DR. SHAPIRO: Correct. Thank you very much, Eric. I appreciate the comment. Dr. Shore thank you very much as well. I really appreciate it as well as you’re help in previous Commission meetings. I’d like now to turn to Dr. Rosenstein, who will talk about a subject which can be described in many ways--psychiatric symptom provoking studies is the way it’s listed in your agenda. Dr. Rosenstein, thank you for coming.

DR. ROSENSTEIN: Well, thank you very much for this opportunity. It’s truly a privilege for me to be here to talk with you about psychiatric symptom provoking studies. The reason this research strategy is so interesting and challenging to me is that it kind of commands our attention on the scientific and ethical questions that are really at the heart of research with vulnerable subjects. What justification is there for intentionally producing distress in research subjects? What are our standards for adequate informed consent? How are risk-benefit determinations made? What is the nature of the subject/investigator relationship? Psychiatric symptom provoking studies bring each of these issues into bold relief and I think it provides an opportunity to really understand what the issues are in a way that sometimes gets confused.

My main message today is that like so many of the complex issues that your Commission is considering, judgments about these studies depend critically on why they are done, when they are done, with which subjects, under what circumstances. In short, it’s the context of this research paradigm that is most important in the evaluation of the ethical aspects of them.

Before I get into the body of my talk, what I’d like to is just mention very briefly what my history is with this, how I got interested in the area. And I also want to publicly credit Dr. Frank Miller for--if you’ll excuse the pun--for provoking me into taking a more serious look at the bioethics of this research paradigm. Dr. Miller is a bioethicist and a philosopher, and currently serves on the NIH Clinical Center Ethics Committee and the NIMH IRB with me, and I had a collaboration with him for the last six or seven years. About four years ago he came up to me and wanted to have a discussion about some of the studies that he had seen on the IRB that were troublesome to him. And to make a long story short what he was bothered by was the concept of the person in the white coat taking someone who was sick and making them feel worse. And I think it’s important at the outset to acknowledge that for many people, upon first hearing about these studies, there is a similar emotional reaction to this, and so it’s--I think that there are very legitimate ethical questions and my hope is to put these studies in some kind of perspective and to offer some suggestions about how to evaluate them. I believe that you’ve been given a copy of the manuscript that we wrote, and I believe that that’s why I’m here today.

Let me begin with an overview of what I want to say. First I’ll offer a definition of what psychiatric symptom provoking studies are, and then briefly mention some examples of challenge paradigms in other fields of medicine. I’ll then address the scientific justifications for doing this type of research. Then I want to move on to what I mean by the contextual and timing influences and, in particular, how they impact on the ethical considerations. And then finally I want to end with some specific recommendations to investigators, subjects, and IRBs.

So, one definition of psychiatric symptom provoking studies is as follows: The administration of a psychological, physiological, or pharmacological challenge to elicit psychiatric symptoms for the purpose of identifying their neurobiological causes and consequences. Now, I think it’s particularly important to keep in mind that the point of these studies is to model psychiatric disorders so that we can learn more about them. There is not very much inherent value in demonstrating that you can make someone anxious, make someone sad, make someone psychotic or more psychotic. However, I think there is compelling scientific value in identifying relationships between symptom expression and underlying neurobiology.

A second point I want to underline is...these symptoms are not unintended side effects. I think a lot of people, including investigators, can get confused about that. Any time a drug or other intervention is presented, there’s potential for both intended, usually beneficial effects and unintended adverse effects. In symptom provoking studies, the whole point is to elicit symptoms that are characteristic of the condition under study. In this respect, psychiatric symptom provoking studies really did extend out of a tradition in medicine, I believe, of turning on or stressing the system in order to learn more about that system under dynamic conditions.

Now, I’m aware that this next slide may be controversial because I’ve listed some of the other challenge paradigms in medicine. But I hope you’ll give me an opportunity to at least offer my opinion about the goodness of fit and the limitations of such analogies.

Just very quickly--one you’ve, I’m sure, heard about is cardiac stress testing to uncover angina. We’re all familiar with that. Glucose tolerance testing. Tests of new analgesics. Electrophysiologic stimulation. Instead of putting a sock around the heart and stimulating different areas of the heart in order to identify areas that--where there may be a focus of arrhythmia, and then trying out different antiarrhythmics to try to suppress that focus. Sometimes flashing lights in front of someone with a seizure disorder is used to identify an area in the brain where there’s a seizure focus. Certain kinds of bronchodilators have been used in testing pulmonary function.

