Controversy in Counting and Attributing Events in Clinical Trials

WHICH events should be counted in a randomized clinical trial, and when they are counted, which treatment should be blamed? That a controversy exists is clear from reading "Letters" columns after publication of randomized clinical trials1 2 3 4 5 6 7; correspondents and authors debate whether all or just some events should have been analyzed and whether the original treatment should have been blamed for events occurring after trial therapy was stopped or switched to the alternative limb.This controversy flourishes both among and between clinicians and methodologists,8 and its key elements are, with rare exceptions,9 , 10 either skimmed over in "how-to" publications or treated . . .