- Email: [email protected]
Sense and nonsense:
Schizophrenia Research 35 (1999) 219–225
Sense and nonsense: an essay on schizophrenia research ethics William T. Carpenter *, Robert R. Conley Maryland Psychiatric Research Center, University of Maryland School of Medicine, PO Box 21247, Baltimore, MD 21228, USA Received 1 September 1998; accepted 5 September 1998
Abstract In this essay, the authors select topics from the current debate on the ethics of schizophrenia research. Accepting competent and voluntary informed consent as essential for most projects, the authors discuss the relation of diagnosis to decisional capacity, the respective roles of psychosis and cognitive impairments in decisional capacity, and whether impairments in capacity can be remediated. The roles of investigator, external agent, patient subject, and family or surrogate in the informed consent process are reviewed. A lack of understanding of the treatment of persons with schizophrenia has distorted and inflamed public discussion of issues such as ‘withholding treatment’. A standard, based in common sense, is proposed: for viewing protocols; for allowing autonomy and altruism despite diagnostic class; and, for a meaningful discussion of what is meant by and what should be done about ‘risk without direct benefit’ protocols. © 1999 Elsevier Science B.V. All rights reserved. Keywords: Schizophrenia; Research ethics; Informed consent; Research safety
1. Introduction The Toronto Conference on Research Ethics provides an excellent opportunity to consider certain issues in ethics as applied to research with severely mentally ill persons. It is important that views from many vantages be considered in the discussion of these issues. The content of this essay will be focused on issues in schizophrenia research. This is an area of much heat and misunderstanding in the public discussion. The discussion of a few selected issues is offered in the hope that it will
* Corresponding author. Tel: 410-455-7101; Fax: 410-788-3837; e-mail: [email protected]
increase rational consideration of optimal research procedures and diminish the influence of recommendations based on a misunderstanding of the actual circumstances of clinical investigation. The discussion will be divided into sections on informed consent, on the idea of withdrawing medication, and on the several perspectives in which research protocols are judged.
2. Informed consent The conduct of clinical investigation is a noble enterprise and a foundation of scientific medicine. Much human suffering and enhancement of quality
0920-9964/99/$ – see front matter © 1999 Elsevier Science B.V. All rights reserved. PII: S0 9 2 0 -9 9 6 4 ( 9 8 ) 0 0 12 8 - 5
220
W.T. Carpenter, R.R. Conley / Schizophrenia Research 35 (1999) 219–225
of human life emanate directly from clinical investigations. The clinician investigator has many responsibilities, but foremost among them is the protection of the safety, welfare, and dignity rights of persons who participate as subjects in investigations. Informed consent is at the heart of the processes that protect the autonomy, dignity, and safety of potential research subjects. Many factors influence an individual’s ability to become informed and to exercise autonomy. Foremost among these influences may be social cognition and emotional factors. Yet another factor, decisional capacity at the conscious cognitive level, is also required and has been the principal focus in the current debate regarding the quality of informed consent in schizophrenia research. The following comments are presented as a dialectic within six dichotomies which are often presented in the informed consent debate regarding schizophrenia.
disorders, procedures are required for distinguishing between potential subjects with adequate capacity and those with diminished capacity. Procedures would then be developed for selecting among the capacitated and excluding the incapacitated and/or for enhancing capacity in those with suspect capacity. This task is vitally important in schizophrenia, but no more so than for other conditions in which some proportion of subjects have impaired decisional capacity. A current illustration of the importance of the distinction between decisional capacity and diagnosis is the report and recommendations of the National Bioethics Advisory Commission released for comments on 1 July 1998. Beginning with the problem of decisional capacity for informed consent, the recommendations are related exclusively to mental disorders and make restrictive recommendations for individuals based on group membership (diagnosis) rather than decisional capacity (NBAC Website: http://www.bioethics.gov).
