A simple strategy convenient for processing of RIA data and quality control in the laboratory on a...

The experience with a computer strategy of mathematically simple models for the description of standard curves and continuous quality control calculations with outlier diagnostics for the processing of RIA data in laboratory is discussed. To get over the difficulties associated with the approximation of the nonlinear shape of standard curves, the combination of weighted iterative linear logit-log regression and point-to point logit-log interpolation is proposed. Monitoring of assay performance is based on three large pool sera determined in triplicates. The control data are checked by statistical tests for the rejection of single outlying observations, followed by calculation of objective criteria on within- and between-assay precision by means of analysis of variance for a one-way classification on the basis of which actual triplicates and whole assays can be assessed. Combined with quality control chart technique and the construction of a rough three-point precision profile the empirical knowledge of an experienced assayist should be included in the RIA interpretation as can be easily organized by way of a dialogue-program on a desk-top-calculator.