9.4 GOLOMB B. A. (1996). Using placebos. Nature 379(6568): 765

SIR - I pointed out earlier (Nature 375, 530; 1995) that because placebos are unregulated and have not been shown to be inert, an apparent positive, negative or null effect of a drug in a placebo-controlled trial could issue instead from a negative, positive or same direction effect of the placebo. J. H. Schoemaker (Nature 377, 98; 1995) concedes the point, but believes it prudent to ignore this problem, because facing it might cost money.

Efforts to address the placebo problem need not be costly, and some might even save money. Focus groups could be convened to devise low-cost suggestions. Guidelines on what constitutes acceptable placebos could be developed. Journals could insist that all constituents of placebo and drug be named by weight. One or two placebo agents might be adopted as standards, and their effects examined in detail so that these effects might be adjusted for in analysis. Selected well-studied flavours and coatings could become standard for use with both drug and placebo, actually reducing costs by obviating the need to match the flavour and appearance of the placebo to each drug.

Schoemaker asserts that "in the worst case", if a drug is credited for effects it does not produce because "by chance" there had been negative effects of the placebo, "this may not be harmful". Yet it may be enormously harmful if the result influences clinical practice. In addition, it is naive to suppose that only by chance can a drug be falsely credited because of a deleterious placebo. Pharmaceutical companies currently determine the composition of placebos in trials of their own drugs; as billions of dollars may be at stake in the outcome, production of the 'inert' comparison agent by the drug company that manufactures the treatment drug under study represents a clear conflict of interest, particularly worrying in the absence of regulations, testing of placebo agents for inertness, or even a statement of the placebo composition in published journal articles. These concerns might be mitigated if there were formal regulations about placebo constituents and if production of placebos were removed from the hands of the company manufacturing the study drug.

NATURE - VOL 379 - 29 FEBRUARY 1996