Analysis of Longitudinal Ordinal Data with Drop-Outs

Incomplete data are common problems in longitudinal studies. For incomplete longitudinal binary data, Fitzmaurice et al. (2001) discussed the impact on bias of the different estimating equation approaches where incompleteness follows a MAR (missing at random) process. They pointed out that GEE (generalized estimating equations) method proposed by Liang and Zeger (1986) has manifest bias as MAR drop-out rate increases. The main purpose of this article is to explore the performance of two group sequential tests based on GLMMs (generalized linear mixed models) and GEE models for analyzing longitudinal ordinal data under a variety of drop-out processes.