Between-examiner reproducibility in manual differential leukocyte counting

In manual-visual differential leukocyte counts, the counting itself may be different depending on the examiner. We estimated the between-examiner (B-E) reproducibility in differential leukocyte counting. During 2 months, daily, after performing manual-visual differential leukocyte counts, two slides with blood films were retained. Next day these two blood films were re-examined by different examiners. For each type of leukocyte the B-E reproducibility was estimated from the 58 pairs of data obtained by four technicians. The B-E coefficient of variation for each type of leukocyte was: basophilocytes, 263.2%; eosinophilocytes, 68.8%; lymphocytes, 32.5%; metamyelocytes, 69.6%; monocytes, 55.0%; myelocytes, 132.5%; and neutrophilocytes, 6.6%. For each type of leukocyte the coefficient of variation is the sought estimate of the average reproducibility due to the "human factor" as a whole, without regard to any examiner in particular. Knowledge of this component of day-to-day imprecision, and its follow-up, may be used to decide corrective actions (training of the examiner, etc.) and to set critical differences to interpret the significance of changes in serial results.