Recheck the outcome-level risk-of-bias judgement, apply the matching certainty downgrade, and record the reason in the Summary of Findings table so the rating and its justification stay aligned.
GRADE rates certainty at the outcome level using explicit criteria, and RoB 2 produces a risk-of-bias judgement for a specific randomized-trial result across defined domains. A serious risk-of-bias concern that was not carried into the certainty rating is a gap between two linked judgements, so the repair starts by confirming both judgements refer to the same outcome and the same body of evidence.
- Confirm the risk-of-bias judgement. Open the RoB 2 assessment for the outcome in question and check that the domain judgements and written justifications support the overall result of low risk, some concerns, or high risk. Expected result: a documented judgement you can cite for the downgrade.
- Apply the downgrade to the certainty rating. If the risk-of-bias judgement is serious enough to lower certainty, reduce the outcome's certainty rating by the appropriate level and state risk of bias as the reason. Expected result: a revised rating with a named reason.
- Update the Summary of Findings table. Enter the revised certainty rating and the downgrade reason in the table for that outcome, following the Cochrane GRADE Summary of Findings guidance. Expected result: the table, the rating, and the risk-of-bias assessment agree.
- Check the downstream text. Search the manuscript, abstract, and conclusions for the old rating and revise any statement that relied on it. Expected result: no remaining sentence contradicts the revised rating.
Document the change with the outcome name, the prior rating, the revised rating, and the risk-of-bias justification, so reviewers can trace the decision. GRADEpro workflows support building and updating Summary of Findings tables, which helps keep the rating and its reason in one place.