The Use of Placebo-Controlled Clinical Trials for the Approval of Psychiatric Drugs

Abstract

This is Part 2 of a two-part series on the ethical issues surrounding the use of placebo arms in clinical trials for psychiatric drugs. Part 1 discussed the ethical argument from a statistical, population-based perspective. Part 2 explores the ethical issues of placebo-controlled studies as they relate to individual psychiatric patients who may participate in them. Many patients who are candidates for psychiatric clinical trials would receive poor treatment for their mental illness under standard treatment conditions. Industry-sponsored clinical trials often provide treatment resources otherwise not available to patients at a more intense level of care than the local standard. Moreover, study design features, such as those developed at University of California, San Diego (presented herein), can mitigate the risks of placebo arms. With this in mind, clinical trials represent an ethical option for many patients with chronic mental illness.

Introduction

In Part I, I reviewed the scientific reasons why inclusion of a placebo arm in clinical trials of new psychiatric drugs is considered by many an ethical imperative since it reduces the chance that large numbers of patients will be treated with drugs mistakenly deemed to be efficacious due to faulty interpretation of study results.¹ That rationale for including a placebo arm was based upon the “greater good” argument—the benefit derived by the population at large who are candidates for treatment with the investigational medication. As a complement to that population-based argument, this article will explore the ethical issues of placebo-controlled studies as they relate to individual psychiatric patients who may participate in them.

Case Vignette

JS was a 27-year-old man with a five-year history of schizophrenia who was brought to a university hospital emergency department by his parents with an acute psychotic exacerbation. According to his parents, JS had been living on the streets, refusing his parents’ repeated attempts to have him come back home, take his medication, or follow up with his outpatient county psychiatrist. That morning, JS called his parents in a highly psychotic state seeking their assistance to stop the voices, which he stated were “driving me crazy.” His parents took the opportunity to bring their son to the university hospital. The consulting psychiatrist who was seeing JS in the emergency room conveyed to his parents his plan to restart their son on an antipsychotic medication and transfer him to the county psychiatric hospital which, in that county mental health system, was the only facility funded to care for indigent psychiatric patients. JS’s parents were very displeased with this plan because during two previous exacerbations, JS was seen at the county facility’s emergency psychiatric unit, restarted on antipsychotic medication, and released after a 24-hour observation period. JS’s parents felt that their son’s poor adherence and outcome following those interventions could have been prevented with a more comprehensive intervention. The parents inquired about the possibility that their son could be admitted to the university hospital’s psychiatric facility as a participant in a clinical trial for a new schizophrenia medication, which they learned about from a flyer in the emergency room. However, when the consulting psychiatrist discussed this option with JS, he declined to participate because the study required participants to be inpatients for the first three weeks. JS’s parents were extremely disappointed when they learned their son declined, as they felt very strongly that participation in the study was his best treatment option. After a heart-felt discussion with his parents, JS agreed to consent to the study and was admitted to the inpatient unit to undergo screening. JS was enrolled and started on blinded study medication but after 10 days of being in the study he was not sufficiently improving based upon the formal discontinuation criteria in the protocol. He was, therefore, disenrolled from the study due to lack of sufficient efficacy. A meeting was held with JS and his parents in which the principal investigator of the study presented two options: stabilizing JS in the hospital using approved antipsychotic medication with all costs covered by the study sponsor or enrolling him in the open-label extension of the clinical trial in which he would receive the active investigational medication unblinded, with flexible dosing at the discretion of the investigator (no chance of placebo). He could continue in this open-label trial until the study medication became approved by the United States Food and Drug Administration (FDA) or its development was discontinued by the study sponsor. JS and his parents chose the open-label extension option and he was started on the study medication in the hospital. After 14 days, his psychosis was much improved and he was discharged with continued study follow up as an outpatient. He remained in the study, and was seen regularly at the university hospital for nearly two years until the study medication was approved. During this time, JS’s symptoms were well controlled and his functioning was improved sufficiently that he was able to obtain employment and move into an independent living facility.

Are Placebo-Controlled Trials Ethical?

