Placebo Responses in an Arthritis Trial

Abstract

An injection trial, using two steroid drugs and a placebo, was carried out on a group of 39 patients, all of whom had rheumatoid or osteo-arthritis affecting the knees. This paper analyzes the results of the placebo injections, together with those of a further trial, using orally-administered placebo tablets. Twenty-five per cent showed objective improvement with steroid injections and 18% with placebo injections, a difference which is not statistically significant with small numbers. The order in which injection courses were given had a marked effect on the results, the first course producing far greater objective improvement than either of the subsequent courses irrespective of the substance administered. In the tablet trial, fewer claimed improvement with placebo, and more claimed deterioration than in the injection trial. The discrepancies between claims for dummy tablets and placebo injections are discussed. Altogether 71% of the patients showed a placebo response of some kind. Placebo responders were found to show higher scores for neuroticism and introversion. Negative responders showed higher scores for social dominance. It is shown that placebo responders form a continuum coincident with a curve of normal distribution, and constant non-responders do not show side-effects. A suggestion is made regarding the conduct of clinical trials.