AbstractBackgroundExternal quality assessment (EQA) programs for general chemistry tests have evolved from between laboratory comparison programs to trueness verification surveys. In the Netherlands, the implementation of such programs has reduced inter-laboratory variation for electrolytes, substrates and enzymes. This allows for national and metrological traceable reference intervals, but these are still lacking. We have initiated a national endeavor named NUMBER (Nederlandse UniforMe Beslisgrenzen En Referentie-intervallen) to set up a sustainable system for the determination of standardized reference intervals in the Netherlands.MethodsWe used an evidence-based ‘big-data’ approach to deduce reference intervals using millions of test results from patients visiting general practitioners from clinical laboratory databases. We selected 21 medical tests which are either traceable to SI or have Joint Committee for Traceability in Laboratory Medicine (JCTLM)-listed reference materials and/or reference methods. Per laboratory, per test, outliers were excluded, data were transformed to a normal distribution (if necessary), and means and standard deviations (SDs) were calculated. Then, average means and SDs per test were calculated to generate pooled (mean±2 SD) reference intervals. Results were discussed in expert meetings.ResultsSixteen carefully selected clinical laboratories across the country provided anonymous test results (n=7,574,327). During three expert meetings, participants found consensus about calculated reference intervals for 18 tests and necessary partitioning in subcategories, based on sex, age, matrix and/or method. For two tests further evaluation of the reference interval and the study population were considered necessary. For glucose, the working group advised to adopt the clinical decision limit.ConclusionsUsing a ‘big-data’ approach we were able to determine traceable reference intervals for 18 general chemistry tests. Nationwide implementation of these established reference intervals has the potential to improve unequivocal interpretation of test results, thereby reducing patient harm.
Establishment of variation source and age-related reference interval models for 22 common biochemical analytes in older people using real-world big data mining