Continuum analysis method applied to blood count reference intervals.

Conventional methods for extracting reference values from a sample of a purported normal population have a tendency to include some paranormal individuals and to obscure the existence of marginal or subclinical disorders. The suggestion is offered of defining the normal range of an analyte by extrapolation of the portion of its dispersion that fits a continuum. The method softens the present dichotomous character of the reference range. It narrows the normal range of some hematology parameters, suggesting that some paranormal individuals are excluded.