A Bayesian Modeling Approach for Safety Data Analysis in Drug Development

• Activity Number: 128
• Type: Contributed
• Date/Time: Monday, August 2, 2010 : 8:30 AM to 10:20 AM
• Sponsor: Biopharmaceutical Section
• Author(s): Yu Gu*+ and Ke Zhang and Liqiang Yang
• Companies: Florida State University and Pfizer Inc. and Pfizer Inc.

Abstract

Safety planning and evaluation becomes a primary focus in drug development and occurs throughout a product's life-cycle. The use of modeling-based quantitative approaches plays an important role in the identification of potential safety issues and the assessment of the relation between important factors and safety response. We propose a Bayesian logistic mixed effect model to analyze AE data. System organ class is also considered in the model; therefore, the information within the same system organ class can be borrowed to help model the rate of AEs. The multiplicity issue in AE data analysis is easily solved by the Bayesian framework. In addition, the additional important risk factors can be easily added in the model. The Bayesian posterior predictive information can also help study teams do safety planning for the late-phase drug development.