CSP 2017 Online Program

A New Method to Assess Measurement Agreement in Machine Readings (303468)

Keywords: statistical assessment, concordance correlation coefficient, Bland-Altman plot, bias

In a clinical study measurements are often taken using old and new machines at different time epochs. Due to budgetary and time constraints, the principal investigator should determine whether the machines are compatible to each other based on a pair of single measurements taken from a small subgroup of participating patients. In this study we investigated the behavior of the concordance correlation coefficient (CCC) and Bland-Altman plots via computer simulations. We found that CCC tends to produce high scores even when there is a significant amount of bias, and Bland-Altman plots are rather conservative. Here we propose a new and simple procedure to address the shortcomings of these methods and illustrate the procedure using real world examples.