All Times EDT
Keywords: Estimand, MMRM, SLOPE
Many clinical trials collect data on a quantitative outcome at multiple time points, with a primary focus on comparing the mean slopes of the change from baseline outcome over time between treatment groups. A commonly used analysis approach is based on a longitudinal mixed model where the dependent variable is the change from baseline outcome at each time point and the fixed effects include time (as a continuous covariate), treatment group, a time by treatment group interaction term, and some baseline covariates. One alternative approach is to estimate each patient’s slope and to then perform an ANCOVA to test the difference in the mean slopes between groups. In this talk we will discuss the estimands addressed by these analysis approaches, including our concerns with hypothetical estimands. We will also present the pros and cons of these two approaches via simulation studies and examples of clinical trials.