CTSI Blogs

The Research Ethics Blog by Bernard Lo, MD

Protecting Participants in Clinical Trials

Meta-analyses of clinical trials of antidepressants have found a trend towards increased suicidality in children and adolescents receiving selective serotonin reuptake inhibitors (SSRIs). These meta-analyses raised concerns about conflicts of interest because published trials showed that the drugs were effective and safe in these circumstances, but when unpublished studies carried out by the manufacturers were also considered, the drugs were shown to be ineffective and possibly unsafe. The report Conflict of Interest in Medical Research, Education, and Practice provides other examples of failure to publish negative results or serious adverse events in industry-sponsored clinical trials. See summary (PDF 304KB). Here we discuss another ethical issue: How to protect participants in clinical trials if their condition worsens?

It is no surprise to clinicians that some patients experience worse depression despite treatment. Moreover, because they are depressed, participants may not appreciate that their condition is worsening or be able to seek help. For example, many persons with mental illness have poor access mental health care. At various stages of a clinical trial, steps might be taken to protect such participants.

During review of the trial: Many trials of antidepressants are designed by drug companies and do not undergo external scientific peer review. What should be responsibility of the IRB be with regard to the potential for suicidality in trial participants?

At enrollment: What exclusion criteria for participants might reduce the risk that patients may become suicidal or, if they do so, have a poor outcome? How should concerns about poor access to psychiatric care be addressed in the exclusion criteria?

In real time during the trial: What procedures should the protocol specify for indentifying and responding to participants who become suicidal? What “safety-net” measures should be in place? How should the development of suicidality be handled as an endpoint? (Note that the usual endpoints in depression clinical trials are change in symptom scales and the percentage of patients in remission.)

After cases are identified: If cases of suicidality are identified, how should the medical monitor or the Data and Safety Monitoring Board be notified? What steps should they take when evaluating such cases? Note that there are several issues to consider. One is whether the response of the investigators was adequate to minimize the risks to participants, and another is whether such cases are more common in one arm of the trial.

However, such protections for participants may have undesirable consequences for the trial. IRBs rarely carry out in-depth scientific review. Excluding persons at increased risk for suicide may compromise the power, efficiency, or generalizability of the trial? (Note that antidepressants may have greater effectiveness in persons with more severe symptoms.) Putting in place safety-net measures may increase the complexity and cost of the trial. In your view, what steps should researchers and sponsors take in clinical trials of antidepressants to protect participants whose condition worsens during the trial?

Comments

Why can't suicidality be a cause for withdrawal from the trial? Is there evidence to suggest that those with severe depression become suicidal before getting better on medication? It seems like they should withdraw and be treated for their suicidal ideations.

The intro to this case raises a very significant biomedical research issue in that trials run by Big Pharma companies are, in many cases, profit driven and thus have few vestiges of academic ideals like the pursuit of unbiased understanding. One the byproducts this approach to clinical research is that negative results and adverse events are sometimes downplayed or unreported in hopes of salvaging a potential source of revenue for the company. I think oversight mechanisms need to be in place to expose companies that intentionally conceal studies that undermine therapies that they intend to bring to market. In terms of the more central issues in this case: when serious adverse events, i.e. suicidality, are potentially evoked by active treatment, the safety of the study participants needs to be the investigators first priority. In particular, suicide ideation questionnaires and/or psychological evaluations need to take place before, during, and after the trial to identify changes in mood and risk of self harm. While I think excluding any subject with a history of suicidal thoughts would be most conservative, this would indeed limit the generalizability of the study, especially when regular assessments of mental state could limit much of the study risk. Within the framework of regular subject assessments, the Data and Safety Monitoring board needs to be notified when any subject exhibits increased suicide ideation from week to week. Subsequently, based on circumstance, judgements should be made as to whether a subject needs to be removed from the trial and this decision should be based upon predetermined parameters. Again, this approach may yield decreased sample size and power, but these are necessary ethically based precautions that are more important than attaining a statistically significant result.

I very much agree with what this blogger has posted. Ethically based precautions should clearly trump potential financial and/or statistical gains. And an issue like suicidality should not be taken likely given that a life is at stake. Exclusion criteria should be included and strict monitoring established to ensure that the chances of this tragic adverse effect are minimized as much as possible.

