Comparison of the global statistical test and composite outcome for secondary analyses of multiple coronary heart disease outcomes

Abstract

Multiple outcomes (or multiple endpoints), such as mortality and recurrent myocardial infarction, are increasingly common in clinical trials and are often of interest in secondary analyses. Traditionally, a clinical trial protocol is built around a single event as its primary outcome, with several secondary outcomes specified, the analyses for which lack the same level of power. To accommodate all the relevant outcomes and to increase the power of the comparison in trials evaluating the efficacy of treatments for coronary heart disease, investigators often chose to construct a composite outcome. The more conventional composite outcome fails to account for the relative importance and the relationship (correlation) among its components. The purpose of this work is to demonstrate the usefulness of the Global Statistical Test, which considers the correlation between multiple outcomes, as an alternative strategy for these situations and to demonstrate its effect on hypothesis testing and power analysis issues in comparison with the traditional composite outcome analysis. Data from the cardiovascular clinical trial Systolic Hypertension in the Elderly Population are used as an example.