Real world evidence (RWE) is clinical evidence about how a medical product works, including its benefits and risks, drawn from data collected outside of traditional clinical trials. It comes from analyzing information generated during routine healthcare: doctor visits, insurance claims, disease registries, and increasingly, wearable health devices. The concept has become central to how drugs and medical devices are evaluated after they reach the market, and it’s playing a growing role in regulatory approvals themselves.
Real World Data vs. Real World Evidence
The distinction between real world data (RWD) and real world evidence (RWE) matters, even though the terms are often used interchangeably. Real world data is the raw material: information about patient health and healthcare delivery collected routinely from a variety of sources. Real world evidence is what you get when you analyze that data to draw conclusions about how a treatment performs.
Think of it this way: a hospital’s electronic health records contain millions of data points about patients, their diagnoses, and treatments. That’s RWD. When researchers analyze those records to determine whether a cancer drug works as well in everyday patients as it did in a clinical trial, the findings they produce are RWE. The quality of the evidence depends on both the analytical methods and the reliability of the underlying data, which is why regulators evaluate both layers separately.
Where the Data Comes From
Real world data is pulled from sources that already exist in the healthcare system rather than being generated by a purpose-built study. The most common sources include:
- Electronic health records (EHRs): clinical notes, lab results, prescriptions, and diagnoses captured during routine care
- Medical claims and billing data: records submitted to insurers that track diagnoses, procedures, and medications at a population level
- Patient registries: organized databases that collect information about people with a specific disease or who receive a particular treatment
- Digital health technologies: data from wearable devices, sensors, and health apps that continuously monitor metrics like heart rate, activity levels, or blood glucose
Each source has strengths and gaps. EHRs capture rich clinical detail but often contain unstructured notes that are difficult to analyze at scale. Claims data covers enormous populations but tells you what was billed, not necessarily what happened clinically. Registries are purpose-built and well-organized but may only capture a slice of the patient population. Wearable devices generate continuous, objective measurements but are still being validated for use in formal regulatory studies.
How RWE Differs From Clinical Trials
Randomized controlled trials (RCTs) remain the gold standard for proving a drug works. By randomly assigning patients to receive a treatment or a placebo, and by blinding both patients and doctors to who gets what, RCTs minimize bias and isolate the effect of the drug itself. No other study design controls for unknown confounding factors as effectively.
But RCTs have real limitations. They enroll carefully selected patients who may not resemble the broader population that will actually use the drug. They’re expensive, logistically complex, and take years to complete. For rare diseases, running a large randomized trial can be impractical or even impossible. And once a drug is approved, it’s often unethical to run a new placebo-controlled trial just to study it in a different patient group.
RWE fills these gaps. Observational studies using real world data can examine how treatments perform in populations that were underrepresented in trials: older adults, people with multiple chronic conditions, racial and ethnic minorities. They can track outcomes over much longer time periods than a typical trial allows. And they can answer questions about comparative effectiveness, showing how two approved treatments stack up against each other in everyday practice rather than against a placebo.
The tradeoff is rigor. Without randomization, observational studies are vulnerable to selection bias and confounding. Patients who receive one treatment over another may differ in ways that affect their outcomes, and those differences can be difficult to fully account for. Researchers use sophisticated statistical methods to adjust for these imbalances, but the results generally carry less certainty than a well-run RCT. Transparency in methodology and adherence to established reporting guidelines are critical for RWE to be taken seriously.
How Regulators Use RWE
Both the FDA and the European Medicines Agency (EMA) have built formal frameworks for incorporating real world evidence into regulatory decisions, though their approaches differ in some respects. The FDA defines real world data broadly enough that even clinical trials can generate RWD if the data is collected through routine healthcare systems. The EMA takes a narrower view, restricting the definition to data collected outside of traditional clinical trials.
In practice, RWE is used most heavily in two regulatory areas: expanding the approved uses of existing drugs (label extensions) and monitoring safety after a product reaches the market.
A recent analysis of FDA approvals found that roughly 25% of drug label extensions between 2022 and 2024 included available real world evidence. The proportion was consistent across those three years, hovering between 23% and 28%. Oncology dominated, accounting for nearly 44% of submissions that included RWE, followed by infectious disease and dermatology. Nearly half of the real world evidence studies submitted addressed both safety and efficacy together.
The EMA published updated guidance in 2025 on using real world data in non-interventional studies, and maintains catalogues of vetted data sources that companies can draw from when conducting these studies. International harmonization efforts are underway to align terminology and standards across regulatory agencies worldwide.
Post-Market Safety Monitoring
One of the most established uses of RWE is tracking drug safety after approval. Clinical trials typically enroll hundreds to a few thousand patients. Rare side effects that occur in one out of every 10,000 or 100,000 users simply won’t show up in a trial of that size. They only become visible when millions of people start taking the drug in real clinical practice.
By analyzing electronic health records, insurance claims, and registry data, researchers can detect safety signals that emerge over months or years of widespread use. This kind of surveillance has been instrumental in identifying unexpected risks with approved medications, leading to updated warnings, restricted use, or in some cases, market withdrawal. It’s the primary reason regulators began investing in real world data infrastructure in the first place, and it remains the area where RWE has the longest track record and broadest acceptance.
Challenges and Limitations
Despite its promise, real world evidence faces persistent challenges that limit how far it can go in replacing or supplementing traditional research. The biggest issues fall into three categories: data quality, data access, and methodological rigor.
Data quality is an ongoing concern. Health records are created for clinical care, not research. They contain missing values, inconsistent coding, and large volumes of unstructured text (like physician notes) that are hard to analyze systematically. A survey of stakeholders in oncology found that despite 92% recognizing the potential of real world data to generate useful evidence, concerns about poor data quality and unstructured information remained significant barriers.
Access is another obstacle. Health data is fragmented across hospitals, insurers, and health systems that use different formats and don’t always share easily. Privacy regulations add necessary but complex requirements for how patient data can be linked and used. Even when data exists, getting permission to use it for research can take months.
Methodological variability is perhaps the most fundamental challenge. There is significant variation in the quality and methods used across observational studies, which directly affects whether results are valid and reproducible. Two research teams analyzing the same dataset with different approaches can reach different conclusions. This is why regulatory agencies and professional organizations have invested heavily in standardizing how real world evidence studies should be designed, conducted, and reported.
Where RWE Fits in the Bigger Picture
Real world evidence is not a replacement for randomized controlled trials. It’s a complement. The strongest regulatory decisions draw on both: RCTs to establish that a drug works under controlled conditions, and RWE to confirm it works in the messier reality of everyday healthcare. Some newer study designs, called pragmatic trials, blend the two approaches by randomizing patients within routine clinical settings and collecting outcomes through existing health records rather than dedicated trial infrastructure.
For patients, the practical impact is that treatments are increasingly evaluated not just on how they perform in ideal trial conditions but on how they perform for people like you, in settings like yours. For the healthcare system, RWE offers a way to generate evidence faster and at lower cost, particularly for questions that traditional trials struggle to answer. The technology to collect and analyze this data is advancing rapidly, with automated systems pulling structured information from health records and wearable devices contributing a continuous stream of patient-generated data that didn’t exist a decade ago.