Evaluation of the clinical chemistry tests analytical performance by using different models and specifications

Abstract

Abstract Background The importance of managing analytical quality in clinical laboratories is known. Goal-setting models are critical for analytical quality management, along with correctly implemented error models. However, the methods used to determine analytical performance and more importantly, the relevant analytical quality goals are open to discussion. Our aim was to compare the analytical performance characteristics of routine clinical chemistry tests with different goal-setting models which was proposed by various establishments. In addition, to provide a perspective to Turkish total analytical error (TAE) circular letter that compulsory to calculate from 2016. Materials and methods This study was performed by the data obtained from the internal and external quality control of clinical chemistry tests which were measured by Roche Cobas c501 biochemistry analyzer. TAE calculated with TAE% = 1.65 ×(CV%) + Bias% formula. Nordtest uncertainty model was used in the calculation of measurement uncertainty (MU). In this context, total analytical error was evaluated with biological variation (BV), RCPA, CLIA and Turkish allowable total error (ATE) goals. Measurement uncertainty was evaluated with only permissible measurement uncertainty (pU%) goal. Results In our study, RCPA goals are the most stringent, followed by the BVEuBIVAS, BVRicos, pU%, CLIA and finally the ATETurkey goals coming in last. In cumulatively, BVEuBIVAS goals were 18.3% lower than BVRicos for evaluated parameters. Conclusion The balance between applicability and analytical assurance of goals should be well ensured when determining goal-setting models. Circular letter (2016/18) creates awareness to the analytical quality management but still open to development. Biological variation dependent total allowable error model never designed to be used as benchmarks for measurement uncertainty and it is not methodologically appropriate for assessing measurement uncertainty which was estimated by the Nordtest method. Also considered that, the use of “permissible MU” is more methodologically appropriate in the evaluation of measurement uncertainty.