|
Activity Number:
|
249
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 8, 2006 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #306442 |
|
Title:
|
Sample Size Re-estimation for Time-to-Event Studies
|
|
Author(s):
|
Erinn Hade*+ and Soledad Fernandez and David Jarjoura
|
|
Companies:
|
The Ohio State University and The Ohio State University and The Ohio State University
|
|
Address:
|
320 W. 10th Ave., Columbus, OH, 43210,
|
|
Keywords:
|
sample size re-estimation ; survival ; failure rate ; time to event
|
|
Abstract:
|
Sample size re-estimation after an initial start-up period in clinical trials typically requires a planned interim analysis and re-estimation of the variance from the initial collected data. Authors have examined issues related to variance re-estimation and the timing of interim analyses, including options of blinded or unblinded data and the effect of unblinding on the bias of final estimates. Published work on sample size re-estimation has been closely examined for Gaussian and binomial outcomes. Some have explored time to event outcomes through a sequential testing design. We discuss issues related to sample size re-estimation for time to event analyses without the use of a sequential design and focus on studies of diseases that have low failure rates in both treatment and control arms. We present examples and simulations that show re-estimation can be critical with low failure rates.
|
