Group sequential methods have been used for a correct application of interim analysis, which is conducted to allow for possibly early termination of a alinical trial for ethical, economical and administrative considerations. The classical group sequential methods are applied for cross-sectional data and the boundaries can be easily computed due to the property of independent increment structure (IIS) between the successive test statistices. owever, it does not hold for longitudinal data. For analyzing longitudinal ordinal data, group sequential methods based on generalized linear mixed models (GLMM) and generalized estimating equations (GEE) models are proposed. The performance ofthese two approaches are compared with respect to their type I error rate and power bysimulation studies. The proposed methods are demonstrated by a real data set with ordinal responses.
