Group sequential analysis op a clinical trial with repeated measurements

Recently some work has been done in group sequential analysis for comparing two treatments with respect to the rate of change in a repeated measures design with staggered subject entries. We relax the involved assumption of linearity for the time-response curve and propose a group sequential procedure assuming multivariate normality with an ante-dependence structure for the repeated measurements obtained at protocol scheduled visits. The exit probabilities are determined through an a-spending function proposed by Lan and DeMets. A numerical example is given to illustrate the methodology.