OR WAIT 15 SECS
Jill Wechsler is ACT's Washington Editor
A proposed policy on Data Monitoring Committee functions and operations raises important issues about the conduct of studies and the evaluation of data.
To enhance its ability to protect the safety of clinical trial subjectsand to ensure the accuracy of data developed to support New Drug ApplicationsFDA set out to clarify its policy on the use of Data Monitoring Committees. Also known as Data and Safety Monitoring Boards, DMCs oversee safety issues and track treatment effects during the course of a clinical study.
FDAs draft Guidance for Clinical Trial Sponsors on the Establishment and Operation of Clinical Trial Data Monitoring Committees (issued November 2001) describes the roles, responsibilities, and operations of DMCs. The agency is seeking public comment on how the policy might be revised and implemented. To this end, it held a public meeting 27 November 2001 to hear from affected parties. The guidance applies to clinical studies of drugs, biologics, and medical devices, whether sponsored by industry or government. FDA hopes to finalize the policy by the end of this year.
Many members of the research community consider the FDA draft document overly prescriptive, particularly for investigators at academic research centers, which increasingly participate in clinical trials for FDA-regulated products. Although agency officials insist that a guidance is nonbinding and merely reflects the agencys current thinking, researchers and sponsors maintain that the proposal will become the gold standard for both commercial and academic research. With FDA and the Department of Health and Human Services (HHS) Office for Human Research Protections (OHRP) shutting down research organizations that fail to comply with the rules, few study sponsors are likely to ignore agency guidance in this area. OHRP director Greg Koski commented at the November meeting that FDAs draft guidance is not a fait accompli but aims to stimulate discussion. He sees this effort to define the way DMCs are constituted and interact with institutional review boards (IRBs) as part of the larger task of remodeling the federal system for monitoring clinical research. Developing guidance on DMCs, he says, should impose a greater level of uniformity on government and commercial clinical research.
Study sponsors have used DMCs for years, primarily to oversee large, randomized multicenter clinical studies supported by grants from the National Institutes of Health (NIH) or sponsored by the Veterans Administration. The aim is to have an independent panel of experts review accumulating data on study safety, efficacy, and conduct, particularly for trials with mortality or major morbidity endpoints, to provide data for objective interim assessments. The DMC advises the sponsor on the continuing safety of trial participants and the validity and scientific merit of the study.
FDA regulations require DMC use only for clinical trials involving emergency research where informed consent is not feasible. Such studies usually involve life-threatening situations, and oversight by an independent data monitoring committee aims to provide added protections to subjects. References to DMC operations also appear in several guidance documents on clinical trial conduct developed through the International Conference on Harmonisation (ICH). The ICH Guideline for Good Clinical Practice suggests the use of such independent panels to assess the progress of a clinical trial and its safety data in order to recommend whether the sponsor should continue, modify, or stop a trial. Other ICH policies describe DMC composition and operating procedures.
In the last few years, sponsors of commercial trials have used DMCs more often, particularly for studies with major endpoints, explained Susan Ellenberg (director, Office of Biostatistics and Epidemiology, Center for Biologics Evaluation and Research), who chairs the FDA working group developing the agencys DMC guidance. An increase in industry and government research collaboration, plus heightened public concern about bias and conflicts of interest in clinical trials, have stimulated DMC oversight.
The trend accelerated noticeably following publication of a 1998 report on institutional review board operations from the Office of the Inspector General (OIG) of the HHS. The OIG recommended that DMCs be used more frequently to oversee certain clinical trials and to review adverse events occurring during a trials progress. To encourage this, the OIG called for HHS to more clearly specify DMC requirements and operations.
In response, FDA formed a committee with representatives from CBER, CDER (Center for Drug Evaluation and Research), CDRH (Center for Devices and Radiological Health), and the Office of Compliance (OC), which conducts field inspections of clinical sites to support applications filed with FDA. The panel published the draft guidance on 19 November 2001. It describes models for DMC operations and functions.
A major concern for sponsors is that FDA and NIH are moving toward recommending DMCs for a broader range of clinical investigations. Although the government requires that sponsors monitor safety issues and report adverse events related to all trials, until now no DMC has been needed in most cases. FDAs draft guidance encourages increased use of DMCs for Phase 1 and Phase 2 studies that raise issues about risk and objectivity. Ellenberg said at the November meeting that a DMC might be useful if an early trial presents important safety concerns, but that this would be a rare occurrence for most Phase 1 trials.
The key issue is where FDA and HHS officials will draw the line. Today, all NIH-funded interventional studies undergo review by a DMC as well as an IRB. While the guidance does not aim to lay out any single model for DMC operation, it outlines three key factors that should determine when a sponsor decides to form an outside review panel: risk, practicality, and scientific validity.
