What Is Real-World Evidence
- Zhifei Zeng
- 6 days ago
- 5 min read
From Clinical Trials to the Real World
Data collection has always been an indispensable part of the drug development process. Traditionally, after a drug is approved for market release, structured data collection from real-world settings has primarily focused on pharmacovigilance for safety monitoring. However, in recent years, both the academic community and regulatory agencies have increasingly recognized that relying solely on clinical trial data is insufficient to comprehensively reflect a drug's efficacy and usage in real-world healthcare settings. By systematically collecting and analyzing real-world data (RWD), researchers can obtain critical insights beyond safety, providing valuable references for clinical practice and policy-making.
RWD refers to various types of data related to patient health status and healthcare services collected in routine clinical practice. Its sources are diverse, including electronic health records (EHR), medical claims data, disease or product registries, and patient-generated data from mobile devices and digital health technologies. Real-world evidence (RWE) generated from these data can complement traditional clinical trial evidence, providing a more comprehensive understanding of drug efficacy, safety, and disease management. Compared to the highly controlled data generated by randomized controlled trials (RCTs), RWD is often large-scale and complex in form, presenting challenges in collection, management, and analysis. With the development of artificial intelligence and big data analysis methods, the potential of RWD is gradually being unlocked, providing an important supplement to address the limitations of RCTs and explore the efficacy and safety of drugs.
Examples of Real-World Data (RWD)
Among various sources, healthcare databases are considered one of the most important forms of RWD. EHR systems record daily clinical practice and laboratory test information, which, after systematic analysis, can be used to explore drug efficacy, safety, quality of life, and the natural history of diseases. For example, the U.S. FDA's “Sentinel Initiative” integrates national databases to enable real-time monitoring of medical product safety; the European Health Data and Evidence Network promotes cross-border health outcomes research through a unified data model.
Registers are also critical data sources. They collect clinical data on specific diseases or treatment populations in a prospective manner, reflecting the characteristics of larger, more representative populations. They are frequently used in epidemiological research and drug safety monitoring. For example, the European Cystic Fibrosis Society Registry monitors disease progression and drives care improvements, while the UK's National Cancer Registration and Analysis Service provides core evidence for cancer trends and public health policies.
Additionally, medical claims databases primarily collect medical service billing and insurance payment information, covering inpatient, outpatient, and pharmacy services, and are widely used for drug utilization and long-term efficacy assessments. The Medicare and Medicaid databases in the United States, as well as the Ontario Pharmacy Evidence Network (OPEN) in Canada, provide important support for health policy development and service optimization.
How to generate real-world evidence (RWE) from real-world data (RWD)
RWD can be transformed into RWE through a variety of experimental and observational study designs, each suited for different research purposes.
Cohort studies are widely used to assess disease incidence, risk factors, natural history, and treatment outcomes, and can be either retrospective or prospective—for instance, the XANTUS study validated the real-world efficacy and safety of rivaroxaban in stroke prevention.
Cross-sectional studies collect information from a population at a single point in time, making them efficient for evaluating disease prevalence, treatment patterns, or medication adherence.
Case-control studies, typically retrospective, are particularly useful for identifying risk factors for rare diseases or those with long latency periods, allowing simultaneous assessment of multiple variables.
Registry studies, based on prospectively collected long-term data, help monitor disease trends, as shown by Taiwan’s cancer registry analysis, which revealed key epidemiological features of AML in Asian populations.
Claims and EHR database studies also provide valuable insights: claims data are often used to assess healthcare resource utilization and costs, while EHRs support evaluation of treatment effectiveness and long-term outcomes—for example, analyses from the US Renal Data System demonstrated a lower bleeding risk with apixaban compared to warfarin in dialysis patients with atrial fibrillation.
Finally, pragmatic clinical trials bridge the gap between RCTs and real-world practice, conducted in routine care settings to better reflect clinical reality, such as studies showing the lack of significant efficacy of erythropoietin in acute kidney injury. Together, these diverse methods highlight the flexibility and breadth of RWE research in addressing unanswered clinical questions.
The Challenges of Using Real-World Evidence (RWE)
Real-world evidence (RWE) is becoming increasingly important in medical research, but it is not easy to transform the chaotic data from everyday medical practice into reliable conclusions. Researchers need to overcome a number of obstacles.
First is the issue of missing and imbalanced data. Hospitals and clinics do not collect data in a standardized manner, and some important information may be omitted, resulting in “missing pieces of the puzzle.” If not handled properly, these gaps can lead to biased research results.
Second is the issue of bias. Unlike clinical trials, treatments in real-world settings are not randomly assigned but decided jointly by doctors and patients. This introduces bias. For example, patients with more severe conditions often receive more intensive treatments, making it difficult to directly compare study results.
Third is the challenge of privacy and data access. Patient health information is highly sensitive, and laws strictly regulate its use. While this is intended to protect patients, it also makes it difficult for researchers to obtain sufficient and high-quality data. Currently, new methods are being explored, such as “de-identified” data or digital health identities, to advance research while protecting privacy.
Another issue is duplication and inconsistency. A patient's information may be scattered across electronic health records, insurance claims, pharmacy records, and even health apps. Without careful matching, the same patient may be counted as multiple different individuals in a study. Researchers are therefore attempting to develop secure identity verification systems.
Cross-regional and cross-system differences also pose challenges. Different hospitals, countries, and databases use varying formats, languages, and standards, making global data integration and analysis more complex.
Finally, regulatory standards are still evolving. Agencies like the US FDA and European EMA are establishing new rules to ensure the collection and use of RWE are more transparent and reliable. In the future, more public RWE registration systems will be launched to enhance the quality and credibility of research.
Conclusion
RWE is becoming increasingly important in the healthcare sector. By leveraging data from everyday clinical practice, researchers, companies, and policymakers can gain a better understanding of how treatments work in the real world (rather than just in clinical trials). Despite ongoing challenges, technological advances and closer collaboration among stakeholders are helping to unlock the potential of RWE. Through continued efforts, RWE can play a key role in ensuring that patients have faster, safer, and more cost-effective access to effective medications.
Sources
Assessed and Endorsed by the MedReport Medical Review Board
