Analysis Population

Collecting data on humans in clinical trials is very complicated, for scientific, logistic, ethical, and financial reasons. Analyzing such data is no less complicated, and it is seldom a simple exercise to determine which data to analyze, let alone how to analyze them and report the findings. In this article, we seek to describe some issues faced by the data analyst in a clinical trial. We define several terms that are often used to describe succinctly a data set created according to certain principles. We identify several issues and themes that frequently develop when discussing the selection of data for analysis. We also provide some principles that we have found to be useful approaches to dealing with the difficulty of selecting data for analysis in a complex but real setting. Finally, we provide an illustrative example of how an analysis dataset was actually chosen.