Clearly, some of these examples are very well- established, validated, and safe diagnostic tests that are used in clinical practice. I’m not going to suggest that the psychiatric symptom provoking studies are at that point yet. I understand that difference. Some of these have had profound benefits in terms of changing clinical practice. Some have had less--you know, or more modest clinical utility, and some of these are purely investigational. For instance, in the endocrinology world, the standard test to provoke the hypothalamic pituitary adrenal axis is CRH. Well, it looks like in the investigational setting interleuken 6, when administered, is a much more potent stimulus for the HPA. Now, I don’t know whether it’s going to develop into a clinical test or not, but there are certainly other examples in medicine where provocation is used to learn more about the underlying condition, and hopefully we’ll at some point have more direct clinical utility.

So, again, the point is not to represent nontherapeutic psychiatric symptom provoking studies in the same light as validated diagnostic tests. Some may or may not develop into that direction. The point is that this is the way of thinking—again, in the tradition of medicine, where you stimulate a system under dynamic conditions in order to learn more about it. It’s not unique to psychiatry.

I think a reasonable question is: Is there something fundamentally unique about triggering psychic distress, and I think that’s a debatable issue and I would argue that to do some of these studies, this is the only way you could do it, but I think that there are legitimate claims that there’s something a little bit different about the psychiatric distress, although other people would argue that it’s not fundamentally different than making somebody short of breath or having angina or some other symptom expression of what they suffer from. I’m sure there’ll be discussion about that.

For returning to psychiatric challenges, the scientific literature contains a great variety of symptom provoking studies. Some employ purely psychological provocation such as public speaking or mood induction, but clearly the predominant paradigm is to use the medication to induce the symptoms of the disorder under investigation. For instance, intravenous infusion of lactate or inhalation of carbon dioxide has been used to precipitate panic attacks within the clinical research setting. Tryptophan depletion has been used in affective disorders and a number of other conditions. You’ve heard about amphetamine, methylphenidate, ketamine to probe perceptual or cognitive symptoms characteristic of psychosis. Interestingly, scopolamine is an anticholinergic agent that impairs memory and is currently being evaluated for its potential as an early predictor of dementia of the Alzheimer’s type. This is, I think, a good example because--and we now have agents that can be used to delay the onset of the memory impairment in Alzheimer’s disease, and I think that it would be wonderful to know who is at risk early on so that interventions could be started sooner rather than later.

I think it might be helpful to go over in some detail one symptom provoking paradigm, and that’s the tryptophan depletion study. In this paradigm, which has been used in a number of different conditions, mostly in major depression, the basic procedure is that if someone presents with a major depression and then is treated successfully with a selective serotonin reuptake inhibitor, a type of antidepressant, SSRIs --and I name that because the proposed mechanism of action is thought to be modifying availability of the neurotransmitter serotonin in the central nervous system. Patients will--subjects will come in and be treated with an SSRI, and if there’s a resolution of their symptoms then under double-blinded conditions with a placebo-controlled paradigm, an amino acid drink is given. Either a placebo, which is in the red line--it’s kind of hard to see where it doesn’t change tryptophan levels—or a specific amino drink which is designed to acutely lower tryptophan in the blood and hence in the CSF. What the point is, is to see whether there is a time-limited and brief-I’ll talk about the scale in a minute--increase in depressive symptoms that then might be correlated with a variety of different neurologic measures, either hormones that are measured or perhaps some kind of brain metabolism scan.

This is the Hamilton depression rating scale, and just to put this in perspective, you know, numbers in the teens are usually kind of consonant with what we might see in a mild to moderate depression. I should mention that the tryptophan depletion study, at least in depression, seems to be a fairly specific probe in that individuals who have responded to an SSRI are much more likely to have a temporary increase in their symptoms than individuals who have responded to an antidepressant that has a different mechanism of action such as dizipramene, which is a selective inhibitor of norepinephrine.

