Measurement uncertainty in manual differential leukocyte counting

One of the most frequently requested examinations in the clinical laboratory is the differential leukocyte count. Despite many technological improvements, thousands of differential leukocyte counts are made by microscopic examination of a blood smear, counting 100 leukocytes and expressing the fraction of the specific leukocyte types as percentages (rounded to integer values) of the total leukocyte count. Although in the clinical laboratory it is not usual practice to report measurement uncertainties, currently the ISO 15189 standard considers measurement uncertainty as a very helpful element for a comprehensive interpretation of any measurement result.

The estimation of the measurement uncertainty of each differential leukocyte count result was carried out according to international guidelines. The sources of standard uncertainty taken into account were: pre-metrological variation, random distribution, between-examiner reproducibility, and rounding to an integer.

In this example, a sample of blood with a concentration number of leukocytes of 3,5x10(9)/L (1)) is taken into consideration. For each differential leukocyte count result, the standard uncertainties corresponding to each source of measurement uncertainty, as well as the combined and the expanded uncertainties, were estimated with information from the literature.

The procedure presented here to estimate the measurement uncertainty of differential leukocyte count results is appropriate to fulfill the requirements of the ISO 15189 standard related to measurement uncertainty. Knowledge of this uncertainty is helpful in interpreting sequential results obtained in the same patient.