| Study |
Bias |
| Randomisation process |
Deviations from intended interventions |
Missing outcome data |
Measurement of the outcome |
Selection of the reported results |
Overall |
| Authors' judgement |
Support for judgement |
Authors' judgement |
Support for judgement |
Authors' judgement |
Support for judgement |
Authors' judgement |
Support for judgement |
Authors' judgement |
Support for judgement |
Authors' judgement |
Support for judgement |
| Subgroup 1.2.1 Three months |
| Manasa 2018 |
Low risk of bias |
The authors mentioned that "Randomization was done using the computer‐generated random number tables and the envelope method." The baseline characteristics were comparable between groups in terms of age, duration, preoperative VA, MLD, and MHI. |
Low risk of bias |
It is unclear whether or not participants were masked to their assigned intervention. Carers or people deliverying the interventions couldn't be masked to the participants' intervention due to the trial context. It is unlikely that there would be deviations from the intended intervention as the intervention is a surgical technique done by physicians as previously assigned. It is likely that participants were analyzed based on group that they were previously assigned. |
Low risk of bias |
Although the number of losses to follow‐up was not explicitly specified, it seems like 9% of all were dropped out. Such number is acceptable. |
Some concerns |
The authors mentioned that "best‐corrected VA (BCVA) was measured using the standard ETDRS chart and converted to logMAR for statistical analysis." Such a method of measurement was appropriate and applied similarly on both groups. However, it is unclear whether or not outcome assessors were aware of the intervention assigned to the participants which is of some concerns. |
Some concerns |
There is no pre‐specified analysus plan available to assess. The authors did not report "change in visual acuity" but instead reported the absolute values at the follow up visits, which is of some concerns. |
Some concerns |
The RCT was judged to have some concerns of bias in some domains. |
| Velez‐Montoya 2018 |
Low risk of bias |
The selection of the ILM peeling technique for each individual case was randomized with a simple block randomization technique (3 × 3). None of the surgeons were aware of the ILM peeling technique before the surgery. They only became aware of the appointed technique after randomization. The baseline characteristics of participants across groups were similar in terms of age, time of evolution, and macular hole measurements at baseline. |
Low risk of bias |
Although the authors did not clearly state whether or not participants were masked throughout the study, it is likely that they were not aware of their treatment assignment. Physicians delivering the interventions were aware of surgery techniques as they could not be masked. However, It is unlikely that there would be deviations from the intended intervention as the intervention is a surgical technique done by physicians who were aware of the appointed technique after randomization. The authors did not clearly specified which analysis were used to estimate the effect of assignment, but It is likely that participants were analysed according to their assigned groups. |
Low risk of bias |
The authors did not mention the number of losses to follow‐up nor evidence that the result was not biased by missing outcome data. However, it is unlikely that losses to follow‐up, if there was any, could be related to participants' health status. |
Low risk of bias |
The author mentioned that "the assessment of the BCVA in Snellen lines [later converted to its logarithm of the minimum angle of resolution, (logMAR) equivalent for statistical purposes]". Such a method of measurement was applied similarly to both groups. Also, they mentioned that "all BCVA assessments and OCT tests during the follow‐up visits were done by a different physician or technician, who was blinded to the ILM peeling technique and randomization of each patient." |
Some concerns |
There was no protocol available to compare. The authors did not report "change in visual acuity" but instead reported the absolute values at the follow up visits, which is of some concerns. |
Some concerns |
The RCT was judged to have some concerns of bias in one domians. |
| Subgroup 1.2.2 Six months |
| Kannan 2018 |
Some concerns |
The authors stated that "System‐generated random number were used to recruit the patients into two groups," suggesting that randomization were properly performed. In terms of allocation concealment, the author specified in the trial registry that the method of concealment was "an open list of random numbers," suggesting that allocation was not properly concealed. The baseline characteristics in terms of sex, age, minimum diameter, base diameter, and baseline visual acuity were comparable between groups. |
Low risk of bias |
The authors stated that participants were masked to their assigned intervention. Carers or people deliverying the interventions couldn't be masked to the participants' intervention due to the trial context. It is unlikely that there would be deviations from the intended intervention as the intervention is a surgical technique done by physicians as previously assigned. Participants were analyzed based on a group that they were previously assigned. |
Low risk of bias |
Although the authors did not explicitly specified, it is likely that all participants' data were available. |
Low risk of bias |
The authors mentioned that "the Snellen visual acuity was converted into a logarithm of the minimum angle of resolution i.e. logMAR for statistical analysis." Such a method of measurement were appropriate and applied similarly on both groups. However, it is unclear whether or not outcome assessors were aware of the intervention assigned to the participants which is of some concerns. |
Some concerns |
Although there was a pre‐specified analysis plan available in the trial registry, such details did not provide sufficient information to assess consistency in reporting outcomes. The authors stated in the trial registry that their visual outcome of interest was "the gain in vision". However, the authors reported vision change from baseline but did not specify at which time point and the unit of change was in lines, which was not clear. |
High risk of bias |
The RCT was judged to have a some concern risk of bias in the randomization process and the selection for the reported result. |
| Subgroup 1.2.3 One year |
| Michalewska 2010 |
Low risk of bias |
The authors stated that simple randomization was performed. Although the authors did not specify whether or not the allocation sequence was concealed, it is likely that participants was not aware of their treatment allocation before receiving intervention as the ranodmization was performed immediately before surgery. The baseline characteristics in terms of sex, age, preoperative macular hole diameters, preoperative visual acuity, and duration of the macular hole were similar between groups. |
Low risk of bias |
It is unclear whether or not participants were masked to their assigned intervention. Carers or people deliverying the interventions couldn't be masked to the participants' intervention due to the trial context. Although 7 cases in PPV with ILM flap group were accidentally seperated their ILM remnant from the margin of macular hole, it is likely to be occured by chance. Participants were analyzed based on group that they were previously assigned. |
Low risk of bias |
Although the authors did not explicitly specified, it is likely that all participants' data were available. |
Low risk of bias |
The authors mentioned that "best‐corrected visual acuity was measured by a blinded examiner using Snellen tables." Such a method of measurement was appropriate and applied similarly on both groups. The authors stated that examiners were masked to participants' assigned interventions. |
Some concerns |
There is no pre‐specified analysus plan available to assess. The authors did not report "change in visual acuity" but instead reported the absolute values at the follow up visits, which is of some concerns. |
Some concerns |
The RCT was judged to have some concerns of bias in one domain. |