2.1. Decisional capacity versus diagnosis 2.2. Cognition versus psychosis Many brain disorders are associated with impaired decisional capacity in some proportion of afflicted subjects. A rational approach to the problem presented by informed consent would focus on the general issue of decisional capacity, how this is to be assessed, what procedures are necessary to assure or enhance capacity. What procedures are necessary if subjects lacking decisional capacity are to be entered into research? This rational approach is compromised by singling out psychiatric disorders in general and schizophrenia in particular as diagnoses associated with lack of decisional capacity. Many assert that the mentally ill are a vulnerable population with impaired decisional capacity, while other brain disorders such as strokes are regarded as potential problems for decisional capacity where procedures that evaluate individuals’ capacity become the focus. Social stigmas tend to designate certain psychiatric disorders as ‘impaired’, while ignoring other disorders where the proportion of individuals lacking decisional capacity may be similar. In evolving optimal procedures, schizophrenia research, like research on the elderly and with many other brain
An uninformed line of argument equating psychoses with lack of decisional capacity impedes progress in finding optimal informed consent procedures. How, some ask, can a person with delusions give a rational manipulation to the salient facts in an informed consent procedure? To the schizophrenologist the answer is quite straightforward: persons with delusional systems do a rational manipulation of most facts most of the time, with an understanding and an appreciation of the personal relevance. This is why they use money to pay bus fares, order the food they wish to eat at fast-food restaurants, and put the right shoe on the right foot. The belief that the FBI is manipulating their thoughts affects some things, but not most. The task in informed consent is to learn whether it impedes the understanding, appreciation, and rational manipulation of the information essential to the informed consent process (Appelbaum and Roth, 1981). In investigations at the Maryland Psychiatric Research Center (MPRC ) (presentation by Dr William Carpenter at the APA Annual Meeting, May 1997), we found a range of performance on
W.T. Carpenter, R.R. Conley / Schizophrenia Research 35 (1999) 219–225
a decisional capacity task, with most patients initially scoring in a questionable range regarding informed consent. However, severity of symptoms provided only a modest correlation with this performance. The variance in decisional capacity performance was principally accounted for by cognitive variables rather than by symptom variables. A reasonable conclusion from this finding is that evaluation for capacity for informed consent in schizophrenia would do better to focus on cognitive impairment than on symptom expression. Focus on information processing and learning impairments is the key to successful informed consent education (see below). 2.3. Remedial versus fixed impairments in decisional capacity Informed consent for research (defined as an event where the subject reads information, may ask questions, and signs a form) is a significantly flawed procedure for most humans in most experimental circumstances. Most workers would agree that an informed consent process that includes an ongoing dialogue is essential to optimal understanding. In this context the signing of an informed consent document would represent the culmination of a process or perhaps a significant point of entry following a process that will be continued during the subject’s research participation. In our study, using the MacArthur Competence Assessment Tool for Clinical Research in Schizophrenia (MacCAT-CRS ) (Appelbaum and Grisso, 1996; presentation by Dr William Carpenter at the APA Annual Meeting, May 1997), we found many subjects scoring in a range defined as questionable for decisional capacity. Nevertheless, all participating subjects had signed a consent document for actual research protocol participation and had passed a test documenting fundamental features of their understanding of the pertinent facts in the real consent process. We thought the discrepancy between these two sets of observations might be based on informed consent for real research taking place in an ongoing informed consent educational process, while the MacCAT-CRS evaluation was based on a one-time presentation. To decide if this hypothesis was supported, subjects scoring in the questionable range entered an abbreviated
221
informed consent process where they were given the opportunity to discuss the protocol more than once and in more than one setting. Retesting a week later resulted in the schizophrenia subjects significantly enhancing their performance on the MacCAT-CRS evaluation. They now scored with about the same mean and range of performance as observed by Kovnick in a normal control population studied at the University of Virginia (presentation by Dr Jeffrey Kovnick at the APA Annual Meeting, May 1997). The conclusion is quite straightforward: cognitive impairments such as difficulty in reading comprehension can be addressed with remedial activity. The fundamental cognitive liability may not be altered, but a person’s performance can be enhanced. 2.4. Investigator versus external agent as a responsible party for informed consent Some critics of current procedures believe that clinician investigators should be removed from the informed consent process. External monitors would ascertain decisional capacity and be responsible for assuring adequacy of the informed consent process. Three problems exist with removing clinician investigators from a central responsibility for this process. First, the strongest in the chain of protections for research subjects needs to be the personal assumption of responsibility and integrity of the clinical investigator. To remove the investigator from personal responsibility for assuring the integrity of the research process pins too much hope on procedural assurance and unnecessarily removes a principal component of protection procedures. Other protections are, of course, essential, but no evidence that the process is routinely undermined by the presence of clinical investigators has been presented, and there is much reason to judge this an essential component. The second consideration is the practical problem of imagining an ongoing informed consent process that could be adequately accomplished by external and independent monitors. The third consideration is cost. If substantial funding is diverted into the creation of a new process, less new knowledge will be purchased for the research dollar. The ethical and social costs of this exchange are important to debate.
222
W.T. Carpenter, R.R. Conley / Schizophrenia Research 35 (1999) 219–225
2.5. Patient consent versus family/surrogate consent This essay is based on the proposition that subjects selected for participation in an informed consent process have adequate decisional capacity to act in their own interest. Alternative procedures for children and decisionally incapacitated adults are essential. However, the proposition that legally significant consent for schizophrenic persons could be altered by family or other surrogate deprives individuals of autonomy rights and dignity without a determination that they lack decisional capacity. At a practical level, many adult schizophrenics competent to provide consent either do not have or do not wish to involve a relative or significant other in this process. They, perhaps like the reader, do not relinquish personal autonomy rights lightly. From the investigator’s point of view, having significant others completely informed and encouraging them to participate to the extent desired by the patient subject is welcome, but the proposition that the patients lose autonomy without a finding of decisional incapacity is unwarranted and stigmatizing. 2.6. Withholding treatment Among the most controversial research designs in the present debate are those that involve a drugwithdrawal period, a placebo-treatment period, or a pharmacologic-probe strategy that may increase symptomatic manifestations. A great deal of public discussion is based on a remarkable misunderstanding of the actual data about this type of research, and informed discussion in the media is rare (Hilts, 1994a,b, 1996, 1998a,b; Willwerth, 1993; Weiss, 1998). Consideration of this issue needs to include the following appreciation of treatment in schizophrenia. 2.6.1. Aspects of treatment Treatment involves many elements, including certain psychosocial treatment procedures that have been scientifically validated as efficacious using clinical trials methodology. The discussion of treatment as though it were fully defined by antipsychotic medication is flawed.
