OnCo
ideasIdea

An automated pipeline emulating trials of every common drug against every cancer

Millions of people take common drugs and some get cancer. Running standardised analyses across whole-country records could rank which old drugs deserve a real trial.

Nordic registries, UK CPRD and OpenSAFELY, US VA and Medicare, and Korean and Taiwanese national databases together cover hundreds of millions of people. Target trial emulation with active-comparator new-user designs, pre-registered protocols and negative-control outcomes can screen drug-cancer pairs systematically rather than one paper at a time. The proposal is an open, federated pipeline that runs a common protocol across databases for a defined library of drugs and cancer outcomes, publishes all results (including nulls) with bias diagnostics, and feeds a public prioritisation list for the repurposing fund.

Hypothesis
The pipeline identifies at least five drug-cancer pairs with concordant protective associations across three or more independent databases and negative-control checks within two years, and at least one enters a randomised trial.
Rationale
Single-database observational studies have produced many false leads; federated, pre-registered, multi-database analysis with shared code is the only way to separate robust signals from artefacts at scale.
What would test it
Build the pipeline for 50 drugs and 10 cancers across four databases, publish concordance results, and compare with known positives (aspirin-colorectal) and known negatives (metformin-breast).
Maturity
preclinical evidence
Who has to act
data
Cost to try
Medium ($1M to $50M)
Years to first evidence
3
Bottlenecks it attacks

Connected

3top