|
Activity Number:
|
436
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 5, 2009 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Statistical Learning and Data Mining
|
| Abstract - #305329 |
|
Title:
|
Power Simulation Using Nonlinear Bayesian Prediction Model on Alcohol Dependence Literature Database
|
|
Author(s):
|
Yun-Fei Chen*+ and Haoda Fu
|
|
Companies:
|
Eli Lilly and Company and Eli Lilly and Company
|
|
Address:
|
Lilly Corporate Center, Indianapolis, IN, 46285,
|
|
Keywords:
|
Alcohol Dependence ; literature database ; Nonlinear Bayesian ; Integrated Two-component Prediction ; survival analysis
|
|
Abstract:
|
Alcohol dependence is a relatively new target of drug development. Acamprosate is FDA approved alcohol dependence medication to prolong the periods of abstinence. Acamprosate literature database is used to simulate study patient level attributes for trial design. We used 23 placebo and 16 Acamprosate studies with study duration between 56 to 360 days to predict the percent of abstinence at 6 months for each group. A new method referred as Integrated Two-component Prediction (ITP) model (Fu, Manner, 2008) to estimate inter and intra subject/study variability as well as predict long term out come is applied. Based on the predicted percent of abstinence from ITP model, we simulated the time to first relapse patient data. By altering sample size, effect size and time of assessment using power simulation on survival analysis, an interim analysis strategy is meaningfully constructed.
|
