A former senior mathematical statistician at FDA
Date & Time
October 8, 2021, 11:00 am – 12:00 pm
Abstract: This presentation will compare two rank based methods for analyzing multicenter clinical trials using mixed model. The first rank based method is based on Friedman type rankings (intra-center) and the second rank based method is based on joint ranking (Wilcoxon-type). These two methods are nonparametric alternatives to the mixed model normal theory analysis for the multicenter clinical trials. Both rank-based methods offer a complete analysis, including estimation of fixed effects and their standard errors, tests of linear hypotheses and diagnostics. These methods will be compared with the normal theory method using an example from a clinical trial and by simulations.