Record each outcome in a matching table that pairs the registered measure, timepoint, and analysis with the published report, then flag any discrepancy for the review lead. ClinicalTrials.gov defines structured registration and results fields for study identification, status, design, eligibility, locations, outcomes, and reported results, so those fields give you the registration side of each row.
- Build one row per registered outcome and copy the registered measure, timepoint, and stated analysis into separate columns. Expected result: a fixed registration baseline you can compare against without re-reading the registry entry.
- Add the matching reported columns from the published paper: the outcome as reported, its timepoint, and the analysis actually presented. Expected result: each row shows registration and report side by side.
- Mark each row as matching, differing, or absent from the report, and write a one-line note naming the specific difference (for example, a changed timepoint or a different summary measure). Expected result: a logged result rather than an impression.
- Escalate every differing or absent row to the review lead with a recommendation on which record to cite, and record that decision in the table. Expected result: a documented citation choice for the review.
Where the review assesses risk of bias, Cochrane Handbook Chapter 8 describes RoB 2 as judging a specific randomized-trial result across bias domains using signalling questions and written justifications, producing low risk, some concerns, or high risk. That result-level focus means the outcome row you log is also the unit a bias judgement attaches to, so keep the row identifier consistent between the matching table and any bias assessment.
Keep the table as the audit trail: one row per outcome, the registered fields, the reported fields, the match status, the note, and the escalation decision.