requirement. So I figure that's got to be a typo. And, in fact, what we're talking about is NBAC recommends that not practicable will be understood as the following. Is that okay?

DR. SHAPIRO: Yes. That's what I thought about it. That is either we have something to say here or we don't. If we want to say something, okay, if not we don't have to say anything.

MS. CHARO: And even before we get started on filling in what we think "not practicable" ought to mean, I'd like to add this to the other request about the meaning of human subject in a request to OPRR to simply alert us to any guidance they've ever issued or any contacts they've ever had with IRBs that have called them or written to them saying "What the heck do you mean by this?" so we can find out what precedence there are out there for interpretation of this term. It could be that I just missed it in the guidebook and there is an interpretation, but I didn't spot it.

DR. SHAPIRO: I didn't either.

MS. CHARO: And that would be useful information to supplement whatever we do here.

DR. SHAPIRO: We'll certainly do so. Alta wanted information on OPPR's dealing with the issue of human subjects and relatives.

MS. CHARO: Who have kin, living kin, right.

DR. SHAPIRO: What information they've put together.

DR. MESLIN: What you've just described is not a proposed recommendation. That's a request that we obtain that information to inform us?

MS. CHARO: It's a request for information. That's right, so we can work through what we think is a tentative recommendation today, but I would love to be able to test it against anything they provide.

DR. MESLIN: Yes. I mean, we have, for example, provided the Commissioners the two memos that OPRR provided guidance memos to NIGMS and to NHGRI of about—Melody Lynn knows the date, if she's in the audience she can confirm it—of probably a year ago which answered a number of these types of questions, not the specific one you're mentioning. But you would like to know whether there are other documents of that kind where OPRR has responded to questions of this sort.

MS. CHARO: On these two topics. Right. And if they've already been given to us, I apologize in advance. My filing skills are minimal.

DR. SHAPIRO: It's on these two particular issues as opposed to just other kinds. One dealing with the living kin, and one dealing with this issue of practicable, if there's ever been any tussling, conversations, working out of what on earth this means. David?

DR. COX: But I would just like to make a plea that as we ask HHS or OPRR or anybody for their interpretations of these, when nothing is forthcoming, that we make interpretations and that we aren't stymied by other people's lack of action.

DR. SHAPIRO: It's just to see if there might be some useful information.

Bernie?

DR. LO: I was going to jump to the next thing, of trying to flesh this one out.

DR. SHAPIRO: Yes. Let's go ahead.

DR. LO: This becomes a loophole; becomes a huge tunnel. Clearly, if someone is dead or moved away and you can't readily locate them, I think that would count. What bothers me is in most health care systems every year there is sort of a re-enrollment process where they check and make sure you still have insurance coverage and you've chosen them as the primary provider. It seems to me that in that situation where you're providing ongoing care, there is built-in periodic contact with the person. It seems to me at least the presumption or the strong presumption in that situation it's practical to contact a patient. Now, it may be very expensive from the point of view of the researcher who doesn't have a large grant. But I think we have to distinguish I can't do it because the person's not there or I can't locate them versus I can't do it because I just don't have enough 32 cent stamps and Xerox money.

DR. SHAPIRO: I think, Bernie, I think that point is well taken. But I think it is a critical issue whether economic barriers get sufficiently high to say that means impractical at some level. If it's only economics we're talking about and not logistics, that is, presumably with enough resources you can find a lot of people, even those who moved away and so on, you can trace them down. And so we might want to have some discussion of just what kind of economic burden do you have to exceed in order to sort of satisfy this hurdle.

Alex?

MR. CAPRON: We heard earlier on that there are situations in which health plans, for example, are cooperating with researchers on a prospective basis. And in a certain way we've separated the retrospective and the prospective, but today's prospective becomes tomorrow's retrospective. And it seemed to me that when we are talking about a health care organization that is, in effect, setting itself up in a contractual relationship with a group of researchers, particularly when those researchers are private companies that see a product development coming out of this, that the notion that impracticality is a pretty thin excuse comes in here.

