MethodCompare: An R package to assess bias and precision in method comparison studies.

Area of Science:

• Biostatistics
• Medical Device Evaluation
• Clinical Research Methodology

Background:

• Bland and Altman's limits of agreement are widely used in clinical research to compare measurement methods.
• However, traditional methods can be misleading when measurement error variances between methods are unequal.
• This limitation necessitates advanced statistical approaches for accurate method comparison.

Purpose of the Study:

• To introduce and evaluate a novel statistical methodology for assessing agreement between quantitative measurement methods.
• To address the limitations of existing methods when measurement error variances differ.
• To provide visual tools for comparing new measurement methods against established reference standards.

Main Methods:

• Implementation of a new statistical methodology developed by Taffé (2016) using the R package MethodCompare.
• Development of three novel plots: bias plot, precision plot, and comparison plot for visual evaluation.
• The methodology accommodates multiple reference measurements and at least one new method measurement per individual.

Main Results:

• The MethodCompare package provides a robust framework for agreement assessment, overcoming limitations of Bland-Altman.
• Visualizations (bias, precision, comparison plots) facilitate intuitive understanding of method performance.
• The approach is demonstrated effectively through three simulated examples.

Conclusions:

• The Taffé methodology, implemented in MethodCompare, offers a superior alternative for assessing agreement between measurement methods, particularly when error variances are unequal.
• The visual plots enhance the interpretation of new method performance relative to a reference method in clinical research.
• This approach is valuable for researchers developing and validating new quantitative measurement techniques.