[On the uncertainty of measurements illustrated by the BONN Risk Index].

Every result of a measurement is subject to error. This error is inevitably caused by inaccuracies of the devices used and by environmental influences. For correct evaluation of a measurement result, this error must be estimated as exactly as possible. In kidney stone research and urolithiasis treatment, risk indices are applied to determine a patient's health status [e.g., AP(CaOx), RS(CaOx), BRI]. These operands (indices) combine important physical and chemical parameters of a certain disease pattern in a theoretically or empirically obtained mathematical formula. The index-related total error resulting from the single errors of the individual input parameters might be calculated with the Gaussian error propagation method. This method takes into account that individually determined parameters contribute differently to the total result. Due to this, summation of the various specific errors is insufficient. Instead, they must contribute to the total error depending on their individual weight. In medical borderline cases, knowledge of the value of this error may improve diagnosis since critical observation of an acquired value helps to avoid false interpretation. Using the BONN-Risk-Index (BRI) as an example for determination of the urinary calcium oxalate crystallization risk in native urine, the method of error calculation is demonstrated in detail and the advantages of including discussion of error in method development are discussed.