DECISION AND DETECTION LIMITS FOR LINEAR HOMOSCEDASTIC ASSAYS

This paper focuses on the effect of estimating the calibration paramet ers on the decision and detection limits for linear homoscedastic assa ys. Assay decision and detection limits are defined. For linear homosc edastic assays expressions for these limits are derived when the calib ration parameters are known and unknwon. These expressions, together w ith others previously reported in the literature, are used to calculat e limits for a particular example of a cyclic AMP assay. The various m ethods of computing decision and detection limits are compared in a si mulation study. The results demonstrate the importance of taking into consideration the variation in the estimators of the calibration param eters.