Bias | Berrien 2004 [23] | Blank 2014 [25] | Hanrahan 2011 [26] | Miles 2003 [24] | Wang 2010 [28] | Williams 2006 [27] | Wiliams 2014 [29] |
---|---|---|---|---|---|---|---|
Random sequence generation | Small table of random digits. | Randomised on a 1:1 basis but method unclear. | Computer-generated algorithm | Table of random numbers. | Not reported. | Stratified randomisation, with block size of 10. | Stratified randomisation, with block size of 10. |
Allocation concealment | Randomisation list held by clinical coordinator of HIV program, kept in a locked file. | Not reported. | Person allocating different from the one assessing the inclusion. | Not reported. | Not reported. | Not reported. | Not reported. |
Incomplete outcome data | Lost to follow-up: 5% intervention; 11% control. | Lost to follow up not reported. | Lost to follow-up: 10% intervention; 5% control. | Lost to follow-up at 6Â months: 51% intervention; 58% control. | Lost to follow-up: 14% intervention; 17% control. | Lost to follow-up at 12Â months: 28% intervention; 25% control. | Lost to follow-up at 12Â months: 5% intervention; 22% control. |
Selective reporting | All outcomes reported. | All outcomes reported. | All outcomes reported. | All outcomes reported. | All outcomes reported. | All outcomes reported. | Incomplete reporting of social support and stigma. |
Blinding of participants and personnel | None. | Participants not blinded. | None. | None. | Not reported. | All personnel were blinded except the home visit team. Participants not blinded. | None. |
Blinding of outcome assessment | None. | Research staff blinded. Participants disclosed information, unmasking experimental status. | Data collectors blinded. | Data collectors blinded. | Not reported. | All personnel including interviewers were blinded except the home visit team. | Not reported. |