I agree that sponsors many times intentionally delay for years study findings that are negative and rather publish only positive results, or interpret negative results as favorable or not significant for fear from loss of revenue, after significant investment in years of drug development and clinical research. Another reason for being cautious with publication is fear from misinterpretation by the media and public, exaggeration, media scandalism, which may harmfully effect the introduction or may result in withdrawal of a new drug, that may be effective. However, when it comes to life threatening adverse events, transparency is a must. Therefore it should be common interest that data be transparent in the scientific community, regulators and IRBs, as well as sophisticated statistical methods should be used and experts should be available to assess the risks and benefits of a new drug in a clinical trial.

I agree that sponsors many times intentionally delay for years study findings that are negative and rather publish only positive results, or interpret negative results as favorable or not significant for fear from loss of revenue, after significant investment in years of drug development and clinical research. Another reason for being cautious with publication is fear from misinterpretation by the media and public, exaggeration, media scandalism, which may harmfully effect the introduction or may result in withdrawal of a new drug, that may be effective. However, when it comes to life threatening adverse events, transparency is a must. Therefore it should be common interest that data be transparent in the scientific community, regulators and IRBs, as well as sophisticated statistical methods should be used and experts should be available to assess the risks and benefits of a new drug in a clinical trial.

Clearly, as a fellow blogger pointed out, the main difficulty with industry sponsored research is the financial motivation. Though in an ideal world, pharmaceutical companies would place patient outcomes over their shareholder's profits, we all know that in the real world of capitalism and economic lability, this is not always the case. Thus, I think one way of increasing patient safety in industry sponsored research is for the government to play a bigger role in the initial stages of study design. If a government organization like the FDA was intricately involved in the planning and organizing of industry funded research, then there would be more visibility and accountability for negative patient outcomes. Furthermore, this could improve clinical knowledge by likely increasing the amount of negative publications. However, this approach would come with the disadvantage of increasing redtape and delaying drug development. In regards to suicidality in specific, I believe that it should be a serious adverse event closely monitored at frequent intervals by the DSMB, and that unexpected increased rates of this outcome should be immediately evaluated and potentially lead to trial termination. Suicide after treatment for depression is akin to fatal intracerebral hemorrhage after treatment for acute ischemic stroke. There have been many agents tested in ischemic stroke patients that have been shown to increase recanalization rates; however, only TPA has been shown to have a net positive effect for these patients, as all the other agents have been offset by considerably increased rates of symptomatic hemorrhage. Similarly, SSRI's and other more novel antidepressants may improve depression scales but these benefits need to be clearly weighed against potentially increased catastrophic events such as suicide. Thus, study researchers and reviewers should be careful to pick the most appropriate and relevant endpoints that truly reflect the outcomes of interest.

Establishing exclusion criteria is a way to enroll a target population with the right characteristics into a trial as well as to protect human subjects from potential adverse events. Although these measures may have undesirable consequences for the trial, this is a secondary concern and the safety of human participants should obviously take a higher priority. This is especially true for vulnerable populations like the one addressed in this ethics discussion. However, excluding patients at high risk from a trial is not always the best way to protect patients, yet all patients should be informed of possible adverse events per IRB protocol. What if a new drug was under investigation that could possibly reduce the risk of suicidality, then the investigators may not want to exclude these patients from the study. In terms of access to psychiatric care, this could be built into the follow up of the patient as a possible recommendation if they continue, or move, to a high risk characterization and a risk incurred by the investigators. For example, for patients of continued high risk of suicidality (ie, considered a threat to themselves) a referral to psychiatric care could be built into the protocol. How these patients are counted in terms of loss to follow up, drop out, or an unsuccessful case is up to the investigators

I think one of the main issues of this case is failure to report negative results. This neglect of publishing negative results is potentially very harmful to future patients and practice guidelines based on only positive results. In this particular study that includes suicidality as an outcome - measures should be put in place for close psychiatric follow up and means of hospitalization or medication adjustments for patients who express this outcome. Patients should not be excluded from the trial for severity of depression or current suicidality at the onset of the trial, because ultimately this is one of the most important outcome measures in a trial of depression - one of the outcomes that will ultimately decrease long term mortality from depression. I think as a responsible study design therefore - patients should be followed closely by psychiatrists in order to protect and closely monitor for depression as other patients would be treated in acute episodes of depression. In addition, my impression of new onset suicidality is that it may be a consequence of patients actually starting to feel better - have more energy, motivation, etc to actually pursue an action - so long term mortality and suicide attempts should be recorded closely in study participants because new onset suicidality may be a symptom on a path to recovery or may reflect worsening depression. Nevertheless it is a very important outcome point - and should be reported in all trials of major depressive disorder.

