|
Activity Number:
|
1
|
|
Type:
|
Invited
|
|
Date/Time:
|
Sunday, August 2, 2009 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Biopharmaceutical Section
|
| Abstract - #303006 |
|
Title:
|
Interval Censoring in Time-to-Progression Data
|
|
Author(s):
|
Zhenming Shun*+
|
|
Companies:
|
sanofi-aventis
|
|
Address:
|
2902 Johnson Circle, Bridgewater, NJ, 08807,
|
|
Keywords:
|
Interval censoring ; PFS ; Oncology trials ; Right censoing ; time-to-event data ; type-I error
|
|
Abstract:
|
In clinical trials for cancer treatments, tumor assessment data are likely interval censored. Our research is intended to provide a theoretical explanation and quantification for the relationship of the treatment effects versus censoring intervals. We consider the exponential distribution with an exponential regression model for simplicity and use the interval method to handle the censoring. When the censoring interval is large drug effects for individual treatment groups are inflated; but the treatment difference between treatment groups is deflated. The variability of point estimate is increased due to the censoring interval but the standard error can be reduced by increasing the sample size. Any imbalance in the censoring intervals between the treatment groups will inflate the type-I error but the inflation can be reduced using interval method for censoring handling.
|
