BIOS Seminar: Data Monitoring Committees

With more than 10 years of experience in the design and conduct of phase I to III clinical trials, Dr. Barry has provided biostatistical leadership working for both contract research organizations and academic institutions. Dr. Barry is currently a co-PI of the NIAID/DAIT-funded Statistical and Clinical Coordinating Center, serving as a Scientific Leader of the autoimmune disease group and previously was an Assistant Professor at Duke University, and at Dana-Farber Cancer Institute and Harvard Medical School.

Presenter: William (Bill) Barry, PhD, Senior Research Scientist II, Rho, Inc.

Abstract:

Data monitoring committees (DMCs) have a unique role in trial oversight, and while responsibilities may vary somewhat according to the clinical research setting, some fundamental principles apply broadly. DMCs are used to periodically review the accumulating unblinded safety and efficacy data by treatment arm, and advise the trial sponsor on whether to continue, modify, or terminate a trial based on a benefit–risk assessment. Statistical methods have been developed to control error associated with multiplicity
when conducting interim analyses under fixed and adaptive designs. Plans for execution of interim analyses need to also safeguard against the potential of operational bias. Three scenarios where analysis plans, organization flow, and decision processes are critical to maintaining trial integrity are: (1) early termination in multi-arm studies, (2) adaptive sample size re-estimation, and (3) unplanned changes to the study design. Examples from NIHsponsored and investigator-initiated trials are used to illustrate the complexity of these scenarios and the impact execution had on the trial.