Rebuild the missing certainty column by rating each outcome separately with GRADE criteria, then add footnotes that record the reasons for each rating. GRADE rates certainty at the outcome level using explicit criteria, so the column must reflect per-outcome judgements rather than one rating for the whole review.
- List every outcome in the table and confirm each has its own row before you add the column. Expected result: a complete outcome set that matches the outcomes reported in the review text.
- Rate certainty for each outcome against the GRADE domains, starting from the study design and recording each reason for downgrading or upgrading. Expected result: a rating (high, moderate, low, or very low) with a written justification per outcome.
- Populate the certainty column with those ratings and add footnotes that state the specific reason for each change, such as risk of bias, inconsistency, indirectness, imprecision, or publication bias. Expected result: a reader can trace every rating to its stated reason.
- Cross-check the completed table against the review text so the ratings and the conclusions describe the same certainty. Expected result: any mismatch between table and text is corrected in one place or the other.
- Record the completion in the review's documentation, noting which outcomes were rated and which footnotes were added. Expected result: an audit trail for later updates or peer review.
GRADE guidance and the GRADEpro workflow support building Summary of Findings tables with outcome-level certainty ratings and explanatory footnotes. Cochrane Handbook Chapter 15 covers how certainty and confidence intervals feed into responsible review conclusions, which is the standard to check the finished table against. When a rating is disputed, describe what each source can verify and resolve the difference before publishing the table.