RWD vs. RWE — The Distinction That Determines Everything Else

Before this module can go any deeper, it needs one distinction held precisely, because the two terms get used interchangeably in casual conversation in a way that causes real confusion: Real-World Data (RWD) is any health-related data collected outside a traditional, pre-specified randomised clinical trial — electronic health records, insurance and administrative claims, patient registries, wearable device outputs, pharmacy dispensing records, and patient-reported data from apps and forums, among other sources Lesson 15.3 covers in depth. Real-World Evidence (RWE) is something different: clinical evidence about a product’s usage, benefits, or risks, derived specifically from analysing that RWD.

The relationship is best understood as raw material and finished product. RWD is what exists, sitting in a database, unanalysed. RWE is what a properly designed study, with a defined question and an appropriate methodology, actually produces from that data. And critically, the same underlying RWD source can generate multiple, genuinely different pieces of RWE depending on what question is asked of it — the exact same insurance claims database might simultaneously support a formal PASS study (Module 9’s territory), a drug utilisation study, a comparative effectiveness analysis between two products, and a specific pharmacoepidemiology investigation into a particular adverse event signal, all drawing on the same raw data but producing entirely different evidence depending on the study design applied.

It’s worth being explicit about what RWE is not: a replacement for randomised controlled trials. The two answer genuinely different questions, and neither substitutes for the other in most circumstances. A randomised trial establishes causal efficacy under tightly controlled conditions in a carefully selected population — exactly the rigor that makes a trial’s findings so trustworthy within the population it actually studied. RWE characterises what happens under the messy, uncontrolled conditions of routine clinical practice, across the full diversity of the population that actually uses a medicine — precisely the breadth a trial’s controlled design can’t provide, at the cost of the causal certainty a trial’s randomisation delivers.

That tradeoff is exactly why regulators don’t treat all RWE as automatically authoritative. The FDA’s Real-World Evidence Framework and EMA’s guidance distinguishing interventional from non-interventional RWD use both exist specifically to define what "regulatory-grade" RWE actually requires — pre-specified protocols, validated outcome definitions, appropriate confounding control, transparent methods. A poorly designed RWE study, drawing on genuinely real data, can still produce a misleading or unreliable conclusion — which is exactly why the rest of this module spends so much time on study design (Lessons 15.4-15.5) and confounding (Lesson 15.6) rather than treating "we used real-world data" as a sufficient credential on its own.