Dispersion models and longitudinal data analysis

Dispersion models provide a flexible class of non-normal distributions with many potential applications in biostatistics, accommodating a wide range o f continuous, discrete and mixed data. Starting with Liang and Zeger's gene ralized estimating equation method, we review some recent applications of d ispersion models in longitudinal data analysis, including state space model s based on the Tweedle class of exponential dispersion models. In medical a pplications the latent process of a state space model may often be interpre ted as an unobserved potential morbidity process, which is modelled as a fu nction of time varying covariates. By allowing a multivariate response vect or of 'symptoms', the model integrates several response variables mixed typ es into a single model. For growth curve models, the latent process reflect s the 'true' growth. Copyright (C) 1999 John Wiley & Sons, Ltd.