Abstract:

In the recent article "Is the placebo powerless?", by Hrobjartsson and Gotzsche (2001), readers are led to believe that the placebo effect is a myth. Indeed, the accompanying editorial by Bailar (2001) suggested that despite the negative data many still want to believe in the placebo effect. In essence, saying the placebo effect is part of our cultural lore, our common home. This suggestion sparked us to think about the placebo data in a new way. What if where you come from, i.e., cultural beliefs instilled by professional training, has an impact on results? The idea is not original. Rosenthal (1967) demonstrated over 30 years ago that unintended, covert communication by experimenters to study participants can affect subjects' responses and be predicted by experimenter attributes. A direct way to assess this hypothesis is to divide the Hrobjartsson and Gotzsche data by the first author's degree (Ph.D. vs. M.D.) and the direction of the effect size (trend for placebo effect vs. not). Ph.Ds were significantly more likely to find placebo effects than M.Ds. Analysis of all studies revealed that 73% of Ph.Ds found Pro-placebo effects and 27% of PhDs found Cons-placebo effects (p<0.01). In case of M.Ds, 57% found Pro-placebo effects and 43% found Cons-placebo effects (p>0.10). While the pattern for Ph.Ds was consistent across medical categories (e.g., psychiatry, internal medicine), the pattern of M.Ds shifted toward that of Ph.Ds in research areas that included more M.Ds as first authors (e.g., surgery). It appears that there may be some sort of self-fulfilling prophecy in action that which we all know is true, that there's no place like home.

Keywords: placebo, meta-analysis, psychosocial influences