So, then, the next question is why do a study like this? And I think there are a number of important justifications to keep in mind. One is that one of the best ways to illustrate that a particular biological factor actually causes something is to block it. This paradigm allows in a sense one way to look at whether you can first elicit a symptom and then have a very specific way to block it, and I’ll show you an example of another paradigm in just a moment where that was what was used. The tryptophan depletion paradigm is an example. Again, it’s made tremendous contribution, I think, in our understanding of the surrogate nerve mechanisms and mood disorders as well its implications for treatment. Clearly, if we can identify some sensitive and specific diagnostic tests for a variety of different psychiatric disorders, it would pave the way for intervening much earlier under different treatment approaches. While I will acknowledge that we’re not here yet with any of these psychiatric symptom provoking studies, this is a very reasonable hope that if someone comes to see a psychiatrist right now with a major psychotic episode or major depression, there’s precious little information that we have to decide which of multiple agents are really better for that individual person. The ultimate goal, I think, for many of these approaches is individualized therapeutics to predict who’s likely to respond and who’s not and whether there are underlying biologic predictors of different response or even drug toxicity.

Let me give you an example of one challenge paradigm that at least starts to approach this condition and I’ll try to run through this quickly. What you have here is a scale turned upside down of sadness, and these are self-ratings of--mood ratings of a woman who suffered from severe premenstrual syndrome. Every month just prior to and through her menses she had profound sadness with suicidal thinking, and this was predictable every month, right here. She was enrolled in a study at the Clinical Center in which she received--it was a clinical trial where received, in this bar here, lupron, which is a synthetic GRH analog. It essentially shuts down ovarian functions temporarily, and what you see then is that in the period after she received the lupron her symptoms went away. And, here’s where the provocation comes in. Under double-blind conditions, she received--and this was randomized--either estrogen or progesterone, and what you can see is an exacerbation of her symptoms. Again, this is under carefully monitored clinical setting. There was a brief increase of symptoms here--less so with progesterone. Clearly, the study was not designed as the diagnostic test for her; it was designed to learn about the role of reproductive hormones in a well-characterized patient population under carefully controlled and monitored conditions. Nonetheless, what happened with this individual was there were recommendations at the end of her participation to her gynecologist with respect to what reproductive oral contraceptive agent we tried and which concentrations of estrogen and progesterone. For other women who didn’t have such a clear response to suppression of the ovarian cycle, there have been a number of women who had previously considered pherectomy as a treatment for this and then no longer considered that. And so at least it’s possible to envision a time when there might be an individualized characterization of someone’s response. Again, we’re not there yet.

So, really, the question I think for this group is: Under what circumstances are psychiatric challenge studies acceptable or not? And, this is where I think context and timing become particularly important.

Some studies simply can’t be done without doing symptom provoking studies. For example, the brain metabolism studies through looking at provocation and kind of the neurobiological effects of different drugs with suspected mechanism of action that are related to the underlying path of physiology require a symptom-free and on-symptom state. You spend a lot of time, I know, considering a clinical state of subjects in clinical research, and this is clearly one of the most important aspects because it goes directly to informed consent. And, it’s obviously one thing to ask someone who’s in remission from a syndrome if they’re willing to do a study that may make them feel worse temporarily and quite another to approach someone who’s in an acutely decompensated state with the same request.

This is something that Dr. Miller and I have been thinking more about lately, which is how are symptom provoking studies as individual protocols connected to a larger research program? I think that there’s a lot of potential for blurring of some of the important distinctions, and I’ll say something more about that in a minute. But the basic idea here is: Are these add-on challenge studies that are being done when someone is going to be in a drug washout phase for other reasons, or are there design aspects of the study that are put in place specifically to allow the possibility of symptom provoking study? I think that the fact that most of these studies can be best characterized as nontherapeutic makes it particularly important to keep the individual protocol distinct from how it fits in with an overall program, which may be fairly characterized as expected-benefit participation.

The last point here is one that I also think hasn’t been given as much attention in the literature as certainly I think I should, and that is: What’s the relationship with the investigator? Is this a longstanding relationship, or is this a brand new and somewhat anonymous relationship? Is the investigator asking someone that they’ve known for many years to do something that they believe is mild and short duration? Or, is this a request that is not taking place within the context of an ongoing relationship?

