Real-World Evidence in FDA Regulatory Strategy: Scientific Foundations and Health Economic Implications

The integration of real-world evidence (RWE) into regulatory pathways reflects a paradigm shift in biomedical innovation. Historically confined to post-market safety surveillance, real-world data (RWD) are now increasingly used to support pre-approval decisions, label expansions and even new indications for drugs and biologics. The U.S. Food and Drug Administration (FDA) has formalised this evolution through its Real-World Evidence Program and a series of guidance documents aligned with the 21st Century Cures Act.

However, the scientific and evidentiary standards for regulatory-grade RWE remain stringent. For real-world studies to influence FDA decision-making—particularly for efficacy claims or HTA adoption, they must meet high thresholds of internal validity, data provenance, transparency and causal inference.

This article explores the FDA’s position on RWE, the underlying methodological frameworks and the relevance for economic evaluation and international health technology assessment (HTA) alignment.

Real-world data (RWD) refers to information on patient health status or healthcare delivery collected through routine sources such as:

• Electronic health records (EHRs)
• Medical claims and billing systems
• Product and disease registries
• Digital health technologies (e.g., apps, sensors)
• Laboratory and imaging repositories

Real-world evidence (RWE) is defined as clinical evidence regarding the usage, safety or potential benefits and risks of a medical product, derived from the analysis of RWD.

This distinction underscores that RWD is the raw input whereas RWE is the analytically validated output and not all RWD leads to acceptable RWE.

In 2018, the FDA released its Framework for the Real-World Evidence Program [FDA, 2018], setting expectations for sponsors using RWD to support regulatory decisions under section 505(c) of the FD&C Act or for biologics under the Public Health Service Act.

Core Criteria for Acceptable RWE:

1. Fit-for-Purpose Data Quality
   • Completeness, accuracy, and reliability of RWD must be systematically assessed (e.g., structured vs. unstructured fields in EHRs).
   • Traceability and auditability of datasets is critical, especially for interoperability and data mapping.
2. Transparent Study Design and Causal Inference
   • Protocol pre-registration and prespecified statistical analysis plans are essential to mitigate data dredging.
   • Use of methodologies such as propensity score matching, inverse probability weighting or instrumental variables is encouraged where RCTs are infeasible.
   • FDA emphasises use of target trial emulation techniques to minimise bias [Hernán & Robins, 2016].
3. Regulatory-Grade Analytical Validation
   • Endpoints must align with validated clinical outcomes or surrogate measures accepted in prior FDA submissions.
   • Sensitivity analyses are required to test assumptions under missing data, confounding, or misclassification.
4. Data Standardisation and Submissibility
   • Use of CDISC/SDTM formats and structured vocabularies (e.g., MedDRA, SNOMED CT) ensures consistency and machine readability for FDA reviewers.

The growing use of RWE in FDA approvals has downstream effects on payer acceptance, HTA evaluations and economic modelling strategies.

1. Cost-Consequence and Budget Impact Models

Real-world datasets often contain utilisation metrics, such as hospitalisation rates, outpatient visits or procedural volumes, that can feed into cost-consequence frameworks. These models are often favoured in early HTA or payer dossiers (e.g., NICE MedTech evaluations or HAS early dialogues).

• RWE can validate assumptions around resource use in budget impact analyses (BIA) and calibrate unit costs in different settings (e.g., ICU, oncology infusion clinics).
• In decentralised healthcare systems (e.g., U.S., Germany), payers increasingly request local RWE to support formulary decisions and provider contracting.

2. External Controls in Rare or Rapidly Evolving Conditions

FDA has endorsed externally controlled studies—including historical controls or synthetic comparator arms—when randomisation is infeasible or unethical, such as in:

• Rare diseases (e.g., Duchenne muscular dystrophy, amyloidosis)
• Rapidly progressing conditions (e.g., glioblastoma, acute sepsis)
• Pediatric populations or advanced cancer

These models reduce time to market and may lower trial costs by 30–50%. However, acceptance by HTA bodies remains inconsistent: while France and Germany may require comparative effectiveness data, the U.K. (via NICE) increasingly considers these models under the HTEP and digital medtech pathways.

3. Longitudinal RWE for Lifecycle HTA

FDA has emphasised the need for long-term post-market RWE, not only for safety but also for effectiveness durability, adherence patterns, and treatment sequencing. These are highly relevant to:

• Value-based reimbursement contracts
• Health economic models with Markov states or survival-based outcomes
• Multinational HTA dossiers under EUnetHTA 21 guidance (2025)

Examples of RWE Regulatory Use in Practice

These examples highlight both the feasibility and regulatory utility of structured, high-integrity RWE platforms.

Strategic Recommendations for Sponsors and Developers

• Early engagement with FDA via INTERACT or Type B meetings is critical when planning to submit RWE as part of a regulatory package.
• Build partnerships with data custodians (e.g., Flatiron, TriNetX, OMOP-CDM sites) to ensure data governance and linkage strategies are robust.
• For global trials, align RWE with HTA-specific requirements, including ISPOR Good Practices for RWD/RWE and jurisdictional standards (e.g., HAS, IQWiG, NICE DSU).
• Leverage real-world datasets to populate economic models, not just for costing, but also for event probabilities, adherence and patient behaviour.

Real-world evidence is transitioning from a secondary consideration to a scientific asset with regulatory and economic weight. The FDA has developed a detailed and evolving framework to guide its integration, and life science innovators who wish to leverage RWE must adopt methodological rigor, ensure data transparency, and recognise its economic implications.  For stakeholders across regulatory, clinical, and HTA ecosystems, the message is clear: RWE is no longer optional—it is foundational.

Selected References

• FDA. Framework for FDA’s Real-World Evidence Program. 2018.
• Hernán MA, Robins JM. Using Big Data to Emulate a Target Trial When a Randomized Trial Is Not Available. Am J Epidemiol. 2016.
• FDA. Considerations for the Use of RWE to Support Regulatory Decision-Making. 2023.
• ISPOR. Real-World Data Task Force Reports. 2023.
• Tufts CSDD. Cost Reductions Through External Control Arms. 2021.
• NICE DSU. Real-World Evidence in Cost-Effectiveness Models. 2022.