Introduction
In early 2026, the U.S. Food and Drug Administration (FDA) signaled a meaningful shift in how real-world evidence (RWE) can be used. This change has the potential to reshape both premarket and post-market strategy—yet many manufacturers are still underestimating its practical impact.
Recent developments indicate a growing willingness to accept high-quality real-world data not only for post-market surveillance, but also in support of premarket submissions and label expansions.
This shift reflects an evolution in the FDA’s long-standing RWE framework, building on earlier guidance such as the 2018 Real-World Evidence Program Framework and the 2017 guidance on the use of RWE for medical devices. Manufacturers can use real-world data to demonstrate safety and performance where data integrity, methodological rigor, and documentation are clearly established.
This article examines what has changed in the FDA’s approach to RWE, where real-world evidence can now be applied across the regulatory pathway, and what manufacturers should do differently when planning evidence generation strategies. In practice, many teams continue to treat real-world evidence as a supplementary dataset rather than a strategic component of evidence generation—despite clear regulatory signals to the contrary.
What Has Changed in the FDA’s Approach to Real-World Evidence
Recent developments in the FDA’s real-world evidence (RWE) framework signal a shift from cautious acceptance toward more practical integration of RWE into regulatory decision-making. While earlier guidance established the scientific principles for evaluating real-world data, current updates suggest a broader and more flexible application across the product lifecycle.
The FDA’s Real-World Evidence Program Framework (2018) and the guidance Use of Real-World Evidence to Support Regulatory Decision-Making for Medical Devices (2017) laid the foundation for incorporating RWE into regulatory submissions. However, these documents were often interpreted conservatively, with real-world evidence primarily used for post-market surveillance or supportive analyses.
More recent regulatory signals, including updates to FDA communications and policy direction, indicate a reduced barrier to using RWE in both drug and device applications. In particular, there is increasing acceptance of real-world data to support:
- Label expansions and new indications
- Post-approval study requirements
- Select premarket submissions, where appropriate data quality and study design can be demonstrated
This represents a shift in emphasis. Rather than asking whether real-world evidence can be used, the focus is now on whether the data are sufficiently reliable and relevant to answer the regulatory question.
For manufacturers, this change has practical implications. Real-world evidence is no longer limited to a supplementary role but may be considered a viable component of an overall evidence strategy.
In practice, this means that high-quality registry or electronic health record (EHR) data may now be considered in contexts where randomised clinical trials were previously expected.
The expanded role of real-world evidence (RWE) is best understood by examining where it can be applied across the regulatory pathway. While RWE has traditionally been used to support post-market activities, current FDA guidance and recent developments suggest a broader role across both premarket and lifecycle decision-making.
Post-Market Surveillance and Long-Term Safety
Post-market surveillance is where RWE has traditionally lived—and where it’s most mature today. Data collected through registries, electronic health records (EHRs), and post-approval studies can be used to monitor long-term safety and performance in real-world clinical settings. In this context, RWE supports ongoing benefit–risk evaluation and may inform updates to safety labeling or risk management strategies.
Label Expansions and New Indications
Real-world data can also support label expansions by providing evidence of how a product performs in patient populations or clinical scenarios not fully represented in premarket trials. For example, claims databases or registry data may be used to demonstrate effectiveness in broader or more diverse patient groups, particularly where conducting new randomised trials is not feasible.
Premarket Submissions: A Shifting Landscape
The most significant recent development is the increasing acceptance of RWE in premarket submissions. While randomised controlled trials remain central to establishing safety and effectiveness, the FDA is showing greater openness to high-quality real-world data in specific contexts.
For example, a manufacturer preparing a 510(k) submission for a Class II device may use registry or EHR data to demonstrate performance in a contemporary patient population, particularly where the predicate device is based on older clinical data. Similarly, in De Novo or PMA submissions, real-world evidence may be used to supplement clinical trial data or address specific evidence gaps, where the data source and study design meet FDA expectations.
This means that real-world evidence is no longer limited to a supporting role after approval. It may now be considered earlier in the evidence generation strategy, particularly where traditional trials are impractical, incomplete, or insufficient to capture long-term outcomes. In practice, FDA pushback is often linked to insufficient justification of data sources or lack of alignment between the study design and the regulatory question being addressed.
Practical Use of RWE Across the Regulatory Pathway
Key Classification Documents:
MDCG 2021-24: Guidance on
classification of medical devices MDCG 2022-5: Guidance on borderline
products MDCG 2019-11: Guidance on
qualification and classification of software
What the FDA Expects: Data Quality, Study Design, and Documentation
As expectations around real-world evidence (RWE) continue to evolve, the FDA continues to emphasise that acceptance depends on whether the data are sufficient to support the intended use. The regulatory question being addressed—whether related to safety, effectiveness, or label expansion—determines the level of rigor required.
In practice, one of the most common issues in RWE submissions is overestimating the reliability of electronic health record data without adequately addressing missing data, inconsistent coding, or unclear endpoint definitions.
Data Reliability and Relevance
Real-world data must be both reliable and relevant to the intended use. This includes clear data provenance, consistent data capture, and sufficient completeness to support meaningful analysis. Data sources such as registries, electronic health records, and claims databases should be well-characterised, with an understanding of how data are collected, coded, and maintained.
Study Design and Control of Bias
Because real-world evidence is typically derived from observational data, study design is critical. The FDA expects manufacturers to address potential sources of bias, including confounding, selection bias, and missing data.
Appropriate methodological approaches may include predefined study protocols, use of comparator groups, and statistical techniques such as propensity score matching or sensitivity analyses. Transparent study design is essential to ensure that findings are scientifically valid and reproducible. Regulatory feedback frequently highlights gaps in study design, particularly where confounding and selection bias have not been addressed at the protocol stage.
