Mathematics - Colloquium
Friday, April 26, 2013
3:00 PM-4:00 PM
ABSTRACT: In clinical evaluation of a drug, disease progress is often monitored through indicators with natural ordering depending on his/her disease status. In those investigations, data obtained are usually short individual time series of ordinal scores indicating the severity of the disease. The severity is usually judged through change in plasma level subsequent to dosing. The purpose of this study is to perform exposure–response (ordinal) modeling from pharmacokinetic view-point for epileptic patients received the drug Divalproex. As some patients withdrew prematurely in the late phase clinical trials, we propose a joint model for the dropout hazard and one-compartment PK model for longitudinal ordinal score that can be used to describe the change in the concentration of plasma . Incidentally the PK parameters appear to be time dependent so the procedure for estimating them becomes nonstandard. We propose a flexible Monte Carlo EM based two step estimation method and study its performance through a simulation study. In the context of Divalproex data , we explore the impact of dropout model on the ability of joint exposure–response relationship to predict observed longitudinal data patterns.
Suggested Audiences: Adult, College
Last Modified: April 3, 2013 at 3:08 PM