Antipsychotic medication represents a component of standard treatment usually recommended on a continuous basis. However, this treatment is a partial technology with very substantial shortcomings. The shortcomings include: a. Actual treatment effects. Efficacy is for the psychotic symptomatic components of the illness. Primary negative symptom and cognitive deficits have not been documented as treatment responsive. However, these aspects of schizophrenia are the principal contributors to longterm morbidity, poor functional outcome, and diminished quality of life. Publicly asserted analogies to withholding insulin from diabetics or immunosuppressant drugs during organ transplants represent a gross misunderstanding of schizophrenia treatment. b. Actual medication status in standard care. In standard clinical care, non-compliance with antipsychotic medication is a severe problem ( Weiden and Olfson, 1995). Standard treatment, therefore, is not the idealized following of the doctor’s recommendation with a robust symptom reduction and prevention of relapses, but the imperfect administration of a partial technology addressing one aspect of the illness. The potential therapeutic benefits of antipsychotic drugs are very substantial, despite these major shortcomings. Nonetheless, greater clarity as to the actual implications of medication discontinuation would enhance consideration of the risk/benefit evaluation necessary to consider the ethical implications of placebo-controlled studies and medication-withdrawal studies. 2.6.2. Safety data a. Safety. There are very substantial safety data from random assignment prospective studies where medications have been withheld from one group and given to the other. These data are reviewed elsewhere (Carpenter et al., 1997; Gilbert et al., 1995), and found negligible evidence for any lasting disadvantage even in the face of substantial contemporaneous symptomatic disadvantage. This led us to conclude that the key feature was not the question of withholding treatment. Rather, in study designs with a
W.T. Carpenter, R.R. Conley / Schizophrenia Research 35 (1999) 219–225
medication-free period, inclusion and exclusion criteria must be designed for safety, and ‘clinical rescue’ procedures must be in place for rapidly detecting difficulties and providing effective intervention in order to provide a risk/benefit result that is justified as clinical prudence. b. Medication-free does not mean no medication treatment. The original drug/placebo studies determined whether antipsychotic drugs were efficacious, and later whether efficacious medication had implications for long-term relapse prevention. In these early studies, placebo-treated patients were permitted to advance to a serious degree of worsening in relapse-prevention studies or to prolonged experience of psychosis in acute-treatment studies. Now that the therapeutic efficacy has been documented, it is essential that medication-withdrawal and placebosubstitution studies have an active medication intervention strategy to minimize risk. A placebo group need not be a no-treatment group. As mentioned above, many other elements of treatment may be present, and often research studies provide enhanced clinical monitoring and other clinical therapeutic techniques in the context of limitations placed on the antipsychotic drug component of treatment. The second consideration is the use of antipsychotic drugs in a targeted, rapid intervention (Carpenter and Heinrichs, 1983) manner to minimize the length of time a person remains psychotic or to minimize the extent to which an exacerbation progresses. In this regard patients in a placebo control study are either receiving continuous neuroleptic treatment or a rapid intervention, antipsychotic drug treatment that is generally less effective (in the short term) than continuous medication, but is more effective than no antipsychotic drug treatment. In this approach, it appears that the patients receiving placebo are not at increased risk for longer term disadvantage.
223
of these it is important that considerations meet a commonsense standard. Instances to consider include: 3.1. Protocol-unique manipulation versus the overall package of care Consider a treatment study that involves a 1-month evaluation period, a 1-month fixed-dose stabilization period, and a 2-month controlled experimental period followed by a 6-month open label administration of the experimental treatment. The principal, manipulation is comparing an experimental drug with placebo during the controlled treatment phase. Should the risk/benefit analysis focus on the placebo versus experimental drug phase, or on the entire protocol package? A patient considering participating will necessarily consider the many months of participation and how this might compare with what they would anticipate during the 10-month period in a standard care setting. It is essential that they specifically consider the few weeks of placebo controlled treatment. But any prospective subject will consider the over-all package. What should the ethicist consider? Patient subjects may choose to take some risk if assigned to placebo, either because they have confidence in the treatment intervention rescue procedures or because they feel the risk is justified in relation to their hope that the experimental medication (which they will receive in the openlabel period) will offer some advantage to them. The close monitoring provided and the psychoeducational programs associated with the research clinic may result in a more substantial compliance during the year than would be expected in standard care. The investigator would not put this forward as a benefit, and the ethicists on the IRB may not consider it, but no one can stop the referring clinician, family member, or patient from thinking about the general implications of research-based clinical care.
