Marginalized Models for Longitudinal Count Data



Heagerty (2002) and Lee and Daniels (2007) have proposed marginalized transition
models for the analysis of longitudinal binary data and ordinal data, respectively.
In this talk, we will propose similar models for longitudinal count data.
We will also propose a model to accommodate overdispersed count data.
Fisher-scoring algorithms are developed for estimation. Methods are
illustrated with a real dataset (Epileptic seizure data) and are compared with other
standard methods.