I wonder how much of that effect is due to subjects not understanding what a placebo. Would the effect remain if the study population was other physicians or scientists?
I think it's relatively unreasonable to assume that when the study says that participants are informed that they're taking a sugar pill that the participants don't understand what that means, nevermind assuming that no effort at all was made by the administrators to address this fairly obvious issue. If the entire point of the study is testing whether or not patients still benefit when they understand that they're not taking any medication, it seems self-evident that they would be required to ensure their test subjects understand that they're not taking any medication!
From the article, the guy who ran the study apparently still has reservations on this point: "And he can even be disparaging of his own work, wondering, for instance, whether the study in which placebos were openly given to irritable bowel syndrome patients succeeded only because it convinced the subjects that the sugar was really a drug."
