Limits Of Agreement Method

Barnhart HX, Yow E, Crowley AL, Daubert MA, Rabineau D, Bigelow R, Pencina M, Douglas PS. Selection of tuning indices to evaluate and improve the reproducibility of measurements in a central laboratory environment. Stat Methods Med Res. 2016;25 (6): 2939-58. doi.org/10.1177/0962280214534651. The five statistical methods for assessing conformity with repeated measurement data are described below. The five main methods are all based on linear models of mixed effects and are therefore based on similar (if not identical) assumptions. If the assumptions of the mixed model are not valid, the correspondence index is also not calculated on the basis of this model. In Table 1, you will find a list of assumptions and general modelling techniques that can be used to evaluate them.

Several different methods have been proposed in the literature to assess the concordance of continuous data, of which the concordance coefficient of concordances [3, 4] and concordance limits [5] are the most widespread. The probability of coverage [6], the overall deviation index [6, 7] and the coefficient of individual compliance methods [8, 9] were also described. All five methods can be calculated on linear models of mixed effects. With a focus on practical application and interpretation, the aim of this study is to show how these five approaches can be applied to the same convergence problem and to highlight the strengths and weaknesses of each method, allowing researchers to decide which methods should be used in their own studies. Reviews of evidence of compliance have already been published in the literature of Barnhart et al. (2007) [2], Obuchowski et al. (2015) [10], Barnhart et al. (2016) [11] and Barnhart (2018) [12]; with the last three titles with examples of real value to compare match indices.

However, the examples cited come almost exclusively from the fields of quantitative imaging and nuclear laboratory research. In this article, we extend the methodological work already done to the field of analysis of unbalanced clustered data in applied clinical research, particularly in the field of respiratory rate measurement in COPD patients. In addition, we focus specifically on the linear implementation of the mixed effects model of the methods and not on the more general approach used in the above-mentioned documents. In particular for compliance limits, this method is not taken into account in previous verifications. Emphasis is placed on the fact that mixed effects modelling is increasingly used in clinical research and has advantages over fixed effect methods (e.g. analysis of variance (ANOVA)) for various reasons described in Brown (2015) [13]. In particular, i) missing or unbalanced data pose fewer problems for analysis and (ii) conclusions can be drawn on the basis of a larger patient population [13]. .

. .