Platelet counter accuracy and precision determine whether a point-of-care result can be trusted for clinical decisions or for PRP preparation. This page summarises how the accuracy and precision of the PC100 platelet counter were evaluated against Sysmex automated hematology analyzers, in human whole blood (20-600 platelets/nL) and in platelet rich plasma (250-3600 platelets/nL). The full method comparison is described in the evaluation of the analytical performance of the PC100.
How platelet counter accuracy and precision are assessed
Accuracy (trueness) describes how close a result is to the reference method and is expressed as bias. Precision describes how reproducible repeated results on the same sample are and is expressed as the coefficient of variation (CV). A platelet counter needs both: a device can be very precise but consistently biased, or unbiased on average but scattered. The PC100 evaluation followed the CLSI EP9 guideline for method comparison, using Passing-Bablok regression to quantify bias and duplicate measurements to quantify CV.
ACCURACY OF THE PC100 FOR WHOLE BLOOD COUNTS
Accuracy is a measure of closeness of the agreement between the quantified analyte values obtained by the tested method/device and the true values of the gold reference method/device. Systematic error or inaccuracy is quantified by the average bias, which is usually expressed as the constant and proportional bias from a regression procedure. Following the CLSI EP9 method comparison guideline, the passing-bablok linear regression method was used to find the best mathematical fit between the quantified platelets obtained by PC100 and the reference device.
There is a close agreement between the quantified platelet values of PC100 and Sysmex® XN-9000 Automated Hematology System. Results of regression analysis showed that the PC100 Platelet Counter can quantify blood platelets with an average bias (inaccuracy) of -3.362 (-2.312%) across the concentration range of 20-600 platelet/nanolitre. Accuracy is thus above 95%.
PRECISION OF the PC100 FOR WHOLE BLOOD COUNTS
Precision is a measure of closeness of agreement between quantified analyte values obtained by repeated measurements of the same sample/object under a specific set of conditions. Random error or imprecision is commonly expressed as the coefficient of variation (CV) calculated from the results of a set of duplicate measurements. Using duplicates for each sample, the CV% was calculated for the platelets of blood samples counted by PC100.
The PC100 Platelet Counter counts platelets with an average precision of 97.5% in the range of 20-600 platelet/nanolitre.
ACCURACY OF THE PC100 FOR PRP COUNTS
There is close agreement between the quantified platelet values of the PC100 Platelet Counter and Sysmex® XP-300 Automated Hematology System. Results of regression analysis showed that the PC100 Platelet Counter can quantify PRP platelets with an average bias (inaccuracy) of (2.0%) across the concentration range of 250-3600 platelet/nanolitre. Accuracy is thus above 95%.
Interpreting CV at low and high platelet concentrations
The CV of any cell counter increases at low concentrations, because fewer cells are counted per measurement and the statistical (Poisson) counting error becomes larger. Near the lower end of the whole-blood range (20-50 platelets/nL, relevant for thrombocytopenia) a somewhat higher CV is therefore expected for every method, including laboratory analyzers. At PRP concentrations thousands of platelets are counted per slide, so counting statistics are favourable; here platelet counter accuracy depends mainly on correct sample preparation and dilution.
English 
German 
French 
Spanish