RWEReal-World Evidence

The two terms are often used interchangeably and they should not be. Real-World Data (RWD) is data relating to patient health status or the delivery of health care routinely collected from a variety of sources. Real-World Evidence (RWE) is the clinical evidence about the usage and potential benefits or risks of a medical product derived from analysis of RWD. RWD is the input; RWE is the output.

Common RWD sources include electronic health records, medical and pharmacy claims, product and disease registries, patient-generated data from in-home or wearable devices, and other data sources informing health status (mortality registries, social determinants of health datasets). None of these are collected for the purpose of supporting a regulatory submission — that is the central methodological challenge of RWE.

RWE entered US drug regulation through §3022 of the 21st Century Cures Act (Public Law 114-255, December 2016), which directed FDA to evaluate the potential use of RWE for two specific purposes: supporting approval of a new indication for an already-approved drug, and supporting or satisfying post-approval study requirements. The legislation did not contemplate RWE replacing the traditional NDA RCT package for a novel drug.

FDA published its initial RWE Framework in December 2018 and has since issued a series of guidance documents: 'Considerations for the Use of Real-World Data and Real-World Evidence To Support Regulatory Decision-Making for Drug and Biological Products' (final, August 2023); 'Data Standards for Drug and Biological Product Submissions Containing Real-World Data' (final, December 2023); and ongoing guidances on EHR/claims data, registries, and sponsor responsibilities.

For medical devices, FDA's 'Use of Real-World Evidence to Support Regulatory Decision-Making for Medical Devices' (final August 2017, updated 2023) has been operational longer, reflecting that device decisions historically draw on more post-market and registry data than drug decisions.

On the EU side, EMA's DARWIN EU (Data Analysis and Real World Interrogation Network) became operational in 2022 and is now central to EMA's RWE workflow. It coordinates federated queries across a network of contributing data partners (national health databases, insurance datasets) without centralising patient data.

RWE is now part of the regulatory toolkit for several decision types, but it is rarely the sole evidence.

• Label expansion for already-approved drugs — new indications, new populations, new dosing — where an external control arm constructed from RWD complements a single-arm trial.
• Post-marketing safety and effectiveness — RWE is the dominant evidence for routine pharmacovigilance and PMR/PMC commitments.
• Rare-disease and orphan products — where a traditional RCT is infeasible and a natural-history dataset functions as the comparator.
• Device post-market surveillance — registry-based studies are routinely accepted by FDA and EU notified bodies.
• Pragmatic effectiveness studies that complement a registered RCT — capturing how the product performs outside the controlled trial environment.

RWE does not replace the pivotal RCT for first-in-class drug approvals. FDA's 2023 guidance was explicit that the strength of RWE-based conclusions depends on the quality of the underlying data, the comparability of the comparator population, and the ability to address confounding — and that these challenges become progressively harder for novel-product first approvals.

ALCOA+ — attributable, legible, contemporaneous, original, accurate, complete, consistent, enduring, available — applies to RWE source data, but the assessment is harder than for protocol-driven trial data. An EHR entry is contemporaneous and attributable by virtue of clinical workflow, but its accuracy for a specific regulatory question depends on coding practice at the site, completeness of the chart, and whether the relevant outcome was even captured.

FDA's 2023 RWD/RWE guidance and the December 2023 data standards guidance set expectations for data-source provenance, traceability from source through analysis dataset, and the relevance and reliability assessment that sponsors must include in the submission. Sponsors are expected to characterise the data source's strengths and limitations for the specific regulatory question — not just submit summary tables.

An external control arm is a comparator group constructed from RWD or from historical clinical trial data, rather than from a concurrent randomised arm. They are now accepted by FDA in specific contexts — rare diseases, oncology with strong natural history, settings where a randomised comparator is unethical — but require explicit justification of the comparability assumption, statistical adjustment for measurable confounding, and sensitivity analysis for unmeasured confounding.

FDA's 'Considerations for the Design and Conduct of Externally Controlled Trials for Drug and Biological Products' (draft, February 2023) lays out the expectations. The comparator must be representative of the target population, outcome ascertainment must be comparable, and time-period bias (treatment in 2020 vs control in 2010) must be addressed.

1. Define the regulatory question — label expansion, post-marketing commitment, pragmatic effectiveness study — before selecting a data source.
2. Conduct a data-source feasibility — does the EHR/claims/registry actually capture the population, exposure, comparator and outcome of interest?
3. Pre-specify the analysis in a protocol (PASS for EMA pharmacovigilance studies, fit-for-purpose protocol for FDA) before looking at outcomes.
4. Document the data-source provenance, transformations and lineage end-to-end — the analysis dataset must be traceable back to source.
5. Engage the regulator early — FDA's RWE Sub-Committee and EMA's DARWIN EU both prefer pre-submission interaction.