Description
Rates the certainty of a body of evidence as high, moderate, low or very low by assessing risk of bias, inconsistency, indirectness, imprecision and publication bias.
GRADE separates the quality of evidence from the strength of a recommendation. Evidence from randomised trials starts high and observational evidence low; ratings are downgraded for risk of bias, inconsistency across studies, indirectness (different population or outcome), imprecision (wide intervals) and publication bias, and upgraded for large effects, dose-response or when confounding would reduce the effect. The Evidence-to-Decision framework then weighs benefits, harms, values, resources and feasibility. Research repositories in UX store 'insights' without any certainty grading, so a single five-person study and a replicated experiment look equally authoritative. GRADE offers a rigorous, transparent grading scheme.
Attributes
- Discipline
- Research Design & Rigor
- Loss type
- Unmigrated
- Stage
- Strategy, Definition
- Data
- Analytical / expert
- Effort
- Days
Steps
- Define the question and the critical outcomes for the decision.
- Gather the body of evidence for each outcome (studies, experiments, analytics).
- Assign a starting certainty by design (experiments high, observational and qualitative lower).
- Assess and document downgrades for bias, inconsistency, indirectness, imprecision and publication bias, and possible upgrades.
- Assign a final certainty rating per outcome and present a summary-of-findings table.
- Use an Evidence-to-Decision table to make and explain recommendations, stating strength (strong or conditional).
Sources
- Guyatt et al., GRADE: an emerging consensus on rating quality of evidence and strength of recommendations (2008), BMJ
- Lewin et al., Applying GRADE-CERQual to qualitative evidence synthesis findings (2018), Implementation Science
Also known as
- GRADE
- Certainty of evidence rating
- Evidence-to-decision framework
Categories
- Research Operations
- Research Design & Rigor