Recommendations for estimating measurement uncertainty in medical laboratories (ISO 15189 and ISO 20914)


The ISO 15189 standard for medical laboratories is in the middle of its revision process.1
The working group, structured in drafting groups for individual parts of the standard, receives some guidance from the ISO CASCO top: use as a template and follow a process like the one used for the revision of ISO/IEC17025:2017, drafting by small groups, revision of the text by the whole project group, revision by WG1 of TC ISO 212, inclusion of ISO 22870 (POCT)2 in the new ISO 15189, references to ISO 15190 (health and safety)3, ISO 22367 (risk management)4 and ISO/TS 20658 (pre- examination processes)5.
The principles of the revision were reiterated and clarified in 2020 by WG1 of ISO/TC 212, confirming the normative reference ISO/IEC 17025, the directive to be less prescriptive, risk-based and linked to patient care, with inclusion of relevant aspects of POCT, taking into account other ISO documents, to avoid redundant repetition, and to synchronise relevant clauses in several ISO documents, including ISO TS 20914 (uncertainty)6 and the molecular diagnostic standards package developed by ISO TC 212 WG 4 (ISO 20166, ISO 20184, ISO 20186ISO 21474, ISO 23118 and others)7.
Thus, the need arises to review ISO 20914 for measurement uncertainty in medical laboratories and to set some guidelines for its application, with particular attention to "niche" areas that objectively present difficulties, as well as to examinations, present in all laboratory areas, with results that are not quantitative, i.e. qualitative, consisting of nouns or adjectives, both those with a rank or intensity relationship (ordinal) and those having only the identity or diversity relationship (nominal). The computational tools of laboratories should be able to handle the new concepts of uncertainty, as they already do with internal quality control.8 SIPMeL has already produced recommendations for information systems9 and for quality control10, addressed both to laboratories and to manufacturers of systems, devices and reagents. Also in this case, in the framework of the necessary close collaboration between laboratories and manufacturers, SIPMeL collects the main indications for the application of ISO standards, of the main literature and of the CLSI guidelines for the estimation of measurement uncertainty and its use in the validation and verification of examination procedures.
These Recommendations do not address the statistical and stochastic theoretical bases of the proposed solutions, for which reference is made to the original documents in the citations.