ABSTRACT
In cancer trials, it is unethical to use a placebo control when approved and effective therapies are available. A response to this problem in the investigation of a new test therapy is to replace the placebo control by an established therapy (which is referred to as active control agent or standard therapy) and to demonstrate that the test therapy is not inferior to the active control agent in the sense that the effect of the test therapy, when compared with the efŒ- cacy of the active control agent, is not below some non-inferiority margin. In practice, there may be a need to develop a new therapy that is non-inferior (but not necessarily superior) to an established efŒcacious therapy; for example, the new therapy is less toxic, easier to administer, and/or less expensive. As a result, a clinical trial for the establishment of non-inferiority of a test therapy as compared to an active control agent has become very popular in drug research and development. Clinical trials of this kind are referred to as active-controlled trials and statistical tests for establishing non-inferiority are called non-inferiority tests. An overview of design concepts and important issues in these trials is provided by D’Agostino et al. (2003).
