Simple pattern-mixture models for longitudinal data with missing observations: analysis of urinary incontinence data

In longitudinal studies each subject is observed at several different times. Longitudinal studies are rarely balanced and complete due to occurrence of missing data. Little proposed pattern-mixture models for the analysis of incomplete multivariate normal data. Later, Little proposed an approach to modelling the drop-out mechanism based on the pattern-mixture models. We advocate the pattern-mixture models for analysing the longitudinal data with binary or Poisson responses in which the generalized estimating equations formulation of Liang and Zeger is sensible. The proposed method is illustrated with a real data set.