[Planning of psychotropic drug trials. Statistical evaluation. Hypotheses and methods].

Any reasonable evaluation presupposes a precise question. Furthermore it is necessary to eliminate all factors of disturbance, which are able either to simulate or to mask a result of evaluation. The most important of these factors is the psychiatrist himself. Therefore, experimental designs have to taken in account the factor psychiatrist. Criteria of the effects of any drug must be relevant with regard to comparison of two or more drug effects. The reproduciability of criteria is very desirable, but a wish-fulfilment under the most circumstances. Because criteria should be independent and mutually exclusive, all scales like the Hamiltons one are not suitable for evaluation. Finally one should use as few criteria as possible. Models of evaluation are described for the most frequent questions in connection with clinical trials in psychopharmacology.