Behavioral research and the double-blind placebo-controlled methodology: challenges in applying the biomedical standard to behavioral headache research.

The randomized, double-blind, placebo-controlled experimental design has prevailed as the "gold standard" in biomedical research, intended to control potential bias in patient/group assignment, investigator allegiance, patient expectations, and nonspecific therapeutic effects. Properly executed, such designs ensure a study's internal validity and allow differential group outcomes to be attributed to the active treatment. These controlled trials generally yield more conservative outcomes than open trials and case reports and establish efficacy in pharmaceutical research. In meta-analytic reviews, studies are often assigned quality scores based in part on the degree to which they meet this scientific standard. Applying the biomedical research design standards for blinding and placebo control to clinical trials evaluating behavioral and other nonpharmacologic headache treatment nearly always is either infeasible or simply not possible. Only rarely is blinding meaningfully achievable in administration of behavioral or psychological therapies. Various "psychological placebo" control conditions have been forwarded in behavioral studies (eg, sham treatments, pseudomeditation), but these controls are incapable of emulating an inert control condition comparable to that of the pill placebo in pharmacologic research, and they are best reserved for studies examining the mechanisms whereby an intervention produces improvement. This article reviews the conceptual and procedural challenges in applying the standard pharmaceutical clinical trials research design to behavioral headache research as well as implications for meta-analyses across studies of various treatment modalities.