3. Perspectives on research ethics
3.2. Altruism
There are several vantages from which the ethics of particular protocols can be addressed. In each
Many humans wish to contribute to society by participating in biomedical research. This wish
224
W.T. Carpenter, R.R. Conley / Schizophrenia Research 35 (1999) 219–225
may include hope for some future benefit, or benefit of future knowledge for their own family. But altruism is legitimate, and few among us want to be deprived of autonomy in such personal decisions. It is important that decisionally capacitated persons not be deprived of altruism as a legitimate motivation for research participation based on membership in a mental illness diagnostic group. In the study of chronic disease there is another consideration, where a potential subject’s perspective may be quite different from that put forth by the ethicists and the investigator. A 25-year-old person with primary negative symptoms and cognitive impairments participating in an experimental trial of a treatment for this component of the illness may be told that there is no potential for direct benefit since the treatment, even if effective, cannot be continued for more than 8 weeks. Any long-term use of the medication would depend on safety data unlikely to be generated until the hypothesis of efficacy is supported in short-term experiments. Nonetheless, if the patient feels that this type of research will increase the likelihood of effective treatment being available 10 years from now, they may wish to participate in the hope of future benefit, though no hope of benefit is presented as an aspect of protocol participation. The patients who participated in the initial drug/placebo maintenance therapy trials were soon to have the next 30 years of their treatment influenced by the scientific evidence gained in the initial trials. Those who advocate exclusion of the mentally ill from any study involving risk without direct benefit (regardless of the individuals’ decisional capacity) seem ready to deprive individuals of their personal autonomy without meeting a standard of common sense or respect for dignity rights. 3.3. High-risk research without direct benefit Of the many issues swirling in the current debate on research ethics in schizophrenia, this will be last touched upon in this essay. The assertion that where there is a question of decisional capacity a person should not be allowed to participate in research if it does not offer direct benefit, or if there is an increment above minimal risk, is an ill-
founded proposition because of the practical restrictions on the definition of terms. First, regarding no-benefit research: this, of course, refers only to the specific experimental manipulation per se and not to the general proposition. If one tests an hypothesis with an imaging technique, the prospective subjects will be told there is no direct benefit on their disease process from the imaging technique applied. However, from the potential subject’s perspective, there may be enhanced quality of care, more detailed assessment, a range of interesting aspects in which they would like to participate, the hope that they will be better informed about the disease by way of their participation, and the hope that such basic research will eventually lead to better treatments, all of which humans will consider while the ethics debate has an exclusive focus on the benefit of the experimental manipulation per se. This context may be too narrow to be valid from the perspective of a potential subject. Regarding risk, most discussions go on as though we all know what we mean by risk, but in many circumstances the designated risk level bears no relation to common sense. Many IRBs consider persons with schizophrenia to be a vulnerable population and, therefore, all research is judged to be above minimal risk. Administering a drugabuse questionnaire constitutes high-risk research, despite procedures for maintaining confidentiality and the absence of any instances of a person being prosecuted or embarrassed by information becoming public. The mere theoretical potential for the material being harmful or embarrassing, is sufficient to rate the protocol as high risk. Common sense would help in resolving this type of semantic issue. But serious consideration is being given to ending high risk/no direct benefit in ‘vulnerable’ populations defined by a mental illness diagnosis. To date, in schizophrenia research, no imaging study or genetic study offers direct benefit. All are ‘risk’ protocols. The draft report of the National Bioethics Advisory Commission recommends cessation of this type of study in the mentally ill (NBAC Website: http://www.bioethics.gov)! The issues in the current research ethics debate are important and complex. Evolving the best
W.T. Carpenter, R.R. Conley / Schizophrenia Research 35 (1999) 219–225
procedures is dependent on a realistic appreciation of clinical care during research participation.
References Appelbaum, P.S., Roth, L.H., 1981. Clinical issues in the assessment of competency. Am. J. Psychiatry 138, 1462–1467. Appelbaum, P.S., Grisso, T., 1996. MacArthur Competence Assessment Tool—Clinical Research Version. University of Massachusetts, Worcester, MA. Carpenter, W.T., Heinrichs, D.W., 1983. Early intervention, time-limited targeted pharmacotherapy in schizophrenia. Schizophr. Bull. 9, 533–542. Carpenter, W.T., Schooler, N.R., Kane, J.M., 1997. The rationale and ethics of medication-free research in schizophrenia. Arch. Gen. Psychiatry 54, 401–407.
225
Gilbert, P.L., Harris, J., McAdams, L.A., Jeste, D.V., 1995. Neuroleptic withdrawal in schizophrenic patients: a review of the literature. Arch. Gen. Psychiatry 52, 173–188. Hilts, P.J., 1994a. Agency faults a UCLA study for suffering of mental patients. New York Times, 10 March. Hilts, P.J., 1994b. Medical experts testify on tests done without consent. New York Times, 24 May. Hilts, P.J., 1996. Conference is unable to agree on ethical limits of research: psychiatric experiment helped fuel debate. New York Times, 9 January. Hilts, P.J., 1998a. Scientists and their subjects debate psychiatric research ethics. New York Times, 19 May. Hilts, P.J., 1998b. Psychiatric unit’s ethics faulted. New York Times, 28 May. Weiden, P.J., Olfson, M., 1995. Cost of relapse in schizophrenia. Schizophr. Bull. 21, 419–429. Weiss, R., 1998. Volunteers at risk in medical studies: complex research projects strain system of safeguards. Washington Post, 1 August, p. A01. Willwerth J., 1993. Tinkering with madness. Time, 30 August.
