Secondary use of external quality assessment data - estimating inter-assay variation in LOINC-coded datasets

The secondary scientific use of routine laboratory data will increasingly rely on LOINC as a semantic standard, particularly within the European Health Data Space (EHDS). LOINC enables the aggregation of large data sets from clinical care for research purposes, particularly in epidemiology. However, current approaches to semantic standardisation largely neglect metrological aspects - in particular, the considerable limitations of analytical standardisation for many fundamental analytes and the resulting scatter of result values across different assays that use common LOINC codes. This incomplete harmonisation leads to statistical uncertainty that must be taken into account when quantitative conclusions - such as diagnostic thresholds for analytes - are derived from aggregated, LOINC-derived data sets. In this opinion piece, we propose using the extensive global data pool generated by external quality assessment (EQA) programs to finally annotate LOINC codes with a sound and useful uncertainty metric. This represents secondary scientific use of EQA data that is analogous to and supports the secondary use of routine diagnostic data from patient care for research. With a proof-of-concept analysis, we demonstrate the feasibility of this approach, which offers a wide range of design options. We suggest that consortia of EQA providers, coding institutions, scientific societies, and the IVD industry could advance precision research through this concept. It is noteworthy that the proposed annotation strategy - linking semantic test codes to uncertainty metrics based on EQA data - is not limited to LOINC as a semantic coding system.