Is there an alternative to the randomized controlled trial?

The advantages of the concurrently randomized controlled trial as a device to assess the efficacy of drugs intended for use in the field of psychiatry are reviewed. The importance of "assay sensitivity" is discussed and illustrations of errors that can arise when inferences about efficacy are based upon comparisons that find no difference between experimental and standard treatments are presented. These illustrations are derived from data collected in two different pre-marketing commercial drug development programs. Also discussed is the argument that controlled trials are inappropriate devices to test drug efficacy because of their alleged lack of sensitivity to drug effects.