Time-Variant Response Patterns and Modeling Precise Assessment Time in Confirmatory Clinical Trials

• Activity Number: 129
• Type: Contributed
• Date/Time: Monday, August 2, 2010 : 8:30 AM to 10:20 AM
• Sponsor: Biopharmaceutical Section
• Author(s): Yoko Tanaka*+
• Companies: University of Pittsburgh Graduate School of Public Health
• Keywords: confirmatory clinical trials ; time-response patterns ; time-variant PK/PD patterns ; precise time assessment ; GEE models

Abstract

In confirmatory clinical trials where an optimal treatment dose group and a placebo group are repeatedly observed, one challenge is to capture precise assessment time, and another is to model time accordingly, i.e., on a continuous time scale. To show the latter, we simulated 3 time-response patterns, quadratic, Emax, and logistic curves. From these respective continuous time-response patterns, data values at 7 time points were sampled around a constant visit window for 3000 participants. Normally distributed baseline and time-dependent covariates were simulated. To demonstrate the impact of continuous versus discrete time (based on visit indicator), we applied General Estimating Equation (GEE) models including continuous or discrete time. Discrepancies in parameter estimates and statistical inference were evaluated and showed bias when using different time scales.