Performance goals for the laboratory testing of antithrombin, protein C and protein S.

To achieve a reliable analytical quality for both monitoring and diagnostic testing, laboratories need to fulfil the widely accepted analytical performance goals based on the biological variation of the analytes of testing. Not only is the short-term analytical performance, which regularly is assessed by internal quality control procedures, of importance, but also the long-term analytical performance. To assess the long-term analytical performance, data obtained from an external quality assessment programme can be used. In this study we have used the evaluation model designed by the ECAT Foundation for the assessment of the longterm analytical performance, including imprecision, bias and total analytical error. The model was applied to the data from 136 different laboratories for the assay of antithrombin (activity), protein C (activity and antigen) and protein S (activity, total and free antigen). The imprecision (median; range), reflected by the long-term analytical coefficient of variation (LCV (A) ), was the lowest for antithrombin (7.6%; 2.6 - 43.8%) and the highest for protein S activity (17.2%; 4.3 - 88.6%). For bias and total error the same pattern was observed (antithrombin: 3.8%; 0.3 - 17.1% and 9.1%; 3.4 - 34.3%, respectively; protein S activity: 12.8%; 3.1 - 34.8% and 24.5%; 9.9 - 87.0%, respectively). For the majority of the laboratories (70 - 85%) the imprecision contributes considerably more to the total error than the bias. However the effect of the bias on the analytical quality is not negligible. Assays for antithrombin, protein C and protein S are mainly used for diagnostic testing. About 70 - 100% of the laboratories can fulfil the desirable performance goal for imprecision. The desirable performance goal for bias was reached by 50 - 95% of the laboratories. In all cases the highest numbers of laboratories fulfilling performance goals was obtained for the protein C variables. To improve the analytical quality in assays of antithrombin, protein C and protein S it is highly recommended that primarily imprecision (non-systematic failures) be suppressed. However the effect of the bias (systematic failures) on the analytical quality should not be neglected. A useful tool for determining the imprecision (LCV (A) ) and bias is the long-term analytical performance evaluation model as used by the ECAT Foundation.