Statistics and Its Interface

Volume 5 (2012)

Number 4

Futility stopping in clinical trials

Pages: 415 – 423

DOI: https://dx.doi.org/10.4310/SII.2012.v5.n4.a4

Authors

Pei He (Stanford University, Stanford, Calif., U.S.A.)

Tze Leung Lai (Stanford University, Stanford, Calif., U.S.A.)

Olivia Yueh-Wen Liao (Stanford University, Stanford, Calif., U.S.A.)

Abstract

Early stopping due to futility, also referred to as a go/nogo decision, during interim analysis has become an important feature of clinical trial designs. Current methods for futility stopping in literature are mostly based on conditional power or predictive power in conjunction with the theory of stochastic curtailment or group sequential design. They have certain drawbacks that have been noted in literature. Herein we describe a new approach to futility stopping in clinical trial designs and the statistical theory underlying this approach. Simulation studies and theoretical analysis show the advantages of the approach in both parametric and nonparametric problems.

Keywords

conditional power, efficient, group sequential testing, survival endpoint

2010 Mathematics Subject Classification

Primary 62L05. Secondary 62F03.

Published 16 November 2012