That is to say, it is not just on the reenrollment, but if you're putting people's tissues into a database which you then plan in effect to market, it seems to me reasonable to say you have some obligation to do the kinds of things that I think we heard the breast cancer group was doing, that of having a regular recontact of the people whether or not they're still in your health plan. But the every six months or every year, "We've got your tissues here. We want to stay in touch with you. Please confirm that this is your address. If this has been forwarded to you, please give us your new address. If we don't hear from you, that means we're taking your tissue out of the bank, which is a great cost to science and we really encourage you to continue to allow us to do the research." And whatever arrangements.

But it does seem to me we're talking about potential harms to people. We wouldn't care about this if there weren't some potential harms to people, and the ethics of protecting people against that harm unless they've consented to it. On the other hand, dollars and cents. And dollars and cents should be just as relevant in this aspect as they are to the fact you have to buy expensive reagents and hire Ph.D.s to do the research and so forth. No one says we ought to be able to do research by using slave labor or something. I think we ought to keep that in mind when people wring their hands and say, "But if you put up that rule and say that if you've sent a postcard and you've gotten no answer back, then that's enough. That it's impractical to re-contact the person because it would cost more to be more active and trace them down and do a little more digging. And if we don't do that, we lose the tissue." All right, well, we lose the tissue because you want to protect people against being used in situations where information is developed about them which could be harmful to them and they haven't consented.

And it's just a cost matter. So I think we have to elaborate on this practicality in some detail, and I'd set a fairly high level for impracticality before I would be comfortable.

DR. SHAPIRO: Diane?

DR. SCOTT-JONES: I agree with Alex on this. I think the situation is exactly the same as that of longitudinal research where the burden is exactly the same—you have to recontact the persons who participated in the study and ask them to continue to participate. And it does cost money to follow them. There are some people who die and you have sample attrition. But it's just the same. I think we may be over-emphasizing this notion that it's impractical to continue to contact people. And there are firms that specialize in doing just that for longitudinal studies; they find very clever ways to contact people and to get their continued enrollment in the study.

DR. SHAPIRO: Just to pursue this particular point. David, I know you. It's one thing to conduct a longitudinal study and an organization required to do that. It's another to come up with a problem and you want to go to a tissue bank, use some tissue or some other material of one kind or another, and then, not because you want to do a longitudinal study, because you want to do a study now on a particular issue, and then you're faced with the same problem of having to contact the people in order to protect them from harms.

And the question then is, I would think, that at some level economic burdens become high enough because they're a proxy for practical. That is that it's so hard to find these people and you have to put so much resources into it, there must be some burden which—

MR. CAPRON: But the choice to use existing tissue samples rather than currently collected ones where the people are easily contacted is itself perhaps—

DR. SHAPIRO: Correct. That's a good point.

MR. CAPRON: This is an easily, readily available source. And we're not talking about a situation where you go to the tissue bank and say give me a bunch of anonymous samples, I don't need to know anything more about them. We're saying give me this, give me their medical records or heavily abstracted stuff from their medical records.

This is a choice. And if we're going to shrug and say, well, but it's just a matter of inconvenience to get that far and you could do it if you spent more money, or the way Diane says, you get the same firm to work and trace down this person and say are you willing to do this, if we're going to shrug it off, why are we worried about any of this? Why are we saying there's any concern? Just use the samples, don't worry about consent. And I don't think that's our attitude.

DR. SHAPIRO: Right. I agree.

Bernie?

DR. LO: I think we also have to look at the burdens relative to prospective benefits. So it seems to me, in addition to everything that's been said so far, it's not clear to me why for most studies you can't just use the samples for which you have recontacted and gotten consent. So that if you think there's an overwhelmingly important scientific reason why because those people may not be representative of all the samples, then you have to make that argument. But it seems to me, again, there's a counter. If you're so concerned about the sort of generalizability of a population, then you need to go back to a population drawn sample, not just all the people that happen to end up in your hospital.

