[The number of subjects or patients in clinical pharmacological studies]

A technique permitting without preliminary informations the assessment of the minimal number of test persons or patients to be studied in clinicopharmacological investigations dealing with multiple variables (vital functions, laboratory data) is reported. The procedure is applicable to investigations using data recorded at different time intervals and concomitant variables (basic values at the start of the investigation, age, and body weight). The minimal number of test persons or patients is estimated from the results of a formal examination. In this examination is tested whether a multivariate analysis of a therapeutic effect using all data only once and simultaneously is based on sufficient degrees of freedom of the error variance.