Apply RoB 2 to one result by answering each signalling question with a written justification tied to the paper or protocol, then judging that result low risk, some concerns, or high risk. The Cochrane Handbook describes RoB 2 as assessing a specific randomized-trial result across bias domains using signalling questions and written justifications.
Scope the assessment to one result
RoB 2 assesses a specific result, so fix the outcome, the effect measure, and the analysis you are judging before you start. Keep the assessment on that result; a separate result needs its own assessment.
Work through the domains in order
- Answer each signalling question for the result and write a justification that cites the paper or protocol passage supporting your answer. The expected result is a documented answer for every question in the domain.
- Use the answers to reach a domain-level judgement for that result. The expected result is a domain judgement you can defend from the justifications you wrote.
- Combine the domain judgements into one overall judgement for the result: low risk, some concerns, or high risk. The expected result is a single outcome-level judgement, not a judgement of the whole trial.
Keep reporting and appraisal separate
The CONSORT 2025 statement is a reporting guideline and checklist for randomized trials; it supports complete reporting and is not a study design or quality appraisal tool. Use CONSORT to check whether the trial reports what you need, and use RoB 2 to judge the result.
Write the justification as the record
Each justification should name the domain and question, state your answer, and point to the paper or protocol text that supports it. When the paper and protocol disagree or a detail is missing, record the answer as unresolved and note the specific passage you checked. That written record is what makes the final low risk, some concerns, or high risk judgement traceable for readers and reviewers.