Name: Tie-Hua Ng
Title: Ph. D.
Institution: U.S. Food and Drug Administration
Street: 1401 Rockville Pike, HFM-217
Authors: Tie-Hua Ng
Title: Issues of Simultaneous Tests for Non-Inferiority and Superiority
Abstract: Dunnett and Gent (1996) proposed simultaneous tests for noninferiority and superiority and argued that no multiplicity adjustment is necessary because of the intersection-union principle. Their argument seems indisputable. However, using the same argument, we would be able test a sequence of nested hypotheses without multiplicity adjustment but that would ?almost? have the same ?effect? as a post-hoc specification of the null hypothesis of a specific treatment difference. When two nested hypotheses are tested such as in the simultaneous tests for noninferiority and superiority, there seems to be no problems. However, as the number of nested hypotheses increases, the problems become apparent.
References: 1. Dunnett, C. W., and Gent, M. (1976), ?An Alternative to the Use of Two-Sided Tests in Clinical Trials,? Statistics in Medicine, 15, 1729-1738. 2. European Agency for the Evaluation of Medicinal Products, Committee for Proprietary Medicinal Products (CPMP). Points to Consider on Switching Between Superiority and Non-inferiority, 2000. http://www.emea.eu.int/htms/human/ewp/ewpptc.htm 3. Morikawa T, Yoshida M. (1995), ?A useful testing strategy in phase III trials: Combined test of superiority and test of equivalence,? Journal of Biopharmaceutical Statistics, 5:297-306. 4. Phillips A, Ebbutt A, France L, and Morgan D, (2000), ?The International Conference on Harmonization Guideline ?Statistical Principles For Clinical Trials?: Issues in Applying the Guideline in Practice,? Drug Information Journal, 34:337-348.