Sense and nonsense: an essay on schizophrenia research ethics William T. Carpenter *, Robert R. Conley Maryland Psychiatric Research Center, University of Maryland School of Medicine, PO Box 21247, Baltimore, MD 21228, USA Received 1 September 1998; accepted 5 September 1998
Abstract In this essay, the authors select topics from the current debate on the ethics of schizophrenia research. Accepting competent and voluntary informed consent as essential for most projects, the authors discuss the relation of diagnosis to decisional capacity, the respective roles of psychosis and cognitive impairments in decisional capacity, and whether impairments in capacity can be remediated. The roles of investigator, external agent, patient subject, and family or surrogate in the informed consent process are reviewed. A lack of understanding of the treatment of persons with schizophrenia has distorted and inflamed public discussion of issues such as ‘withholding treatment’. A standard, based in common sense, is proposed: for viewing protocols; for allowing autonomy and altruism despite diagnostic class; and, for a meaningful discussion of what is meant by and what should be done about ‘risk without direct benefit’ protocols. © 1999 Elsevier Science B.V. All rights reserved. Keywords: Schizophrenia; Research ethics; Informed consent; Research safety
1. Introduction The Toronto Conference on Research Ethics provides an excellent opportunity to consider certain issues in ethics as applied to research with severely mentally ill persons. It is important that views from many vantages be considered in the discussion of these issues. The content of this essay will be focused on issues in schizophrenia research. This is an area of much heat and misunderstanding in the public discussion. The discussion of a few selected issues is offered in the hope that it will
* Corresponding author. Tel: 410-455-7101; Fax: 410-788-3837; e-mail: [email protected]
increase rational consideration of optimal research procedures and diminish the influence of recommendations based on a misunderstanding of the actual circumstances of clinical investigation. The discussion will be divided into sections on informed consent, on the idea of withdrawing medication, and on the several perspectives in which research protocols are judged.
2. Informed consent The conduct of clinical investigation is a noble enterprise and a foundation of scientific medicine. Much human suffering and enhancement of quality
0920-9964/99/$ – see front matter © 1999 Elsevier Science B.V. All rights reserved. PII: S0 9 2 0 -9 9 6 4 ( 9 8 ) 0 0 12 8 - 5
220
W.T. Carpenter, R.R. Conley / Schizophrenia Research 35 (1999) 219–225
of human life emanate directly from clinical investigations. The clinician investigator has many responsibilities, but foremost among them is the protection of the safety, welfare, and dignity rights of persons who participate as subjects in investigations. Informed consent is at the heart of the processes that protect the autonomy, dignity, and safety of potential research subjects. Many factors influence an individual’s ability to become informed and to exercise autonomy. Foremost among these influences may be social cognition and emotional factors. Yet another factor, decisional capacity at the conscious cognitive level, is also required and has been the principal focus in the current debate regarding the quality of informed consent in schizophrenia research. The following comments are presented as a dialectic within six dichotomies which are often presented in the informed consent debate regarding schizophrenia.
disorders, procedures are required for distinguishing between potential subjects with adequate capacity and those with diminished capacity. Procedures would then be developed for selecting among the capacitated and excluding the incapacitated and/or for enhancing capacity in those with suspect capacity. This task is vitally important in schizophrenia, but no more so than for other conditions in which some proportion of subjects have impaired decisional capacity. A current illustration of the importance of the distinction between decisional capacity and diagnosis is the report and recommendations of the National Bioethics Advisory Commission released for comments on 1 July 1998. Beginning with the problem of decisional capacity for informed consent, the recommendations are related exclusively to mental disorders and make restrictive recommendations for individuals based on group membership (diagnosis) rather than decisional capacity (NBAC Website: http://www.bioethics.gov).
