Clinicians are usually inclined to reduce data obtained from diagnosis to a simpler form such as "yes or no" answers in order to obtain useful parameters for daily practice. It should be realized, however, that the diagnostic process very rarely exhibits "black and white" situations. Rather the evaluation of numerous "grey levels" is imperative. The diagnostic process, therefore, remains always incomplete and inaccurate, and it represents an evaluation of probabilities rather than certainties. For this reason, the diagnostician should realize and understand the mathematical relations between the information provided by diagnostic tests and the clinical situation actually present at the time of the test. This will allow one to convert diagnostic results into therapeutic procedures with a high degree of confidence. Since results from diagnostic tests are quite often used to decide on treatment, complex data are reduced to simple dichotomy, such as presence or absence of disease, normal or abnormal conditions, etc. In order to react to diagnostic tests in an ordinal, dichotomous manner, the clinician has to choose a particular level of a test at which he initiates treatment without having the assurance that this level represents the one and only standard at which treatment has to be initiated.