32

Data Management Committees and IRBs/Ethics Committees

Over the years, in addition to Investigational Review Boards (IRBs), the concept of a separate and additional independent group to monitor the safety of clinical trials has developed. This group or function is known under several names, including Data Monitoring Committee (DMC), Data Safety Monitoring Committee (DSMB), Data Safety Board, Clinical Trial Safety Monitoring Committee/Board, and others. This was codified as a draft guidance by the U.S. Food and Drug Administration (FDA) in 2001 (Web Resource 32-1) and by the European Medicines Agency in 2003 (Web Resource 32-2).

As stated in the FDA document: “A DMC is a group of individuals with pertinent expertise that reviews on a regular basis accumulating data from an ongoing clinical trial. The DMC advises the sponsor regarding the continuing safety of current participants and those yet to be recruited, as well as the continuing validity and scientific merit of the trial.”

The EMA definition is similar: “A group of independent experts external to a study assessing the progress, safety data and, if needed, critical efficacy endpoints of a clinical study. In order to do so a DMC may review unblinded study information (on a patient level or treatment group level) during the conduct of the study. Based on its review the DMC provides the sponsor with recommendations regarding study modification, continuation or termination.”

The sponsor or creator of the DMC can be the holder of the IND or equivalent, a company or government agency, or any individual or group to whom the sponsor delegates authority for decision-making, including the study steering or executive committee, the contract research organization (CRO), or the principal investigator. The presence of a DMC, however, is additive to the safety precautions in the trial. All legal and regulatory obligations in all jurisdictions must still be carried out by the investigator (patient protection, AE reporting, etc.) and sponsor (expedited reporting, signaling, etc.).

The DMC must be independent, meaning that no member has any personal basis for preferring the trial outcome to be in one or the other direction, and no member has any ability to influence the trial conduct in a role other than that of a DMC member.

The committees should contain at least three members and include such expertise as

• Clinical medicine (appropriate specialty)
  
• Biostatistics
  
• Biomedical ethics
  
• Basic science/pharmacology
  
• Epidemiology/pharmacovigilance—drug safety
  
• Clinical trial methodology
  
• Legal
  
• Patient advocate/community representative

Ideally, there should be geographic representation, especially in international trials, demographic representation on the committee relevant to the trial (race, gender, age), personalities amenable to consensus development, reliability and time to attend meetings, and prior DMC experience. The appointment to the committee should be made by the sponsor, and, for government-sponsored trials, members should be acceptable to the health agency and the investigators. Expenses and honoraria should be paid by the sponsor, and the members should be independent and free of conflicts of interest. The DMC should not have any representation from the industry sponsor, study investigators, or individuals who stand to gain or lose financially from the study outcome, such as major consultants or investors in the sponsor or a competitor.

A DMC is needed in general for (FDA criteria):

• Large, randomized multisite studies that evaluate treatments intended to prolong life or reduce risk of a major adverse health outcome such as a cardiovascular event or recurrence of cancer
  
• Any controlled trial of any size that will compare rates of mortality or major morbidity
  
• When DMC review is practical
  
• When DMC review helps ensure the scientific validity of the trial
  

  EMA criteria:

  
  
• In case of life-threatening diseases, usually the implementation of a DMC is indicated from an ethical point of view.
  
• Certain patient populations (even if trial is in a noncritical indication): pediatric and mentally disabled patients.
  
• Prior knowledge or strong suspicion that a treatment under consideration has the potential to harm patients (even though it will be eventually more effective than other available treatments).
  
• Preplanned interim analyses for early stopping (either for futility or for positive efficacy) or in case of complex study designs in which a possible modification of the study design based on unblinded interim data is intended. In such a situation, the use of an independent DMC gives more credibility to the process.
  

