Application of traceability concepts to analytical quality control may reconcile total error with uncertainty of measurement.

In this issue of CCLM, a contribution from Prof. James West-gard (1) rekindles the debate, started 10 years ago by Dr. Dybkaer (2, 3), who opposed total error and the uncertainty of measurements performed in clinical laboratories. In agreement with the ISO 15189 standard (4), both scientists agree in principle that laboratories should know the uncertainty of their results for assessing whether the methods employed are suitable for clinical use. However, the point is how they should determine the uncertainty in practice, given that the standard does not clarify the mechanism. The concept of uncertainty was introduced in the 1990s due to the lack of consensus on how to express the quality of measurement results. Although laboratorians may readily understand the meaning, its determination may be difficult in actual practice. The Guide to the Expression of Uncertainty of Measurement (hereafter referred to as GUM) approach (5) establishes a general model for evaluating and combining all relevant sources of uncertainty of measurements , and has been fully endorsed by metrology institutions and suppliers of reference materials (6). The GUM concept should also be used in accredited laboratories that perform reference procedures (7, 8). However, the application of GUM in clinical laboratories is not straightforward and has encountered many practical problems and objections (9, 10). Examples in the literature, that are few in number, demonstrate the need for determination and fortitude to provide a correct estimate of the uncertainty of each component using the bottom-up GUM approach (11, 12). However, omitting a significant contribution to uncertainty results in the calculated uncertainty being too low, with poor credibility of the figures (9). Thus, if the uncertainty concept should meet more widespread practical acceptance, the complexity of the evaluation process for uncertainty must be markedly reduced so that clinical laboratories can estimate this important characteristic preferably using the information that already exists. One of the basic concepts behind the theory of uncertainty is that the procedures employed in clinical laboratories should be, in principle, unbiased (2). It is not fortuitous that in the ISO 15189 standard, the requirement referring to uncertainty (item 5.6.2) is just before that (item 5.6.3) requiring verification of trueness and traceability of results by application of metrological concepts in order to minimise procedure-dependent bias (4). This requires the development of a reference measurement system and the reliable transfer of measurement values from the highest hierarchical levels of the system to …