DR. SHAPIRO: I’d like to get our meeting back together. We are running a little bit behind time, but it was well worth it because of the discussion with our guests.

The next item on our agenda is to continue our discussion regarding the use of human biological materials, the report that we’re in the process of creating in that area. We have done a lot of hard work on that issue in the last two meetings as we’ve talked back and forth on just how we’re going to approach this issue and what kind of structure we should give our reports. We will have some guests to address us on some aspects of this issue tomorrow as part of our discussion in the morning. That’s on your agenda. And while we had some discussions in Los Angeles on this issue and some very helpful discussions, there were aspects of this that we didn’t fully consider at that time we need to make sure that we address as we go ahead today.

Let me begin our discussion by just turning to Kathi, who has prepared some materials for us. I think, Kathi, you want to go through at least one aspect of that to begin with and get reaction from Commissioners on that issue. Let me turn to Kathi Hanna.

DR. HANNA: I first want to go over what I think in the next day and a half will be helpful to me at least to proceed to put a report together by the May meeting. So I’ll go down the list first and then I’ll ask if we can concentrate on one thing this afternoon.

The first is, and this is what I’m going to ask if we can get started on this afternoon, the definitions that are at the very conclusion of the document following the reconfigured boxes. If we can decide or agree on what the different terminology is and what we mean when we are using these different categories, if I understand that we are in agreement on those terms, then that will make it much easier to put the whole document together. So I’d like to get started on that this afternoon.

And if we can get to some agreement on these definitions, then we can turn to the boxes. I’ve tried to reconfigure the table slightly. So we can talk about that and whether that’s useful or not. And I’ve also tried to indicate where I think we have completed a box and where we still need to talk about what the recommendations are going to be. So that’s task two.

I think if we can start on completing those boxes and, therefore, recommendations for prospective collection of biological materials and we can reach some agreement on that, then it would be fairly easy to turn to previously collected samples and make sure that everyone is in agreement on what has been concluded in those areas.

I think it would be helpful also tomorrow to have a discussion about risks and what people perceive as risks so I have a good understanding of when we talk about things like minimal risk, or risk to the individual, or risk to groups, that we all understand what we mean by that. I think if you’re going to try and include some guidance to institutional review boards, then we need to be very clear about what we think risks are.

We still have to grapple with the issue of community. I’ve left it out of this table for now because I think what I was hearing was that any time that community was implicated a whole different set of considerations kicked in regardless of whether it was previously collected, prospectively collected, in a clinical setting, research setting, whatever. So I think we need to talk about that. We need to talk about how one identifies community, and what we want to tell IRBs or investigators about what they have to do there.

And then in more kind of procedural tasks, you’ve all received the paper by Alan Buchanan. I think it would be helpful to get some feedback on that; what you find in that to be useful, and where you think that what he has written supports some of the recommendations and conclusions that you’re making, and how we can integrate that into the paper.

Then the last thing is that Chapter 2, which doesn’t appear in this briefing book because it’s relatively unchanged since the last time, we are going to change that to also include a section on use of human biological materials to represent some of the examples that David Korn provided in his paper to give a sense of the historical perspective on use of samples, and include some more information on what we see as the future in terms of use of tissues.

There is also a memo in front of you today which is really based on information that was forwarded to us by Steve Holtzman when he talked to some scientists at his company about what they see as the future of the kind of research they’re doing and what their data needs are going to be.

So that’s my list. I’m sure that everyone else has their own list of what they think we need to get done. But I would find it useful if we can agree on terminology and definitions so that for the rest of the discussion we all know what we’re talking about.

DR. SHAPIRO: Kathi, I think that’s very helpful. Obviously, there are a number of different agendas we could order, but that sounds like a very systematic and useful one. Let me just play it back and make sure that I’ve understood it so I have some sense in my head of how we will structure this discussion.

You want to begin with some agreement on definitions. You’ve provided some start to that in the last page of the document that we all received. You then want to proceed to the boxes and its various dimensions, which I won’t try to describe right now, dealing, first of all, with prospective collection and then dealing with the historical collections that are available, historical in the sense of already collected. And then you would like to proceed to understand a little better about what we may all mean by risks and harms, both the individual, to groups, and so on, and what guidance therefore that we might give IRBs and so on, that set of issues. Then to the community collectivity issue, which we all talk about and the more we talk the more complicated, vaguer, and wispier that idea gets. But we need to at least pin that down and decide what it is we want to do with that. And then deal with the material that Alan has sent, and what we think about that. And then Chapter 2, which you may want to talk about if we get that far down the agenda.

That seems very appropriate to me. I’m sure there’s more than one systematic way to go at this, but that seems like one.

Tom, does that seem sensible to you?

DR. MURRAY: Yes. I think that covers my list very well. I have some questions and I know Carol wants to make some comments.

DR. GREIDER: Only if we’re going to be going into some of the substance.

DR. SHAPIRO: Yes. Well, I hope we’re going to get started right now. I just want to go down and see if that’s okay as an agenda to proceed.

DR. MURRAY: That’s fine.

DR. SHAPIRO: All right. Let’s start then with the definitions. To remind all of you, it’s on page 5-15.

DR. HANNA: Would it help, I have an overhead.

DR. MURRAY: Yes. It certainly would be helpful for the people in the audience to know what we’re talking about—if that’s something we want to happen.

PROFESSOR CAPRON: May I ask a question precisely on this. As we work on documents now which are marked as working documents and so forth, are we making these available—I didn’t check the table outside—for people in the public as we go along who are sitting here trying to understand what we’re talking about?

DR. SHAPIRO: Let me turn to staff to find out. I just don’t know.

MS. HYATT-KNORR: So far, we have not made the drafts available.

PROFESSOR CAPRON: At some point we need to have a discussion of this. We got ourselves in terrible trouble on the cloning report, as it turned out, that we were having discussions and I was not aware until our last meeting that the people who were coming to hear the discussions, including the press, were trying to figure out what we were talking about from the oral comments on something they hadn’t read. We’re a public body. When we have staff drafts that are going through the process before meetings, of course those have to be worked out by the staff. But once we have a public meeting and are meeting in public session to discuss a document, it has to be understood by everyone that they shouldn’t take this to be our final document or say it’s our conclusions, as the top line on every one of these says. But I don’t understand why they aren’t available to the people who come to this meeting or anyone else who wants to see them while we’re working on them. We’re supposed to be a public Commission. We can only benefit by having people understand what we’re doing and not trying to parse it together like something from a Delphic Oracle.

DR. SHAPIRO: Thank you.

PROFESSOR CAPRON: That is not intended as rabble-rousing.

DR. GREIDER: Can I respond quickly to this comment?

DR. SHAPIRO: Yes.

DR. GREIDER: Although I agree in principle with everything that you just said, Alex, there are a number of things that I feel as a Commissioner which I have gotten as a draft document that I’ve never seen before in my life. I would feel very uncomfortable if these were handed out as a draft document to the public as something that NBAC is working on when I’ve never seen them before.

