ABSTRACT
Statistical hypotheses and testing to rule out a prespecied increased risk of an adverse event is statistically similar to that in the determination of noninferior efcacy. Examples include establishing the safety of a test treatment compared to placebo or establishing the safety of a test compound compared to an active control, both with the objective of ruling out an important increase in the rates of adverse events. Less common, but possible, is the comparison of a test compound to an active control with inference desired on the event rate of the test compound compared to a putative placebo. Because the design and analysis are dependent on the objectives, and the objectives can vary, it is vital to prespecify and dene the study objectives.
