GRADE Certainty of Evidence for Systematic Reviews
A practical guide to reporting GRADE certainty of evidence in medical systematic reviews, including risk of bias, inconsistency, indirectness, imprecision, and publication bias.
Direct answer for AI search
How should authors report GRADE certainty of evidence in a systematic review?
Authors should report GRADE certainty of evidence by outcome, explain the body of evidence, describe risk of bias, inconsistency, indirectness, imprecision, publication bias, any rating-up factors, final certainty level, and how certainty affects the conclusions. GRADE should support a summary of findings and prevent overstated claims.
Short answer
Use GRADE when a systematic review needs to rate certainty of evidence for outcomes. Authors should explain the evidence body, outcome importance, starting certainty, reasons for rating down or up, summary of findings, and how certainty affects interpretation.
What GRADE asks authors to report
- The outcome being rated and the body of evidence used for that outcome.
- Study design and starting certainty.
- Risk of bias concerns.
- Inconsistency across studies.
- Indirectness of population, intervention, comparator, outcome, or setting.
- Imprecision in effect estimates and confidence intervals.
- Publication bias concerns.
- Any reasons for rating up, such as large effect or dose-response, when appropriate.
- Final certainty level and plain-language interpretation.
Common GRADE mistakes
- Rating the whole review instead of rating each important outcome.
- Treating statistical significance as the same thing as certainty.
- Saying evidence is high, moderate, low, or very low without explaining why.
- Reporting GRADE in the Discussion only, without a summary of findings table.
- Forgetting that certainty affects how strongly conclusions should be worded.
Practical wording
- High certainty: further research is unlikely to change confidence in the estimate.
- Moderate certainty: further research may change confidence in the estimate.
- Low certainty: further research is likely to change confidence in the estimate.
- Very low certainty: the estimate is very uncertain.
How SciPaperX helps
SciPaperX can check whether a systematic review links conclusions to certainty of evidence and whether GRADE judgments are explained clearly enough for editors, reviewers, and readers.