A systematic assessment of key design and performance characteristics of drug exposure registries requested by the U.S. Food and Drug Administration

Purpose: The purpose of the study is to evaluate contributions to postmarket safety assessments and identify potential factors for enhancing implementation and utilization of registries in regulatory decision-making.

Methods: Registry documents (e.g., protocols, reports) submitted to the FDA were identified up to January 2016 through an extensive, systematic review of internal records and resources. We characterized nonpregnancy drug exposure registries based on prespecified design elements, performance, and regulatory impact.

Results: A total of 65 registries were identified: 56 registries were open and 9 closed. Among open registries, 20% were pending, 14% delayed, and 16% ongoing less than ≤3 years. Most registries (82%) examined safety issues that originally arose from clinical trials; most frequent safety issues investigated included infections, gastrointestinal dysfunction, and liver toxicity. Although 74% of registries ascertained baseline health conditions and monitored concomitant medication use, fewer (45%) considered drug exposure duration or dosage. Thirty-seven percent of non pending registries had enrollment below sample size expectation. Seventeen registries published findings in journals/conference proceedings, 13 from open registries. Three closed registries generated results that contributed to product label changes. High-performance registries scored higher in design metrics related to sample size considerations (76% versus 62%) and adequate analysis plan (53% versus 35%), and interim report submission (76% versus 65%). There was a significant difference in proportion of registries with clear primary objectives between high versus not high performing registries (100% versus 78%).

Conclusions: This study suggests clear objectives, patient accrual/retention efforts, adequate analysis plans, and interim reports contribute to the performance of drug exposure registries.