Generalized linear random effects models with varying coefficients

摘要

In longitudinal studies where subjects are measured repeatedly, the effect strength of covariates may vary with time. Additionally, observations taken at one individual have to be considered as dependent. These two aspects are accommodated in a model which allows for time-varying coefficients and include subject-specific effects to account for heterogeneity of subjects. The proposed estimation procedure is based on a localized version of the marginal likelihood approach in generalized linear models with random coefficients. The efficiency of the estimation procedure is demonstrated in a small simulation study and an application to real data is given.