|
Activity Number:
|
529
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 5, 2009 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Biopharmaceutical Section
|
| Abstract - #304239 |
|
Title:
|
Estimating Shelf Life Using Quantile Regression with Random Batch Effects
|
|
Author(s):
|
Michelle Quinlan*+ and Walter Stroup and James Schwenke
|
|
Companies:
|
University of Nebraska-Lincoln and University of Nebraska-Lincoln and Boehringer Ingelheim Pharmaceuticals, Inc.
|
|
Address:
|
340 Hardin Hall North, Lincoln, NE, 68583,
|
|
Keywords:
|
quantile regression ; shelf life ; random batch effect
|
|
Abstract:
|
Shelf life is the time period a product remains capable of acceptable performance. Traditionally, shelf life is estimated by modeling the mean response. However, other rationale requires focus on modeling a percentile of the response distribution. Quantile regression provides a methodology for doing so. Quantile regression involves minimizing an asymmetrically weighted sum of absolute errors and is implemented by methods including simplex and interior point. Theory for quantile regression with fixed effects has been developed. However, there is no analogous methodology for quantile regression with random effects. Ad hoc methods with random effects are being investigated. An example and discussion of SAS® code for quantile regression with random batch effects are presented for estimating shelf life from multi-batch stability data.
|
