Outcome-switching (changing the what outcome you measure after an experiment has started) can invalidate statistical methods leading to false positives and problems with trial replication. Yet, most scientific journals have not reliably reported when this has been happened.
Compare-Trials.org tracks outcome-switching with public registries and then encourages journals to issue corrections, yet some major journals such asAnnals of Internal Medicine have resisted this form of transparency. Scientifically valid results require scientifically valid methods, and it is concerning when major medical journals are not holding themselves to modern standards of scientific rigor.
Toward Health 