PROFESSOR CAPRON: Well, we aren’t working on them. It says right at the top this is a draft document developed for, not by, for the National Bioethics Advisory Commission. It does not represent conclusions and should not be cited or referenced, I guess cited or referenced as conclusions. I mean, obviously, anybody can quote from it. Right now we’re going to have this up on the board, and someone who is really good at shorthand can take that down and write about it or go to a class and talk about it.

DR. SHAPIRO: All right. Your point is well-taken. Let’s just get on with the work here.

PROFESSOR CAPRON: Well, are we resolved that we’re going to do something in the future?

DR. SHAPIRO: We’ll try to do something. I’m not resolving anything right now until we get a chance to talk to staff about it.

Let’s go to the definitions.

DR. GREIDER: First, I’d like to apologize to my fellow Commissioners because I wasn’t aware that we were going to be dealing with this today and I haven’t thought about it as carefully as I would like to before discussing it. But I would like to turn to some of the definitions, and I absolutely applaud the idea of trying to put some information into these boxes since we’ve been trying to do that for a number of meetings. But what I’m going to give you is some quick impressions that I have having looked at these definitions. I would like to be able to give more definitive suggestions and hopefully I can do that tomorrow.

And so let me give you just a couple of quick impressions that get to the issue of these definitions, and that is these definitions Kathi took from the reorganized table where she renamed a number of the categories that we had in the past. And one of the distinctions that I felt that we had made in our other meetings that I thought was a big step forward was to refer to the way in which tissues were used rather than the tissues as entities sitting somewhere. Zeke had suggested to us that we use the language of tissues that are used in an identifiable manner, or used in a coded manner, or that sort of issue. And we’ve now changed back to the sort of historical definitions that many of the previous people had before us, which is referring specifically to the tissues themselves.

And since this comes out in the definitions, I would like to see some rewording of the actual terminology. For instance, the first one, unidentifiable sample. That I think should be something along the lines of a sample that is used in an unidentifiable manner. And coded but untrackable samples, I think a sample which coded so that it is untrackable. Something along those lines. I’m sorry I don’t have exact language, and I’m happy to help work with it. But I feel the direction that we had moved in, and I don’t see that reflected in the language as has been rewritten.

DR. SHAPIRO: David?

DR. COX: This is a very contentious point, what you’re bringing up now, Carol. The genetics working group I think very rightly realized that how things are going to be used is a very useful way of looking at things. But at our last meeting, we had a discussion that I’d call the "red face test," and that is, if you have things that you can identify and you say that they can’t be identified, people are going to look at us and wonder what the hell we’re talking about because it does not pass the "red face test."

If you have something that you can identify and you say you can’t identify it, that’s garbage. So that we can incorporate, in my view, the concept of something in terms of how it’s being used, but I can’t describe how strongly I feel that if a tissue can be identified, that we cannot define it as being not able to be identified. That’s a very strong position, I realize. But I put it up just that strongly so we can have a really clear discussion about it.

DR. SHAPIRO: Tom, then, Alta then Alex.

DR. MURRAY: We made progress last time and I’d rather not have to go over everything from last time. I think actually Kathi’s definitions capture key distinctions. Now, I may want to label differently and others may want to provide a different label to them. I want to just make two points I guess here.

One is, in response to David’s point, no one is failing the "red face test" in anything that we’re proposing to do here. An advance I think we made at the last meeting was to distinguish between those situations in which—I think we had four categories. One category was in their current state of storage there is no identifying information. There’s insufficient information with these tissues to know from whom they came. That’s category one that Kathi calls it "unidentifiable" here. We can decide what labels to use.

The second category is, in their state of storage, the tissues have identifying information with them. The tissues go forward with perhaps some concurrent information but insufficient to identify the person and there is no backwards link, there’s no code or any other thing that would permit you to go back and tell who it came from. So that’s her second category. She calls it "coded but untraceable." I used to phrase it something like "identities unrecoverable."

Number three is, again, the information has gone to the researcher without specifically identifying information but with something. There may be some kind of—Kathi calls this one "coded but traceable"—there may be a way to get back. Perhaps there might even be a code held in the Grand Cayman Islands or something, but somebody could potentially make the link. We want to be clear about that.

And the fourth one is where we know. The information that actually goes to the researcher includes sufficient information to identify it.

Furthermore, in samples that are not identifiable; we mean at least two things: One is there isn’t specifically identifying information going with them or available; and we also mean, number two, that the context in which you get this information wouldn’t permit you to identify, and we used examples about that. Sometimes you can identify because you know the nature of the repository, there are only three cases, and you can figure out who it is. That would not count as a "not identifiable" tissue.

I hope we can still work with those four basic categories. I think Kathi has captured them here. Kathi wants to say something.

DR. HANNA: Yes, I just want to add that I can’t take credit for the terms "coded but untraceable" and "coded but traceable." Those actually came from Zeke.

DR. MURRAY: Okay. So we’ll either praise or blame Zeke today for the particular terms.

[Laughter.]

DR. MURRAY: The second point I want to make is a very quick conceptual point. Philosophers of language worried about this decades ago, and that is, do we want to make use of commonly available terms but then make them fit around the concepts that we have which do not necessarily coincide with the current use of these terms? In that way, do we want to reform the language, and then hope that people will understand what we’re saying? Or, alternatively, do we want to introduce new terms into the discourse that we think have two advantages: (1) they better capture precisely what we mean, and (2) they’re unlikely to be confused by being assimilated into the old uses of the term? We need to make a call on that. I don’t have any particular view standing here today as to which way to go on that. But we just need to be conscious that we need to make that decision.

PROFESSOR CHARO: First, I feel like the discussions around our characterization of the situations have been particularly heated, in part because many of us are guessing what the substantive rules are going to be for each category. And so by manipulating the nature of the categories, we feel like we might be manipulating the outcome of the rules. If we imagined that a sample that’s understood to be used in an "unidentifiable" fashion will be free of most of the kinds of regulatory constraints that flow from an IRB, then we have a big investment in what goes into the category of "unidentifiable."

I would like to urge that we keep these things very separate because with this set of definitions, which I find very useable, we can fiddle on the exact descriptions within them, but I find very useable, we don’t have anything in there yet about the manner of the use by the various kinds of investigators in terms of both identifiability and the actual risk that it poses to the individuals from whom the samples were taken, since identifiability is not going to be synonymous with risk in all case. And so I would like to urge that we not go back to the effort to define based on use, stick with these, but to be very attentive to the fact that the tentative contents of these boxes is still entirely up for grabs.

We’ve yet to actually even fill in for the clarification of all the way the current regulations operate in each of these boxes. And, in fact there was something that was distributed today that will make that easier to do, to give people a starting point for what exists and what would have to be changed, added to, deleted, etc. But that there’s plenty of room for blood to flow over the actual rules that will apply without having it to flow over these definitions.

DR. SHAPIRO: Alex, then Bernie.

