Statistical Tests Involving Multiple Endpoints

Abstract

The statistical analysis of data from preclinical and clinical studies usually involves multiple tests of many endpoints. Because of the conduct of a large number of statistical tests the chance of false positive findings could increase considerably. This talk will describe the use of global test and p- level adjustment approaches to assessing treatment effects. The commonly used single-step unweighted p-level adjustment approach for testing individual endpoint hypotheses will be generalized to a weighted adjustment and subset hypotheses.