Background: A scatter plot combined with correlation and regression analysis to assess the agreement between analytical methods that measure the same quantity has a long tradition in clinical chemistry. However, only when linearity is absolute this approach may be useful to estimate whether the methods under comparison produce the same errors in measurement. The aim of this study is to provide a new original regression-based tool which does not suffer from the problems of the classical regression/correlation method and that is also easy to use and to interpret.
Methods: To evaluate the performance of the new method, the results of some analytical determinations were compared by the new procedure, the Bland-Altman test, Passing-Bablok regression and Mountain plot.
Results: All the tests used to compare the data allowed superimposable interpretations, both when methods under comparison were in agreement and in disagreement.
Conclusions: The understanding of the new test is truly intuitive and it is reduced to the appraisal of how much the equation of the best-fit line (Y=X) is able to predict the value of the Y. The goodness of this prediction can be evaluated by the coefficient of determination, whose value is closely connected to the differences between the methods under comparison. The result of our simulation is that quite high agreement between the methods occurs for determination coefficient equal or greater than 0.98.