|
Activity Number:
|
163
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 9, 2004 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Biopharmaceutical Section
|
| Abstract - #301447 |
|
Title:
|
Influence of Withdrawal Criteria and Treatment Time-response Profiles on Dropout Process in Clinical Trials: A Simulation Approach
|
|
Author(s):
|
Naum M. Khutoryansky*+ and Andrew Liu
|
|
Companies:
|
Novo Nordisk Pharmaceuticals, Inc. and Novo Nordisk Pharmaceuticals, Inc.
|
|
Address:
|
100 College Rd. West, Princeton, NJ, 08540,
|
|
Keywords:
|
longitudinal clinical trials ; incomplete data ; dropouts ; dropout rate ; simulation
|
|
Abstract:
|
We propose a simulation approach to evaluate the effect of different withdrawal criteria and treatment time-response profiles on the dropout process and dropout rate in longitudinal clinical trials, particularly in diabetes clinical trials of phase III. The goal of simulation is to estimate in advance (before the trial) the average dropout rate for feasible combinations of withdrawal criteria and the treatment titration schedule. The probability of dropping out of the trial is presented as a sum of two probabilities associated and not associated with the effectiveness of the treatment and withdrawal criteria. The first probability is obtained by simulation and the second one is estimated using data from previous similar trials. The simulated clinical trials resemble diabetes clinical trials with repeated measures design where there is a measured continuous variable (e.g., fasting plasma glucose) that is used both as an efficacy parameter and as the indicator for withdrawal if it exceeds a certain threshold (constant or variable in time). Some illustrative examples based on those simulations are given to investigate the sensitivity of the dropout rate to the threshold variations.
|