When thoughtful people consider the question of whether it is ethical to conduct placebo-controlled trials for psychiatric disorders, such as schizophrenia, many tend to view the issue as a unidimensional consideration: treatment with placebo (clinical trial) versus treatment with a medication that has established efficacy (standard treatment). An implicit assumption that goes hand-in-hand with this perspective is that most patients with schizophrenia do not have the capacity to evaluate the relative risks and benefits of participating in a clinical study and thus are not capable of supplying legitimate informed consent. In the vignette above, which is based upon an actual situation that occurred in a clinical trial that I supervised, caring and well-informed parents lobbied strongly to have their acutely symptomatic son enrolled in a study that they knew included a placebo arm. They did so because they understood that participating in the clinical trial afforded their son care options that would not be available to him otherwise and that participating in the study offered advantages that they believed outweighed the risk represented by the possibility of being randomized to placebo. This vignette underscores the complexity involved in the treatment of patients with chronic mental illnesses and how it may be too simplistic to distill the debate about the ethics of placebo-controlled research into a comparison between treating cognitively compromised patients with placebo versus treating them with efficacious medications. Such a simplistic distillation ignores crucial, real-world details that need to be factored heavily into any bona-fide risk-benefit assessment. In the sections that follow, I will discuss some of the important factors that are often overlooked by critics of placebo-controlled studies. Since criticism of placebo-controlled studies is most strongly voiced in regard to patients with schizophrenia, I will focus on key issues related to this population.

Table 1 lists study features that should be considered when making a risk-benefit consideration about a patients participation in a clinical trial with a placebo arm.

Table 1.

Assessing the risk/benefit of a patient’s participation in a placebo-controlled clinical trial.

WHAT TO ASK	WHAT TO CONSIDER
What is the statistical risk of receiving placebo?	This can range from 50% in a single dose Phase II study to 20 or 25% in the typical Phase III study, depending on the number of nonplacebo treatment arms.
How experimental is the study medication?	Consider whether the investigational drug is a novel mechanism or one that is well established. Also, consider how much safety/efficacy data already exists for the investigational agent.
What degree of medical, psychiatric, and social care resources does the patient have without participation in the study?	The lower the level of resources a patient has, the more resource advantages a placebo-controlled trial may offer.
Does the study have an open-label extension that the patient can be enrolled in after placebo phase? Can the subject be enrolled even if he is discontinued early from the placebo-controlled study (e.g., due to lack of improvement)?	Open-label extensions studies offer patients the opportunity to receive psychoactive medication with intense monitoring for long periods with no risk of placebo.
Does the study provide for hospitalization during the initial period?	Hospitalization during the initial period provides a safe, vigilant environment to start patients on study medication and often provides a higher level of care than many study participants would otherwise receive.
Is there a washout? How long?	Some studies allow study medication to be started without a washout period or with a very short washout period.
Does this patient have treatment and adherence challenges under standard treatment conditions that may be alleviated in a study setting?	Patients who historically demonstrate poor treatment adherence may benefit from the paternal structure of a clinical trial and the increased motivation due to compensation.
How intense is the clinical monitoring in comparison to the likely level of monitoring the patient would receive if not participating as a study subject?	Clinical studies typically involve more frequent and in-depth follow-up visits than the standard of care, but a patient with access to good mental healthcare may benefit from intense clinical monitoring less than an indigent patient or someone with very limited access to mental healthcare.
Does the study provide resources to adequately stabilize any subject on standard treatment at the end of their participation?	Industry sponsors typically agree to cover costs involved in transitioning study participants back to standard treatment, including any needed hospital stay.
What is the reputation of the site conducting the trial and the institutional review board approving it?	Not-for-profit institutional review boards, such as those operating in academic institutions, tend to be more conservative in what studies they will approve and what ethical and safety features they require

Psychosis Does Not Imply a Lack of Capacity to Consent

It is a common yet erroneous assumption that certain psychiatric conditions, including schizophrenia, render people cognitively incapable of providing truly informed consent. Research into this subject has clearly revealed that while patients with active psychotic symptoms often have decrements in their consent capacity, they nevertheless frequently have the ability to understand the nature of clinical research protocols, including the potential risks involved, and can make rational informed decisions.² Furthermore, many schizophrenia patients who have difficulty adequately understanding the information presented to them in a standard informed consent approach may experience enhanced understanding when relatively simple educational remediation interventions are implemented into the consent procedure.³ Because psychiatric disorders such as schizophrenia can compromise capacity, it is the practice of quality research sites to objectively determine the ability of potential research volunteers to understand to what they are being asked to consent. For example, the University of California, San Diego, Brief Assessment of Capacity to Consent (UBACC) is a validated, brief instrument developed to screen for capacity and to identify potential patients who may require educational remediation.⁴

