Assessment of Agreements in Linear and Generalized Linear Mixed Models
Study of measuring agreement is intend to evaluate whether the readings from one rater/ measurement agree with those from other raters/measurements. In this dissertation, we are going to present a general method to assess agreement for a large variety of data with repeated measurements using linear and generalized linear mixed models. In the first place, a set of agreement statistics, including mean square deviation, concordance correlation coefficient, precision and accuracy coefficients, is presented for evaluating the intra-, inter-, and total-rater agreement in the multiple-rater and multiple-replications cases. Secondly, likelihood-based approaches are developed to estimate all the agreement statistics. Asymptotic properties of these estimates are also discussed for different data structures. Furthermore, our method has the merit of handling missing values and covariates naturally, and a new set of restricted agreement statistics is proposed in order to capture the true random variations and between-instrument effects adjusted for the covariate effects. Simulations for both linear and generalized linear mixed models are conducted to show the accuracy and effectiveness of our approaches. In the end, two industry datasets are evaluated using our approach. One is the cardiac function measurements used to determine the agreement between impedance cardiography and radionuclide ventriculography estimates, and the other one is an antihypertensive patch dataset given by FDA for assessing individual bioequivalence.
SubjectAssessment of Agreement
Concordance Correlation Coefficient
Generalized Linear Mixed Models
Restricted Maximum Likelihood