PROFESSOR CAPRON: I agree with Alta. I was going to suggest that we consider recognizing that we are most concerned about—I do think there is something to Zeke’s point that Carol is pressing, is samples as they are used in research and the samples can be unidentifiable either because they were collected without personal identifiers, or because they have been provided to the researcher without any coding that links them to the identifiers in the storage base. But it may be advantageous to say that there are those two differences.

I want to urge that we drop the second sentences from the first four definitions. These I do not believe they are being used in definitive fashion. It would be possible to say that’s an implication and that belongs in commentary on the text as a reason for distinguishing or as a consequence of having this category. But if people are agreeable, I would like to urge that we do that.

I would also suggest an alternative wording, as I am concerned about the kinds of things that David has raised, for the "unidentifiable sample." If I can read this and if it makes sense, then I’ll get it typed out and provide it to Kathi or whatever. "’Unidentifiable sample’ means that no personal information is attached to the sample through which it can be linked to the person from whom it was obtained." And I say that because the ultimate point that was made to us is if you do a genetic test on the person, and on the sample, after the fact, you could, in fact, say, oh, this turns out to have been his sample, but you didn’t know it at the time. But no personal information as opposed to the biological information is attached to the sample. If that’s agreeable, I’ll provide it.

DR. SHAPIRO: My understanding of what you said, Alex, it is perfectly consistent with what’s here and is a little clearer. I agree with you.

PROFESSOR CAPRON: It was just meant as a clarification, not a substantive change.

DR. SHAPIRO: I agree with you.

We have a lot of people who want to say something and I’m having a hard time restraining myself. But David, you were the next, then Bernie.

DR. COX: I just wanted to address a point that Tom made with respect to language. I really applaud that statement. I think that we can use language in a very powerful way to help clarify the situation but that we have to be really careful that it clarifies and doesn’t leave opportunity for people to use it in an ambiguous fashion. That’s what I’m most concerned about. And so I’m not wedded at all to the terms that people have used in the past. But my point is, and I actually completely concur with the direction that Alex just went, if we can make these definitions sort of bulletproof, then it is very difficult to have those definitions be misinterpreted, and then follow with what Alta said in terms of how we use those definitions in a creative way.

DR. SHAPIRO: Thank you.

Bernie?

DR. LO: I agree that it’s really important we get definitions that are clear and make sense. I have a couple of suggestions with the first four. One, it seems to me the issues that are pertinent to whether a given sample falls into one of these four categories are, first, how is it stored, with or without identifiers; how is it passed on to the researcher, with or without identifiers and what kind of identifiers; and third, if it’s in some sort of code, what kind of code, who holds the key to it, and how easy is it for the researcher to crack.

I think the third and fourth, we’re trying to draw a bright line and it’s really a matter of probability. That if you really work hard enough and ask enough people, you probably can crack something in three, but it’s going to be a lot more trouble in four. So I think that it’s almost I see sort of a grid with several, that these are the rows and columns of yes, noes.

And, finally, I’m concerned about—I’m not sure why we’re saying "traceable" and "identifiable". Are we meaning them to be different, or do we really mean coded but identifiable with a lot of effort but not easily on the part of the researcher? So if we’re going to use two different terms, I’d like to see some clarification of why we say identifiable in some and traceable in others.

DR. SHAPIRO: Okay. I’ve got a number of people here.

Carol, first.

DR. MURRAY: Could I just say I don’t think we should be wedded to the terms "coded" and "traceable" here. They’re Zeke’s candidates to express what we want to express but we should use the term that strikes us as most effective.

DR. LO: Right. I was just saying that to use both traceable and identifiable without explaining creates confusion for me.

DR. SHAPIRO: Carol?

DR. GREIDER: I just want to respond to the idea, getting back to this sort of meat issue about whether we talk about the sample or whether we talk about how the sample is used. I want to assure my colleagues that I don’t have any particular agenda here that I’m pushing, and I haven’t decided that there’s a particular box things should go in, or that sort of issue. It really is just an issue about trying to be as honest as I possibly can about what we’re really talking about.

When I look at this table I ask the question, is this really a sample we’re talking about or is it how we use the sample? And it’s really that simple a question that is my concern. I’m generally happy with the four areas and I think we made a lot of progress at the last meeting, and I agree that we had muddled some issues that we unmuddled at the last meeting by making these four categories. But I still come down to it and I look at this number three, "coded but untraceable sample," that doesn’t really refer to a sample, it refers to how you’re using the sample. And so I just have a very hard time thinking about something that exists physically versus how you’re using it, and that’s my only agenda that I have.

DR. SHAPIRO: We do need to talk this out. I do think there is a distinction and I do think, my own sense of this, Carol, is that the distinction Alta made is actually very helpful. It may require that we change this table in some way so the table wouldn’t look the same any longer, that’s a discussion you have to have, but I think it really is useful. I think we have to get the definitions right. A number of suggestions were made. Bernie’s suggestion that maybe we don’t need two words, maybe one will do; that is, we can use "identifiable" and "unidentifiable" rather than "unidentifiable," "untraceable," as if this were a new concept being introduced. I think that maybe helped. We have to work on other things in these definitions. But I do think that distinction is actually a helpful one that moves us forward.

DR. DUMAS: The distinction between how the sample is used and what we mean by certain terms that we use?

DR. SHAPIRO: That’s right.

DR. DUMAS: And I think we need to keep those two things separate.

DR. SHAPIRO: That’s what I understood to be Alta’s point.

DR. DUMAS: Okay.

PROFESSOR CAPRON: Excuse me. Could I add in response. Would it be helpful to say that if we are going to look at this as our major concern as the sample in the hands of the researcher, then a sample can be unidentifiable because there was no personal information attached to it, or because it has been provided in a way which is coded but not linked with or traceable to the original. In other words, to recognize that you come into the point of research from two different ways. Alta still says no.

PROFESSOR CHARO: I’m shaking my head because what happened in—

PROFESSOR CAPRON: I know, this is the—

DR. SHAPIRO: Larry had his hand up first.

PROFESSOR CAPRON: That’s absolutely right. I apologize, Larry. You’re right. I apologize.

DR. MIIKE: Maybe it’s the advantage of me being at the last meeting by phone with a terrible connection—But I agree with Carol. I’ll tell you why. Alex says changing of the definition of number one makes all of these compatible with how it’s used. And it really is how it’s used. Because you may have a repository but you’re going to take a little bit of that and you’re going to code it or you’re going to make it unidentifiable, and that’s what we’re worried about, that little piece, there is or is not a connection to the original repository.

So it is the way in which it is used. That is, once you change this little front definition that means the samples were collected, we accept Alex’s definition, all of these define the use. I think it’s more important to do it that way than to talk about how it is originally stored.

DR. SHAPIRO: I’m not sure that I’ve understood what you’ve said. I really think that the issue, to the extent that I’ve understood it, is that we can describe the state of the sample without knowing how someone is going to use it.

