Reflections of an IRB Chair

Abstract

In summary, the local IRB is now and probably always will be the sole protection afforded the research participant. It is incorrect to perceive local IRBs as having jurisdiction only over the consent process. Multiple levels of review of a cooperative group- or company-sponsored protocol prior to IRB submission do not guarantee that there are no serious design flaws in the study that would render the results invalid or not statistically credible. The most serious error an IRB can make (on a par with failing to obtain proper consent) is to approve a study that cannot possibly test the hypothesis in a credible fashion.