Overview
Evidence synthesis without quality grading produces recommendations that treat a single small RCT and a large well-conducted cohort study as equivalent evidence. GRADE (Grading of Recommendations Assessment, Development and Evaluation) provides a systematic framework for rating evidence certainty across five domains — risk of bias, inconsistency, indirectness, imprecision, and publication bias — and for converting that certainty rating into recommendation strength. Understanding that evidence quality and recommendation strength are distinct concepts is foundational: strong recommendations can be made from low-quality evidence when benefits clearly outweigh harms.
The Evidence Synthesis Framework applies GRADE across all five domains, shows the upgrade/downgrade logic transparently, and produces recommendation statements that correctly distinguish strong from weak recommendations based on evidence quality and benefit-harm balance.