MEASUREMENT OF INTERRATER AGREEMENT WITH ADJUSTMENT FOR COVARIATES

The kappa coefficient measures chance-corrected agreement between two observers in the dichotomous classification of subjects. The marginal probability of classification by each rater may depend on one or more confounding variables, however. Failure to account for these confounde rs may lead to inflated estimates of agreement. A multinomial model is used that assumes both raters have the same marginal probability of c lassification, but this probability may depend on one or more covariat es. The model may be fit using software for conditional logistic regre ssion. Additionally, likelihood-based confidence intervals for the par ameter representing agreement may be computed. A simple example is dis cussed to illustrate model-fitting and application of the technique.