Towards standardization of external quality assessment schemes by using bias values based on biological variation

Abstract

The precision and trueness of current instruments and methods in clinical laboratories is much better than in the past. However, the z-score and other comparison variables that are currently used in external quality assessment programs are based on relative data. Thus, they may change from program to program and also change over time within the same program and consequently may not be useful or cost-effective. We therefore devised a test-specific decision limit for accepting or rejecting test results based on a combination of the data of within- and between-subject biological variations. We then applied these limits to a group of tests performed in our laboratory and compared our results with those of external quality assessment programs. In addition, we combined external and internal quality control data on the same graph and prepared a two-dimensional graph for different levels of control sera. Inspection of all results of control sera on this new graph was more useful for decision making. We concluded that the z-score is not reliable for comparisons of test results in external quality assessment. As a substitute for this currently accepted practice, we assert that control limits based on biological variation are more reliable, and can be useful in the evaluation of both external and internal quality assessments and their combination on this new graph.