Unified Framework of Meta-Analysis, Composite Likelihoods, and Bayesian Methods for Evidence Integration in Pediatric Extrapolation Development Archetypes

• Activity Number: 374 - Statistical Approaches for Evidence Integration and Considerations to Communication in Product Labeling
• Type: Topic-Contributed
• Date/Time: Thursday, August 12, 2021 : 12:00 PM to 1:50 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Margaret Gamalo-Siebers* and Junjing Lin
• Companies: Pfizer Inc. and Takeda Pharmaceuticals
• Keywords: extrapolation; hierarchical models; composite likelihood; Bayesian models; priors

Abstract

Contingent on the indication being pursued, streamlining the pediatric drug development program is facilitated with the awareness of the mechanism of action of the drug, the feasibility of extrapolation, and the strategic goal within a landscape to ensure timely access to drugs. With these considerations, there are a common development archetype, e.g., whether a drug belongs to an established class where trials in adults and children have been conducted and that the “line of reasoning” for efficacy between adults and children have been established. Interestingly, these development archetypes induce distinctive designs or analytical strategies as they have different levels of evidence in support of extrapolation. This presentation will focus on evidence integration within these archetypes through multiple analytical strategies and how they are related to each other in terms of the questions they answer. Examples of these methodologies range from the meta-analysis, hierarchical models, and composite likelihoods. The accuracy of the estimates from these models is balanced with safety and appropriately communicated in drug product labels for effective and safe use in children.