  Some trials should almost always have a DMC:

  
  
• If more than one investigational drug is being used in a trial
  
• Trials where early stopping for efficacy is considered
  
  
  
  • Treatment reduces mortality or major morbidity
    
  • Treatment reduces toxicity, cost, or other important secondary factors while maintaining efficacy against mortality/major morbidity
  
  
• Trials raising special safety or ethical concerns
  
  
  
  • Early AIDS vaccine trials
    
  • Gene therapy trials
    
  • Trials in especially vulnerable populations
  
  
• International and multicenter trials
  
• Phase III confirmatory trials
  
• Phase IIb test-of-concept trials
  
• Trials where review by independent experts would optimize patient safety and the scientific integrity and credibility of the trial
  
• Politically sensitive or highly emotional trials
  

  When a DMC is generally not needed:

  
  
• Not required or recommended for most clinical studies (FDA)
  
• Trials at early stages of product development, for example, phase I (FDA)
  
• Trials addressing lesser outcomes, such as relief of symptoms, unless the trial population is at elevated risk of more severe outcomes (FDA)
  
• A clinical study that can be performed in a short time frame that does not allow for appropriate preparation of information for a DMC (EMA)
  
• Clinical studies in noncritical indications where patients are treated for a relatively short time and the drugs under investigation are well characterized and known for not harming patients (EMA)

The DMC will have several functions during a trial, including the rapid identification of any safety problems; of logistical problems, such as inadequate accrual, undesirable distribution of baseline characteristics, excess dropouts, or noncompliance; of the continued feasibility of the trial as designed; and of whether the trial objectives have been met and the trial terminated early.

The committee must maintain full confidentiality, as they will usually receive unblinded data. Health agencies typically expect that confidentiality of the interim data will be maintained even if the DMC interacts with the sponsor or trial investigators to clarify issues relating to the conduct of the trial, potential impact on the trial of external data, or other topics.

To these ends, the DMC will review and approve the study protocol, assess the study conduct, evaluate accumulating data for both safety and efficacy, recommend termination or continuation of the study, recommend modifications of the study (including the informed consent), and recommend additional safety or efficacy analyses if appropriate.

The DMC must have a written charter that is drafted and approved by the sponsor and DMC members. The health agency may in some cases also be involved in the drafting and approval. The charter must include the schedule and format for meetings, including unscheduled meetings if a safety problem occurs, the format for presentation of data, the specification of who will have access to interim data and who may attend all or part (closed sessions) of the DMC meetings, procedures for assessing conflict of interest of potential DMC members, the method and timing of providing interim reports to the DMC from the sponsor or data-gathering group, the definition of a quorum for decision-making, and details on how the committee will meet (webinars, teleconferences, etc.; see DAMOCLES Study Group, Lancet 2005;365:711–722).

A key component (though not “officially” a member) of the DMC is the statistician or statistical group, who prepares, analyzes, and presents interim trial results to the DMC. The statistician may be from the company or sponsor or may be from an external CRO or independent consulting firm. The statistician will usually have the blinding key and will unblind study results (see below).

Each DMC meeting usually has an initial open session, followed by a closed session and then a wrap-up open session. At the initial “open” session with DMC and the sponsor, nonconfidential matters are discussed, including the clinical trial program, the status of recruitment, baseline patient characteristics, ineligibility rate, accuracy and timeliness of data submissions, and other administrative data and plans. The closed session is attended by only the DMC members and the (sponsor or outside) statistician who prepared and is presenting the interim analyses to the DMC. After the closed session, a wrap-up open session may be held with the sponsor to relay any recommendations from the DMC. The DMC has the option of conducting an “executive” session with no participants other than DMC members (that is, no sponsor representatives at all), though this is usually not needed.

There has been controversy over whether the DMC should see blinded or unblinded data. Most feel that DMC members should have access to unblinded data to ensure their ability to make accurate risk–benefit assessments. If this is the case, then printed reports of unblinded interim analyses should be supplied to the DMC members several days (at least) before the meeting to allow for adequate review of the information. Sometimes a partial unblinding is done, whereby treatment codes are given for each treatment group (e.g., drug X, drug Y, rather than the names of the drugs). The sponsor should supply the data in a reviewable form. For large studies, some sponsors provide the DMC with listings of individual patients’ data, often in a very non-user-friendly way. When there are hundreds or thousands of patients each with multiple labs, exams, and so forth, totaling millions of data points over thousands of pages, it is hard to “eyeball” the data and find trends or issues. Excellent software is available for examining trends and outliers. Many DMCs also want to see the CIOMS I/MedWatch forms for all serious cases or at least all expedited cases.