Risk of Placebo Is Often Overestimated

Though industry-sponsored clinical trials sometimes include only one treatment arm and one placebo arm, it is much more common for such studies to include 4 or 5 treatment arms, only one of which is placebo. Thus, in most industry studies, the chance of being assigned to a placebo is 20 or 25 percent. Moreover, results of decades of schizophrenia studies demonstrate that subjects randomized to placebo will, on average, significantly improve during the course of the study, though probably not as much as the typical subject assigned to an active treatment arm. In fact, improvement among schizophrenia subjects receiving placebo in research trials is so common that it represents a significant scientific hindrance to the ability of clinical studies to demonstrate efficacy for bioactive treatments.⁵ An analysis of the FDA database revealed that one-quarter of the studies comparing placebo to approved antipsychotics fail to show superiority for the active drug.⁶ Thus, on average, a subject randomized to receive placebo can be expected to exhibit an improvement in his or her symptoms during the course of the clinical trial. The reason for the high rate of clinical improvement among schizophrenia patients who receive placebo in clinical trials has been the source of consternation for clinical scientists.⁵ While the cause of this placebo improvement is the source of much speculation,⁵ the structural factors inherent in industry studies, such as enhanced clinical monitoring, may be important contributors.

Clinical Trials Often Provide More Intensive Monitoring Than Standard-of-Care Conditions

It is no secret that in the United States, a large majority of patients suffering from serious mental illness such as schizophrenia receive inadequate care. The community mental health system, the sole point of care for a majority of this population, is beleaguered. Chronically under-resourced relative to the mandate they are given, public mental health systems across the country wage a quixotic struggle to meet the basic mental health needs of their clients. Exacerbating the inadequacy of public treatment resources are the barriers to treatment presented by severe psychiatric disorders themselves. Because patients with schizophrenia tend to be poorly adherent and have difficulty accessing care, even when they are motivated, a large percentage of them go untreated for long periods of time, often for the majority of their adult life. The emergency room too often becomes the primary venue of treatment for this population. Not only do the psychiatric symptoms of the patients with schizophrenia go under-treated, but so do the medical conditions this population is prone to developing; these undertreated medical conditions then contribute to their shortened life expectancy.

In contrast, the level of care in a clinical trial is, by its very design, intensive. Industry-sponsored schizophrenia studies, for example, typically begin with a required inpatient period of 3 to 6 weeks, during which the patients are rigorously screened for eligibility through standardized diagnostic assessments and a medical workup. The cost for this inpatient component, like all other study components, is fully covered by the pharmaceutical sponsor. Few of the patients hospitalized as part of an industry trial would receive inpatient monitoring for that duration, if at all, under standard treatment conditions. The parents in the vignette in this article understood that their son would not have the opportunity to be stabilized in an inpatient setting if he was not enrolled in the research study. For many patients, the screening medical workup, which invariably includes bloodwork, physical examination, and electrocardiogram (ECG), is something many schizophrenia study participants have never previously received,even during s standard psychiatric hospitalization. On many occasions, in the course of studies I have overseen, the medical screening process has revealed active medical conditions, such as diabetes, hyperlipedemia, hypertension, and cardiac or liver function abnormalities, that were not previously recognized. Regardless of whether those revealed medical abnormalities make a subject ineligible for the study, the principles of good clinical practice⁷ obligates an investigator to inform a study subject about any significant medical finding and facilitate the appropriate follow-up tests or medical care. It follows that many study subjects get started on necessary diabetes treatment or antihypertensive medication as a result of consenting to participate in a clinical study.

After the required inpatient period, patients who are deemed to be doing adequately well can continue the study as outpatients. During this outpatient period, follow-up visits are typically weekly—much more frequent than the typical community standard. Patients who do not show up for a study visit will invoke a vigorous effort by the study team to locate them, which is not typically the case in community clinics. This is driven, in large part, by the pragmatic reality that every study subject is a source of vital data and his or her continued participation and involvement in the study is highly valued by the sponsor and investigative team alike. Study patients are often more adherent with study follow-up visits and with assigned medication, possibly because they receive more social interaction and positive reinforcement than they would under typical treatment conditions. A large part of subjects’ motivation for adherence comes from the fact that they are compensated financially for their participation. While this may be viewed as a cynical motivation, it does increase their attendance at follow-up visits and allows study patients to be closely monitored and maintained on study medication to an extent that would not occur under regular treatment conditions. The frequent follow-up visits in a clinical trial allow for an early detection of clinical deterioration. Patients whose symptoms seem to be worsening can be identified before their condition deteriorates. This allows for study psychiatrists to implement the necessary treatment changes that would forestall a full exacerbation. Contrast this with the very common scenario in which patients treated within a community mental health system who begin to experience a clinical deterioration go unnoticed because of the infrequency of their follow-up visits. Such deteriorations often do not get addressed until they require intensive inpatient treatment, often under involuntary conditions.

