Why Clinical Trials Aren’t the Finish Line

Every module in this course from Module 1 onward has treated the moment of drug approval as a milestone inside a continuous lifecycle, not an endpoint — Module 3 built that framing explicitly. This module gives that idea its most concrete, methodologically rigorous form: clinical trials, however well-designed, produce safety data that is systematically incomplete the moment a drug reaches real-world use, and Real-World Evidence is the discipline that fills that specific, structural gap.

The incompleteness isn’t a flaw in trial design — it’s close to unavoidable. Trials necessarily exclude, or dramatically underrepresent, the very old, the very sick, pregnant women, children, and patients managing complex multi-drug regimens, for sound ethical and methodological reasons. That means the safety data available at approval says relatively little, with any real statistical confidence, about exactly those populations — and almost nothing about rare adverse events (too infrequent for a trial’s sample size to reliably detect), long-duration effects (beyond a trial’s follow-up window), or drug interactions that only become visible once a medicine is used alongside the full complexity of other treatments a real patient is actually taking.

Real-World Evidence closes that gap by converting the data routine healthcare already generates — prescriptions written, diagnoses coded, laboratory tests run, hospital admissions recorded — into structured evidence about what a medicine actually does, in actual patients, over actual durations of use, inside the full complexity of real clinical practice rather than a trial’s controlled conditions. And this has genuinely moved from a marginal, academic pursuit to mainstream regulatory infrastructure: EMA’s DARWIN EU platform runs regulatory-grade federated studies specifically to support PRAC signal assessments (Module 7’s territory); FDA’s Sentinel system, now scaling through its NextGen iteration, processes safety data across more than 100 million patient lives; and CDSCO is actively building infrastructure to integrate hospital electronic health record data with PvPI through a dedicated National RWE Hub.

This module works through that infrastructure in the order a working PV professional actually encounters it: what real-world data sources exist and what each is actually good for (Lesson 15.3), the specific pharmacoepidemiology study designs used to analyse that data rigorously (Lessons 15.4-15.5), confounding as RWE’s central methodological challenge (Lesson 15.6), the data standardisation layer that makes multi-database studies possible (Lesson 15.7), the actual global networks running this infrastructure today (Lesson 15.8), and finally, concretely, how all of it integrates back into the signal detection and confirmation discipline Module 7 already built (Lesson 15.9).