Abstract
In order to estimate the probabilities that a patient would belong to subgroups created by the diagnostic tests most used by European urologists four hundred and twenty-one consecutively referred prostatism patients were studied. It was demonstrated that all the qualities described by these tests were distributed in such a way that the presumed accuracy of the tests might result in a 20 per cent variance in the number of treated patients. A simple self administered home flow test, which was significantly correlated to the maximum flow rate, was shown to be stronger correlated to the symptoms of the patients compared to any other quality. A stochastic table for prostatism was constructed, which may be used for estimations of the influence different decision making may have on the outcome of treatments in this kind of patients.