Methods and Applications of Statistics in Clinical Trials, Volume 2: Planning, Analysis, and Inferential Methods

Chapter 12

Futility Analysis

Boris Freidlin

12.1 Introduction

Interim monitoring of outcome data has become a well-accepted component of randomized clinical trials (RCT); a trial can be stopped early for efficacy if a treatment arm appears definitively better than another arm. Most clinical trials are designed to demonstrate benefit of experimental vs. standard treatment and are thus implicitly addressing a one-sided hypothesis. The one-sided nature of the question provides rationale for both efficacy and futility monitoring. In technical terms, futility monitoring refers to a statistical procedure for stopping the trial early if it appears that the experimental arm is unlikely to be shown definitively better than the control arm if the trial is continued to the final analysis. In the context of evidence-based clinical research, the primary goal of a phase III trial is to provide data on the benefit-to-risk profile of the intervention that is sufficiently compelling to change medical practice. From this perspective, a futility boundary should be interpreted as the point at which convincing evidence exists to resolve the original question—that is, to convince the clinical community that the new treatment is not beneficial.

The advantages of early futility stopping are obvious in terms of minimizing patient exposure to ineffective, potentially toxic experimental treatments as well as in terms of optimizing the use of resources [1]. However, these potential advantages should be weighed ...