OnCo
ideasIdea

An open library pairing completed cancer trials with real-world emulations

Build a public collection where every major cancer trial has a matching analysis done on hospital data, so we learn exactly when real-world evidence can be trusted and when it cannot.

RCT-DUPLICATE calibrated real-world methods against trials mostly in cardiometabolic disease; oncology, with its time-varying treatments and surrogate endpoints, needs its own calibration set. The proposal funds a standing programme that emulates every phase 3 oncology trial with a real-world counterpart in federated data, publishes agreement metrics (direction, magnitude, confidence interval overlap), and maintains a public map of which question types (first-line comparisons, sequencing, dose) are emulable.

Hypothesis
Across 100 oncology trials, emulation agreement will be high for first-line drug-versus-drug comparisons with hard endpoints and poor for questions involving progression-based endpoints or treatment switching, giving regulators an evidence-based rule for when RWE can substitute for a trial.
Rationale
Without a calibration library, every RWE submission is argued case by case. A map of emulability converts the debate into a lookup table.
What would test it
Fund the first 30 emulations across breast, lung and prostate; publish agreement metrics; test whether the resulting rule predicts agreement in the next 20.
Maturity
early clinical
Who has to act
research
Cost to try
Medium ($1M to $50M)
Years to first evidence
3
Bottlenecks it attacks

Connected

4top