On binary longitudinal mixed models in adaptive clinical trials

Abstract

In an adaptive clinical trial research, it is common to use data dependent design weights to assign individuals to treatments so that more study subjects are assigned to a better treatment. These design weights must be exploited for the consistent estimation of the treatment effect. In an adaptive longitudinal clinical set-up, the repeated responses of an individual will, however, be affected by the design weights as well as individual random effects and certain fixed time effects. In this article, we provide an estimation approach that takes the variability of the individual random effects and the longitudinal correlations of the repeated responses into account, and produces consistent and efficient estimate for the treatment effect. The performance of this approach is examined through a simulation study.