DR. MIIKE: No, no, no. If you look at any of these "coded but untraceable" or "coded but traceable," that is after you have applied something to the samples for the researchers to use. It is not sitting there coded but untraceable or coded but traceable; it’s the application to the stored sample that makes them then fall into these categories.

DR. SHAPIRO: Alta?

PROFESSOR CHARO: By the way, I want to reserve at least a moment to talk about the other definitions on review and consent. I feel like another source of the debate here may stem from the fact that we’re dealing with two possessors of a piece of tissue—the repository and the investigator—in what is now I think the paradigm we’re all holding in our heads, which is of a repository that sends samples to investigators. And to complete the paradigm, I think we’re imagining a sample of tissue large enough that they can cut off slices and send individual portions. That’s one. Let me just hold that one in my head for the moment for the purpose of discussion, okay.

So that when the repository sends something to an investigator, and, for example, may send it to an investigator with no personal information attached to it and with no code attached to the slice being sent to the investigator that lets the investigator individually, with the cooperation of the repository, recreate the links from that slice, to the larger sample, to the original donor, we’re trying to capture in the definitions the state of identifiability here from both the perspective of the repository and the perspective of the investigator. And I think it’s the effort to capture both simultaneously that is leading to such debate.

It’s possible that where Carol and maybe Larry, because I also wasn’t quite sure, but where Carol is coming from is that the sample can be identified at the moment at which it is in the investigator’s hands as identified or unidentified. It’s got a name, it’s got a code, it’s got nothing. And the status of the slice that the investigator has might be different than the status of the larger block of tissue back in the repository, right?

Dr. MURRAY: That’s the key definition.

PROFESSOR CHARO: That’s key. And maybe this is where you’re calling it the use of the investigator. But, in fact, it’s actually the status of the slice. It’s also the status of the sample. It is not a use. It’s the sample itself. You’ve now made two samples out of one and they’ve got different identifiability status.

DR. MURRAY: We’re all there. All right.

PROFESSOR CHARO: I also feel like I was the last one to figure it out. But what I’d like to end with though is that I don’t believe that we can fully appreciate the parameters of risk that are created and the parameters of benefit that are created by virtue of degrees of identifiability and degrees of traceability unless we constantly see these things in the context of the repository and the investigator simultaneously. I don’t think that talking about the sample at the moment that it’s in the investigator’s hands without reference to the larger block of tissue back in the repository can let us fully understand the significance of that sample in the investigator’s hands, because of the example that was raised last time of the investigator has 100 samples that are totally unidentified to him or her but if all 100 universally show the BRCA1 gene, the investigator could, in fact, let the repository know you gave me 100 BRCA1 gene samples, and the repository, by virtue of the fact that there really is still information left, could, in fact, identify— I defined the paradigm. I’m only working on one hypothetical at a time. Could we go back to those original 100 people. And that’s why I think that it is still worth the effort to try to capture in our definitions the collective effect of these differing identifiability statuses, of not uses but of configurations of tissue and bits and pieces of the tissue which collectively represent what you’re calling a use.

DR. SHAPIRO: Okay. Tom, Carol, then Bernie.

DR. MURRAY: Alta, I want to affirm what you just said and say that that is why when I talk about tissue for which identity might be recoverable, there are two ways. One is that there exists some kind of intermediary code, and the second is context. And your example of all the 100 samples out of a bank of whatever, that’s a case in which the context might make some sort of identifiable.

PROFESSOR CHARO: Right. But here would come under what’s category two, because it came to the investigator without a code and, under the current regulations, would be considered exempt from IRB review.

DR. MURRAY: That may be true.

PROFESSOR CHARO: Right? And that’s exactly why it’s worth keeping the category separate, because we may want to revisit is the current regulatory status of this particular phenomenon adequate. And it may be that it is because these situations will be rare enough that it’s not worth worrying about.

DR. MURRAY: But in a sample where there were a 200 tissue sample repository and 100 went forward and they all had the mutations in BRCA1, then the context would say I suppose you have a sufficient—well, I don’t know, actually. I want to think about that a little bit more.

DR. SHAPIRO: Carol?

DR. GREIDER: So I agree with everything that you just said, Alta. The reason that I was focusing on the second half, if you have how they are in the repository and how they are in the investigator’s hands, was because that’s going all the cases that you’re going to look at where you’re going to try and be protecting people, what the IRB is going to be looking at is going to be that group. But I don’t have anything against trying to broaden our definitions and include some kind of a definition that would take both of those into account. But I don’t think that just dividing these into the four definitions that we have does that. I think that we get back into a muddled situation in this using these definitions here.

So if we can come up with additional definitions or additional language that does incorporate both of those things, that’s fine. I’m not particularly wedded to the use of terminology. But I just have a very hard time sitting here imagining a particular sample in somebody’s hands being considered identified or nonidentified when, in fact, it is identified back somewhere else. All I am saying is that I don’t think that all of our definitions take both things into account simultaneously.

DR. SHAPIRO: Let me ask a question here, and I do have others, Bernie, then Eric. If you look back, Alta, you’re using your structure, the initial primordial description of the sample that exists in some bank somewhere or will exist in some bank, they’re either identified or they’re not identified, right. They either have identifying information or they don’t. There are no other possibilities here the way that’s defined.

PROFESSOR CHARO: Right.

DR. SHAPIRO: If they are unidentified, then they have no possibilities but to remain unidentified. And no matter how you use it, when you use it, what it’s for, you just have unidentified samples passed from repository to investigators and you have a series of decisions to make of what the appropriate use of that is.

If it’s identified, falling in the second category, then there are some alternatives regarding what Bernie would call probabilities, that get set up here, how hard it is. They get passed on to researchers and then the question is, what kind of wall do you build between the researchers in the first stage and second stage of this? Does that capture what you said?

PROFESSOR CHARO: Yes. That absolutely could be done and one could describe the various ways in which tissue becomes identifiable through demographic information, through name and address, etc.

DR. SHAPIRO: Right. It’s not only name and address, there’s other kinds of things.

PROFESSOR CHARO: And all of these things would then be associated with probabilities that one could actually get back to an individual.

DR. GREIDER: So there are two kinds of tissue—identified and not identified.

PROFESSOR CAPRON: At the first stage.

DR. SHAPIRO: Repository tissues are of two types. The research tissues are of four types.

PROFESSOR CAPRON: Correct.

DR. CASSELL: Make the distinction.

PROFESSOR CAPRON: That’s right.

DR. SHAPIRO: What I was going to suggest is that if you sort of look at this table, we’re not at the table yet, I apologize, Kathi, for skipping off the definitions for a second, the first one is "identifiable." You’ve thought of another column in here, right, which is sort of the repository stage or something.

DR. SHAPIRO: You would have "identified," "unidentified," only two. Then the "unidentified" one would break into a series of things.

DR. GREIDER: No, the "identified" one.

DR. SHAPIRO: Bernie?