2.1. Decisional capacity versus diagnosis 2.2. Cognition versus psychosis Many brain disorders are associated with impaired decisional capacity in some proportion of afflicted subjects. A rational approach to the problem presented by informed consent would focus on the general issue of decisional capacity, how this is to be assessed, what procedures are necessary to assure or enhance capacity. What procedures are necessary if subjects lacking decisional capacity are to be entered into research? This rational approach is compromised by singling out psychiatric disorders in general and schizophrenia in particular as diagnoses associated with lack of decisional capacity. Many assert that the mentally ill are a vulnerable population with impaired decisional capacity, while other brain disorders such as strokes are regarded as potential problems for decisional capacity where procedures that evaluate individuals’ capacity become the focus. Social stigmas tend to designate certain psychiatric disorders as ‘impaired’, while ignoring other disorders where the proportion of individuals lacking decisional capacity may be similar. In evolving optimal procedures, schizophrenia research, like research on the elderly and with many other brain
An uninformed line of argument equating psychoses with lack of decisional capacity impedes progress in finding optimal informed consent procedures. How, some ask, can a person with delusions give a rational manipulation to the salient facts in an informed consent procedure? To the schizophrenologist the answer is quite straightforward: persons with delusional systems do a rational manipulation of most facts most of the time, with an understanding and an appreciation of the personal relevance. This is why they use money to pay bus fares, order the food they wish to eat at fast-food restaurants, and put the right shoe on the right foot. The belief that the FBI is manipulating their thoughts affects some things, but not most. The task in informed consent is to learn whether it impedes the understanding, appreciation, and rational manipulation of the information essential to the informed consent process (Appelbaum and Roth, 1981). In investigations at the Maryland Psychiatric Research Center (MPRC ) (presentation by Dr William Carpenter at the APA Annual Meeting, May 1997), we found a range of performance on
W.T. Carpenter, R.R. Conley / Schizophrenia Research 35 (1999) 219–225
a decisional capacity task, with most patients initially scoring in a questionable range regarding informed consent. However, severity of symptoms provided only a modest correlation with this performance. The variance in decisional capacity performance was principally accounted for by cognitive variables rather than by symptom variables. A reasonable conclusion from this finding is that evaluation for capacity for informed consent in schizophrenia would do better to focus on cognitive impairment than on symptom expression. Focus on information processing and learning impairments is the key to successful informed consent education (see below). 2.3. Remedial versus fixed impairments in decisional capacity Informed consent for research (defined as an event where the subject reads information, may ask questions, and signs a form) is a significantly flawed procedure for most humans in most experimental circumstances. Most workers would agree that an informed consent process that includes an ongoing dialogue is essential to optimal understanding. In this context the signing of an informed consent document would represent the culmination of a process or perhaps a significant point of entry following a process that will be continued during the subject’s research participation. In our study, using the MacArthur Competence Assessment Tool for Clinical Research in Schizophrenia (MacCAT-CRS ) (Appelbaum and Grisso, 1996; presentation by Dr William Carpenter at the APA Annual Meeting, May 1997), we found many subjects scoring in a range defined as questionable for decisional capacity. Nevertheless, all participating subjects had signed a consent document for actual research protocol participation and had passed a test documenting fundamental features of their understanding of the pertinent facts in the real consent process. We thought the discrepancy between these two sets of observations might be based on informed consent for real research taking place in an ongoing informed consent educational process, while the MacCAT-CRS evaluation was based on a one-time presentation. To decide if this hypothesis was supported, subjects scoring in the questionable range entered an abbreviated
221
informed consent process where they were given the opportunity to discuss the protocol more than once and in more than one setting. Retesting a week later resulted in the schizophrenia subjects significantly enhancing their performance on the MacCAT-CRS evaluation. They now scored with about the same mean and range of performance as observed by Kovnick in a normal control population studied at the University of Virginia (presentation by Dr Jeffrey Kovnick at the APA Annual Meeting, May 1997). The conclusion is quite straightforward: cognitive impairments such as difficulty in reading comprehension can be addressed with remedial activity. The fundamental cognitive liability may not be altered, but a person’s performance can be enhanced. 2.4. Investigator versus external agent as a responsible party for informed consent Some critics of current procedures believe that clinician investigators should be removed from the informed consent process. External monitors would ascertain decisional capacity and be responsible for assuring adequacy of the informed consent process. Three problems exist with removing clinician investigators from a central responsibility for this process. First, the strongest in the chain of protections for research subjects needs to be the personal assumption of responsibility and integrity of the clinical investigator. To remove the investigator from personal responsibility for assuring the integrity of the research process pins too much hope on procedural assurance and unnecessarily removes a principal component of protection procedures. Other protections are, of course, essential, but no evidence that the process is routinely undermined by the presence of clinical investigators has been presented, and there is much reason to judge this an essential component. The second consideration is the practical problem of imagining an ongoing informed consent process that could be adequately accomplished by external and independent monitors. The third consideration is cost. If substantial funding is diverted into the creation of a new process, less new knowledge will be purchased for the research dollar. The ethical and social costs of this exchange are important to debate.
222
W.T. Carpenter, R.R. Conley / Schizophrenia Research 35 (1999) 219–225
2.5. Patient consent versus family/surrogate consent This essay is based on the proposition that subjects selected for participation in an informed consent process have adequate decisional capacity to act in their own interest. Alternative procedures for children and decisionally incapacitated adults are essential. However, the proposition that legally significant consent for schizophrenic persons could be altered by family or other surrogate deprives individuals of autonomy rights and dignity without a determination that they lack decisional capacity. At a practical level, many adult schizophrenics competent to provide consent either do not have or do not wish to involve a relative or significant other in this process. They, perhaps like the reader, do not relinquish personal autonomy rights lightly. From the investigator’s point of view, having significant others completely informed and encouraging them to participate to the extent desired by the patient subject is welcome, but the proposition that the patients lose autonomy without a finding of decisional incapacity is unwarranted and stigmatizing. 2.6. Withholding treatment Among the most controversial research designs in the present debate are those that involve a drugwithdrawal period, a placebo-treatment period, or a pharmacologic-probe strategy that may increase symptomatic manifestations. A great deal of public discussion is based on a remarkable misunderstanding of the actual data about this type of research, and informed discussion in the media is rare (Hilts, 1994a,b, 1996, 1998a,b; Willwerth, 1993; Weiss, 1998). Consideration of this issue needs to include the following appreciation of treatment in schizophrenia. 2.6.1. Aspects of treatment Treatment involves many elements, including certain psychosocial treatment procedures that have been scientifically validated as efficacious using clinical trials methodology. The discussion of treatment as though it were fully defined by antipsychotic medication is flawed.
