Table 2

Categories of evidence

CharacteristicsConfidence
HighBased on consistent results from well-performed randomised controlled trialsFurther research is very unlikely to change the estimate of the effect
ModerateBased on randomised controlled trials where there is evidence of bias or from other well-conducted study types (eg, well-executed observational studies)Further research is likely to have an impact on the estimate of the effect
LowBased on observational evidence or from controlled trials with several serious limitationsFurther research is likely to have an important impact
Very lowBased on case studies or expert opinionEstimates of effect are far from certain and more research is needed