Futility approaches to interim monitoring by data monitoring committees

Abstract

Clinical trials involving participants with high risk of serious events or being exposed to a new intervention with potential serious risk are typically monitored during the course of the trial. Often, this monitoring activity is conducted by an independent group of experts, often referred to as a Data Monitoring Committee (DMC). The DMC responsibility includes monitoring for early evidence of a harmful effect or convincing evidence of a benefit. If the data are consistent and overwhelming, the DMC may recommend that a trial be terminated early. Trials may also be terminated early if the initial goal of demonstrating a benefit cannot in all likelihood be attained, or is futile. These are complicated issues and should be sorted out in the protocol and the DMC charter prior to the start of the trial before data begin to accumulate. Several statistical methods have been developed to assist a DMC in determining when a negative or harmful trend is substantial enough to render the trial continuation futile. These will be briefly summarized. However useful these methods might be, they alone are not adequate to make the decision and the DMC must take into account other issues to make a judgment.