This notion of the quid pro quo--I hope there’ll be some discussion about this because I think that there is essentially a more or less explicit quid pro quo in much of clinical research and I personally don’t think that there’s anything in and of itself that’s wrong with that. And, I think that--my hope is that increasingly, the quid pro quo can be brought into the light of day and discussed in a more straightforward fashion, and subjects can make a decision about whether they’re willing to do something knowing that it may make them feel worse temporarily and is not going to benefit in exchange for participation in an overall program that may in fact be to their benefit. I’m sure there’ll be discussion about that.

Now, I know that this then kind of leads directly into what the fundamental ethical issues are, and we spend a lot of time talking about decisionmaking capacity in the role of how to assess informed consent. I’m not going to comment on that right now. I do want to mention, though, that it is important to highlight this notion of therapeutic misconception and certainly it is described by Appelbaum and his colleagues. This is a common and significant problem in research subjects. I would just like to make a point here that this is not a one-way street. Investigators are certainly susceptible to a therapeutic misconception, and within the context of research programs, it might be a considered expected benefit. I think it’s possible to blur the distinctions between procedures done with the research study that are intended to benefit a subject vs. those procedures performed solely for the scientific information to be gained. So, I think that when any intervention of no expected benefit to an individual is considered, investigators, subjects, and IRBs must be crystal clear on that point.

With respect to risk-benefit assessment, here’s another area I think that frequently gets blurred, which is, people talk about something being beneficial to society or to the field of medicine or psychiatry as opposed to beneficial to an individual and I think that point is clear enough but I think that the bottom line is that everyone involved in these studies really needs to understand two things: One is that the purpose of the study is not to help; the purpose is to learn more about the underlying condition. The second is that--and this is also different than saying that this study may not be of benefit to you, which is typically how the language reads in a number of different consent forms. There’s different meaning there. And the second thing everyone needs to understand is this: The symptoms are expected; they’re not unintended side effects.

So, let’s then turn to kind of what some of the real risks are and I think that it’s fair to say from the overwhelming majority of published studies of symptom provoking paradigms is that symptoms have been experienced before on a number of occasions by subjects so that I think it’s fair to say that there is less distress associated with the 750^th panic attack than with the first one. Now, that may not always be true, but certainly that’s been my experience as an investigator, that the familiarity somewhat offsets the magnitude of symptoms that are sometimes seen.

Again, the idea here is to model the primary disorder, not to precipitate a full-blown episode of illness or significant clinical decompensation. Nonetheless, there are clearly published reports here of severe and more prolonged reaction. To my knowledge--and I’d be curious if anyone has any other information--I’m not aware of any published reports of kind of catastrophic response to symptom provoking studies in a sense separate from some of the things you’ve heard about with respect to washout studies, but I’ll be anxious to hear what you say about that.

I know that there’s also been some concern about the possible negative effect on the longitudinal course of the primary illness, and I’d have to say that at least to my satisfaction, the data are simply not available. There has been a clinical analog that’s been postulated for kindling sensitization with respect to repeated episodes of major affective disorders or depression. But, again, simply provoking studies are not intended to induce a full-blown major depressive episode or a relapse of a major psychotic episode. These are intended to be temporary and mild, for the most part.

I also just want to mention here that even though the intention is to induce the symptoms, there are unintended benefits that are frequently observed by subjects participating in studies. And, I’ve heard from many participants--for instance with panic disorder—that having a panic attack in a setting of a carefully monitored clinical research situation, is very validating, rather than carrying around a notion for many years that there is a inability to cope with anxiety or fundamental weakness in character. For some people to experience having a panic attack with lactate under blind conditions but not with saline can be reassuring that the fact there may be some underlying neurobiology to this. I won’t belabor some of the indirect benefits that can happen, but I think it’s important to note.

So then, what’s a reasonable approach for evaluating whether a given study is ethically permissible or not. I would argue that the only way we can do this is to look at the overall context of the study and that there are better and worse ways to proceed, and I think that it’s important to look at what happens when clinical research goes well, not just when it goes poorly. And, I’ve a better and worse case scenario here on the left side, on the left and right side. So, obviously, it depends on how things are done. Let’s assume you have a capacitated subject who is approached with consent on the front end of the study who understands this is not therapeutic and likely to induce brief distress. There’s careful clinical observation and follow up. I’m much more comfortable with that than if someone with obviously compromised capacity who is recruited during the middle of a washout or when acutely symptomatic, that if the study is presented as beneficial; if there’s some coercion involved but there’s poor follow up. So it’s really, in my opinion, not a question of the ethics of inducing symptoms but really the overall ethics.