Risk involved. Key factors include use of mortality or major morbidity as primary or secondary endpoints, whether the study involves a highly novel treatment that lacks prior information on clinical safety, enrollment of a highly vulnerable study population (children, elderly subjects, seriously ill patients, for example), and likelihood of adverse events. The guidance states that trials that are large, of long duration and multicenter raise more possibility of safety issues because of the greater overall exposure and the likelihood of unanticipated adverse events.
A DMC also may help deal with ethical issues that may arise if a product performs very well and the sponsor needs objective advice on discontinuing a control arm. A favorable or unfavorable early result might ethically require study termination before planned completion.
Feasibility of DMC input. On the practical side, a sponsor has to consider whether a trial is sufficiently large and long-term to allow time for a DMC to analyze safety issues. Added concerns are whether the sponsor can afford to underwrite a DMC program and whether enough experts are available to serve on a panel.
Need for DMC to ensure scientific validity. A DMC is particularly useful when there are serious concerns about bias and lack of independence by sponsor or investigators. Changing information or circumstances may require the sponsor to modify the protocol during a trial.
Blinding and bias
The issue of access to unblinded interim study datawho should have access and whenhas emerged as a prime point of debate about the proposed policy. Study sponsors need to be able to revise a trial protocol as it proceeds and as new information and situations arise. But to modify a clinical trial protocol without introducing bias and while avoiding premature disclosure of study results, the sponsor should have access only to blinded data, points out Jay Siegel (director, Office of Therapeutics Research and Review, CBER).
Thus, an important role for a DMC is to provide objective advice on the need to modify or halt a trial based on interim safety and efficacy data. The DMC also can assess adverse events that may be part of the disease process or common in the study population and therefore not likely to appear related to the clinical trial. Such analysis may prompt the DMC to recommend reducing a dose to avoid toxicity or changing a consent form to warn of an emerging safety concern.
Through this process, the DMC reduces the need for a sponsor to access interim comparative data, which may oblige the company to disseminate that information to study participants and investigators and compromise its ability to manage the trial appropriately. If individuals recommending changes in inclusion criteria, endpoints, or trial size have knowledge of accumulating data, such changes would inevitably impair the credibility of study results, the guidance states.
A related issue is whether the sponsors own statistician should analyze interim unblinded data. This may raise concerns about how separate the employee is from decision-makers. To avoid conflicts of interest, some experts call for increased use of independent statisticians to review interim data.
The decision by a sponsor to access interim data is a major step and should be discussed with FDA in advance, advises CDER director for medical policy Robert Temple. Early access is necessary if a DMC recommends terminating a study, and FDA should be informed of such a situation. And if early termination arises from data indicating strong benefits from treatment, then FDA wants to know if safety data is adequate, if the benefit has long duration, if subgroups have been assessed sufficiently, and if secondary endpoints are considered, Temple explains.
In addition to using DMCs to monitor a single large trial, sponsors might establish another type of DMC to review an entire developing database, Temple suggests. He describes a DMC Type II to review safety concerns and study design issues arising from a series of smaller, possibly short-term, trials that evaluate relief of symptoms as opposed to treating serious conditions. It may be useful for an outside group to monitor the developing safety database from such studies to make recommendations for future study design. This approach may be especially helpful in dealing with high-risk populations where pooled data may reveal special issues. This broader DMC could sift early study data to be sure that Phase 3 trials cover the desired range of severities and appropriate doses, combinations, and study duration. Moreover, when the investigator has a financial interest in the product or when there are ethical issues involving a study, a DMC may enhance the credibility of research efforts.
SIDEBAR: Data Monitoring Committee Guidance
Once a sponsor decides to use a DMC, the FDA draft guidance provides specifics on how such a panel should operate. An important objective is to clarify the relationship of the sponsor to the DMC, noting that an independent panel will have the sponsor or its representative participate only in open meetings.
Panel composition. A DMC should have members with expertise in clinical medicine, biostatistics, ethics, pharmacology, epidemiology, and clinical trial methodology, as well as patient or community representatives. Geographic representation, demographics, and other experience may be relevant. Candidates should be assessed to avoid conflicts of interest. The sponsor selects a panel chair based on prior DMC experience and ability to communicate and build consensus. Commitment to the position for the duration of the trial is important.
Charter or written operating procedures. These should specify meeting schedule, format for data presentation, data access of members, meeting attendees, assessment of conflicts of interest, and methods and timing for issuing reports. An important point is to clarify what statistical methods and assessments will be used and what thresholds would warrant recommending study termination.
Confidentiality and reporting. Access to coded data, timing and distribution of reports, and importance of strict confidentiality are stressed.
Meeting structure. Specifies formats for open, closed, and executive sessions and who may attend, and what issues and data will be discussed.