Determination of assigned values in control specimens for internal accuracy control and for interlaboratory surveys. Evaluation of 200 different lots with identical experimental design: experiences and conclusions
Hansert, E.; Stamm, D.
Journal of Clinical Chemistry and Clinical Biochemistry 18(8): 461-490
1980
ISSN/ISBN: 0340-076X PMID: 7411027 Document Number: 156055
In West Germany there are guidelines for a basic program for internal and external quality control of quantitative clinical chemical analyses prepared by the German medical society in connection with the Calibration Act. These guidelines specify that for internal accuracy control and for interlaboratory surveys, control specimens are to be used with assigned values and assigned intervals that have been determined by particularly well-qualified, independent laboratories serving as reference laboratories. The assigned values are often different from the best estimate (by measurement) of the true value. They are influenced by the matrix of the specimens and by the method used, and they must be determined under routine conditions. The protocol for obtaining the analytical results under carefully defined conditions is described in detail. The statistical treatment of the analytical results and the author's observations on the structure of the data and the properties of the dispersion and location parameters for the individual laboratories are discussed. These observations led to the development of a procedure for selecting assigned intervals and assigned values in which no special assumptions are made about how the data are distributed. This procedure is described in detail. The importance of the authors' observations for the evaluation of analytical results in the clinical laboratory and for the statistical treatment of such data is discussed. Comparative studies were carried out using the same control specimens with other designs for assigned value determination (e.g. consensus value principle). The advantages of the design presented here are discussed.