Article Text
Statistics from Altmetric.com
One of the most widely used tools for assessing and communicating scientific uncertainty is Grading of Recommendations Assessment, Development, and Evaluation (GRADE), a system for rating the quality of evidence and grading strength of recommendations in healthcare. More than 100 organisations around the world—WHO included1—are using GRADE or have endorsed it.
In GRADE, a quantitative assessment of uncertainty is qualitatively communicated, so that a result obtained as a CI relative to a threshold is expressed as a finding in which assessors have low, moderate or high certainty, or certainty described with other such qualifiers. What these correspond to in quantitative terms, and how decision-makers interpret them, is our issue here. We confine our attention to GRADE’s decision rules for systematic reviews, and do not comment on the problem of multiple outcomes in guideline recommendations.
In a recent guideline article,2 GRADE introduced an idea that appears to undermine sound statistical reasoning in systematic reviews: the idea is that a result …