Evaluating agreement between two analytical methods in clinical chemistry

Abstract

A test and a reference analytical method are usually compared for agreement based on paired data obtained from several independent subjects. Bias between two methods can be classified as constant and proportional. Bias is modeled as an average bias between two methods (constant bias) and proportional bias related to individual measurements. A weighted least square approach is followed for estimating the parameters. Estimates of constant and proportional biases are tested individually and their significance can be used to explain the sources of disagreement between two methods and help deciding a remedial strategy. The proposed model can also be used to determine an optimum combination of subjects and runs per subject and facilitate the allocation of the resources.