Strategic use of Big Data: implementing reference intervals for serum folate and serum cobalamin.

Introduction: Defining trustworthy reference intervals (RIs) for serum folate (FOL) or serum cobalamin (VITB12) is a difficult task. The purpose of this study is to use an indirect approach from the laboratory information's system to indirectly generate RIs for FOL and VITB12.

Materials and methods: A retrospective observational study was performed at a tertiary-care laboratory's hospital during 12 months. All FOL and VITB12 tests were measured using a Cobas8000 e801 system (Roche Diagnostics, Mannheim, Germany). The RIs were calculated using a non-parametric approach. The RIs established in the present study were verified by calculating the fraction of RIs that fell outside the new RIs, in two validation cohorts sampled using the direct and indirect method.

Results: A total of 19,214 (FOL) and 27,420 (VITB12) results were obtained. The RIs were 4.5 nmol/L (90% confidence intervals (CI) 4.4-4.6) to 38.4 nmol/L (CI 38.3-38.5) for FOL and 140 pmol/L (CI 139-141) to 659 pmol/L (CI 657-660) for VITB12. The verification included 8,798 FOL results and 7,365 VITB12 results. For both magnitudes was acceptable since only 0.1% of FOL and 0.02% of VITB12 results fell outside the RIs. Finally, the RIs were verified using a direct method with twenty individuals. For FOL 20/20 cases and 19/20 of VITB12 cases fell within the estimated RIs.

Conclusions: In summary, the use of an indirect data approach has enabled us to calculate RIs for FOL and VITB12. The RIs obtained in our study are lower than those proposed by the manufacturer for both FOL and VITB12.