DR. LO: We’re reinventing history here, because the old, old terminology was "anonymous samples" which in the repository were not identified and, therefore, never could be identified except for going back to the DNA. Then to follow Harold’s model, if it’s identifiable in the repository, you still have a bunch of options. You could totally strip it, anonymize it so that the researcher could never go backwards. You could code it, and you could code it in two ways, either a way that makes it very, very difficult for the researcher to trace it backward, although the keeper in the repository might be able to go forward, or you could code it in a way where the researcher in a sense makes up the code. Or you can keep it completely identifiable all the way through.

DR. LO: But the old-fashioned language of anonymous versus anonymized really tried to capture the duality of what we now have in our unidentified samples, which speaks to what Alta and Harold picked up on. So I don’t know if we’re going back. I think we all want to do something new and throw out what everyone else has said before, but why?

DR. LO: The older language may actually be better.

DR. SHAPIRO: I want a rescue, if possible. Eric?

DR. CASSELL: Just picking up on what you said and what I was thinking as Alta was talking is in the report you make it clear that the repository has responsibilities, that they have ethical responsibilities in regard to the tissues in their keeping. I just got an e-mail about Cornell’s repository in which they say, "Provide researchers with a reliable, ample supply of procured human tissue and all the data for interpretation and experimental--blah--but protecting the patient’s right to privacy and confidentiality." And if we just classify that’s their responsibility, then after that you’ve divided it up and then you’re off and running with the other stuff. But let’s make it clear there are two different things.

DR. SHAPIRO: Kathi?

DR. HANNA: I just want to draw everybody’s attention to the two paragraphs on page 5-4, lines 11 through 25. If you would take a look at that tonight and let me know specifically—

DR. HANNA: This is on page 5-4, lines 11 through 25. This is where I had written it, in a simple way, samples fall into two categories, identifiable and unidentifiable, and then I talk about how within the identifiable category—I’m sorry, you have to continue on to the next page—there are three subcategories. So if you would take a look at that and see if that makes sense. We can always change the words but I think what we’ve just talked about is there.

And I want to just make the point that for the person who donated the tissue, whether it is in category two, three, or four might not make any difference to them. So I just think we keep talking about the pathologist or the investigator, there is a third party. In the first category, obviously they can’t be identified; so we don’t know who they are, they don’t know that research is being done. In categories two, three, and four, when you get to discussions about protections, the third party who is the patient or the research participant is a third party. And so I’m just reminding you.

DR. MURRAY: Do you think two is equivalent to four from the patient’s point of view?

DR. HANNA: No, no. I’m just saying that if it’s identifiable at first, the sample can become two, three, or four depending on how it’s used in research. So when you talk about protections, you’re trying to protect, I’m assuming, the donor of the sample, correct? So, I’m just reminding you.

DR. MURRAY: Right. And virtually all the samples we’ve identified are in their original state they are identifiable. Category one, tissues in their original state which are unidentifiable, apparently are rather rare.

DR. HANNA: Right. I’m just saying that for two, three, or four, you can have one sample that can fall into any of those categories depending on how you use it in the research.

DR. SHAPIRO: Excuse me. Bette, I’m sorry.

MS. KRAMER: Mine is a point of information more than a comment. Going back over the months of discussing these definitions, I think we were at this point several meetings ago when David, and correct me if I’m wrong, you made a fairly strong statement that in your opinion any sample that was coded in such a way that it was untraceable, that’s what we were then calling it, could, in fact, be traced, that a researcher hot on the trail of something would trace it. And you raised that caveat for us and I think that’s when we began to reexamine some of these definitions.

Do you find these definitions acceptable, or do you have that same concern?

DR. COX: Absolutely. In fact, starting off by talking about these as being—the way Harold and the way basically everyone has been saying it, it’s not just one person, is that you have samples that are either unidentifiable or they are identifiable from the point of view of the repository, from the big slab of tissue, right? I mean, that’s the viewpoint that I’m looking at this from, not from the slice, but from where the slice comes. Then, if it’s not identifiable, just the way Harold said it, you can’t identify it. The person doesn’t know it’s theirs, you don’t, nobody knows it’s theirs. And as Tom correctly states, that’s a small subset of all the things we’re talking about.

Then we have this other stuff that is identifiable and then you can do different stuff with that. I think that’s where we can be creative and that’s where all the discussion is. But I’m looking at it from that perspective. And if we go from those definitions, then I’m a really happy camper. But what we do with that, that’s now what we....So I think it’s a three-way street here.

DR. MIIKE: Harold, can I ask David a follow up question. Given that you have unidentifiable/identifiable to begin with, and then we have these three ways of generally distinguishing the identifiable, following up Betty’s question, do you feel comfortable treating those differently, or do you think they are all going to be identifiable in one way or another?

DR. COX: That’s what I’d like to have most of our discussion about. I don’t have a firm position about that, but I think that’s where the meat of the discussion is and that what we can’t do is not have that discussion by already putting things in one category or another. That’s what I’m so heated about because I want to make sure that we have that discussion. And if we already preclude the discussion by how we define things, then I’m very concerned about that.

DR. SHAPIRO: Carol, then I can’t restrain myself from making another comment. But, anyhow, Carol?

DR. GREIDER: I really like the way that Harold phrased this in terms of there being two categories, "identifiable" and "nonidentifiable," and then subcategories off of the identifiable. I think that will get us all in the same camp. And if we’re all in agreement, I mean if David and I are in agreement here, then we’ve made huge progress.

Given that then, getting back to the definitions, I would suggest that the first definition could be "unidentifiable sample" and then all three of the following ones should start with the term "identifiable." Perhaps my suggestion for number two would be "identifiable, but used in"—blah, I’m not sure of the terminology here. And the third one would be "identifiable, but used"—this way. And the fourth one would be "identifiable, but used"—this way. That’s a suggestion. But that all three of those other ones should start with the term "identifiable" if that’s what they actually are.

DR. SHAPIRO: I’m not sure what I would say either takes us backwards or forwards, but I’ll give it a shot anyway.

DR. SHAPIRO: It probably doesn’t move us one bit. I agree, we’ve got these two categories initially, they’re broken down. I think the identified sample really, as it exists in the repository—that it’s identifiable, it can then be passed forward in actually four different categories, not three. Because it could be passed forward anonymously; it could be anonymized. Passed forward with no information.

DR. SHAPIRO: Okay. It can be passed forward, all of these are passed forward with various levels of information.

DR. GREIDER: But there’s disagreement about that.

DR. SHAPIRO: Okay. Just a minute. The issue is the information content with what you pass forward.

DR. SHAPIRO: And they differ by the level of information content. And that makes it either harder or less hard, as Bernie said, to get your way back. It all depends on the amount of content. Now some can be passed forward from its identified sample with no information content; that is, no information that enables any further identification. And then it is a question of how many different categories of information you want to identify here for purposes of operationalizing a set of rules. You’ve got at least two; that is, you could pass them forward completely identified, you can pass them forward completely unidentified. And then the question is do you want any further distinctions regarding the level of information and what that means.

PROFESSOR CAPRON: I lost how that’s four rather than three.

