Composite endpoint

A composite endpoint bundles two or more outcomes into a single measure, so a participant is counted as having an event if any one of the components happens. Researchers use composites to raise the number of events and improve statistical power, and to capture a treatment's overall effect. The catch is that components often differ greatly in importance, and a mild, common component can drive the result while the serious ones show little change. When reading a study, look at each component separately: check whether the components are of similar importance and frequency, and whether the treatment's effect is consistent across them or concentrated in the least serious one. For example, a trial reporting a reduced rate of death, heart attack, or hospitalization combined may owe most of its benefit to fewer hospitalizations, with no real change in deaths. A composite that improves only because of its softest component tells you less than the headline suggests.