Core Grading of Recommendations Assessment, Development, and Evaluation tutorial: Rating certainty of evidence in outcomes from pairwise systematic reviews and meta-analyses

Abstract The Grading of Recommendations Assessment, Development, and Evaluation (GRADE) framework has been the international standard for assessing certainty of evidence (CoE) and informing clinical recommendations for over two decades. Over time, the GRADE guidelines have expanded from overarching principles to increasingly specialized methodological issues, and continue to evolve through bountiful publications. In this tutorial, we summarize the Core GRADE approach and present an updated, streamlined guidance for rating CoE for outcomes from systematic reviews that compare a single intervention with a single comparator. The framework retains four CoE categories (high, moderate, low, and very low) and a structured rating process based on five domains for rating down (risk of bias, indirectness, inconsistency, imprecision, and publication bias) applicable to randomized and observational studies, and two domains for rating up (large magnitude of effect and dose–response relationship) applicable to observational studies. To ensure consistent and transparent CoE judgments across outcomes, review authors should apply the Core GRADE approach, explicitly report how CoE is rated for each outcome, and provide the rationale for any decisions to rate down or up.