DR. SHAPIRO: It’s just a question of the amount of information. It may be some are passed without information, some with full information, that’s two.

DR. DUMAS: Some without information but the information can be obtained.

DR. SHAPIRO: That’s right, that’s the third.

PROFESSOR CAPRON Let me understand then. If the request is for samples from people who have had breast cancer, biopsies from people who have had mastectomies for breast cancer, that’s with information automatically, as opposed to just send me any ten tissue samples you have.

DR. SHAPIRO: And I think for me the hard question I can’t quite get my head around yet is how many distinctions in there are worth making for generating guidelines regarding use. I don’t have an answer to that. I just have a question in my head but I don’t have a way of dealing with it yet.

Tom, then Bernie.

DR. MURRAY: I may do this more than once in our conversation about tissue samples, which is to just try to state as concretely and succinctly as I can advances that I think we’ve made in our understanding. Alex I thought offered something very helpful when he said, look, when it comes to thinking about "identifiable," we’re talking about personal information that may accompany the sample. I think that’s an advance. It gets rid of the confusion with genetic fingerprinting and such. That’s good. Everybody I hope agrees with that, that we should use that as a sort of key phrase in our various definitions.

Alta I thought had two very important points that I want to stay with and try to articulate, if I can read my handwriting. Number one, that I think she has shown us that even if one guarantees that no individually identifiable information goes forward and there’s not even a code a links back the samples, that there are at least some circumstances imaginable, whether or not they’d actually ever happen, under which one can imagine an individual actually having some risk of harm from having their samples forwarded; her example of 100 samples going forward, all of them having the BRCA1 mutation, and we know the 200 people who were in that bank, that means there’s a 0.5 probability of any individual who contributed tissue to that bank has the BRCA1 mutation. But I just want to make the point that there’s a distinction to be made here. The distinction is between figuring out what category to put it in and just descriptively in terms of is it identifiable or not. That does not exhaust the possibilities for being at risk. That’s the distinction in there. Okay?

DR. MURRAY: So I would say that in the case that Alta described, that person is not identifiable and their tissues are not individually identifiable. That does not mean that they are therefore protected from all possible or imaginable forms of risk. In practice, they might be, but that’s a sort of question one might pose to an IRB, to say, look, does this kind of guarantee of lack of identification actually provide a guarantee, provide protection against imaginable risk.

The other point that Alta made that I thought was very important, one that I’m going to try to remember, and that is as we just discuss the categories that are meaningful, let’s try to avoid immediately reading into them our policy recommendations, because then we’re going to want to jiggle with the categories to make them follow. Let’s get the categories straight. I think these are good categories. I like the categories. I’d like us to work with them, realizing that within them we may make some varying recommendations, including there might be two different sets of samples belonging in the same category, we might have somewhat different advice.

The last point I think was progress was Harold’s point, who phrased succinctly what many of us have been sort of trying to say, and that is you’ve got to think the samples as they exist in the repository--and I think, Kathi, for all of us, we’re going to need a definition of a human biological material. And I’m going to suggest, at least tentatively, that we sort of use that and define that, sort of make that our keystone for what is held in the repository, that’s the human biological material, and that we try to reserve the word "sample" for its use by researchers. At least maybe we need the phrase "sample as used by a researcher," or "sample as forwarded to a researcher."

So there’s two of the former, right, to either have that personal information or they don’t, and then there are four of the latter. I hope we can all agree on those. I think those conceptual tools actually take us quite a bit forward.

Tom, it seemed to me that Harold had introduced a complication, however. I’m just not sure what we do with the four. I think our definitions have to do exactly what you say, we have to have the human biological material referring to repositories, and we have our two definitions of identifiable and unidentifiable, and then we say as used in research, material passed on as a sample can be of several different types.

The complication that he introduced was, we can talk about linked codes, unlinked codes, and identifiable or unidentifiable, but his suggestion is that sometimes it’s not the coding information that’s important, it’s the biological information that’s provided along which is a form of identification because it narrows down the group from whom the tissue could have come.

My concern is not with Machiavellian people trying like investigative reporters or something trying to ferret out where did this semen come from. The question in all of these cases is for what may in a research context seem like good and sufficient reasons, are people in a position to link research findings with an individual or a small group of individuals who have a high probability of being the person identified where that information may be unwanted or harmful in some way and at least wasn’t something that the individual may know has been found out about them. Those are the kinds of concerns that arise.

And so, if it is information, not just the code—in other words, you’ve said I want these people like this, and they provide it with codes which are not linked to their samples, they couldn’t tell you, number one, that I gave you is Mr. Jones, number 475 in my repository, but the ten samples, once you find out about them, if you came back and said we’ve just got to find these people because they’ve got something that’s so urgent, they could say, well, at least I can narrow it down because I do know from the information that you asked me, the characteristics, that it must have been these ten people. That’s an example of something that is unlinked, an unlinked code but it’s still traceable.

DR. MURRAY: Well, that would be then number three. That would be identity recoverable.

DR. CASSELL: Potentially recoverable.

DR. MURRAY: Potentially recoverable.

PROFESSOR CAPRON: I just hope we can capture what Harold has brought out because—

DR. SHAPIRO: Exactly that. I was worried about biological information as opposed to the name, address, or that kind of personal information. And it may be an unnecessary complication.

PROFESSOR CAPRON: No, I have a sense that it’s probably the very complication that is most likely if the researcher has narrowed down her research to certain characteristics.

DR. SHAPIRO: Okay. Bernie, then David.

DR. LO: Let me try and follow up, Harold, on your suggestion that if we look at tissues that are identifiable in the repository, there are several different ways they can be passed on to a researcher as samples. I’m just pointing out we have two extremes. One is where all the information is passed on, and one where none of the information is passed on. Now Hertz has a very good ad that says there’s Hertz and there’s not exactly. So passing on all the information and not exactly passing on all the information but coming pretty close.

And that’s what typically is done in research is you never use, the IRBs don’t let you use the patient’s name on all the documents. You’ve got to give them a code. But I may have a code that 001 is me, 002 is Lori, 003 is Eric, and it’s transparent, it’s easy to crack. But it protects against just having papers lying on your desk and people peeking. But for all intents and purposes, it is very transparent, the code. It seems to me those two we should put at one extreme saying it’s really going to be awfully easy for anybody in an investigator shop to crack the code.

Then you have the other extreme, samples where, short of doing investigative reporting or a full day in a file, you’re not going to be able to trace back.

Then in the middle are samples that are coded in ways that don’t make it impossible to ever trace that but make it a lot tougher than is the case with ordinary coding as it’s understood in the research enterprise. And sometimes you can crack the code because this elaborate encryption scheme with the matching pass words to someone in Antarctica or something, you can go to Antarctica and bribe the guy and get the code. Or, you may be able to deduce it through this kind of elaborate piecing of information the way Alex characterizes investigative reporters.

The point for our purposes is it takes a lot of effort, it’s not something I’d just pick up the phone and call someone and say "Hey, tell me who 003 is because I need to get more information." There’s an inherent check in the process of getting more information because it’s just much more complicated and more difficult, but I still can do it.

