Following public consultation, the EMA has finally published its Reflection paper on use of real-world data in non-interventional studies to generate real-world evidence for regulatory purposes, providing clearer guidance on regulatory expectations and methodological standards. Key points include:

• Regulatory applications: clarification on when Real World Data (RWD) studies may support regulatory submissions.
• Artificial Intelligence (AI): if AI is used for data processing, methodologies, potential bias, and impact on results must be detailed in the study protocol.
• Primary vs. secondary data collection: new guidance addresses feasibility, selection and information bias, data accuracy, and ethical considerations when RWD is collected through primary data collection rather than secondary use.
• Patient involvement: greater emphasis on engaging patients and advocacy groups in study design and governance.
• Governance, ethics and data privacy: reinforced alignment with GDPR and EU data protection standards.
• Data quality and reliability: Stronger focus on validation of data coding algorithms and reference to the HMA-EMA Data Quality Framework.
• Statistical analysis: clearer distinction between sensitivity and supplementary analyses, with recommendations to pre-specify these in protocols and discuss them early with regulators.

The  aim is to enhance the scientific and ethical robustness of RWD studies for regulatory decision-making, thereby supporting their broader use in clinical research. This is particularly important in the field of rare diseases, where limited data and recruitment challenges hinder the generation of new evidence.