Learning from a Trial Stopped by a Data and Safety Monitoring Board

One of the reasons we conduct clinical trials to test specific aspects of patient care is to be sure that the interventions we do benefit patients. As clinicians, we are well aware that the outcomes of a treatment may vary substantially from one participant to another; a treatment that works spectacularly well in one person is ineffective in another. If a trial is too small, an average treatment effect favoring an intervention may be consistent with the play of chance. Trials with larger numbers of participants measure average treatment effects more precisely and are more likely to detect real intervention . . .