ABSTRACT
In order to obtain better overall knowledge of a treatment effect in clinical trials, the clinical trialists often collect many medically related or correlated endpoints and test the treatment effect for each. However, the problem of multiplicity arises when multiple hypotheses are tested. Ignoring this problem can cause false positive results. A lot of statistical methods have been
proposed to handle multiplicity issues in clinical trials. However, many commonly used multiple testing correction methods proposed to control family-wise type I errors (FWER) disregard the correlation among the test statistics, for example, the Bonferroni correction and Holm procedure.
