Strong evidence for a benefit is often measured by death or something extremely severe. When using other clinical endpoints cherry-picking is to be avoided. [1

Strong evidence for a benefit is often measured by death or something extremely severe. When using other clinical endpoints cherry-picking is to be avoided. [1] > For a trial to stop early for benefit, the observed effect on the composite end point ideally should be consistent in the individual components. [1] https://www.ahajournals.org/doi/full/10.1161/circheartfailure.111.965707 https://www.ahajournals.org/doi/full/10.1161/circheartfailur...