Evidence and Recommendation Grading
Two-level grading of recommendation grading:
Level A: Strong recommendations: Benefits outweigh the risks
Level B: Weak recommendations: Benefits and risk equivocal or uncertain
Three-level grading of evidence
Level 1: (High-level) evidence [Research results that address clinical outcomes and quality criteria that minimize bias. Examples of studies that may be cited are based on research that is randomized, metadata data analysis, double-blind, a national clinical guideline, multicentered, controlled clinical trial, or a systematic review.]
Level 2: (Moderate-level) evidence [Research results that address clinical outcomes, using scientific investigation, but not meeting Level 1 criteria. Examples of studies that may be cited are case-controlled, qualitative, cohort, cross-sectional, clinical outcomes, nonrandomized, epidemiological, retrospective, comparative, developmental, observational, or pragmatic.]
Level 3: (Low-level) evidence [Research that represents reports that are not based on scientific analysis or clinical outcomes. Examples include case series, case reports, expert opinions, comments, letters, commentaries, and pilot studies.]
Recommendations are based on evidence synthesis and clinical experience, with a clear statement of the evidentiary basis.
A synthesized GRADE system is used for grading recommendations, which considers the evidence quality, risks, and benefits.
