https://orcid.org/0000-0001-5304-0530
![Sample our Mathematics & Statistics journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/5943/TOC-Subject-Sample-2021-Mathematics-Statistics.jpg"})
Abstract
The Hochberg procedure is known to be more powerful than the Bonferroni method but the test statistics need to satisfy certain dependence conditions. Due to the lack of validation of the positive dependence assumption in the Hochberg procedure, the U.S. Food and Drug Administration guidance on multiple endpoints in clinical trials provides conservative recommendations on the Hochberg procedure, and limits its application on a small set of standard test statistics. Based on the demand of using the Hochberg procedure in a more flexible way, we develop a test to validate the dependence assumption in the Hochberg procedure and Benjamini-Hochberg procedure. A simulation study is conducted for power analysis, and a case study in metastatic breast cancer is included to illustrate how the proposed test can be applied to validating the dependence type between the progression-free survival and overall survival.
Acknowledgments
We thank Ajit C. Tamhane for helpful discussion. Some preliminary results of this research article were present at Spring 2021 Seminar Series in the Department of Statistics at Northwestern University, Statistical multiplicity adjustment in clinical trials on May 12, 2021, at Fall 2021 Biostatistics Seminar Series, Division of Biostatistics, Department of Preventive Medicine, Northwestern University, Multiplicity considerations in clinical studies on October 18, 2021, and at 2022 Joint Statistical Meetings (Washington, DC) in Section 57 Recent Advancements in Multiple Comparison Methodology – Topic Contributed Papers, On dependence assumption in p-value based multiple test procedures on August 7, 2022. The authors thank editor and an anonymous referee for suggestions that improved this article.