Robust Regression Techniques for Multiple Method Comparison and Transformation

Abstract

A generalization of Passing–Bablok regression is proposed for comparing multiple measurement methods simultaneously. Possible applications include assay migration studies or interlaboratory trials. When comparing only two methods, the method boils down to the usual Passing–Bablok estimator. It is close in spirit to reduced major axis regression, which is, however, not robust. To obtain a robust estimator, the major axis is replaced by the (hyper-)spherical median axis. This technique has been applied to compare SARS-CoV-2 serological tests, bilirubin in neonates, and an in vitro diagnostic test using different instruments, sample preparations, and reagent lots. In addition, plots similar to the well-known Bland–Altman plots have been developed to represent the variance structure.