Analytical performance specifications in clinical chemistry: the holy grail?
The conference in Milan in 2014 resulted in a consensus statement on quality specifications. An EFLM Task and Finish Group was started to deal with issues that remained unresolved and needed further development including total analytical error methods. Most importantly were questions concerning measurement uncertainty (MU) and flaws in the model for calculating the permissible (or allowable) total error (TE). This is related to performance specifications based on biological variation, the second model of the Milan consensus. Performance specifications are closely linked to the work of Westgard and the concept of TE, that depends on the combined effect of the random and systematic errors of the method which is compared to a permissible TE that is in most cases based on biological variation. Where the first models for performance specifications did not include bias, later models did. The concept of bias is however complicated, both in the estimation of the value, and in the integration of bias in a mathematical model. In many fields outside clinical chemistry, the error concept has been abandoned in favour of the MU model. Bias is commonly excluded from this model, where only the uncertainty of bias is taken into account. There is no general consensus on these issues, and challenges remain to integrate the different concepts, both for the definition of performance, of performance specifications as for quality control procedures.