(A-211) Real-World Data for Evidence Generation Supporting FDA Approvals for Labeling Extension of Drugs and Biologics

Background: With improvements in acceptability of real-world data (RWD) after the enactment of the 21st Century Cures Act, use of real-world evidence (RWE) in the regulatory decision-making process has increased. However, the extent of its use in labeling extensions remains unclear.

Objectives: This study aimed to examine whether RWD was used in supplemental new drug applications (sNDAs) and biologics license applications (sBLAs), the role of RWE, and how it was included in labeling extensions, specifically focusing on submissions for adding a new indication or expanding the existing indicated population.

Methods: We reviewed FDA documents available through Drugs@FDA between January 2022 and May 2024 to identify sNDAs and sBLAs related to labeling extensions. RWE studies identified to support these labeling extensions were extracted from FDA’s prescribing information and review documents. Additional searches were conducted on ClinicalTrials.gov and PubMed to identify relevant RWE likely used in these submissions but not mentioned in FDA’s official documents. The characteristics of the identified RWE were then summarized.

Results: Of the 218 labeling extensions granted, 55 were found to include RWE, with 26 involving more than one RWE study. A total of 88 studies contributed to the RWE, with 4 studies identified in FDA documents and 84 sourced from ClinicalTrial.gov and PubMed. Of these 88 studies, 75% used electronic health records (EHR) as the data source, 5.7% combined EHR with patient-generated health data (PGHD), and 4.6% used registry data. Similarly, of the 55 labeling extensions with RWE, 71% incorporated EHR, followed by EHR combined with PGHD (11%) and EHR with registry data (7%). In both studies and approvals, EHR remained the most commonly used data source to support effectiveness, safety, and both. Notably, a higher proportion used EHR to support effectiveness (~74%) and both (~80%) than to support safety alone (~45%), whereas a higher proportion used PGHD to support safety (~18%). EHR is the most commonly used data source in 9 of the 14 therapeutic areas and remains the most frequently used across years. Furthermore, 36.3% of studies reported methods for handling missing data, 39.8% reported methods for bias assessment, and 21.6% reported both.

Conclusions: RWD is supporting regulatory decision-making for labeling extensions. However, our study identified only 4 RWE studies from FDA documents. FDA does not systematically describe RWE in its documents for supplemental applications; only the approval letter and updated labeling are available. FDA should enhance the transparency of its review to improve understanding of the data quality and methods deemed acceptable in labeling extensions.