Seven items Cochrane recommends authors report in a completed systematic review that used the RoB 2 tool
| What to report | Further details |
| Methods—‘Assessment of risk of bias in included studies’ | |
| 1. Include all the RoB 2 considerations from the Protocol. | Ensure the review includes all the RoB 2 information from the protocol. If there were any deviations from the protocol, these should be detailed in the review’s methods section, along with justification. |
| 2. State the version of the RoB 2 tool that was used. | The riskofbias.info website lists the current version and archived versions of the RoB 2 tool. Ensure that the version used is stated. |
| Results—‘Risk of bias in included studies’ | |
| 3. Refer to results-level RoB 2 tables, which includes the support for judgement for each domain assessment. | Tables summarising the risk of bias for all results that contribute to an outcome should be included in the review. Each result for the outcomes prespecified for risk of bias assessments (likely to be the reviews’ critical and important outcomes) should have a table that includes the risk of bias judgements (high, low or some concerns) and the support each judgement. In certain circumstances, authors may wish to use other figures that best present the risk of bias data, for example, weighted risk of bias bar plots can provide a succinct summary when there are lots of studies in a synthesis. See figure 2 for an example of a risk of bias figure for one result for one outcome created using robvis.20 |
| 4. State how to access detailed risk of bias assessments data (with consensus responses to the signalling questions). | Authors should state how the data can be accessed, for example, via a online supplemental file 1, repository or other. |
| 5. Provide a brief overview of the risk of bias assessments. | Consider overall comments on key aspects of the risk of bias assessments, for example, the quality of randomisation and extent to which blinding was implemented. Consider whether there are important differences in risk of bias by outcome. If risk of bias assessments are very similar (or identical) for all outcomes in the review, a summary of the assessments across studies should be presented here. If risk of bias assessments are very different for different outcomes, this section should be very brief, and summaries of the assessments across results should be discussed with other GRADE considerations in the Discussion (see point 7 below). |
| Results—Describing the effects of interventions | |
| 6. Refer to visual representations of the risk of bias assessments in relation to each result. | Using forest plots with traffic lights is highly recommended (see Figure 2-6 in Williams et al
10). For synthesis without meta-analysis, we recommend that a column is added to any visual representation of the data to highlight the overall risk of bias associated with each of the results in the table/figure (see table 1 in Williams et al 10). Guidance: section 7.6 of the Cochrane Handbook.1 |
| Results—‘Subgroup analysis’ | |
| (If applicable) Discuss any subgroup analysis conducted that relates to the overall risk of bias judgements. | When preparing subgroup analyses it is advisable to use the ‘overall’ judgement of bias, rather than using bias relating to specific domains. Guidance: section 7.6.2 of the Cochrane Handbook.1 |
| Results—‘Sensitivity analysis’ | |
| (If applicable) Discuss any sensitivity analysis conducted that relates to the overall risk of bias judgements. | When preparing sensitivity analyses it is advisable to use the ‘overall’ judgement of bias, rather than using bias relating to specific domains. Guidance: section 7.6.2 of the Cochrane Handbook.1 |
| Discussion—‘Certainty of the evidence’ (previously the ‘Quality of the evidence’ section | |
| 7. Discuss any risk of bias judgements that affect the certainty of the evidence along with all other GRADE considerations. | Along with the other GRADE considerations, highlight any important implications from the risk of bias assessments for each of the outcomes prespecified for risk of bias assessments. These are likely to be the reviews’ critical and important outcomes or those included in the summary of findings table. For example, if the risk of bias assessments results in downgrading the certainty of the evidence for a specific outcome and whether the effects of the intervention may need to be interpreted with caution. Guidance: section 7.5 and Section 14.2.2 of the Cochrane Handbook. 1 4 |
GRADE, Grading of Recommendations, Assessment, Development and Evaluations; RoB 2, Risk of Bias 2.