The impact of laboratory error on the normal range: a Bayesian model

Abstract

Interpretation of clinical laboratory results, aside from clinical considerations, is based on the probability of the result being within a given normal range. This probability is influenced by the degree of error inherent in the analytical method. It would be advantageous to assign a more definite probability to the result of the measurement by combining the error distribution of the result around the true value and the distribution of the healthy population that serves as a reference. Bayesian statistics permits the revision of this prior information into a single probability.