Drug sponsored research on anti-depressants with clinical endpoints related to depression scores must deal with the very real and serious risk of worsening depression and suicidality. The patients enrolled in these trials must be protected and there must be very comprehensive methods set up to ensure safety. I do believe that the individual institution IRBs should take a major role in ensuring patient safety. One way they could do this would be to require drug companies to meet strict reporting criteria before they would agree to enroll patients for the trial. This would also include making sure that enrolled patients have good access to psychiatric care (even if they don't have insurance, for example). With regard to exclusion criteria, on the one hand, it would be important to include patients who have a prior history of suicidality as this would reflect the true population for which the clinical trial results are to be applied (external validity). However, this could potentially put patients at unneccessary risk. Thus, I would include these patients, but make sure that patients are contacted regularly or depression is assessed at short time intervals to ensure that worsening depression is not missed.

It is the responsability of the IRB to ensure that patient safety measures are enacted, particularly when known serious adverse events are associated with our intervention. These measures often add cost and complexity to the study design, and therefore may not be embraced by researchers. In the case of SSRIs with published observational data suggesting increased risk of suicide, it becomes the researcher's responasbility (demanded at the level of the IRB) to provide behavioral health services to participants. Patient blinding may require behavioral health services provided to both placebo and intervention groups, but these complex measures ensure that study convenience and affordability are trumped by ethical considerations of patient safety.

I agree that the IRB must take responsability for patient safety measures. In the case of SSRIs, we have published data suggesting an increased suicide risk. Given this extreme form of a serious adverse event, the IRB must ensure that the study design addressess this issue and makes additional effort to protect their participants. In this case, interval behavioral health sessions should be incorporated into the protocol, and may be best administered to both placebo and intervention groups to ensure patient blinding. This would clearly affect the study in several ways that may not be welcomed by the researchers. Namely 1) the study question would be modified, now comparing an outcome between behavioral health + placebo groups versus behavioral health + SSRI groups 2) the cost of adding behavioral health may exceed available funding 3) patient adherence may plummet when participants are forced to attend behavioral health sessions. While this modified study may have become a complex, expensive trial, the IRB must ensure that our intended studies are improved to meet ethical standards. Posted by Monika Sarkar for Clinical Trials

I agree that exclusion criteria are key in the trial design process. Of course one can not predict whether a subject will or will not attempt suicide, but it is something that can be looked at statistically to see with demographics pose the highest risk. Additional measures should be taken during the trial to ensure all subjects are protected (even if it means excluding certain subjects).

Yes. I agree. Moreover, direct verbal communication is needed with the participant and the staff. This way it can be assessed whether the participant might be suicidal.

I like Monica's idtea of having both the active and control arms include a behavioral intervention to help ensure patient safety. I don't think excluding subjects who may become suicidal is reasonable unless that class of subjects is readily identifiable (e.g., children) and a blackbox warning can be included on the drug label to help preclude its use in these subjects in a real world setting. Instead, I think the DSMB will have to play a very important role in regularly monitoring the safety data.

I like Monica's idea to have a behavioral intervention in both the active & arms to help ensure patient safety. I don't think excluding patients at risk for suicidality is reasonable, unless they are clearly identifiable (e.g., children) and a black box warning can be placed on the drug label to preclude their use in these populations. Thus, ongoing monitoring by the DSMB would be critical to ensure safety in this trial.

I think it's incumbent upon the trialists and DSMB members to incorporate appropriate monitoring of patients for symptoms of worsening depression. This could take the form of questionnaires or, given that depressed patients may not fill out such forms, a better approach would be regularly scheduled checkup phone calls by a study coordinator. Clearly this adds to a study's cost, but prevention of suicide in an antidepressant trial is of paramount importance. Checkup calls would also potentially provide more reliable data points, thus also improving data collection overall.

