Docket Management
Docket: 01D-0489 - Draft Guidance: Clinical Trial Sponsors on the Establishment & Operation of Clinical Trial DMCs
Comment Number: EC -9

Accepted - Volume 1

Comment Record
Commentor Dr. Anne Lindblad Date/Time 2002-02-15 13:01:17
Organization The EMMES Corporation
Category Company

Comments for FDA General
Questions
1. General Comments This document provides definitions of independence for Data Monitoring Committees (DMCs) that, once released, will become a gold standard for quality in the clinical trials community. I believe the definition of independence and portrayal as the optimum composition of a DMC fails to place safety first. The document equates the definition of independent DMC membership with superior quality of monitoring partly because of difficulties in accommodating changes in design and outcome measures mid-trial, rather than concerns for comprehensive monitoring of safety. Many purists would respond that when faced with the dilemma of mid-trial change in outcome definitions, the gold standard should be to start a new trial rather than change the existing trial. I believe the optimal composition of DMCs should be driven by how best to monitor for safety. Safety monitoring requires more than following the protocol, which should describe as fully as possible the planned analyses and outcomes for safety. Analyzing safety outcome data includes looking for and characterizing unexpected events and their biologic plausibility. Doing the latter often requires further hypothesis generation and testing of those hypotheses within the trial data. Such analyses, by their nature, would not be predefined. As data accumulate, insights are offered by the study chairman, statistician, and sponsor, who were involved in the trial design phase and may also be involved in the execution. These insights can benefit the DMC during the presentation and discussions of safety data, and can not be replaced by an independent analysis group or by a DMC whose members typically spend no more than 48 hours a year thinking about a trial and reviewing its data. If changes to a trial design must be considered, an outside, independent body (who has never seen any interim results) could be formed to make recommendations about design modifications to accommodate changes in medicine and unexpected events. The Guidelines state “Unblinding of the sponsor may increase the risk of further unblinding, e.g. participants, potential participants, investigators, thereby potentially compromising objective safety monitoring, equipoise, recruitment, administration of the intervention and other aspects of the trial. Unblinding a responsible sponsor is unlikely to create any additional risk of unblinding study participants and investigators than unblinding the members of the DMC. If the trial is a double-masked trial and both the participants and the persons making the outcome assessments are masked, the potential damage of a leak of interim results by any party is minimized. Issuing a guidance that does not recognize the acceptability and preference of including knowledgeable individuals, such as the study chairman (when that chairman is not the Principal Investigator of an enrolling clinic), study statistician, and a representative from the sponsor (who has ultimate responsibility for safety monitoring under the IND), as ex-officio members of the DMC is to throw away a valuable and much needed resource and to possibly compromise safety. Clearly there are exceptions to an optimum composition. One example would be when there is confounding between efficacy and safety monitoring (e.g., when mortality is the primary efficacy and safety outcome). In such cases the sponsor may decide it is not in their best interest to be present during the review of unmasked data. When we provide guidelines for Data Monitoring Committee composition, safety should come first. The advantages and disadvantages of variations from a plan that optimizes information flow to the DMC should be carefully considered before implementation.




EC -9