Transparency and Documentation
Regulatory acceptance of RWE depends heavily on transparency. Manufacturers must clearly document data sources, study design, analytical methods, and underlying assumptions.
This includes maintaining detailed records that allow regulators to understand how conclusions were reached and to evaluate the limitations of the data. In many cases, early engagement with the FDA may be necessary to align on study design and data acceptability before submission.
For example, data intended to support a label expansion will typically require a higher level of methodological robustness than data used for routine post-market surveillance.
Limitations of Real-World Evidence
Despite its expanding role, real-world evidence (RWE) remains subject to important limitations that must be addressed in regulatory submissions.
Unlike randomised controlled trials, real-world data are observational and may be affected by confounding, selection bias, and incomplete data capture. These factors can introduce uncertainty into study findings if not appropriately controlled through study design and statistical methods.
Data integrity also remains a key consideration. Real-world datasets—particularly electronic health records and claims data—are often collected for clinical or administrative purposes rather than research, which may result in inconsistent coding, missing variables, or limited outcome definitions.
For this reason, the FDA continues to evaluate RWE on a case-by-case basis, with particular attention to whether the data and methodology are sufficient to support the specific regulatory question.
As a result, RWE is most effective when used to complement, rather than replace, high-quality clinical trial data in regulatory decision-making.
A Step-by-Step Approach for Manufacturers Using Real-World Evidence
As the FDA’s approach to real-world evidence (RWE) evolves, manufacturers should take a structured approach to integrating real-world data into their regulatory strategies. The following steps outline a practical framework for applying RWE in line with current expectations.
In practice, integrating RWE into a regulatory submission depends less on data availability and more on structured planning.
Step 1: Identify the Regulatory Question
Start by clearly defining the regulatory objective. This may include supporting a label expansion, addressing a specific safety concern, or filling an evidence gap in a premarket submission. The intended use of RWE will determine the level of data quality and methodological rigor required.
Step 2: Assess Whether RWE Is Fit for Purpose
Evaluate whether real-world data can appropriately address the question. This involves considering the availability, completeness, and relevance of data sources such as registries, electronic health records, or claims databases. Not all regulatory questions can be answered using RWE alone.
Step 3: Select and Characterise Data Sources
Once RWE is deemed appropriate, identify suitable data sources and assess their quality. This includes understanding how data are collected, coded, and maintained, as well as evaluating data provenance, consistency, and potential limitations.
Step 4: Design the Study with Regulatory Expectations in Mind
Develop a study protocol that addresses potential sources of bias and confounding. This may include defining comparator groups, selecting appropriate endpoints, and applying statistical methods such as propensity score matching. Study design should be aligned with FDA expectations from the outset.
Step 5: Ensure Transparency and Documentation
Maintain clear documentation of data sources, study design, analytical methods, and assumptions. Regulatory acceptance of RWE depends on the ability to demonstrate how conclusions were reached and to allow for independent evaluation of the evidence.
Step 6: Engage Early with Regulators Where Appropriate
For higher-risk or more complex applications, early engagement with the FDA may help align expectations regarding data acceptability and study design. This can reduce uncertainty and improve the likelihood of successful integration of RWE into the submission.
Implications for Manufacturers
As the regulatory landscape evolves, manufacturers should consider how real-world evidence can complement traditional clinical development strategies. Through careful planning, integrating real-world data into evidence generation plans may help companies better understand product performance, identify safety signals earlier, and support regulatory interactions.
The Future Role of Real-World Evidence in Regulatory Decision-Making
The role of real-world evidence (RWE) in regulatory decision-making is expected to continue expanding as healthcare data infrastructure and analytical capabilities improve. Increasing availability of large-scale datasets, including electronic health records and registries, is enabling more comprehensive evaluation of long-term safety and real-world performance.
At the same time, advancements in statistical methods and data analytics are improving the ability to address bias and confounding in observational data. These developments are supporting more consistent and reliable use of RWE across a wider range of regulatory applications.
For manufacturers, this suggests that RWE is likely to become an increasingly important component of evidence generation strategies, particularly in areas where traditional clinical trials are limited or impractical.
Conclusion
The FDA’s evolving approach to real-world evidence reflects a broader shift in regulatory science—from cautious acceptance toward more practical integration of real-world data across the product lifecycle.
Recent developments indicate that RWE is no longer limited to post-market use but may play a meaningful role in premarket submissions, label expansions, and long-term safety evaluation. However, acceptance remains dependent on whether the data are appropriate for the regulatory question, with clear expectations around data integrity, study design, and transparency.
Manufacturers who continue to treat real-world evidence as a secondary data source risk falling behind. Those who integrate it strategically will be better positioned to navigate regulatory pathways more efficiently and with greater flexibility—ultimately enabling more relevant and decision-ready evidence.
Frequently Asked Questions
Can real-world evidence replace randomised clinical trials?
Real-world evidence may complement clinical trial data in certain contexts, but it does not replace randomised controlled trials where high levels of evidence are required for establishing safety and effectiveness.
When can RWE be used in a 510(k) submission?
RWE may be used to support a 510(k) submission where it helps demonstrate performance in a contemporary patient population or addresses gaps in predicate data, where robust data and well-defined study design meet FDA expectations.
What types of data are acceptable for RWE?
Common sources include electronic health records, registries, claims databases, and patient-reported outcomes. Acceptability depends on data reliability, relevance, and the ability to support the specific regulatory question.
Latest Articles
ATTNs Ethan Glasser Dover, Delaware, 19901, United States
