Global Assessment Variables

Global statistical tests (GSTs) and associated global treatment effects (GTEs) summarize a treatment's global benefit without the need of performing multiple univariate tests for individual outcomes. This chapter discusses different versions of GSTs. Some of the GST's discussed in the chapter include: rank-sum-type GST; adjusted rank-sum-type GST; likelihood ratio (LR)-Based GST; and Follmann's GST. The main rationale for using a global outcome in a clinical trial is that, by providing a more comprehensive and thus more informative assessment of the patient, it also provides a more clinically meaningful primary endpoint than use of any single variable by itself. Two other approaches for using multiple endpoints in the primary analysis of a clinical trial are the T2 statistic and use of multiple comparison procedures.