|
Activity Number:
|
78
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Sunday, August 2, 2009 : 4:00 PM to 5:50 PM
|
|
Sponsor:
|
Biopharmaceutical Section
|
| Abstract - #304202 |
|
Title:
|
Shelf Life Determination with Common Intercept
|
|
Author(s):
|
Roswitha E. Kelly*+ and Jinglin Zhong and Meiyu Shen
|
|
Companies:
|
FDA and FDA and FDA
|
|
Address:
|
WO21 RM 4656, Silver Spring, MD, 20993,
|
|
Keywords:
|
shelf life determination ; common intercept ; fixed effects model
|
|
Abstract:
|
We explored a special case of shelf life estimation, namely when a product of a given strength is packaged into various package types. In the case where the batch term is crossed with package type, there is only one set of release data for any given attribute. We investigated four cases which deal with this situation: 1) repeating the release data for each package type; 2) assigning the release data randomly to a package type; 3) forcing each regression line through the average of the release data; and 4)deleting the release data from the shelf life estimation. We applied real stability data to each model and employed simulations to establish their respective bias. This talk will present the authors' findings and their current thinking regarding the most appropriate model for shelf life estimation when release data are associated only with the batch factor.
|
