Comparison of multiple testing procedures for clinical trials with two doses and multiple ranked endpoints

Clinical trials comparing two doses of an investigational product versus placebo are becoming more and more frequent in the confirmatory phase of clinical development. These clinical trials generally include a primary endpoint, and several key secondary endpoints ranked hierarchically. Strong control of the family‑wise error rate (FWER) is required in particular if the study is aimed at supporting approval of the investigational product. This paper describes a simulation study comparing different multiple testing procedures (MTP), under a wide range of plausible scenarios. A case study including two doses and three ranked endpoints was introduced to illustrate the MTPs and evaluate their performance. Our findings provide practical guidance for statisticians and clinical researchers designing trials with two doses and multiple endpoints, highlighting which methods offer optimal performance under different anticipated scenarios. The results also challenge some common assumptions about hierarchical testing approaches and demonstrate the robust performance of enhanced mixture methods across diverse dose-response scenarios. This paper also shows that the enhanced mixture method can be simplified in trials that include two doses and hierarchically ranked endpoints.