So, I think in addition to what we've been saying about not allowing economics to sort of just run willy-nilly over concerns about privacy, I think also the argument on the other side that the benefits are overstated. I think we need to look very critically at arguments saying what the scientific loss would be if we restricted ourselves to samples where people have been contacted and get some sort of meaningful consent.

DR. SHAPIRO: Diane?

DR. SCOTT-JONES: I'd like to say that I also agree with what Bernie just said. I think there is always a conflict between the obligation to get informed consent and the possibility that you're reducing the generalizability of your sample because you're not including all of the persons. That's just something that we have to live with. And I agree with Bernie that we may be over-emphasizing that. This is always the problem of research and it is not particular to this kind of research. And that is, if we believe that research participation should be voluntary, then we have some loss on the scientific side that we learn to live with and we learn to manage it by other means.

DR. SHAPIRO: David?

DR. COX: So, Eric, this is one of those one-liners to put on the little card that you strap to people's chests, right? However, as much as I'm in favor of this, this discussion frequently gets extended in the following way, which is impractical. When you can't contact those people, you will extract all of their information out of the study so it's not there. That's not possible to do. So that you may not use the samples anymore, but that train left the station already in terms of how that sample information and data was in the study. And it is absolutely impractical, and I'm open to people showing me how this can be done, to take that information out of already analyzed data.

This is one of the reasons, and this will come up because people will try and obfuscate this basic principle, which is you go back and recontact and use the sample. But they say, well, but how can we ever extract that information out of our data in our study if it's already in there? The answer is, you don't. You just don't put it in anymore. But this is this whole business about giving people the right to withdraw from a study. So that they have the right to withdraw, but it has to be understood that after a certain amount of stuff is done, it really is impractical to take all of that information out of the study.

DR. SHAPIRO: Alta, then Trish.

MS. CHARO: First, I've got to tell you that we've run into exactly that situation on our IRB and I sympathize.

DR. COX: You can see it over and over again.

MS. CHARO: Right. I apologize for throwing a monkey wrench into what seems to be a developing consensus that money doesn't matter, money is no object, consent must be obtained at all cost. But I am not completely, maybe I'm misunderstanding, but what I'm hearing is that we should not allow people to say it's impractical to re-contact for consent simply because it's expensive. If I'm mishearing, that's great because then it's not so much of a monkey-wrench.

I am reminded of David's example about the kind of tiered research that might go on in which the first pass is simply to identify people that you might want to recruit for a more intensive look. So the example you gave was you might want to do a first pass looking for all people that have a hypothetical marker for a gene that is suspected to be associated with prostate cancer. And the goal is simply to identify the subset that have the marker, and then in fact to contact them and specifically ask whether or not they would be willing to be part of a study. And that first pass might be over many, many samples.

Now, in many circumstances, it might be true, that contact is feasible because they are members of health plans that have this kind of reenrollment process, et cetera. Not always, because a lot of the tissue collections that are described here in the early parts of the draft are not maintained by health plans. But it may be possible here. But the expense might be quite substantial. In light of the effect that has on your ability to get the research grants, which, of course, are limited by dollars, there are X number of dollars to go around and if you increase the cost of each one by virtue of requiring contact even for this kind of first pass, you reduce the overall number of grants available or you increase the R&D cost for a pharmaceutical company. You can speculate there is going to be some number of marginal research that won't get done. It will be that last few things.

I understand why this sounds really awful because it's the same argument about minimum wage and how there are marginal jobs that won't get filled because it costs too much. But I'm worried about this because there are, it seems to me, examples of exactly the kind of thing where expense seems to be a very legitimate concern but I don't have any idea of how to peg it. Is it percent of the overall cost of the research? Is it a certain absolute number of dollars? I have no idea how to implement this instinct that there really is an economic argument to be made that at a certain point it may be technically feasible but it is not practicable.

DR. SHAPIRO: David, then Bernie. Excuse me, Trish, you were on the list.

