Achieving accuracy for routine clinical chemistry methods by using patient specimen correlations to assign calibrator values. A means of managing matrix effects

Abstract

Accurate results obtained from routine methods used in the clinical laboratory can be achieved if the methods are traceable to definitive or reference methods or are based on widely accepted methodologic principles. At Eastman Kodak Co (Rochester, NY), we assign calibrator values for each method that is available on Kodak Ektachem analyzers by correlation to a reference method or a methodologic principle using patient specimens. Kodak's Reference Laboratory is responsible for maintaining the accuracy and consistency of the reference method or methodologic principle by standardizing with National Institute of Standards and Technology materials, where available, monitoring performance through statistical process control, participating in proficiency testing programs, and following procedures consistent with the International Organization for Standardization guideline, "General Requirements for the Competence of Calibrations and Testing Laboratories (ISO 25). "Patient samples that span the reportable range are run in replicate on the reference method or methodologic principle over several days. These patient sample results are then used to create a calibration curve based on the relationship of concentration to the raw response of the routine method, where the same patient samples are also run in replicate on the same day. Each day, multiple vials of stable calibrator fluids are included in these tests, along with control materials. Each calibrator is assigned a value of equivalent concentration (in millimoles per liter) by predicting the concentration that would be present at the measured response if the test fluid were a patient sample. These are termed supplementary assigned values (SAVs). The same calibrators can be used to compensate for matrix effects resulting from changes in reagent formulations, by appropriately adjusting the assigned values. For example, when a process improvement for Kodak Ektachem clinical chemistry slides for phosphorus was introduced, an SAV change of 0.39 mmol/L (1.2 mg/dL) was required for one calibrator to maintain accuracy of the patient samples. A comparison of patient samples, using the SAVs specific for each formulation, gave excellent correlation (in millimoles per liter; r2 = 0.995) by the following equation: New Formulation = 1.02 (Old Formulation) -0.016; Sy.x = 0.074. The benefits of this approach to calibration include (1) accuracy is traceable to a reference method or methodologic principle and, wherever available, to reference materials; (2) new calibrator values (SAVs) can be assigned whenever the reagent formulation changes to maintain accurate patient results; and (3) if matrix effects are present, they will not adversely influence accuracy because traceability is based on patient sample comparisons.