Principles and Approaches for the Use and Evaluation of Fit-for-Purpose, Real-World Data Sources

• Activity Number: 78 - Key Steps in Deriving Real-World Evidence from Analysis of Real-World Data
• Type: Topic-Contributed
• Date/Time: Monday, August 9, 2021 : 10:00 AM to 11:50 AM
• Sponsor: Biopharmaceutical Section
• Author(s): Mark Levenson* and Weili He
• Companies: FDA and AbbVie
• Keywords: electronic health records; machine learning; misclassification; bias assessment

Abstract

FDA and industry have initiated numerous efforts to leverage rich data sources from routine medical practice to address the benefits, risks, and usage of medical products. The use of these data sources (real-world data (RWD)) and the evidence derived from them should meet the established rigor and standards required for regulatory decisions. In 2018, FDA release a Framework for its real-world evidence program. As part of the Framework, FDA will consider the fitness for use of the data. The ASA Biopharmaceutical Section Real-World Evidence Working Group is currently considering statistical principles to define and evaluate the fitness of use of RWD. The talk will present an initial landscape review of data integrity, data relevance and reliability, computable phenotypes, and validation and assessment of RWD variables.