| Ia |
Evidence from a meta-analysis of at least three randomized controlled trials (RCTs) |
| Ib |
Evidence from at least one RCT or a meta-analysis of fewer than three RCTs |
| IIa |
Evidence from at least one methodologically sound, non-randomized controlled trial |
| IIb |
Evidence from at least one methodologically sound, quasi-experimental descriptive study |
| III |
Evidence from methodologically sound, non-experimental observational studies, e.g., comparative studies, correlation studies, and case studies |
| IV |
Expert committee reports or expert opinion and/or clinical experience of recognized authorities |
| Recommendation grade |
Positive recommendation |
Negative recommendation |
| A |
“Must” recommendation: at least one RCT of good overall quality and consistency supports the recommendation directly, without extrapolation (evidence levels Ia and Ib) |
“Must not”: recommendation against the measure in question based on level Ia and Ib evidence. |
| B |
“Should” recommendation: well-conducted clinical trials, other than RCTs, support the recommendation either directly (evidence levels II or III) or by extrapolation (evidence level I) if the studies do not directly address the subject in question. |
“Should not”: recommendation against the measure in question based on level II and III evidence. |
| 0 |
“May” recommendation: expert committee reports or expert opinion and/or clinical experience of recognized authorities (evidence level IV) or extrapolation from evidence of levels IIa, IIb, or III. This recommendation grade indicates that no directly applicable clinical studies of sufficiently high quality are available for consideration. |
Recommendation against the measure in question based on level IV evidence or extrapolation from evidence of levels IIa, IIb, or III. |
