A model for homœopathic drug tests including statistical analysis^{[ * ]}

Summary

Homœopathic drug tests provide data on which drug pictures are based. Current drug registration laws require proof of drug-specificity as well.

A brief description is given of Hahnemann's method, which should be kept in mind particularly for tests on new drugs. It is however not suitable for statistical analysis. Problems arise because of placebo symptoms, and the drug-specificity of individual symptoms can only be expressed in probabilities.

Statistical analysis of homœopathic drug tests has to be based on comparison between the totality of symptoms elicited during exhibition of the drug and the totality of symptoms recorded on placebo. Simply counting the incidence of symptoms proved inadequate. The model currently used in conjoint German and Austrian trials is presented for comment and discussion. This also takes account of continued drug action and includes a number of new parameters on which to base comparison.