Table 2.
Assessment of bias.
| Lundgren et al. [30] | Random sequence generation | Randomised but not described | Unclear risk of bias |
| Allocation concealment | Randomised but not described | Unclear risk of bias | |
| Blinding of participants and personnel | Blinding not possible | Not assessed | |
| Incomplete outcome data | Follow-up data in each group incomplete | Moderate risk of bias | |
| Selective reporting | Clear outcomes | Low risk of bias | |
| Other sources of bias | None | Low risk of bias | |
|
| |||
| Creasy et al./Perkins et al. [19, 20] | Random sequence generation | Randomised but not described | Unclear risk of bias |
| Allocation concealment | Randomised but not described | Unclear risk of bias | |
| Blinding of participants and personnel | Blinding not possible | Not assessed | |
| Incomplete outcome data | No loss to follow-up reported | Low risk of bias | |
| Selective reporting | Clear outcomes | Low risk of bias | |
| Other sources of bias | None | Low risk of bias | |
|
| |||
| Gelin et al. [29] | Random sequence generation | Randomised via computer based algorithm | Low risk of bias |
| Allocation concealment | Randomised via computer based system | Low risk of bias | |
| Blinding of participants and personnel | Blinding not possible | Not assessed | |
| Incomplete outcome data | Some loss to follow-up | Moderate risk of bias | |
| Selective reporting | Clear outcomes | Low risk of bias | |
| Other sources of bias | None | Low risk of bias | |
|
| |||
| Hobbs et al. [21] | Random sequence generation | Randomised with 2 × 2 factorial design | Low risk of bias |
| Allocation concealment | Computer generated randomisation | Low risk of bias | |
| Blinding of participants and personnel | Blinding not possible | Not assessed | |
| Incomplete outcome data | Four withdrawals | Low risk of bias | |
| Selective reporting | Clear outcomes | Low risk of bias | |
| Other sources of bias | None | Low risk of bias | |
|
| |||
| Badger et al. [31] | Random sequence generation | Randomised but not described | Unclear risk of bias |
| Allocation concealment | Randomised but not described | Unclear risk of bias | |
| Blinding of participants and personnel | Blinding not possible | Not assessed | |
| Incomplete outcome data | All patients lost to 6-month follow-up | High risk of bias | |
| Selective reporting | Clear outcomes | Low risk of bias | |
| Other sources of bias | None | Low risk of bias | |
|
| |||
| Greenhalgh et al. [26] | Random sequence generation | Detailed description of Stata generated randomisation | Low risk of bias |
| Allocation concealment | Computer generated randomisation | Low risk of bias | |
| Blinding of participants and personnel | Blinding not possible | Not assessed | |
| Incomplete outcome data | Moderate loss to follow-up | Moderate risk of bias | |
| Selective reporting | Clear outcomes | Low risk of bias | |
| Other sources of bias | None | Low risk of bias | |
|
| |||
| Kruidenier et al. [27] | Random sequence generation | Computer generated block randomisation | Low risk of bias |
| Allocation concealment | Computer generated block randomisation | Low risk of bias | |
| Blinding of participants and personnel | No blinding | High risk of bias | |
| Incomplete outcome data | Moderate losses to follow-up | Moderate risk of bias | |
| Selective reporting | Clear outcomes | Low risk of bias | |
| Other sources of bias | None | Low risk of bias | |
|
| |||
| Mazari et al. [25] | Random sequence generation | Sealed envelope used to randomise | Low risk of bias |
| Allocation concealment | Sealed envelope used to randomise | Low risk of bias | |
| Blinding of participants and personnel | Blinding not described | Unclear risk of bias | |
| Incomplete outcome data | Moderate loss to follow-up | Moderate risk of bias | |
| Selective reporting | Clear outcomes | Low risk of bias | |
| Other sources of bias | None | Low risk of bias | |
|
| |||
| Spronk et al./Fakhry et al. [22–24] | Random sequence generation | Computer generated block randomisation | Low risk of bias |
| Allocation concealment | Computer generated block randomisation | Low risk of bias | |
| Blinding of participants and personnel | Blinding not possible | Not assessed | |
| Incomplete outcome data | Prolonged study with some loss to follow-up | Moderate risk of bias | |
| Selective reporting | Clear outcomes | Low risk of bias | |
| Other sources of bias | None | Low risk of bias | |
|
| |||
| Murphy et al. [18, 32] | Random sequence generation | Web based randomisation | Low risk of bias |
| Allocation concealment | Web based randomisation | Low risk of bias | |
| Blinding of participants and personnel | Observers blinded | Low risk of bias | |
| Incomplete outcome data | Prolonged study with some loss to follow-up | Moderate risk of bias | |
| Selective reporting | Clear outcomes | Low risk of bias | |
| Other sources of bias | None | Low risk of bias | |
|
| |||
| Bø et al. [28] | Random sequence generation | Computer based randomisation | Low risk of bias |
| Allocation concealment | Computer based randomisation | Low risk of bias | |
| Blinding of participants and personnel | Observers blinded | Low risk of bias | |
| Incomplete outcome data | No loss to follow-up | Low risk of bias | |
| Selective reporting | Clear outcomes | Low risk of bias | |
| Other sources of bias | None | Low risk of bias | |