1 The design and analysis of clinical trials

This chapter discusses the main features of clinical trials that need to be taken into account by a statistician involved in their planning, analysis and interpretation.

Abstract

Publisher Summary This chapter discusses the main features of clinical trials that need to be taken into account by a statistician involved in their planning, analysis and interpretation. Design, execution, analysis, and interpretation are inextricably interwoven, and each aspect gains by being considered in context rather than in isolation. A simplistic view of the objective in a clinical trial (the selection model) is that the investigator seeks to choose the better, or best, of a set of possible treatments for some specific condition. It is important that the investigator(s) should draw up, in advance, a detailed plan of the study, to be documented in the trial protocol. This should include a summary of the case for undertaking the trial, and of the evidence already available on the relative merits of the treatments under study. Trial analysis includes descriptive statistics, mainly in the form of tables and diagrams, that are needed to show the main baseline—that is, before treatment—characteristics of the patients in the various groups. Statistical inference is required to draw conclusions about treatment effects, for both efficacy and safety.