RCT vs Real-World Evidence: Two Ways to Know If a Treatment Works
How do we know if a treatment really works? There are two fundamentally different ways to find out. One is the randomized controlled trial (RCT), long treated as the gold standard of medical evidence. The other is real-world evidence, built from data collected in everyday clinical practice. The Clarity Clinical Solutions video "RCT vs Real-World Evidence: Two Ways to Know If a Treatment Works" compares the two approaches and argues that both matter. This article covers the same comparison, with the regulatory context filled in.
Why the RCT is the gold standard
An RCT is an experiment. Patients are randomly assigned to receive either the new treatment or a control, usually a placebo or the current standard of care. Randomization makes the groups as similar as possible except for the drug being tested, which lets researchers isolate the treatment's true effect. RCTs answer one question: can this intervention work under ideal conditions?
The design earns its reputation three ways. Randomization removes selection bias, because the groups are comparable at the start. Strict protocols and blinding keep doctors and patients from influencing the results. And pre-planned sample sizes give the trial statistical power to detect real treatment effects. This is why agencies like the FDA and EMA have required RCT data before approving a new drug.
But RCTs have real limits, and the video does not sugarcoat them. A single Phase III trial can cost $20 to $100 million. Recruiting thousands of patients takes years. And the biggest problem: trial populations are highly selected. Strict eligibility criteria can exclude the majority of real-world patients, including the elderly, children, and people with multiple health conditions. The results may not transfer cleanly to the patients who will actually receive the drug.
What real-world evidence is
Real-world evidence starts with data collected during routine clinical care, not in a controlled experiment. The raw material is called real-world data, and it includes electronic health records, insurance claims, data from wearable devices, and patient registries. Analyzed with rigorous scientific methods, that data becomes real-world evidence. RWE answers a different question: does this treatment work in real clinical practice?
The FDA maintains an active program on real-world evidence and defines the terms the same way (FDA on real-world evidence). Sources include hospital and clinic records with diagnoses, medications, and lab results; insurance claims that track procedures, prescriptions, and hospitalizations over time; wearables that record heart rate, activity, and sleep; and registries that follow people with specific diseases for years or decades.
Where RWE shines
RWE covers ground where RCTs cannot help. For rare diseases, there may be only a few hundred patients in the world, not enough to power a traditional trial. For pregnant women and children, who are often excluded from RCTs, real-world data may be the only safety information available. For long-term safety, RWE can follow patients for decades and catch rare side effects that pre-market trials missed. And for comparing two marketed drugs, insurance claims can answer questions no company will pay to study.
Regulators see the value. The FDA has issued guidance on using electronic health records and medical claims data to support regulatory decisions (FDA guidance on RWD from EHRs and claims data), and a separate framework for real-world evidence in medical devices (FDA guidance on RWE for devices).
Where RWE is weak
RWE has serious limitations, and the video names the biggest one: confounding by indication. Sicker patients tend to get different treatments, so the treatment and control groups are not comparable to begin with. Data quality is another issue. Electronic health records are full of missing values, coding errors, and inconsistent measurements. Selection bias means the people who get a treatment are systematically different from those who do not. And most importantly, association does not equal causation. Proving cause and effect from observational data is inherently difficult. The FDA's own framework document walks through these concerns in detail (Framework for FDA's Real-World Evidence Program).
The regulatory shift
The FDA has moved steadily toward RWE over the past decade. The 21st Century Cures Act, signed in 2016, defined real-world evidence and real-world data for the first time in federal law and required the FDA to develop a framework for evaluating RWE in regulatory decisions (FDA on the 21st Century Cures Act). The FDA published its real-world evidence program framework in 2018, began piloting RWE for new indications of approved drugs through a program the video calls Project 351, and has kept issuing guidance on how to submit RWE in regulatory applications. The trend is clear: RWE is becoming a standard part of the regulatory toolkit.
Which is better? It depends
RCTs have the highest internal validity. They are the best tool for establishing cause and effect, so if the question is whether a drug works, the RCT is the answer. But RCTs have poor external validity: the results may not apply to the diverse patients seen in real clinics. RWE has the opposite profile: lower internal validity, high external validity.
The two approaches are not competitors. They answer different questions and work best together. The future is hybrid: pragmatic trials that embed randomization into routine care, target trial emulation studies that mimic a hypothetical randomized trial, and AI tools that improve confounder selection and causal inference. The video's summary is worth repeating: RCTs establish efficacy, RWE measures effectiveness, and the smartest approach uses each for what it does best.
The bottom line
RCTs tell us whether a treatment can work under ideal conditions; real-world evidence tells us whether it actually works in practice. RCTs are expensive, slow, and narrow, but they remain the standard for regulatory approval. RWE is faster, cheaper, and more representative, but it struggles with confounding and causal inference. Choosing one over the other is the wrong move. The two together give patients, doctors, and regulators the full picture.
This article is based on the Clarity Clinical Solutions video "RCT vs Real-World Evidence: Two Ways to Know If a Treatment Works." Watch it here: RCT vs Real-World Evidence: Two Ways to Know If a Treatment Works
References
- Clarity Clinical Solutions — "RCT vs Real-World Evidence: Two Ways to Know If a Treatment Works" (source video). https://www.youtube.com/watch?v=PpeVDGKu_Fc
- FDA — Real-World Evidence (science and research program page). https://www.fda.gov/science-research/science-and-research-special-topics/real-world-evidence
- FDA — Real-World Data: Assessing Electronic Health Records and Medical Claims Data to Support Regulatory Decision-Making (guidance). https://www.fda.gov/regulatory-information/search-fda-guidance-documents/real-world-data-assessing-electronic-health-records-and-medical-claims-data-support-regulatory
- FDA — Use of Real-World Evidence to Support Regulatory Decision-Making for Medical Devices (guidance). https://www.fda.gov/regulatory-information/search-fda-guidance-documents/use-real-world-evidence-support-regulatory-decision-making-medical-devices
- FDA — Framework for FDA's Real-World Evidence Program (2018 report). https://www.fda.gov/media/120060/download
- FDA — 21st Century Cures Act (regulatory information page). https://www.fda.gov/regulatory-information/selected-amendments-fdc-act/21st-century-cures-act