So let me just finish up with some of the recommendations that Dr. Miller and I put in our paper, and kind of just run through these.

I think the first and foremost point I want to make here is that the science must be sound—without valuable information, obviously the protocol shouldn’t go any further.

Here is where design issues come in and are particularly important. I think that whenever possible washout placebo phases should be minimized, but in many people’s minds there are conditions where they’re essential, and when they are necessary for clinical trials when those situations using provocations under drug-free states should be contemplated. I think that there would probably be agreement that performing psychiatric symptom provoking studies, to take someone who’s been stable on medication and then take them off medication for the sole purpose of doing psychiatric symptom provoking studies, would be inappropriate.

The investigators and IRBs need to be more attentive to inclusion and exclusion criteria. Who are the subjects that are going to be recruited for the study? Are these going to be subjects who have a first episode of an illness, or just going to be subjects who are treatment- refractory. Those on meds of those off meds. Subjects with a strong history of suicide or violence should obviously be excluded for high- risk studies unless there are really compelling reasons and added protection. Obviously, decisional capacity is a critical factor here as well as any consideration of whether there would be surrogate decisionmaking in this setting.

A balance needs to be struck here between picking a challenge procedure that is going to give you measurable outcome variables, something that is strong enough to produce symptoms but not so strong as to induce more distress than is necessary, and that I think isn’t in a sense looking at the least effective dose of something that I don’t think IRBs are always thinking about when they review these studies. I’ve already talked about informed consent somewhat.

And, finally I just wanted to say that my feeling is that the standard of clinical care and clinical research just doesn’t have to be up to speed; I think it has to be better. And I think in many of these studies more can be done to provide for added protections, subject monitoring, follow up. And, my hope is that what we can do is we can devise a way to continue in what in my opinion is an important approach to research but in a way that pays attention to the issues and takes what we’re doing right and can do better and minimizes what we could--what we’re falling short of. So, thank you very much.

DR. SHAPIRO: Well, thank you very much. We’re very much appreciative. Eric?

DR. CASSELL: Well, I have a number of comments. The challenge study that you used as examples in managing although short-lived. There are therapeutic actions that can be taken if they get out of control, there all meant to lead to a treatment. It’s an effective treatment, and they are physical symptoms. The symptom that you’re provoking is not quantitatively the same as angina, for example. And if you don’t--if you’ve worked with it so long that you don’t know that anymore, that would be a distressing thing in and of itself. That happens sometimes, but those are different in tone, they’re different in severity, they’re different in the endpoint, and they’re different in the point of the whole thing. So that’s the first thing.

I want to pick up on quid pro quo for just a moment. It certainly indicates that a patient that you’ve cared for a long time is more willing to work with you in new ways than a patient you don’t know. And they do that out of trust. I can remember a patient that we wanted to put in an experimental protocol who wouldn’t even consent. "I know you wouldn’t let anything happen to me doc." And that’s quid pro quo. But the quid pro quo goes both ways, in fact. You can’t let anything happen because the longer the relationship, the more responsible you are to that person that nothing happen to them.

And then, finally, you’ve given an impression of a degree of control. You don’t say the probability is the symptom will be short; the probability is the reaction will be damped down. That’s all you could really say, isn’t it. You can’t say absolutely. So, in each one of these things you give the impression that the symptom will be short; it won’t be severe. That’s not the case. The probability is.

DR. ROSENSTEIN: In the overwhelming majority of cases that is the case.

DR. CASSELL: Well--but "overwhelming" is a funny word. I don’t know what that means numerically, but I do know what probability means. Probability means that sometimes it’s not the case.

DR. ROSENSTEIN: That’s true, but we have data on it.

DR. CASSELL: Exactly right. And when you set up ethical guidelines and things, they’re not for when everything goes well; that’s for when things go badly. We know everybody’s good and true. It’s, "Have they protected their subjects should things go badly?" So on at least three counts, your argument, for me at least, I find unpersuasive. I understand why you want to do it--it’s the protection of human subjects that’s the question.

DR. SHAPIRO: David?

