Real-world data (RWD) are playing an increasingly important role in clinical research, reshaping how medicines and medical devices are evaluated throughout their lifecycle.
Historically, clinical research has relied on the randomized controlled trial (RCT), considered the gold standard for demonstrating efficacy and safety.
However, as medical practice evolves, diseases become more complex, and operational constraints intensify, the limitations of RCTs have become more apparent.
In this context, real-world data and real-world evidence (RWE) are gaining prominence as complementary sources of knowledge.
According to the Haute Autorité de Santé (HAS), real-world studies contribute to public decision-making by informing reimbursement, defining conditions of use, positioning products within prevention, diagnostic or therapeutic strategies, and assessing their efficiency in price negotiations (HAS, Real-world studies for the evaluation of medicines and medical devices, June 2021)
Why real-world data are becoming essential in clinical research
Real-world data provide insight into how treatments and medical devices are actually used in routine clinical practice, beyond the controlled environment of clinical trials.
They offer a complementary perspective, closer to patients and healthcare pathways, capturing variability in adherence, comorbidities, and real-life use. For health authorities, real-world data help bridge the gap between experimental efficacy and real-world effectiveness, particularly when assessing long-term outcomes and broader populations.
What exactly are real-world data?
Real-world data refer to health data collected outside interventional clinical trials, in routine care settings.
They may originate from multiple sources, including: electronic health records (EHRs), medico-administrative databases (such as the SNDS in France), patient registries and cohorts, data generated by medical devices or digital health tools, patient-reported outcomes and experiences (PROMs, PREMs). Unlike data from RCTs, real-world data are not produced in a strictly controlled environment.
This is both their strength — as they reflect real clinical practice — and their main methodological challenge.
When are real-world data useful across the product lifecycle?
Complementing initial clinical evaluation
In some situations, conducting a conventional randomized trial may be difficult or even impossible:
rare diseases, highly targeted subpopulations, ethical constraints, or timelines incompatible with patient needs. In such cases, real-world data may be used to provide complementary comparative evidence, for example through external control arms.
However, the joint report by the Agence de l’Innovation en Santé (AIS) and F-CRIN stresses that these approaches must remain complementary and rely on rigorous methodology.
Monitoring effectiveness and safety in real-life conditions
After marketing authorization, real-world studies play a crucial role in documenting real-life effectiveness, long-term safety, and patterns of use in broader populations than those included in pivotal trials. The HAS has formalized this role in its methodological guide on real-world studies, applicable to both medicines and medical devices (HAS, Real-world studies for the evaluation of medicines and medical devices, June 2021).
Informing regulatory and market access decisions
Regulatory agencies now explicitly recognize the value of real-world evidence, provided that its generation meets expected methodological standards.
The European Medicines Agency (EMA) integrates RWE into certain regulatory decisions, particularly in post-authorisation contexts or when randomized evidence is limited (EMA guidance on real-world evidence).
Methodological requirements that cannot be overlooked
he use of real-world data in clinical research raises a central challenge: bias control.
Selection bias, confounding, and information bias can significantly affect the validity of results. To address these issues, specific methodological frameworks have been developed, including target trial emulation.
This approach consists of explicitly defining the hypothetical randomized trial that would ideally answer the research question, then emulating this design using observational data. Methodological reference papers published in the New England Journal of Medicine and JAMA demonstrate that, under strict conditions, target trial emulation can produce results consistent with randomized trials:
Concato & Corrigan-Curay, NEJM, 2022 https://www.nejm.org/doi/full/10.1056/NEJMp2200089
Wang et al., JAMA, 2023 https://jamanetwork.com/journals/jama/fullarticle/2803839
These studies emphasize the importance of pre-specification, transparency, and rigorous analytical choices.
An increasingly structured regulatory framework
In France, the HAS insists on the need to: pre-specify study protocols, document data quality, and precisely justify the use of non-interventional studies. At the European level, the EMA reiterates that real-world data are not a systematic alternative to randomized trials, but rather a complementary tool, suitable in well-defined contexts.
Conclusion: a powerful tool, to be used with discernment
Real-world data have become a strategic lever in modern clinical research.
They help better understand real-life use of health products, complement randomized trials, and strengthen post-marketing evaluation. However, their value depends above all on data quality, methodological rigor, and regulatory anticipation.
The challenge is therefore not to replace clinical trials, but to build an integrated approach, combining experimental data and real-world data, to support more relevant, patient-centered evaluation.
See article on the same subject : Real-World Data (RWD): towards a new clinical evidence paradigm for medicines and medical devices
