Factor | Consequence |
---|---|
Study design | ⊕⊕⊖⊖ |
Risk of bias | ↓a |
Inconsistency of results | → |
Indirectness of evidence | → |
Imprecision | → |
Publication bias | → |
Large magnitude of effect | ↑b |
All plausible confounding would reduce the demonstrated effect or increase the effect if no effect was observed | → |
Dose-response gradient | → |
GRADE quality | ⊕⊕⊖⊖ |