DR. COX: I would like to applaud a point in your presentation and that’s the first one, which is talking about the importance of the scientific design. I don’t know a whole lot about ethics, but I know a lot about science, and to simply have a hypothesis isn’t in my view sufficient measure for good science. To have a hypothesis that then has follow up consequences to it and things that you can do subsequently, that’s a good scientific design. Now, I’m being vague, but I think you get my drift. It’s an example of I wonder how a car runs so I’ll go in and I’ll hit it with a hammer. Yep, it doesn’t run anymore. That doesn’t tell me a lot about the engine. And, so that--that in situations like this where you can’t actually predict what’s going to happen, and sometimes it really screws people up. Looking at what the scientific validity of the design is extremely important. So, tell me about that. Who does that? Is it the people that do that kind of research? Is it independent review? Because this I think is a key, key point. I hear what you’re saying. You’re saying sometimes it’s really important to do this. So, who decides when it’s important on a scientific basis?

DR. ROSENSTEIN: I think it’s a very important question, and I can only speak to what happens at the Clinical Center. That’s my only direct experience. And, what happens at the Clinical Center is that the IRB considers that. And, in response to a question that was asked earlier, there has been a shift in the last few years in terms of the scrutiny of the science in the IRB. I think for a long time the only considerations were with respect to subject protections; and certainly there’s been a culture change recently, which is, this needs to be very compelling to all of the scientists and laypeople around the table before we’re going to go ahead with this. And there have been numerous suggestions about design changes, and about collapsing different phases so as to minimize placebo phases.

DR. COX: Well, I ‘ll just say that while I see that as plausible, having the IRB make these kinds of scientific decisions—I am not very enthusiastic. I’m just trying to be supportive of what you’re trying to do, which is to have some vehicle open for this kind of research in really critical situations. It would seem like to have a special group of people who weren’t necessarily invested in that kind of research themselves, that this would be a very helpful thing without new regulations, but really paying careful attention to scientific design.

DR. ROSENSTEIN: I think that’s fair. I wasn’t quite finished. There’s also a separate scientific review of each of the protocols that goes on. Now, I can’t say that that’s an independent group. It’s certainly at the clinical center, and I take that point there. The scientific group you is not solely contemplated in the IRB.

DR. COX: I in no way mean this to be disrespectful, and if it comes out that way I don’t want it to be, but often times in different fields, when other people don’t agree with what you do you start your own journals and have your own people do the reviews. And I think that particularly in situations like this--and it’s true in genetics, it’s true in psychiatry, it’s true in a lot of things--to have, particularly in scientific evaluations for this purpose to have broad-based input I think is extremely important, and I think in many of these areas we tend just toward the opposite because it’s very critical, and so it’s very easy to get people who think like we do that to convince ourselves that this is really good science. So, those are my comments.

DR. SHAPIRO: Thank you. Alex....

PROF. CAPRON: I first wanted to get some clarification from you as to why you’re talking about symptom provoking studies. You excluded washout studies. Because looking in terms of most of the points you made about having to have a group of subjects who are manifesting the particular symptom in order to study it and to see what levels and so forth and so on, the argument that we’ve heard from people who do that kind of work sounds very much like your argument. Is there an a priori for making this distinction between drug challenge and washout?

DR. ROSENSTEIN: I think there are a number of them. I mean, I kept them separate because I think they are separate, and part of the problem is that they get blended together because the challenge studies that in my opinion invoke the strongest responses and concern are those that are done in drug washout states, but that is far from the universe of symptom provoking studies, and so I wanted to--I think it’s important for this group as well as for IRBs and investigators to keep separate what the specific protocol is.

PROF. CAPRON: Well, let me come around to my endpoint. At the end of your presentation I thought I saw the reason why you made the distinction. It’s precisely because the greatest criticisms have been named at washout studies and washout studies have involved long-term consequences and suicide. And you’re--so, your distinction didn’t seem to me it was driven by an initial set of scientific characteristics but rather by trying to distinguish a subset of this field by provoking systems through drug challenges with people who perhaps aren’t the ones that are controls--you felt it was a better form of research.

DR. ROSENSTEIN: No, I didn’t mean to imply that. Obviously there are greater concerns with respect to informed consent and risk in some circumstances than others. I don’t think one can make global statements about whether a given research paradigm is justifiable or not in a washout setting. I think there are a number of valid scientific justifications for learning and doing imaging studies when receptor occupancy is none as opposed to fully occupied.

