Influence of assay method differences on multiple of the median distributions: maternal serum alpha-fetoprotein as an example

Abstract

A straightforward statistical explanation is provided to show how differences between assay methods can affect the distribution of the multiples of the median (MoM). Evaluation of the impact of assay method differences reveals that the upper tails of the MoM distribution are not affected to the same degree as the lower tails of the distribution. The disparities in MoM distributions due to assay method differences result in various sensitivity/specificity combinations for different assays having the same fixed MoM cutoffs. Disparities do not exist if risks are calculated with use of the distributions for affected and unaffected populations that are based on a center's own assay method. Applying published risk tables, however, can affect the accuracy of the risk estimates. We used maternal serum alpha-fetoprotein as an example of an assay with an established history of reporting results in MoM values; however, the concepts presented apply equally well to any assay for which results are reported in MoMs.