Recommendations for generating real-world evidence with regulatory-grade relevance and reliability.

Real-world data (RWD) and real-world evidence (RWE) are increasingly incorporated into regulatory decision-making to complement randomized controlled trials (RCTs), particularly in settings where conventional trial designs are infeasible, such as rare diseases, rare molecular subtypes, and post-marketing evaluation. Across regulatory authorities, the use of RWD/RWE is consistently framed by the fit-for-purpose principle, centered on two foundational concepts: relevance and reliability. However, the absence of operational guidance has created uncertainty regarding how academic registries and healthcare databases can be designed or upgraded to meet regulatory expectations. Drawing on practical experience with disease registries in Japan and their regulatory applications, this review proposes a purpose-oriented framework clarifying the levels of relevance and reliability required according to specific regulatory objectives. We categorize considerations into three use cases: (1) drug development for rare diseases and rare molecular subtypes; (2) Pharmacovigilance; and (3) evidence generation for clinical questions that cannot be sufficiently evaluated through randomized controlled trials (RCTs). For each category, we outline essential requirements related to study design, data elements, quality management, governance, statistical planning, and operational feasibility. We further compare regulatory expectations across the U.S. Food and Drug Administration (FDA), the European Medicines Agency (EMA), the Pharmaceuticals and Medical Devices Agency (PMDA), and the International Council for Harmonisation (ICH). While international convergence is evident at the level of principles, operational thresholds and implementation frameworks differ. By articulating relevance and reliability as a shared regulatory language and emphasizing purpose-driven design and early regulatory engagement, this review provides practical recommendations for generating regulatory-grade RWE that meaningfully informs regulatory decision-making while complementing conventional clinical trials.