A Review of Recent Research on the Type II Error Probability of a Group Sequential Test of Efficacy and Futility with Applications

• Activity Number: 534
• Type: Contributed
• Date/Time: Wednesday, August 7, 2013 : 10:30 AM to 12:20 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Thomas Dobbins*+
• Companies: Merck & Co., Inc.
• Keywords: group sequential test ; type II error probability ; tests of futility ; type II error spending ; adaptive trial ; conditional power

Abstract

For a clinical trial incorporating a group sequential test that allows early stopping for efficacy or futility (GSTEF), the total type II error probability (T2EP) is composed of the T2EPs of tests of futility and efficacy, and the primary hypothesis concerns efficacy. However, in contrast to GSTE, or a fixed sample test, the T2EP of tests of efficacy of GSTEF is neither specified nor known. Recently, Dobbins (J. Biopharm. Stat. 23:378-393, 2013) obtained a partitioning of the T2EP of GSTEF that isolates the T2EP of tests of efficacy, and showed that the T2EP of GSTEF is equal to the T2EP of the tests of efficacy, i.e., the T2EP of GSTE, plus the net additional T2EP due to incorporating tests of futility. This result has several important benefits for GSTEF with T2EP spending method: it allows insight on the consequence of maintaining total T2EP for T2EP of tests of efficacy; it allows setting T2EPs of tests of futility and efficacy directly, it allows incorporating tests of futility as an add-in to GSTE design, and it allows a link to the conditional power method of testing for futility. These benefits are illustrated with an example.