Maybe those are sort of the gross distinctions we can make and just realize there are lots of different kinds of encryption that fit in that intermediate category. And we actually, frankly, haven’t really thought through how you might construct the system that has those characteristics. But we would should be able to deduce, given that you have such a sample and coded, what the ethical implications are for the type of research.

DR. SHAPIRO: David, then Alta.

DR. COX: So it’s really following up on what Bernie said. I’m putting myself in the perspective of the repository now with a sample that’s identifiable and thinking about different things I can do with that sample. So one of the things is that I can give it to people so that they will have it be coded but be coded in one of these extreme ways because they never want to have any identifiers, they aren’t interested in it, and that they would just like the sample with information that will not make it identifiable.

I, as the repository, can do that in one of two ways. I can give it in a coded way that there was no possible way that I can even figure out what it was. Now there may be some way that somebody could do it, but my intention is to do it in such way that I don’t know what’s going on. Not only does the people I give it to not know what’s going on, but I don’t know what’s going on. That doesn’t mean that somebody at some time couldn’t go and figure out what’s going on, but it’s going to be really hard.

Let me just say, I used to think that this was a pretty straightforward issue, but what I’ve been convinced is that just the pieces of information that are given out are sometimes more traceable than one might think. So I’d just like to make that caveat. As an example, a piece of information in terms of family structure, if you know it came from a certain city and you can get from that city the names and addresses of all the people and whose related to who, it’s very easy to identify that even though I may have made it so that I thought it was very non-coded. But my intention was to scramble it so I didn’t know it and the other person didn’t know it.

But then there can be another way you can do it. I can scramble it in a way that the person I’m giving it to doesn’t know it, but I can easily figure it out. And I’d like to make that distinction from the point of view of the repository and that whether I still can figure it out or not, because I think that’s really important from the point of view of everybody—from the point of view of the repository, from the point of view of the researcher, and from the point of view of the patient. So why is it important if I can still figure it out?

Well, from the point of view of the researcher, it’s very important because as a researcher, if I found something really interesting, I know that person in the repository can go back and get it. From the point of view of the patient it’s important because whatever is done on that sample, the repository still knows it’s mine, can figure that out. And from the point of view of the repository it’s in the context of both of those people that I still can figure it out.

So it’s not saying what the risks are, what the harms are, but me, as a person in the repository, that’s a very different scenario for me. And I’d like to make that distinction in terms of how hard it is to do it. Sort of what the intention is. Because I think even if my intention is to make it so that I don’t know and the other person doesn’t know, now we’re talking about situations of, well, how good a job have I done. But it’s what that intention is.

DR. SHAPIRO: Alta?

PROFESSOR CHARO: You know, David, I never like to disagree with you because you’re so much smarter about this stuff, actually. But I fear that in some ways we’re making this too hard because even under current regs in ordinary experiments that don’t involve tissue samples where you’re doing research that involves having peoples’ identities obscured with codes that provide links, a problem emerges about whether or not those codes could be broken, a problem emerges about whether the cell size is small enough that based on demographic information that you are planning to use publicly the identities could be reconstructed. And with regard to biological information, which Harold was mentioning, I remember hearing my colleague, Norm Foss, when we struggled with this at UW say that this is all very temporary because in the future with forensic DNA analysis every sample will be identifiable through direct testing against a national database in which every person’s forensic DNA signature has been entered.

And so I wonder if we are failing to remember that there’s going to be room for an IRB at various stages to accommodate these issues. One could maintain the kind of traditional distinctions among things that have the name and address versus things that have some kind of coded link versus things that have no codes and remember that in the substantive content that we give to the rules that will follow all the various categories there will be a role for the IRB particularly at the repository to make a judgment about whether or not the request for X number of samples meeting certain demographic characteristics poses a threat of a breech of personal identity by virtue of the cell size or the nature of the demographic information.

But we don’t need to necessarily handle all these problems at the level of the definitions, but that they can be left at the level of the substantive protections we build in through the IRBs. And that might simplify our definitional struggles over whether we’re going to define based on whether there’s a code versus defining based on the kind of degrees of difficulty of obtaining personal identity, which to me seems like it’s going to be impossible since it will be a continuous spectrum of ease; it will never have any bright line distinctions.

DR. COX: No, I wasn’t suggesting that be in the definition. I’m completely with you.

PROFESSOR CHARO: Oh, good. So I’m not disagreeing with you. I feel safer.

DR. SHAPIRO: Thank you. Tom?

DR. MURRAY: As I sit here today, once again, I am very, very impressed by how smart my fellow Commissioners are—

DR. CASSELL: And stubborn.

DR. MURRAY: No. I wasn’t even thinking that, Eric. At how wonderful they are at thinking of distinctions and complexities, of which we need to be mindful. But I had come I guess to a similar conclusion as Alta; namely that we shouldn’t try to cram it all into our categories. To my view, the categories are just a kind of tool to let us take the next step, which is to figure out what the recommendations and policies, including IRB things, ought to be. Whatever set of categories we end up with, I think these actually work fine, have to be able to do two things for us. Number one, they have to be sensible in the sense that we sort of know where to put things. Give me an example, can I fit it into one of these four, more or less unambiguously, can I figure out which it goes into? I think these make the test about as well as any set we’re going to come up with.

So then you have the second sort of criteria by which to judge the categories, and that is do they provide a useful first step at organizing our intuitions about what policy implications ought to be and implications for IRB. Not do they exhaust it, because they clearly won’t, but just do they provide us a useful kind of organizing way of getting to make the next step, which is going to be tougher actually.

Those are the only two criteria I think by which to judge the categories. I happen to think these work well. I’m going to ask you to decide whether you want to do that or not tomorrow morning. We really should make a call because I think we need to.

And I want to actually, Harold gave me a kind of silent consent a minute ago, assign a little homework. Kathi asked us to read just about one page. That won’t take long; I hope everybody can do that assignment. Alta had distributed a memo from Gary Ellis which also won’t take more than a few minutes I think to review, just to give us a sense of—it is dated 17 April 1996 to OPRR staff from G.B. Ellis. It’s a frequently asked questions sheet here with responses. That’s going to be quick.

PROFESSOR CHARO: With little pictures.

DR. MURRAY: Yes. And I think we actually need graphics to make some of the points. I’m wanting to think in terms of pictures. I think that will help when we actually do the report. The third homework assignment is a little tougher. I hope you’ve had a chance to read Alan Buchanan’s excellent essay because Alan really cuts through a lot of jargon and gets us to think about what actual interests are implicated on both sides in these various decisions we’re going to make. Even if you just skim Alan’s paper to get a sense, this will help us tomorrow because I think we are tripping ourselves up on issues which are not irrelevant but may not be at the heart of it. And so we need to focus on what specific interests we think we need to be protecting that are implicated in certain of these situations, what interests in terms of research need to be respected and advanced, and how are we going to weigh them. So, if you’ll accept them, those are my suggested homework assignments for tonight.

