Automated haematology: construction of univariate reference ranges for blood cell count and size

The theory of reference ranges advocated by the International Committee for Standardization in Haematology (ICSH) was tested experimentally for an aperture impedance and a flow cytometric system. The reference sample was drawn from volunteer blood donors and the sampling procedure carefully standardised. Subjects were partitioned for gender and smoking habits, and their ages noted. Reference ranges were constructed from Gaussian statistics, after objective evaluation of the data, based on an assumed occurrence of one false positive measurement in every twenty samples. Different reference ranges were obtained on the two different systems for some haematological parameters, and partitioning also produced differences in calculated reference ranges. We conclude that strict adherence to the ICSH recommendations is almost impossible. Nevertheless, it is important that all laboratories should construct their own local reference ranges for their specific analytical systems before cost effective clinical decisions can be made.