A Simulation Study to Evaluate Dose-Response in Dose-Titration Clinical Trials: A Dynamic Linear Mixed Effect (DLME) Modeling Approach

• Activity Number: 256
• Type: Contributed
• Date/Time: Monday, August 2, 2010 : 2:00 PM to 3:50 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Xu Steven Xu*+ and Min Yuan and Julia Wang and Partha Nandy
• Companies: Johnson & Johnson and Fudan University, China and Johnson & Johnson and Johnson & Johnson
• Keywords: dose response ; titration design ; feedback ; simulation ; dynamic model ; selection bias

Abstract

Assessing dose response from dose-titration clinical trials is often problematic due to the selection bias caused by the feedback from the response process to the dosing process. We investigate the performance of a DLME in identifying dose response from titration studies using simulations. The simulations demonstrate that DLME estimator is unbiased (< 0.5%). A nominal 95% confidence interval for the dose-response based on the DLME estimator covered the true value in ~85% of the simulations, showing that the asymptotic confidence interval has reasonable coverage. Compared to the situation, where dose titration is randomly assigned and not related to the response (no feedback), the efficiency of the DLME estimator only reduced by 12% when the feedback is present. In conclusion, DLME models may provide a modeling option in identifying dose response relationship in dose-titration designs.