Antipsychotic medication represents a component of standard treatment usually recommended on a continuous basis. However, this treatment is a partial technology with very substantial shortcomings. The shortcomings include: a. Actual treatment effects. Efficacy is for the psychotic symptomatic components of the illness. Primary negative symptom and cognitive deficits have not been documented as treatment responsive. However, these aspects of schizophrenia are the principal contributors to longterm morbidity, poor functional outcome, and diminished quality of life. Publicly asserted analogies to withholding insulin from diabetics or immunosuppressant drugs during organ transplants represent a gross misunderstanding of schizophrenia treatment. b. Actual medication status in standard care. In standard clinical care, non-compliance with antipsychotic medication is a severe problem ( Weiden and Olfson, 1995). Standard treatment, therefore, is not the idealized following of the doctor’s recommendation with a robust symptom reduction and prevention of relapses, but the imperfect administration of a partial technology addressing one aspect of the illness. The potential therapeutic benefits of antipsychotic drugs are very substantial, despite these major shortcomings. Nonetheless, greater clarity as to the actual implications of medication discontinuation would enhance consideration of the risk/benefit evaluation necessary to consider the ethical implications of placebo-controlled studies and medication-withdrawal studies. 2.6.2. Safety data a. Safety. There are very substantial safety data from random assignment prospective studies where medications have been withheld from one group and given to the other. These data are reviewed elsewhere (Carpenter et al., 1997; Gilbert et al., 1995), and found negligible evidence for any lasting disadvantage even in the face of substantial contemporaneous symptomatic disadvantage. This led us to conclude that the key feature was not the question of withholding treatment. Rather, in study designs with a
W.T. Carpenter, R.R. Conley / Schizophrenia Research 35 (1999) 219–225
medication-free period, inclusion and exclusion criteria must be designed for safety, and ‘clinical rescue’ procedures must be in place for rapidly detecting difficulties and providing effective intervention in order to provide a risk/benefit result that is justified as clinical prudence. b. Medication-free does not mean no medication treatment. The original drug/placebo studies determined whether antipsychotic drugs were efficacious, and later whether efficacious medication had implications for long-term relapse prevention. In these early studies, placebo-treated patients were permitted to advance to a serious degree of worsening in relapse-prevention studies or to prolonged experience of psychosis in acute-treatment studies. Now that the therapeutic efficacy has been documented, it is essential that medication-withdrawal and placebosubstitution studies have an active medication intervention strategy to minimize risk. A placebo group need not be a no-treatment group. As mentioned above, many other elements of treatment may be present, and often research studies provide enhanced clinical monitoring and other clinical therapeutic techniques in the context of limitations placed on the antipsychotic drug component of treatment. The second consideration is the use of antipsychotic drugs in a targeted, rapid intervention (Carpenter and Heinrichs, 1983) manner to minimize the length of time a person remains psychotic or to minimize the extent to which an exacerbation progresses. In this regard patients in a placebo control study are either receiving continuous neuroleptic treatment or a rapid intervention, antipsychotic drug treatment that is generally less effective (in the short term) than continuous medication, but is more effective than no antipsychotic drug treatment. In this approach, it appears that the patients receiving placebo are not at increased risk for longer term disadvantage.
223
of these it is important that considerations meet a commonsense standard. Instances to consider include: 3.1. Protocol-unique manipulation versus the overall package of care Consider a treatment study that involves a 1-month evaluation period, a 1-month fixed-dose stabilization period, and a 2-month controlled experimental period followed by a 6-month open label administration of the experimental treatment. The principal, manipulation is comparing an experimental drug with placebo during the controlled treatment phase. Should the risk/benefit analysis focus on the placebo versus experimental drug phase, or on the entire protocol package? A patient considering participating will necessarily consider the many months of participation and how this might compare with what they would anticipate during the 10-month period in a standard care setting. It is essential that they specifically consider the few weeks of placebo controlled treatment. But any prospective subject will consider the over-all package. What should the ethicist consider? Patient subjects may choose to take some risk if assigned to placebo, either because they have confidence in the treatment intervention rescue procedures or because they feel the risk is justified in relation to their hope that the experimental medication (which they will receive in the openlabel period) will offer some advantage to them. The close monitoring provided and the psychoeducational programs associated with the research clinic may result in a more substantial compliance during the year than would be expected in standard care. The investigator would not put this forward as a benefit, and the ethicists on the IRB may not consider it, but no one can stop the referring clinician, family member, or patient from thinking about the general implications of research-based clinical care.
