The Posterior Probability of Passing a Compendial Standard

• Activity Number: 411
• Type: Topic Contributed
• Date/Time: Tuesday, August 6, 2013 : 2:00 PM to 3:50 PM
• Sponsor: Biopharmaceutical Section
• Author(s): David LeBlond*+ and Linas Mockus
• Companies: CMCStats and Purdue University
• Keywords: Bayesian Methods ; Compendial Standard ; Unit Dose Uniformity ; Operating Characteristics ; ASTM E2703

Abstract

Compendial tests, such as unit dose uniformity (UDU), represent standards against which product batches may be accepted or rejected. A lower bound on the probability of acceptance for such tests can be used as a batch acceptance or process validation criterion. Conformance to the lower bound can be determined from test data using sampling theory arguments (e.g., ASTM E2709). Such probability estimates reflect product quality and should also be useful to both regulators and manufacturers for risk/ manufacturing/ cost assessment as well as life-cycle management. However, sampling theory probability estimates are not always ideal for such uses. We found that in general Bayesian Inference facilitates sound decision making based on probability estimates and leads to more intuitive conclusions. Methodology for estimating the posterior probability of passing the UDU test is described and its advantages compared to traditional sampling theory approaches are illustrated.