Table 1.
Risk of bias assessment
| SIGN criteria | Roy and Lopez | Ernat et al. | MacGregor et al. | Knox et al. | Taanila et al. | Seay et al. | Zack et al. | Monnier et al. |
|---|---|---|---|---|---|---|---|---|
| 1.1 Appropriate and clearly focused question | Y | Y | Y | Y | Y | Y | Y | Y |
| 1.2 Groups are comparable in all respects | NA | Y | Y | Y | NA | Y | CS | NA |
| 1.3 Reports participation rates of each group | Y | NA | NA | NA | Y | NA | NA | Y |
| 1.4 Likelihood that subjects had outcome at time of enrolment taken into account in analysis | Y | Y | Y | Y | Y | Y | Y | Y |
| 1.5 Reports dropout/withdrawal rates | Y, 4.6% | NA | NA | NA | Y, 31% | NA | NA | Y, 3.8% |
| 1.6 Compares full participants with those lost to follow-up | N | NA | NA | NA | NA | NA | NA | Y |
| 1.7 Outcomes are clearly defined | Y | Y | Y | Y | Y | Y | Y | Y |
| 1.8 Assessment of outcome is made blind to exposure | NA | CS | NA | NA | NA | CS | CS | NA |
| 1.9 Recognise that knowledge of exposure status could have influenced assessment of outcome | NA | CS | NA | NA | NA | CS | CS | NA |
| 1.10 Method of assessment of exposure is reliable | Y | CS | CS | CS | CS | CS | CS | Y |
| 1.11 Evidence that method of outcome assessment is valid and reliable | N | N | N | N | Y | N | CS | Y |
| 1.12 Exposure/prognostic factor assessed more than once | NA | NA | NA | NA | NA | NA | NA | NA |
| 1.13 Addresses main potential confounders | CS | Y | Y | Y | Y | Y | N | Y |
| 1.14 Confidence intervals provided | Y | Y | Y | Y | Y | Y | Y | Y |
| 2.1 Overall assessment of study based on risk of bias, clinical considerations, and evaluation of methodology | Acceptable | Acceptable | Acceptable | Acceptable | Acceptable | Acceptable | Acceptable | Acceptable |
Y, Yes; NA, not applicable; N, No; CS, Can’t say