Advancing Real-World Evidence AI in Medicines Regulation

The adoption of Real-World Evidence AI is reshaping how the European Medicines Agency evaluates medicines, according to the European Medicines Agency’s annual report. By moving these technologies from pilot projects to essential regulatory tools, the EMA can now make better-informed decisions about safety, efficacy, and timely availability of treatments. This approach sits at the heart of the EU Medicines Agencies Network Strategy to 2028 and informs new pharmaceutical legislation.

The Data Analysis and Real-World Interrogation Network (DARWIN EU) has grown to 30 data partners across 16 countries, covering approximately 180 million patients. This expanding infrastructure pulls high-quality real-world evidence directly from routine healthcare settings, providing regulators with timely, representative insights that complement traditional clinical trials.

New Guardrails for AI Integration

Governance has advanced in parallel, with the release of the first Artificial Intelligence Observatory Report, a joint Heads of Medicines Agencies–EMA workplan on Data and AI in medicines regulation to 2028, and shared guiding principles with the US FDA for AI use throughout the product lifecycle. The milestone qualification opinion for AIM-NASH, an AI tool that assists pathologists in scoring liver biopsies for metabolic dysfunction-associated steatohepatitis, signals maturing pathways for regulatory acceptance of AI-driven analytics.

47 Percent Jump in Real-World Studies

Real-world data studies rose 47.5 percent year-on-year. DARWIN EU completed or initiated 67 studies that directly influenced regulatory actions, clarifying safety questions around doxycycline and suicidality risk, glucagon-like peptide-1 receptor agonist profiles, mpox vaccine effectiveness, and medicine shortage impacts.

From AI Experiments to Regulatory Impact

More than 100 artificial intelligence use cases are now under active exploration across the network. Deep learning algorithms for liver disease imaging, machine learning models tracking psoriatic arthritis progression, and digital twin simulations for rare-disease trials illustrate concrete progress. These applications of Real-World Evidence AI are moving beyond proof-of-concept toward routine use without compromising scientific rigor.

Strategic Pillar for 2028 and Beyond

These developments position Real-World Evidence AI as foundational elements of the EU Medicines Agencies Network Strategy to 2028. Structured frameworks now exist to integrate diverse data sources into health technology assessment and regulatory decision-making. For health economics and outcomes research professionals, the maturing ecosystem offers richer opportunities to link clinical development with real-world delivery, enabling more nuanced value demonstrations and faster patient access to innovation.