Studies in which patients are observed on several occasions over a period of time are particularly common in medical research. Most clinical trials, for example, are of this format. The characteristic feature of such studies is that multiple measurements of a response variable are obtained at a set of time points for each participant in the study. There are two main difficulties in the analysis of data from such longitudinal studies: The analysis is complicated by the dependence among repeated observations made on the same experimental unit. There may be a substantial proportion of missing values due to participants not attending at all scheduled visits or perhaps dropping out of the study altogether.