3. Perspectives on research ethics
3.2. Altruism
There are several vantages from which the ethics of particular protocols can be addressed. In each
Many humans wish to contribute to society by participating in biomedical research. This wish
224
W.T. Carpenter, R.R. Conley / Schizophrenia Research 35 (1999) 219–225
may include hope for some future benefit, or benefit of future knowledge for their own family. But altruism is legitimate, and few among us want to be deprived of autonomy in such personal decisions. It is important that decisionally capacitated persons not be deprived of altruism as a legitimate motivation for research participation based on membership in a mental illness diagnostic group. In the study of chronic disease there is another consideration, where a potential subject’s perspective may be quite different from that put forth by the ethicists and the investigator. A 25-year-old person with primary negative symptoms and cognitive impairments participating in an experimental trial of a treatment for this component of the illness may be told that there is no potential for direct benefit since the treatment, even if effective, cannot be continued for more than 8 weeks. Any long-term use of the medication would depend on safety data unlikely to be generated until the hypothesis of efficacy is supported in short-term experiments. Nonetheless, if the patient feels that this type of research will increase the likelihood of effective treatment being available 10 years from now, they may wish to participate in the hope of future benefit, though no hope of benefit is presented as an aspect of protocol participation. The patients who participated in the initial drug/placebo maintenance therapy trials were soon to have the next 30 years of their treatment influenced by the scientific evidence gained in the initial trials. Those who advocate exclusion of the mentally ill from any study involving risk without direct benefit (regardless of the individuals’ decisional capacity) seem ready to deprive individuals of their personal autonomy without meeting a standard of common sense or respect for dignity rights. 3.3. High-risk research without direct benefit Of the many issues swirling in the current debate on research ethics in schizophrenia, this will be last touched upon in this essay. The assertion that where there is a question of decisional capacity a person should not be allowed to participate in research if it does not offer direct benefit, or if there is an increment above minimal risk, is an ill-
founded proposition because of the practical restrictions on the definition of terms. First, regarding no-benefit research: this, of course, refers only to the specific experimental manipulation per se and not to the general proposition. If one tests an hypothesis with an imaging technique, the prospective subjects will be told there is no direct benefit on their disease process from the imaging technique applied. However, from the potential subject’s perspective, there may be enhanced quality of care, more detailed assessment, a range of interesting aspects in which they would like to participate, the hope that they will be better informed about the disease by way of their participation, and the hope that such basic research will eventually lead to better treatments, all of which humans will consider while the ethics debate has an exclusive focus on the benefit of the experimental manipulation per se. This context may be too narrow to be valid from the perspective of a potential subject. Regarding risk, most discussions go on as though we all know what we mean by risk, but in many circumstances the designated risk level bears no relation to common sense. Many IRBs consider persons with schizophrenia to be a vulnerable population and, therefore, all research is judged to be above minimal risk. Administering a drugabuse questionnaire constitutes high-risk research, despite procedures for maintaining confidentiality and the absence of any instances of a person being prosecuted or embarrassed by information becoming public. The mere theoretical potential for the material being harmful or embarrassing, is sufficient to rate the protocol as high risk. Common sense would help in resolving this type of semantic issue. But serious consideration is being given to ending high risk/no direct benefit in ‘vulnerable’ populations defined by a mental illness diagnosis. To date, in schizophrenia research, no imaging study or genetic study offers direct benefit. All are ‘risk’ protocols. The draft report of the National Bioethics Advisory Commission recommends cessation of this type of study in the mentally ill (NBAC Website: http://www.bioethics.gov)! The issues in the current research ethics debate are important and complex. Evolving the best
W.T. Carpenter, R.R. Conley / Schizophrenia Research 35 (1999) 219–225
procedures is dependent on a realistic appreciation of clinical care during research participation.
References Appelbaum, P.S., Roth, L.H., 1981. Clinical issues in the assessment of competency. Am. J. Psychiatry 138, 1462–1467. Appelbaum, P.S., Grisso, T., 1996. MacArthur Competence Assessment Tool—Clinical Research Version. University of Massachusetts, Worcester, MA. Carpenter, W.T., Heinrichs, D.W., 1983. Early intervention, time-limited targeted pharmacotherapy in schizophrenia. Schizophr. Bull. 9, 533–542. Carpenter, W.T., Schooler, N.R., Kane, J.M., 1997. The rationale and ethics of medication-free research in schizophrenia. Arch. Gen. Psychiatry 54, 401–407.
225
Gilbert, P.L., Harris, J., McAdams, L.A., Jeste, D.V., 1995. Neuroleptic withdrawal in schizophrenic patients: a review of the literature. Arch. Gen. Psychiatry 52, 173–188. Hilts, P.J., 1994a. Agency faults a UCLA study for suffering of mental patients. New York Times, 10 March. Hilts, P.J., 1994b. Medical experts testify on tests done without consent. New York Times, 24 May. Hilts, P.J., 1996. Conference is unable to agree on ethical limits of research: psychiatric experiment helped fuel debate. New York Times, 9 January. Hilts, P.J., 1998a. Scientists and their subjects debate psychiatric research ethics. New York Times, 19 May. Hilts, P.J., 1998b. Psychiatric unit’s ethics faulted. New York Times, 28 May. Weiden, P.J., Olfson, M., 1995. Cost of relapse in schizophrenia. Schizophr. Bull. 21, 419–429. Weiss, R., 1998. Volunteers at risk in medical studies: complex research projects strain system of safeguards. Washington Post, 1 August, p. A01. Willwerth J., 1993. Tinkering with madness. Time, 30 August.