The protocol is the study written down before it happens: question, outcome, sample size and analysis plan. Registration makes that commitment public — and permanent.
A study that stays in a drawer distorts the literature. Registration makes every started study findable, so a negative trial that was never published can still be counted — and so the same question is not funded twice by mistake.
It also fixes the primary outcome in advance. Without a public record, a secondary outcome that happened to be significant can quietly become the main result — the change that a reader has no way of detecting.
Since 2005 the ICMJE requires prospective registration for clinical trials, and ABCD follows that rule: a trial without a registration number cannot be published. Funders and ethics committees increasingly ask for the same.
Mandatory, without exception, including surgical trials and trials without a drug. Register before the first participant is enrolled.
Expected. Register on PROSPERO before screening; the protocol should already state the question, the databases, the eligibility criteria and the outcome.
Optional, and increasingly common. A registered cohort or case-control study is stronger evidence, because the reader can check what was planned against what was reported.
Write the protocol anyway. Even in a retrospective study, deciding the outcome and the analysis before opening the database is what separates a study from a fishing expedition.
| Registry | When to use it | Notes |
|---|---|---|
| ReBEC | Clinical trials conducted in Brazil | Free, in Portuguese and English, and a WHO primary registry — the natural choice for a Brazilian trial. Registration is also required by ANVISA for regulated studies. |
| ClinicalTrials.gov | Trials with international reach | The registry most journals and funders recognise. Use it when the study is multinational, when a funder requires it, or in addition to ReBEC. |
| PROSPERO | Systematic reviews with a health outcome | Register before screening starts — PROSPERO does not accept a review whose data extraction is already finished. Scoping reviews are not accepted; use OSF. |
| OSF Registries | Observational, scoping and preclinical studies | Accepts what the clinical registries do not: cohorts, case-control studies, scoping reviews, animal studies. Free and immediate. |
| WHO ICTRP | Searching, not registering | The WHO portal aggregates the primary registries. Use it to check whether a similar trial is already running before you start yours. |
Retrospective registration does not disqualify an observational study, but it does disqualify a trial in journals that follow ICMJE. If it happened, disclose the date openly.
The outcome in the paper must be the outcome in the record. Any change needs to be declared and justified — reviewers do check.
A registry entry still marked "recruiting" three years after the study ended suggests nobody is looking after the data either.
PROSPERO refuses reviews whose data extraction is complete. Register on the day you write the search strategy.
Registration says what you intended to do; the reporting guideline says how to describe what you did. Submit the completed checklist with the manuscript — all of them live on the EQUATOR Network.
Not mandatory, but it strengthens the paper and is easy to do on OSF. Ethics approval, on the other hand, is required for every design.
You can, and you must, but declare the actual dates. For journals following ICMJE, a trial registered after the first patient is normally not eligible for publication as a trial.
For a trial run in Brazil, ReBEC is enough and is a WHO primary registry. Use ClinicalTrials.gov when the study is multinational or when a funder or partner requires it; registering in both is allowed.
It is strongly recommended, and reviewers ask for it. Register before screening begins; if PROSPERO does not accept the design, use OSF.
Written by Francisco Tustumi, digestive system surgeon (FMUSP), associate editor of ABCD.
Developed by Surya MKT