Many Placebo-Controlled, Clinical Trials Provide an Open-Label Treatment Opportunity

Another important factor to consider when evaluating the risk-to-benefit ratio for participation in a clinical study with a placebo arm are what opportunities the study provides for the patient after the placebo-controlled portion of the trial. Frequently, short-term, placebo-controlled studies (e.g., 3–8 weeks in duration) are yoked with longer-term, open-label studies that commonly continue for 6 months to 2 years. Participants who complete the short-term placebo study are eligible to participate in these open-label extension studies. Even participants who do not complete the placebo-controlled phase, for example due to lack of improvement, are often still eligible to enter the open-label study. During these open-label studies, the risk of being assigned placebo is eliminated as all patients receive the active investigational medication. The open-label studies allow the sponsor to collect long-term safety data, but also provide somewhat of a quid pro quo for study participants who have taken the risk of being assigned to placebo.

Well-Designed Clinical Trials Have Mechanisms to Minimize Risk to Subjects

There is no doubt that clinical studies designed with a placebo arm present a potential risk to patients with schizophrenia. A patient randomized to receive placebo may experience a worsening in his symptoms compared to treatment with established antipsychotic medication, although improvement on placebo treatment is likely as discussed earlier. Pharmaceutical companies have a strong ethical and economic incentive to avoid adverse outcomes in the clinical trials they sponsor. As a result, most well-designed industry studies incorporate features that mitigate the placebo risks. Furthermore, it is a routine component of industry-sponsored clinical studies that the pharmaceutical sponsor will cover all reasonable costs involved in safely transitioning study volunteers to standard approved treatments after they have completed the study or if they need to be terminated early for any reason. Thus, the parents of JS in this article’s vignette understood that even if he did not respond adequately to the treatment assigned in the clinical trial, he would be stabilized with standard medication, as needed, in the hospital at the expense of the pharmaceutical company sponsoring the study. It has been my experience that industry sponsors are significantly more generous than public or private health insurances with respect to approving requests by study investigators for hospital days and/or outpatient visits in order to stabilize study subjects on standard treatment after they are disenrolled.

In addition to the risk-attenuating features built into most placebo-controlled studies designed by pharmaceutical sponsors, many institutional review boards (IRBs), particularly those operated by academic research centers, require additional safety mechanisms before approving a placebo-controlled study in a high-risk population such as schizophrenia. At University of California, San Diego, the institutional imperative to foster medical research in an ethical and safe way led the UCSD IRB to form a working group in 2004, which I headed, that was mandated to develop guidelines for the design of all placebo-controlled studies at UCSD in high-risk patients such as those with schizophrenia. The guidelines that emerged from that working group are appended to this article.

Summary

In summary, placebo arms in clinical studies represent a risk to individual study participants since patients randomized to placebo are not likely to experience the degree of symptom improvement they would if treated with a bioactive medication. However, it is inaccurate to assume that, for the typical study candidate with chronic mental illnesses, the alternative to participating in a placebo-controlled clinical trial is effective treatment with approved medication. In reality, many study candidates would not receive effective treatment for a variety of reasons related to the structure of mental healthcare and to the nature of their psychiatric disease. Moreover, most industry-sponsored clinical trials, especially those under the regulation of IRBs with established guidelines for placebo studies, not only offer treatment features that are higher than the local standard of care that exists for the same patient population but also incorporate structural features that manage the potential risk of placebo. Thus, an argument can be made that a thoughtfully designed, placebo-controlled clinical trial offers a valuable option for many patients with chronic mental illness and their loved ones.

Appendix 1. Univerity of California, San Diego, Human Research Protections Program: Guidelines for studies involving placebo-only treatment in high-risk patients with psychiatric illness

A workgroup of the UCSD Department of Psychiatry, at the request of the Human Research Protections Program, reviewed and updated guidelines for the safe and ethical conduct of studies involving High Risk Patients with Psychiatric Illness in October, 2004. For the purpose of these guidelines, the workgroup defined “High Risk Patients” as patients diagnosed with a major psychotic illness such as schizophrenia, bipolar disorder, or schizoaffective disorder. These diagnosis fall into the category of psychiatric disorders that the Department of Veteran Affairs refers to as “Significant Mental Illness”. These guidelines should be followed when preparing research study applications to the UCSD Human Research Protections Program (HRPP).

Guidelines

In general, subjects who are assigned to a potential placebo-only treatment trial should be observed in a closely supervised in an inpatient or residential (in house) crisis center during the high-risk period. If the protocol allows for subjects to be discharged after the high-risk period then the following criteria should be addressed in the UCSD HRPP application.

1. Subjects have been observed as an inpatient for at least 1 week after being randomized to potential placebo-only treatment (no less than 1 week) to determine that they are tolerating study medication well and are reliably compliant with treatment and showing no evidence of worsening of symptoms. This must be reflected in each subject’s medical chart.

