Comparaison de deux méthodes de mesure d'une même grandeur : méthode de Bland et Altman

Abstract

Comparison of two methods of measurements: the Bland-Altman analysis. Bland-Altman analysis for comparison of two methods of clinical measurement is frequently used in scientific publications. This article is more appropriate than the convetional linear regression analysis. This paper gives an overview of the principles for the use of Bland-Altman analysis as well as the specific terminology attached to it. The Bland-Altman comparison analysis is mainly a tool for clinical interpretation. The bias and the agreement limits provide the variation of the values of one technique compared to the other. The difference between the two methods of measurement is plotted against the average obtained with each of the two techniques. Bland-Altman analysis can also be used to check the repetability of a measurement technique within the same subject and to determine a repetability coefficient. With an adaptation of the calculation of the agreement limits, the average of multiple measurements for each subject with two measurement techniques can be used for the Bland-Altman analysis.