PROF. CAPRON: You and I aren’t communicating. I fully accept the notion that just as you could provide scientific rationale for saying, "We can learn good things, important things, and maybe even things that would satisfy scientists from outside your immediate field," as being well designed to answer the question. By doing challenge studies, I think we would hear the same thing from people who "provoke symptoms" or provoke the states that you can image by washing out. I reached the conclusion that you would base your distinction based on severity; that is to say, looking at this range of things which you think have been lumped together you said, "Well, if I say I’m only talking about this part," then I don’t have to worry about certain severe consequences, which leads me to my next question for you which is: Would you set any limits on research aimed at answering important questions about diseases that had otherwise frustrated scientists’ ability to offer understanding what’s going on and methods for treatment, and if so how would you set those limits?

DR. ROSENSTEIN: It’s such a vague question it’s hard to answer.

PROF. CAPRON: Well, I don’t think it’s so vague. Obviously one thing I’m getting to is do you think that there should be limits on washout studies? You’ve excluded them from your definition of symptom provocation.

DR. ROSENSTEIN: It’s a complicated, interesting area. I wasn’t asked to talk about the justification of washout studies and we could do that at another time. Some washout studies--I think there are very compelling scientific justifications for them, others not. Depends on the specific study. There are certainly some regulations that I think would be reasonable, and I mentioned one of them in this context during the study, which is I think that it doesn’t make sense to take someone who’s stable off of medicines for the purpose of provoking symptoms to learn more about it. It just doesn’t--you know, there are other examples, but I guess my point is you have to look at what’s the study, who are the subjects that are going to be recruited, what are the protections, what’s the level of informed consent, are these treatment-refractory subjects or first-episode subjects. You can’t make a judgment about whether they’re acceptable or not outside of the overall context; it’s not just severity of symptoms.

PROF. CAPRON: May I ask one more question? In part you answered one of your own questions, which I took to be a real question, not purely rhetorical. Does anyone know of any indications that there have been any long-term adverse consequences? Would you say that if we looked at--you’re at NIMH, is that right?

DR. ROSENSTEIN: Yes, sir.

PROF. CAPRON: If we looked at NIMH sponsored or conducted studies in the last decade, we would see as a uniform part of such studies a process for long-term evaluation of the subjects of those studies, because an absence of data is significant only if the data has been collected and you don’t find something...not if you don’t look for it.

DR. ROSENSTEIN: I think that’s a fair question. I’m aware of one study in normal volunteers looking at one-year follow up after CO₂ inhalation in triggering panic attacks to see whether there were subsequent panic attacks and there were none, but I think you have a fair point. I don’t think that--.

PROF. CAPRON: You don’t think this has been a uniform part of all study design.

DR. ROSENSTEIN: I don’t think it is, but I don’t know the answer to that question.

PROF. CAPRON: Perhaps Dr. Shore could answer that?

DR. SHORE: I was just thinking in terms of drug washouts, that is going to be one subject of meetings that we’re planning now. I agree with Dr. Rosenstein that it depends on how well a person’s doing, on how long they’re going to be off medication. Are they going to be an in-patient in the hospital? Or are they going to be at home? How much supervision do you have? How well have you explained to family and clinicians involved? That’s why there are yes or no answer to that. I think the field has changed within the past five or ten years and that studies that might have been on an outpatient basis or in-patient studies that might have involved two-month washouts are now talking about two-week washouts and that, as I say, I think that there are some times—for example, when you start a new medication—when you need to be off the medication you were taking previously because we have no clue as to drug interaction when you start the medication that’s just recently been approved.

PROF. CAPRON: I was restricting myself not to the washout studies but just to the challenge studies, and that was partly what you addressed in your panel, and I just wondered whether your panel--I didn’t hear you recite it now. You said that one of the criteria that you were aware was important in looking at past examples, because you were clearly responding in part to problems that have arisen in the past. One failure to follow up with those normal volunteers or those psychiatric patients or children who were given challenges with various drugs to see what not just physiologically what’s happening with that but what their behavioral response has been over the subsequent years and how they interpret that advance a year or a number of years later. Was it something you looked at? Were you satisfied that this was being well done?

