10 items Cochrane’s evidence production and methods directorate recommends authors consider and report when developing a protocol for a systematic review that plans to use the RoB 2 tool
| What to report | Further details |
| Methods section—‘Assessment of risk of bias in included studies’ | |
| 1. State that RoB 2 tool will be used and reference it | Reference Sterne et al.7 |
| 2. State the effect of interest—effect of assignment or effect of adherence | Guidance: Section 1.3 of the detailed guidance (available via riskofbias.info); Section 8.2.2 of the Cochrane Handbook.13 |
| 3. List or refer to the results that will be assessed using RoB 2, including outcome(s), outcome measure(s) and timepoint(s) | Guidance: section 1.3 of the detailed guidance (via riskofbias.info); section 7.3.21, section 8.2.1 and section 8.7 of the Cochrane Handbook.13 |
| 4. (If applicable) State how cross-over RCTs and cluster RCTs will be handled. | Reference the RoB variant for crossover trials and/ or the RoB 2 variant for cluster trials. Guidance: RoB for crossover trials via riskofbias.info and RoB 2 for cluster trials via riskofbias.info NB: Please note, if it is intended from the OUTSET to use ONLY data from the first period of a crossover RCT, then the standard version of RoB 2 can be used as it is. However, there is potential impact of selective reporting of first period of data only when carry over is detected by trialists. Omission of trials which do not report first period data may lead to bias at the meta-analysis level. For details, see section 23.2 of the Cochrane Handbook. 17 |
| 5. State who will assess RoB2 (initials), how many and whether independently and duplicate | Guidance: section 7.3.2 of the Cochrane Handbook,1 item 11 of PRISMA 202014 |
| 6. List the domains of the tool | Guidance: section 1.3 of the detailed guidance (via riskofbias.info); section 8.2.3 of the Cochrane Handbook.13 |
| 7. List the judgement options (high, some concerns, low) and how overall risk of bias is reached, for example, using the signalling questions/tool algorithms | Guidance: section 1.1, section 1.2.1 and section 1.2.3 of the detailed guidance (via riskofbias.info); section 8.2.3 and section 8.2.4 of the Cochrane Handbook.13 |
| 8. State if there are plans to use any tools to manage the assessment of bias using RoB 2 | For example, the RoB 2 Excel tool to implement RoB 2 (available via riskofbiasinfo.org) Guidance: section 7.3.2 of the Cochrane Handbook.1 |
| Methods section—‘Data synthesis’ | |
| 9. State whether the primary analysis will include all eligible studies or only those which have low risk of bias, or low risk and some concerns | This may depend on the number of studies with each risk of bias rating as larger numbers of studies make the analyses more robust. It could also be appropriate to combine results from studies at high risk of bias and use a sensitivity analysis to assess the effects of restricting the analysis to RCTs overall ‘low’ or ‘low/some concerns’. Guidance: section 7.6.2 of the Cochrane Handbook.1 |
| Methods section—‘Subgroup analysis and investigation of heterogeneity’ | |
| (If applicable) Specify if subgroup analysis is planned based on risk of bias | Consider whether overall risk of bias should be used as the basis for any subgroup analysis. Subgroup analyses may be done as a means of investigating heterogeneous results, or to answer specific questions about particular patient groups, types of intervention or types of study (as well as clinical heterogeneity there is methodological heterogeneity). Guidance: section 10.11.218 and section 7.6.2 of the Cochrane Handbook.1 |
| Methods section—‘Sensitivity analysis’ | |
| (If applicable) Specify if sensitivity analysis based on risk of bias is planned | Consider whether overall risk of bias should be used as the basis for any sensitivity analysis. A sensitivity analysis is a repeat of the primary analysis or meta-analysis in which alternative decisions or ranges of values are substituted for decisions that were arbitrary or unclear. With respect to risk of bias, review authors may perform sensitivity analyses to show how conclusions might be affected if studies at high risk of bias, or high risk bias and some concerns, were included or excluded. Guidance: section 10.1418 and section 7.6.2 of the Cochrane Handbook,1 item 11 of PRISMA 202014 |
| Methods section—‘Summary of findings and assessment of the certainty of the evidence’ | |
| 10. State how the RoB 2 assessment will be used to assess the certainty of the evidence/ GRADE/summary of findings | State that the overall RoB2 judgement will be used to feed into the GRADE assessment. Guidance: section 7.3.2 of the Cochrane Handbook,1 item 15 of PRISMA 2020.14 |
| Other considerations | Authors should not make any changes to the RoB 2 tool (Sterne et al 2019).7
State how detailed RoB2 data will be stored and presented—the RoB 2 tool may generate a large amount of data. We recommend that the consensus decisions for the signalling questions are available to readers in the full review so that the rational for each judgements is transparent. This could be via a repository or as online supplemental file 1 to the review. Guidance: section 7.3.2 of the Cochrane Handbook,1 item 27 of PRISMA 2020.14 See this published protocol as an example: Contraception decision aids to improve care and effective method use19 |
GRADE, Grading of Recommendations, Assessment, Development and Evaluations; PRISMA, Preferred Reporting Items for Systematic Reviews and Meta-Analyses; RCT, randomised controlled trial; RoB 2, Risk of Bias 2.