Run a certainty assessment outcome by outcome: rate each outcome against the GRADE domains, record the reason and evidence location for every judgement, then check consistency across outcomes before finalizing the evidence profile. GRADE (Grading of Recommendations Assessment, Development and Evaluation) provides a transparent approach to rating certainty of evidence using explicit criteria and outcome-level judgements, so the unit of assessment is the outcome, not the review as a whole.
- Build the outcome list first. List every outcome the review reports, including harms and patient-important outcomes, and fix the order before rating. Expected result: a fixed outcome set that the evidence profile and Summary of Findings table will mirror, so no outcome is added or dropped mid-assessment.
- Rate each outcome against the GRADE domains. For each outcome, judge risk of bias, inconsistency, indirectness, imprecision, and publication bias, and record whether the rating starts from the study design and moves up or down. Expected result: a certainty level per outcome with the domains that changed it named explicitly.
- Write the reason and evidence location for each judgement. For every domain decision, note why the rating changed and where in the included studies or data the support sits. Expected result: an audit footnote trail that lets a reader trace each rating back to its basis.
- Assemble the evidence profile. Compile one row per outcome with its certainty level, the domain judgements, and the footnotes. Expected result: a complete profile covering every outcome in the review.
- Run the consistency check across outcomes. Compare ratings for outcomes drawn from overlapping studies and confirm that similar evidence received similar judgements, and that differences are explained by the recorded reasons. Expected result: a profile where any variation between outcomes is justified in the footnotes.
Cochrane GRADE resources describe GRADE as a systematic approach for rating certainty in systematic reviews and evidence syntheses, including Summary of Findings table guidance and GRADEpro workflows, which support keeping the outcome-level judgements and footnotes in one structured record. Cochrane Handbook Chapter 15 covers confidence intervals, exact P values, absolute and relative effects, applicability, indirectness, and certainty, and is useful when an imprecision or indirectness judgement needs a stated basis.