Online Program Home
My Program

Abstract Details

Activity Number: 29
Type: Contributed
Date/Time: Sunday, July 31, 2016 : 2:00 PM to 3:50 PM
Sponsor: Quality and Productivity Section
Abstract #319926
Title: Frequentist and Bayesian Simulation Using a Random Coefficients Model to Establish Shelf-Life Specification Limits for a Drug Product
Author(s): Richard Montes* and David LeBlond
Companies: Hospira, a Pfizer Company and CMCStats
Keywords: specification limits ; simulation ; random coefficients model (RCM) ; Frequentist vs. Bayesian ; ANCOVA ; release and stability testing
Abstract:

Setting proper specification limits is a component of the overall control strategy to ensure drug product quality and consistency. Statistical analyses of release and stability data is often used to set specification limits. Ad hoc methods based on a fixed effects model (Analysis of Covariance), as adapted from Allen et al. (1991), have been applied in the pharmaceutical industry. Such methods employ worst-case assumptions so the derived limits may be too conservative. A simulation approach using random coefficients model (RCM) to account for random lot effects and analytical variability more probabilistically than current ad hoc methods is proposed in this paper. The RCM estimates are used to simulate future values evaluated at the product expiry. Quantiles of which at specified coverage are designated as shelf-life specification limits and compared to population quantiles to get a confidence coefficient. Both Frequentist and Bayesian versions are explored. The results of this study inform the choice of analysis options against a range of realistic scenarios so that when applied in actual specification setting situations, acceptably conservative operating performance is assured.


Authors who are presenting talks have a * after their name.

Back to the full JSM 2016 program

 
 
Copyright © American Statistical Association