[The basis of diagnostic and therapeutic decisions (author's transl)]

[Quantifications of perioperative risk]

In this paper standardized and quantitative definitions of perioperative risk and risk factor using probabilities are given. A calculation of risk and risk factors is performed using data from a study on perioperative risk in colon resection and a study on a preoperative risk check in general surgery. The problem of one risk factor, a combination of two risk factors and the use of many risk factors to quantify preoperative risk is discussed. Confidence intervals are recommended as a standard method for presenting statistical results on perioperative risk.

[What is a safe principle for therapeutic decision?]

In surgery (as in other disciplines of medicine) a causal therapy is connected with a causal diagnosis. Many physicians confine themselves to symptomatic measures which are useless beyond self-healing or introduce some kind of poly-pragmasia. Sometimes between those extreme positions there is a need of deliberations "as if" or primarily diagnostic operations. Diagnostics as a whole consist of logical as well as of psychological elements. Approximation is made either by some kind of hypothetico-deductive calculations or using some kind of pattern recognition. Anamnesis and immediate personal examination yield a preliminary diagnosis in about 80%. Yet a diagnosis of this sort has to be verified or falsified by scientific methods. The final decision is a function of both the theory of probability and the theory of utility. The combination is demonstrated by the so-called decision matrix and the so-called decision tree, going from simple models to complicated ones.