Gradation scale and others levels
AEstablished scientific evidence based on studies with a high level of evidence (evidence level 1): high-power randomised controlled trials without major bias, or meta-analyses of randomised controlled trials, decision analyses based on well-conducted studies.
BScientific presumption based on a scientific presumption provided by studies with an intermediate level of evidence (evidence level 2), such as low-power, randomised controlled trials, well-conducted non-randomised controlled studies or cohort studies.
CLow level of evidence based on studies with low levels of evidence, such as case-control studies (evidence level 3), retrospective studies, case series, comparative studies with significant biases (evidence level 4).
AEExpert agreement: where no studies exist, recommendations are based on agreement between work group experts, after consultation with the reading group. A recommendation which has not been graded may nevertheless remain relevant and useful. The absence of grading should, however, encourage authors to conduct complementary studies.
Comments from several experts in the pathology in question, solicited by the work group on recommendations.
These experts do not participate in the work group meetings.