Real-World Evidence

RWE is derived from Real-World Data (RWD) aggregated from heterogeneous sources outside traditional randomized controlled trials (RCTs).

• Electronic Health Records (EHRs): Structured and unstructured clinical data from routine care.
• Claims & Billing Data: Administrative records capturing diagnoses, procedures, and prescriptions.
• Patient Registries: Organized systems collecting uniform data on specific diseases or exposures.
• Patient-Generated Data: Information from wearables, mobile apps, and patient-reported outcomes. The analytical validity of RWE depends critically on the quality, completeness, and linkage of these source systems.

Unlike explanatory RCTs conducted under idealized conditions, RWE studies employ pragmatic designs that reflect real-world clinical complexity.

• Cohort Studies: Follow defined patient groups over time to compare outcomes between exposures.
• Case-Control Studies: Retrospectively compare patients with an outcome to those without to identify risk factors.
• Self-Controlled Designs: Use patients as their own controls to eliminate time-invariant confounding. These designs prioritize external validity and generalizability to broader, more diverse patient populations.

Observational RWE is inherently susceptible to confounding by indication and selection bias. Advanced statistical methods are mandatory to approximate causal inference.

• Propensity Score Matching (PSM): Balances treated and control groups on observed covariates.
• Inverse Probability of Treatment Weighting (IPTW): Creates a pseudo-population where treatment assignment is independent of measured confounders.
• Instrumental Variable (IV) Analysis: Leverages a variable associated with treatment but not the outcome to address unmeasured confounding. The credibility of RWE hinges on transparent pre-specification of these analytical approaches.

For RWE to support regulatory decisions, it must demonstrate fitness for purpose through methodological rigor and data reliability.

• Data Relevancy: The RWD must capture the exposure, outcome, and population of interest with sufficient granularity.
• Data Reliability: Completeness, accuracy, and provenance must be auditable via a clear data lineage.
• Transparency: Full protocol registration, pre-specified analysis plans, and public reporting of results are required. Regulatory frameworks from the FDA and EMA increasingly accept RWE for label expansions and post-market surveillance when these standards are met.

A defining advantage of RWE is the ability to assess long-term safety and effectiveness over extended time horizons impractical for most RCTs.

• Survival Analysis: Kaplan-Meier estimators and Cox proportional hazards models quantify time-to-event outcomes.
• Treatment Patterns: RWE reveals how therapies are sequenced, switched, and combined in routine practice.
• Rare Event Detection: Large, aggregated datasets provide the statistical power to identify infrequent adverse events. This longitudinal depth is essential for pharmacovigilance and understanding chronic disease trajectories.

RWE is not a replacement for RCTs but a complementary evidence source that addresses questions RCTs cannot answer.

• RCTs provide high internal validity for efficacy under controlled conditions.
• RWE provides external validity for effectiveness in heterogeneous, real-world populations.
• Hybrid Designs: Pragmatic trials and registry-based randomized trials merge randomization with real-world data collection. This synergy creates a comprehensive evidence package supporting both regulatory approval and clinical decision-making.

A machine-executable algorithm that identifies patient cohorts with specific clinical conditions from electronic health records. Computable phenotypes combine:

• Structured codes (ICD-10, SNOMED CT)
• Laboratory results and vital signs
• Medication orders and procedures
• Temporal logic (sequence and timing constraints)

These algorithms enable reproducible cohort definitions across institutions.

A statistical technique used in observational studies to reduce selection bias by pairing treated and control subjects with similar estimated probabilities of receiving the treatment based on observed covariates. This method attempts to approximate the randomization of controlled trials, making treatment effect estimates from RWE more credible for regulatory decision-making.

An extraneous variable that correlates with both the exposure and the outcome, potentially creating a false association or masking a true causal relationship. In RWE studies, common confounders include:

• Channeling bias: sicker patients receiving newer therapies
• Indication bias: treatment choice linked to prognosis
• Healthy user effect: patients adhering to preventive care differ systematically

Rigorous study design must account for measured confounders.

A legally binding contract that governs the terms, conditions, and permitted uses of limited or de-identified datasets shared between institutions for research purposes. In federated RWE networks, DUAs establish the governance framework for distributed analytics, specifying what queries are permitted and how results may be published without exposing patient-level data.