The randomized trial reported by O'Rourke and coworkers1 (they call it a randomized study but why not call it what it is?) provides insight into the dilemma faced when trying to evaluate a therapy believed, rightly or wrongly, to be efficacious. The fact that extracorporeal membrane oxygenation therapy for persistent pulmonary hypertension of the newborn was introduced in 1976 but not tested in a trial until years later, and then only after it was already viewed in many quarters as effective, is a sad comment on how therapies are introduced and accepted into the armamentarium of treatments.2,3