Recommendations for Methods and Criteria for Evaluating the Imprecision of Leukocyte (WBC) Differential Count (Proportional).

Abstract

The imprecision coefficient of variation (%CV) of leukocyte (WBC) differential counts (proportion, LDC), the mean proportion of a specific type of white blood cell (%x̅), and the finite number of white blood cells sampled by LDC (n) have a specific functional relationship. The criterion for assessing the imprecision of LDC is based on the %CV, which represents the standard error as a percentage of the mean. This criterion should fall within the prediction interval limits associated with %x̅ and n. An evaluation of the LDC can be based on the functional relationship among these 3 variables. The %CV is calculated from the other 2. Using this functional model, curve fitting is applied to the LDC imprecision data. Hypothesis testing is performed on the parameters of the fitted curve. This aims to identify and exclude entire experimental datasets that, after undergoing the fitting process, do not follow a binomial distribution. An indicator, CVx50E, which is used to evaluate the imprecision of LDC (CVx50E represents the %CV when %x̅ = 50%), is used to compare with the theoretical imprecision indicator of binomial distribution, CVx50B. Subsequently, those experimental datasets that still do not conform to the binomial distribution after the fitting process are excluded. The imprecision of LDC is reassessed using CVx50E. This comprehensive approach to evaluating LDC imprecision is applicable to various instruments and techniques, such as the manual-visual, digital imaging system, and flow-cytometric methods, as well as other measurement procedures where data follow a binomial distribution. This report provides recommendations on how to establish acceptable imprecision criteria for LDC, identify experimental data that do not conform to binomial distribution, and perform imprecision evaluation. The report contains examples of imprecision evaluation of LDC data from the literature.