MS. BACKLAR: This is going on another path. I just wanted to go back to say what David said about when you have somebody, and it relates to what Diane said, if you have people in your research protocol and somebody drops out, you don't drop out the data that you have on them up until that point. It doesn't matter when it's genetic research or it's on the kind of research that I do. So it's exactly the same situation.

DR. COX: But just to make that point, because otherwise it sandbags and it obfuscates this principle.

But Alta brings up a really important point, and this is being faced over and over again by the NIH. Alex said this well. This isn't all large epidemiological samples. These are samples that you will use over and over again that you get more and more information in. So the NIH right now in many different settings is trying to grapple with how do you deal with the expense and the coordination of getting high quality data.

And you know what they're doing? They're contracting it out. They pay for it. The expense doesn't go to the investigator, but it's viewed as this is part of the price of doing business. And it's not part of the funds that the investigators ask for in their grants. It gets contracted out. It is part of the research enterprise that for these subsets of patients and materials that are a national resource, you go back to those people. That doesn't mean, though, that it's every large epidemiological study. So I think that the economics doesn't have to be at the individual investigator level.

DR. SHAPIRO: Bernie, and then Trish again.

DR. LO: Just to follow up on this topic, because I think it is really important and I think it does come up a lot with investigators and IRBs. I don't think that we were trying to say that economics never should be a factor. I'm just saying you have to really define what it means to say it's economically unfeasible.

MS. CHARO: My apologies for misstating what you said.

DR. LO: And I think also you just have to be really creative about how to address that. I think we should call for the NIH to fund center grants to set up the infrastructure. As David was saying, that's part of the cost of doing research. I agree with Alta that it's unfair to put the burden on the individual investigator. But the NIH is looking at prospective increases in budgets and this is one of the things that they may want to consider. To go back to Alta's example, losing by not restricting that first sweep to that subset of patients that you have been able to recontact through any of these other mechanisms, and is their—you know a compelling scientific reason why you just won't be able to get that group. And then the next question is if you can't get it from your own database, why not go to one of these other large databases that has been collected by someone else that has set up infrastructure to be able to recontact people, or has had a tiered consent that we're going to talk about later, so that what we may want to do is to centralize a lot of this into a relatively small number of very well maintained large databases that have the capacity for ongoing recontact. It's really important for a patient to know that their personal sample will be used for research. That's going to be impractical if it's at some outlying community hospital that doesn't do a whole lot of research. We shouldn't be misleading people to thinking their particular sample is really to be used. It will only be used if it somehow it has gotten its way into a database that someone is committed to keeping up. I think, again, I go back to the epidemiology. I mean if you're really wanting to get a representative sample, you've got to get a population-based sample. The samples I have lying around my hospital are biased samples of prostate cancer. And I'm kidding myself if I think that I'm gaining something by looking at those data, rather than a population.

DR. SHAPIRO: I know Diane—I want to call Diane in just a second. But just try to listen carefully to what's being said here. There is an issue of a loophole here which is gaping so large that we're not demanding. We're letting it through without sufficient oversight, in some sense. We're letting them use this excuse when, in fact, if you look at science requirements, they really don't deserve to be let through this loophole, and just become larger, and, therefore, not protective enough of those people whose tissue is being used. And I think as an empirical matter, I think it's probably correct. And then—and I think—so I think we all agree with that. I also think we agree that somewhere out there there is a ceiling where the economic stance for something, not just money, it stands for a whole series of difficulties, which we would declare impractical, but that's some high number, or, at least it's a threshold considerably above what's been—what's currently in use, if I understand what people are saying. And then the question is just how do you deal with the—what's in between those two things and how do we articulate. Okay. I'm just trying to think out loud. I'm sorry.