2. Subjects will be free of active self-harm ideation or behavior for at least 1 week before discharge and this is documented in each subject’s medical record.

3. Subjects will meet objective criteria (e.g. based upon the rating tools being used in the study such as PANSS, BPRS, YMRS, CGI, etc. developed by the PI and approved by the review committee) that the subject is appropriate to be discharged and to continue the study as an outpatient.

4. The subject’s appropriateness to continue participation in the study as an outpatient will be independently assessed by two qualified psychiatrists and/or psychologists not connected to the study and formally documented in the subject’s medical chart.

5. Prior to discharge the subject must identify an individual (close friend or relative) who sees him/her on a daily basis and who consents to act as a contact person. The subject must be willing to be monitored by the contact person and be willing to have the study staff and contact person regularly communicate about the subject’s safety and appropriateness to continue in the study. The contact person will be educated by study staff regarding what symptoms and behaviors to look for which may indicate a worsening in the subject’s illness and represent a potential risk to the subject’s safety. The contact person must be willing to contact the study staff should he/she see evidence of such symptoms or behaviors or if the subject appears not to be compliant with medication, expresses concerning thoughts or plans, leaves his/her stable place of residence, or engages in behavior that is contrary to the study and/or may potentially worsen the illness (e.g. illicit drug use). The contact person must agree that the subject is ready for discharge. This agreement will be formally documented in the subject’s medical chart.

6. Subjects must have an identified stable place of residence where they can be contacted by telephone and where the contact person can monitor him/her. This must be in the Contact Person Consent Form and approved by a UCSD Institutional Review Board (IRB).

7. Prior to discharge, the study staff will provide the subject and the contact person written instructions about the study including the name of the study and the PI, the manner in which study medication/s are to be taken, the dates and location of follow-up visits, and phone numbers where the subject can reach study staff 24 hours per day for questions and emergencies. A copy of these instructions will be placed in the subject’s chart.

8. A plan must be submitted and approved by the IRB that details how the research staff will follow the subject, and must include the following: information on frequency of follow-up assessments, the plan if the subject does not show up for a scheduled visit, and the plan if the subject cannot be reached by study staff or the contact person.

9. An outline must be provided which delineates the criteria to be used in determining that the subject is no longer appropriate for the study. For example, most multi-site pharmaceutical industry-sponsored studies allow subjects to be discontinued if “in the judgment of the principal investigator” continuing in the study would be harmful and/or inappropriate for the subject. This outline must provide specific, objective criteria (e.g. based upon the rating tools being used in the study such as PANSS, BPRS, YMRS, CGI, etc.) to be used in determining whether a subject who is potentially on placebo-alone treatment should be discontinued from a study due to evidence of worsening or lack of improvement in symptoms.

10. A plan must be provided for rescue of a subject who has significantly worsened during the course of the study and is now at greater risk. Elements in this plan should include a mechanism for hospitalizing patients readily, and/or discontinuing patients from study medication and restarting standard treatment. How the costs will be covered for stabilizing such a subject should be addressed. Note that it is UCSD policy in sponsored studies that the sponsor, not the subject or the subject’s insurance provider, covers all reasonable costs for patients who need stabilization treatment.

Additional Follow-up Visit Criteria

The research plan should include the provision that study staff will make frequent follow-up assessments of subjects who are enrolled in placebo-alone controlled studies as outpatients. During these follow-up assessments the following should be documented:

1. Subject’s relevant psychiatric symptoms

2. Subject’s ongoing appropriateness to continue as an outpatient in study based upon criteria 1–10 above

3. Evidence that the subject maintains his/her consent to continue participating in the study and is competent to do so

4. Status of the contact person.

Criteria for Termination of Participation

The research plan should include the provision that the PI will make available for a subject who leaves the study a means for him/her to safely discontinue study medication and receive appropriate follow-up treatment.

The following must also be addressed in placebo-alone, controlled studies:

1. If subjects are being taken off their medication, the consent must clearly state that they will be on no medication (placebo alone and/or washout period) for a portion of the study. The number of weeks that a subject may be on an investigational drug must be specified, and it must be clear that this drug may not control their illness. It must be explicitly stated that there is a good chance that their disease may worsen during the trial.

2. A subject’s capacity to consent must be objectively demonstrated. Study staff must submit a copy of the assessment tool that will be used to determine the subject’s understanding of key elements of the study protocol, including design, risks, expectations and alternatives. Further information on decisional capacity assessment is available at http://irb.ucsd.edu/decisional.shtml