DR. SHORE: Actually, there is some literature on that, but it’s fairly limited. There have been some suggestions that exacerbation of symptoms may have long-term problems. It is certainly apparent that the longer a person remains--with schizophrenia, to which I’m referring--that the longer that a person initially remains psychotic before the person receives adequate treatment, the poorer the long-term outcome. That’s why you don’t see studies anymore that are doing long-term psychotherapy for years comparing that to medication. We already know the answer to that question. The answer is those people do not do well. There have been studies that look at the question of whether somebody with chronic schizophrenia, having a fourth or six episode makes any difference to clinical reports five years later. The answer to that, to the best of my knowledge, is no. I’ll send you the reference to that.

PROF. CAPRON: And how about with the normal subjects?

DR. SHORE: I don’t know systematically that’s been determined. I know that, for instance, one of the reasons that we’re bringing together the meeting is that in the past, when the medications with which we were dealing were things like haloperidol that produces tardive dyskinesia or imipramine or elavil that has severe antipulmonergic effects, it was easier to justify balancing the risk of continuing on the medication with the desire to come off the medication and risk increased symptoms. As you probably know, 70 percent of patients who develop schizophrenia stop medications on their own within one to two years because they don’t like the side effects. We are now re-evaluating it because we have new medications. We have this surge in urgent specific reuptake inhibitors. We have now the agent for antipsychotics. Later this summer, we’ll have an additional one for that category, and that’s why we want to look at the risk/benefit ratio because some people, myself and my institute director included, see that the risk/benefit is now changing for medication discontinuation studies. When the only drugs that we had were quite toxic and not tolerated well by most people, then it was a lot easier to justify discontinuing medication since patients were probably going to do that anyway. Now that we have better medications that people are willing to take for longer periods of time and don’t, as far as we know, produce long-term toxicity-like problems with tardive dyskinesia or other movement disorders, I would agree with you that that equation is changing. That’s why we’re bringing together this meeting.

DR. SHAPIRO: Thank you. Rhetaugh and that last question.

DR. DUMAS: This raises a number of questions in my mind. One is, are we implying that it’s unethical to do symptom provoking studies or wash out studies, and if so how do you rationalize that? Is it possible to assess precisely risk/benefit ratio and if you can’t, under what conditions do you proceed with your research? There are a number of issues, I think, surrounding the studies that are being done with people who are mentally ill and I have a sneaking suspicion that we tend to look at them differently. If a person is able to give their consent—and I’d ask to think you about that because there are a number of studies that are done where the long-term outcome is unknown. You might say that about many of the studies that are done, that the long-term consequence is really unknown. It cannot be assessed really precisely a priori. So what is it about the symptoms provoking studies in psychiatric patients that gives us more concern about their protection, even if they--is it because by virtue of their psychiatric diagnosis we fear that they may not really be giving informed consent?

DR. SHAPIRO: Were you addressing that question to everyone here?

DR. DUMAS: Well, I was raising that question--I think that’s one that we need to address. We either do it now or we may need to do it later, but it just occurred to me that our dialog and debate changes depending upon the category of patients that we’re talking about. Now, we started out--we had been talking about people who have some impairment in decisionmaking, but not all psychiatric patients have the impairment in decisionmaking and yet we tend to discuss these issues as if they did.

DR. SHAPIRO: Well, I think, at least the staff tries to be very careful about that. That we can’t just paint them all with one brush. Let’s see if there’s any response. I really want to get on to the report itself in a few minutes.

PROF. CHARO: I’ll tell you, Rhetaugh, briefly for myself. The reason why I view this differently than other areas of research is that in most areas of research there’s the possibility of harm. And your question is simply is the harm that you’re intending to create within tolerable limits based on the scientific goals you have in light of their ability to participate in the decisionmaking to make it a voluntary experience.

That, to me, is somewhat different. It’s on the same spectrum but it’s somewhat different than the possibility of harm. It’s unusual for us to deliberately do harm to people. And that is why I approach these differently.

DR. SHAPIRO: Other comments now? Because if not, we are running atypically late today, and with apologies to Alex, I’m going to suggest, one, that we take a five minute break—at least our break is on time—but then I want to go directly to a discussion of the report on the session involving persons with mental disorders, et cetera, because we just can’t postpone that any longer. We’ll get to the others as soon as we can and when we can.

Let’s take a five minute break. Let’s try to reassemble here precisely at 3:00. Thank you.