Registry-based randomised trial
A registry-based trial randomises patients who are already being tracked by a national or disease registry and uses the registry, not trial visits, to record what happens to them, making very large trials cheap enough to run.
Overview
Registries already record diagnosis, treatment, recurrence and death for whole populations. A registry-based randomised trial uses that machinery: patients are identified and randomised through the registry or at the point of care, and their outcomes are read from the registry and from linked death and hospital records instead of being collected at dedicated trial visits. The design was pioneered in Scandinavian cardiology, where registries are complete and linkable through personal identifiers, and it allows tens of thousands of patients to be randomised for a small fraction of the cost of a conventional trial. The trial-within-a-cohort variant consents a cohort once to future randomisation, then randomly offers an intervention to some members and compares them with the rest.
The design suits questions where the outcome is hard (death, recurrence recorded by pathology, a new cancer) and the intervention is already in use, so a lighter touch is acceptable: comparing two standard schedules, testing a repurposed generic, or evaluating a screening interval. It struggles where the outcome needs trial-specific measurement, such as scan-defined progression or quality of life, and where the registry does not capture the details needed for safety. Data quality is only as good as the registry, and the trial inherits its lag.
The corpus has no oncology trial run entirely through a registry, which is itself a finding: cancer registries in most countries record too little about treatment, and too late, to support one. The closest structure is CIRCULATE-Japan, where the GALAXY observational cohort of more than 2,000 resected colorectal cancer patients, tested for circulating tumour DNA after surgery, feeds the randomised sub-studies VEGA and ALTAIR. The gap is one of the reasons registries and real-world data are treated as a bottleneck on this site; a registry good enough to run trials through is a registry good enough to answer most of the questions patients ask about how care works outside trials.
- Clinics: notes, scans, prescriptions
- De-identified database
- Structured rows
- Real-world survival
Showing the technology this term belongs to: Oncology EHR and real-world data platforms.
Similar pages
not linked directly; found by shared links- TermCluster-randomised trial
Shares Informed consent, Pragmatic trial, Randomised trial, Seamless, adaptive and Bayesian trial designs.
- TermStepped-wedge design
Shares Pragmatic trial, Randomised trial, Seamless, adaptive and Bayesian trial designs, Phase 1, 2 and 3 trials.
- TermN-of-1 trial
Shares Pragmatic trial, Randomised trial, Seamless, adaptive and Bayesian trial designs, Phase 1, 2 and 3 trials.
- IdeaSimulate eligibility against real-world data before every protocol is locked
Shares Weak real-world evidence and registries, Real-world evidence, Trials enrol too few, too slowly.
- IdeaRegistry-based randomised trials for oncology comparative effectiveness
Shares Weak real-world evidence and registries, Real-world evidence, Trials enrol too few, too slowly.
- TrialBWEL (Breast Cancer Weight Loss, Alliance A011401)
Shares Decentralised trial, Pragmatic trial, Event-free / disease-free survival (EFS, DFS, iDFS, RFS).
- TermOMOP common data model and OHDSI
Shares Oncology EHR and real-world data platforms, Real-world evidence.
- TermAccrual and enrolment
Shares Decentralised trial, Informed consent, Pragmatic trial.