There is a “statistical downside” to multiple reviews of interim data that the sponsor may perform in some studies, namely, that repeated statistical comparisons of event rates between the treated and control groups increases the “false-positive” rate if P values are not adjusted for multiple testing. To put it another way, the more interim data reviews that are done, the more likely that a type I error will be found. At the P < 0.05 level, the probability of nominally significant results occurring with three interim analyses is about 10% greater, and with 10 interim analyses about 20% greater (see McPherson, Statistics: the problem of examining accumulating data more than once. New Engl J Med. 1978;290:501–502). To get around this problem, there are various options available that basically increase the P value required for significance (e.g., requiring a level of 0.018 at each interim and final analysis or 0.001 at each interim analysis and 0.05 at the final analysis). No matter what statistical techniques are used, the DMC must not rely on statistical analyses alone to reach a conclusion. Clinical judgment must also be used to evaluate the safety and efficacy of information.

The DMC will compare AE rates in each treatment arm to see whether there are any major imbalances of concern and which seem to be due to the intervention rather than the disease or confounding factors (e.g., all the patients in some Northern Hemisphere centers come down with a seasonal influenza). The DMC may also want to see specific patient information for critical or important cases (e.g., SUSARs, acute liver failure, all patients where the blind is broken by the sponsor or investigator for safety reasons). Again, this calls for good software and reporting tools to allow the DMC to see the data it needs.

The DMC has three options after each meeting. All meetings are minuted.

1. Continue the trial unmodified.
   
2. Modify the trial:
   
   
   
   • Drop a treatment arm or subgroup.
     
   • Modify the treatment dose/schedule.
     
   • Add additional safety tests (e.g., electrocardiograms) at screening or during the trial.
     
   • Modify consent/investigators’ brochure.
   
   
3. Terminate the trial early due to:
   
   
   
   • Serious safety issues.
     
   • The efficacy question has been answered and the results are truly compelling, with the risk of a false-positive conclusion acceptably low.
     
   • The hypothesis is no longer relevant or the hypothesized benefit cannot be achieved.

There are no rigid stopping rules. Rather, the committee must exercise judgment. General rules of thumb have developed. It is usually appropriate to demand less rigorous proof of harm to justify early termination than would be appropriate for a finding of benefit.

Interestingly, there is an ongoing and probably unanswerable ethical conflict with DMCs, namely, where does their primary responsibility lie?

• Perspective 1: The DMC’s primary responsibility is to the specific patients enrolled in the trial. This implies that the DMC should stop the trial early if further results are unlikely to change the conclusion based on interim data. That is, do not put the individual patients at further risk or subject them to placebo for longer than the minimal time necessary.
  
• Perspective 2: The DMC’s primary responsibility is to the entire “patient horizon” and the practice of medicine in general. In this case, the DMC would stop the trial early only if the results are sufficiently persuasive to effect changes in medical practice based on limited data.

No correct answer fits all cases.

The DMC is usually advisory to the sponsor, and the sponsor may reject the DMC’s comments, though this will obviously produce controversy and multiple ethical issues. The DMC results must be communicated to the health agency and the IRBs in any case.

In practice, for large multiyear, multicenter studies, DMCs have proved very useful when well constituted and well run. Logistics can be costly and complex if there are thousands of patients and several studies going on at the same time. It is likely that more and more DMCs will be used. Some are even considering DMCs or their equivalent in the postmarketing setting, looking at the spontaneous data coming into the sponsor or Marketing Authorization (MA) holder. Some have used DMCs in early phase I trials also.

Stay updated, free articles. Join our Telegram channel

Join