It is best to screen out patients that could be suicidal through certain exclusion criteria. One might have to ask other people who are close to the participant if he or she is not afraid to state their emotions verbally. Direct communication with the clinic staff and the participants is key to protect participants in a clinical trial.

Exclusion criteria is important when assessing if the participant may become suicidal. One should ask the people close to the participant if they think he or she is mentally stable to participate in a clinical trial. Moreover, the key to protect participants is to have direct communication with the clinical staff and the participants. The participant must be comfortable and not afraid to articulate about his or her physical and emotional problems to the staff.

The safety of the research subject is critical in any trial. During the trial, the IRB must review significant information regarding the risk profile of the proposed research study. In my opinion, the Drug Company or sponsor should add supplementary plan or application in the protocol to handle unanticipated adverse effects. The sponsor should appoint an independent monitoring committee to review and report rare/serious adverse events. The monitor must ensure that there is clear understanding of AEs and reporting responsibilities on the part of site PIs and site staff. Furthermore, collection of drug related AEs must continue beyond approval of the drug for marketing and commercial use.

I see three issues: 1. Risk and benefit optimization, to ensure that the benefits of the intervention (SSRI) outweigh the risks; 2. Vulnerable participants in the study due to their age (children, adolescent) and their higher susceptibility to serious adverse events; 3. Conflict of interest (for-profit organization withholding negative findings from public) resulting in research misconduct, undermining the scientific integrity of research and trust of public in clinical research. During the review, IRBs must ensure that the investigators in the protocol will set up screening for susceptible suicidals (via screening or medical history), strengthen informed consent by educating about the life threatening risks of the study, ensure that protocol includes interim monitoring of especially susceptible participants to become suicidal, and that adequate extra medical treatment, psychological care is provided free to them throughout the study. IRBs must ensure greater data transparency, make investigators and sponsors promise to disclose all, including negative data, if unwilling, they must inform participants before enrolling them. At enrollment study volunteers with medical history of previously documented suicidal attempts or thoughts should be identified and excluded to avoid poor outcome, if access to 24/7 psychiatric care (e.g. hotline) can not be provided. During the trial susceptible participants should be followed up more extensively, regular visits or calls by study nurse, online questionnaires should be used, to measure symptom scales (mean depression level changes from baseline), and interim monitoring should be ongoing. After suicidal cases are identified, immediate removal of participant from the trial should be decided by board, and medical help should be provided. It is necessary to break the code (unblind), to see if this was an effect of the active drug and whether this is cumulative in one arm of the study.

I see three issues: 1. Risk and benefit optimization, to ensure that the benefits of the intervention (SSRI) outweigh the risks; 2. Vulnerable participants in the study due to their age (children, adolescent) and their higher susceptibility to serious adverse events; 3. Conflict of interest (for-profit organization withholding negative findings from public) resulting in research misconduct, undermining the scientific integrity of research and trust of public in clinical research. During the review, IRBs must ensure that the investigators in the protocol will set up screening for susceptible suicidals (via screening or medical history), strengthen informed consent by educating about the life threatening risks of the study, ensure that protocol includes interim monitoring of especially susceptible participants to become suicidal, and that adequate extra medical treatment, psychological care is provided free to them throughout the study. IRBs must ensure greater data transparency, make investigators and sponsors promise to disclose all, including negative data, if unwilling, they must inform participants before enrolling them. At enrollment study volunteers with medical history of previously documented suicidal attempts or thoughts should be identified and excluded to avoid poor outcome, if access to 24/7 psychiatric care (e.g. hotline) can not be provided. During the trial susceptible participants should be followed up more extensively, regular visits or calls by study nurse, online questionnaires should be used, to measure symptom scales (mean depression level changes from baseline), and interim monitoring should be ongoing. After suicidal cases are identified, immediate removal of participant from the trial should be decided by board, and medical help should be provided. It is necessary to break the code (unblind), to see if this was an effect of the active drug and whether this is cumulative in one arm of the study.