DR. SHAPIRO: I like them. Larry?

PROFESSOR CAPRON: I understood implicit in Tom’s comments just now was that we are not going to spend further time talking now about the first four definitions. I wanted to ask a question of the drafters, however, about the last definition. It says "general consent means consent to future unspecified research use of the sample." Now, "benefits and risks are specified regarding future research use but not regarding any specific protocol." I don’t quite get that.

If it means the risk of being in a research use, of having a sample used is specified, that seems to me to be already stated in the first sentence. If it means something more than that, I thought the whole message we got from researchers was until you knew what the research was you really couldn’t say anything about what the benefits and risks would be. Having your tissue used for one test might have no research risk, no personal risk, having it used for something else could be very risky. So I don’t know what that sentence means.

DR. SHAPIRO: I have a comment on that, but Kathi?

DR. HANNA: Just my response is the last three definitions are the definitions that have appeared on the boxes since September. So I’m glad we’re finally getting around to talking about them. But these are unchanged from what were on Zeke’s boxes.

PROFESSOR CAPRON: Right. What I’m asking now is, what do we think we’re saying here? This isn’t criticism of a drafter. Now that you hold them out here very usefully, we stop and look at them, when I look at this one I don’t understand what that second sentence means.

DR. SHAPIRO: My own view of that is, and it was a point I guess made by Alan Buchanan that actually makes that in his paper, the second sentence is not operational in any real sense. So I think it could be dropped, is my own view, as Alan suggests.

DR. MURRAY: What Buchanan writes about—blanket consent—I think is relevant to this.

DR. SHAPIRO: Why don’t we spend a few minutes, we were supposed to adjourn five minutes ago, but I just would like to spend a few minutes looking at those bottom four definitions and see if people have some comments, suggestions, etc.

Bernie?

DR. LO: With regard to the administrative review by the IRB, I would suggest an addition—first of all, it may not be the IRB chair, but somebody delegated by the IRB. It is not just the identifiability status, but also the prospect of community harm which comes into the part of the table that was truncated from the previous page.

DR. SHAPIRO: I have no problem with that, except that I do want to get some discussion, now is not the moment, on how we’re going to deal with this whole idea of community and so on. But I have no objection right now. I understand the point.

Alta?

PROFESSOR CHARO: I think I’ve got more trouble with the two phrases that were recurring from the earlier incarnations, "administrative review by the IRB" and "full review by the IRB." They don’t match anything that exists in the current regulatory language and, as a result, I find, as somebody who works with the current regulations, as most people who work with the new things we recommend will be, that they are tremendously confusing.

The current system is one in which investigators make the first cut at determining whether or not they need to even go to the IRB. And if we want to change that, we can change that, but I don’t think you want to call it "administrative review by an IRB."

The second thing is that when it goes to the IRB a decision is made about whether or not the substance of the research is exempt from IRB review. And that’s what I think Zeke meant by "administrative review by the IRB." That’s not administrative, that’s review by the IRB. And whoever is designated in that institution according to their multiple assurance to make that decision makes that decision. It’s often the administrator rather than the chair.

If it’s not exempt, it then goes to the IRB for review. And this is where I think Zeke was imagining the phrase "full review by the IRB." But that obscures the fact that there are levels of rigor in IRB review, including eligibility for review by only a portion of the IRB. And I don’t think he intended to eliminate all of those various nuances in the concept of IRB review. And by putting in a phrase like "full review," it implies that we’re requiring that the entire committee has to be meeting and reading the protocol together and voting on its acceptability, all of which are regulatory rules for certain categories of research.

I would like to suggest we dump this stuff because it’s more trouble than it’s worth, and focus as we go through tomorrow the actual content of what we want for each of these boxes, on what the current regs say and then identify specifically where we would like to change those regs, and, at that point, see whether a change in vocabulary is necessary or whether existing vocabulary is sufficient to convey our meaning.

PROFESSOR CAPRON: Don’t we know, Alta, already that what’s conveyed by the last two is not captured by the present regulations?

PROFESSOR CHARO: What is conveyed by what last two?

PROFESSOR CAPRON: The full versus—

PROFESSOR CHARO: Oh, I wasn’t even talking about, you mean full consent and general consent?

PROFESSOR CAPRON: Right.

PROFESSOR CHARO: Oh, I didn’t even touch on those. No, I was only talking about administrative review and full review.

PROFESSOR CAPRON: I mean the whole notion of expedited, waived, or so forth.

PROFESSOR CHARO: Right. Those are separate from the concept of consent. I found full consent and general consent also to be confusing and nonintuitive ways of conveying what was being attempted to be conveyed. But I was focusing only in my kind of regulatory obsession over here on the regulatory language.

PROFESSOR CAPRON: Well, the consent language is in the regulations.

PROFESSOR CHARO: That’s right. The word "consent" in the regulations does not track this distinction between full and general in any way.

PROFESSOR CAPRON: That’s right. That was my point.

PROFESSOR CHARO: That is true as well, yes.

DR. SHAPIRO: Carol?

DR. GREIDER: I agree with you, Alta, that I think that what Zeke was trying to get at, back to the first two issues, wasn’t very well fleshed out and we didn’t discuss it a lot in the subcommittee. But I do think that there was something that he was trying to get at, which would be some sort of an additional looking at issues in terms of community and that sort of stuff that isn’t in the current regs that I hope that when we fill in the boxes we will incorporate that and use some other language, with the idea being people haven’t thought about community and, if we are going to bring that in, somebody has to determine whether this researcher has completely ignored a major issue and has decided that there’s no implications of this protocol, and that needs to be reviewed by somebody.

PROFESSOR CHARO: My point, Carol, is simply that somewhere lost in the discussions that led to whatever information was being used to generate these boxes, somewhere lost in there was the degree to which some of these things are already handled with the regs, already require certain things, and knowing that will help us to identify where they need to be added to, changed, or, in some cases, certain things deleted. But at least we’ll be working off of a regulatory framework that exists. When the rules get changed it will be virtue of amendment of those existing regs.

DR. SHAPIRO: I agree with that. If you go at it that way, you come up with words like "expedited" and "waiver" which are already defined and we could kidnap where appropriate and change where inappropriate. And so we’ll need a different way to look at that. And that’s a part of the homework that Tom gave may be helpful in that respect, Alta, or whoever provided those charts on the regulations, that really would be helpful.

PROFESSOR CAPRON: Did you mean kidnap or hijack? They’re both Federal offenses, but one of them is worse than the other.

DR. SHAPIRO: I don’t know which one I would choose then. But, in any case, perhaps we ought to come back to these issues tomorrow, and Tom has assigned some homework. And Kathi, maybe you, Eric, and I can get together to talk a little further about that right now.

ADJOURNMENT, DAY 1- Dr. Harold T. Shapiro DR. SHAPIRO: And we can adjourn the meeting. Thank you very much.

Adjournment, 5:13 p.m.