DR. SCOTT-JONES: My point—Bernie started making, and that is that there is an infrastructure, the scientific endeavors, and David was referring to some of this, too, where people are now engaged in discussions of how to use data jointly, how to data share, how to do things to minimize the cost of research. So I think we need to make our recommendations in the context of what is going on right now in the scientific community and many people are discussing now how to minimize the cost of doing research through a number of means. So I think that you know we should put that in mind before we think that it's going to be too expensive to do the kind of things that we're talking about. Okay. Well, we're going to—although we're not through this, I think we're going to call the discussion on this particular report to a close now to get on to other things. We will continue this discussion by E-mail and other issues, that if we have time later today, we're going to pick up some of these. If we have time we can come back to some these issues, so it just depends on how much time we use on the other issues. There are still some very important issues to be thought about here. So I haven't divided this between important and unimportant, but just to try to keep ourselves to some kind of reasonable schedule. So we may very well come back to this later on today, but if not we will do so well before the next meeting using one form of communication or another. I also want to mention something which didn't come up in yesterday's meeting directly, but it's really indirectly in some of the memos that we have received. Namely, there are increasing initiatives, and the level of state laws in this area. Do you know from the memos we received—and some of you are much more knowledgeable than—or other reasons. Those will have to be acknowledged in some way and we'll have to couch our recommendation of the number of areas, given the background that states may choose more restrictively regulation on a whole number of these things, which, of course, they are perfectly free to do with these, as I understand it. We've not had a chance to discuss that, but that will also have to get worked into some of our material. Yes, Bernie.

DR. LO: If I could make a request of the staff with regard to that point? I think, as one of the people who doesn't know a whole lot about this, I think it is important to know what the pertinent state laws are. And so, I would actually like to see actual provisions of Florida and California. And it wasn't clear from this whether Delaware is a bill or a statute. I am actually much more interested in statutes that really are there, as opposed to bills which you know may or may not go anywhere. But the actual language and provisions would be really important, and maybe some comparison of the—I don't know whether it's three or four states that should have pertinent existing statutes.

DR. MESLIN: We can give it to all of the Commissioners. We have such a document available here today, and we can give one to you if that would be helpful.

MS. CHARO: Also, Bernie, just by way, I think you may have received a fax late last week of one memo from one of the staff people that actually did, at least in the fax that I got, have something like 20 pages worth of the actual provisions. So, if you—you might find that you have some of it already. It's easy enough for them, sure, to get you another copy but.

DR. SHAPIRO: I've got the six page version.

DR. MESLIN: The lengthy text that Laurie provided was not faxed to everyone, but the cover memo itself was.

MS. CHARO: And the—well, never mind. If you didn't get it. You didn't get it, right.

MR. CAPRON: One other thing, at some point we need to decide. One thing Laurie does in this memo is to say that if you see a pattern of state legislation or even a lot of legislative activity that isn't yet statutory, but it's still on the bill level and you make some conclusions about the area of public concerns—this is the sociological use of the law, rather than a binding legally. And see I think that's a reasonable thing to do in the way of buttressing our concern.

DR. SHAPIRO: I agree. I agree with that. All right. We are due to take and will take now a break. We will try to reassemble at 10:30, because I know some of you need to check out, do other things, make phone calls, and so on. So, let's reassemble here at 10:30. Now, we'll turn our attention directly then to the what was known in shorthand as the capacity draft. So, thank you very much. We'll reassemble at 10:30.

DR. SHAPIRO: All right. I want to turn now to that part of our agenda that will deal with the draft we've all been studying regarding research involving subjects with mental disorders that may affect their decisionmaking capacity. And, again, I want to thank Jonathan for being here with us for this purpose, and for the help he's given us all along in this. And let me just say something about the agenda, then turn over the chair to Jim, who will take us through this discussion. We will focus on this till 11:30, at which time we will take a pause for public comments. We will then immediately return to dealing with this report, and work on it through lunch. At one o'clock, Mary Claire King will be joining us and she will speak to us approximately in the neighborhood of a half an hour. And then, depending on how far we’ve gotten, we’ll return to this or move on to other subjects. So, let me turn to Jim. Jim.