Two-Arm Non-inferiority Active Control Clinical Trial Or The Clinical Trials that All Statisticians Hate to Design and Analyze but Can't Wait to Talk About

Abstract

Two-arm active control clinical trials were often used in testing a new oncological treatment when the efficacy of the treatment is less crucial. In the last 5 years, due to the ethical restriction, two-arm active controlled clinical trials became the only choice to evaluate a new treatment when a standard treatment is available in the market and placebo became no choice in many clinical trial planning. Many proposals on design, analysis and regulatory requirements were discussed at professional meetings and professional journals. In this presentation, we will examine and discuss the various objectives of the active controlled "non-inferiority" trials. And we will discuss accordingly, the design and analysis consideration for the objectives. In general, all the proposed "new" design and analysis methods can be categorized into two traditional approaches, the historical control study approach and the cross- study comparison approach. The weakness and the required assumptions will also be carefully examined and discussed.