Development and validation of effective real-time and periodic interinstrument comparison method for automatic hematology analyzers.

OBJECTIVES We developed and validated an interinstrument comparison method for automatic hematology analyzers based on the 99th percentile coefficient of variation (CV) cutoff of daily means and validated in both patient samples and quality control (QC) materials. METHODS A total of 120 patient samples were obtained over 6 months. Data from the first 3 months were used to determine 99th percentile CV cutoff values, and data obtained in the last 3 months were used to calculate acceptable ranges and rejection rates. Identical analyses were also performed using QC materials. Two instrument comparisons were also performed, and the most appropriate allowable total error (ATE) values were determined. RESULTS The rejection rates based on the 99th percentile cutoff values were within 10.00% and 9.30% for the patient samples and QC materials, respectively. The acceptable ranges of QC materials based on the currently used method were wider than those calculated from the 99th percentile CV cutoff values in most items. In two-instrument comparisons, 34.8% of all comparisons failed, and 87.0% of failed comparisons were successful when 4 SD was applied as an ATE value instead of 3 SD. CONCLUSIONS The 99th percentile CV cutoff value-derived daily acceptable ranges can be used as a real-time interinstrument comparison method in both patient samples and QC materials. Applying 4 SD as an ATE value can significantly reduce unnecessarily followed recalibration in the leukocyte differential counts, reticulocytes, and mean corpuscular volume.