GRADE guidelines: 3. Rating the quality of evidence

GRADE categorizes evidence quality into four levels: high, moderate, low, and very low.

• Quality of evidence is assessed for a body of evidence rather than individual studies.
• Randomized trials are initially rated as high-quality evidence, while observational studies are rated as low quality.
• Factors such as imprecision, inconsistency, indirectness, and publication bias can reduce the quality rating.
• Several factors can enhance confidence in the effect estimates.
• The GRADE system distinguishes between evaluating evidence quality and making recommendations.

AI simplified