You can create two types of figures from the data entered in the Risk of bias tables:
A risk of bias graph is a plot of the distribution of judgements (Low risk of bias, Unclear risk of bias and High risk of bias) across studies for each risk of bias item.
A Risk of bias summary is a summary table of the judgements for each risk of bias item for each study.
With RevMan 5.3, the Risk of Bias summary table for an outcome can viewed in data tables and the Forest Plot window and be included in Figures based on forest plots.
A new option under Interface preferences allows turning the RoB Summary on by default. It can be toggled for an individual outcome using the Risk of Bias button.
Once you have added a figure of the type Forest Plot you can open the figure, edit the properties and select ‘Show Risk of Bias Summary’ to activate this. The Risk of Bias summary will then be located to the right of the forest plot graph.
Bias domains that apply at study level will be shown for all outcomes, but for outcome level domains (blinding of participants and personnel, blinding of outcome assessment and incomplete outcome data), this feature will only work correctly if the relevant outcome groups have been specified in detail, by selecting the analyses applicable to each outcome group in the Risk of Bias Properties..
A May 2014 search of all published reviews showed that only a small percentage of reviews included outcome groups that were specified in sufficient detail, as this has not been necessary in the past.
RevMan adjusts for that fact in two ways:
If only a single outcome group exists for a domain, that will always be shown even if outcomes have not been added to it.
If multiple outcome groups are defined, but the current outcome is not included in any of these, a warning will appear. If the outcome remains unlinked to an outcome group, the domain will not be displayed on any figures based on the outcome.
The Risk of bias figures are called 'Risk of bias and applicability concerns ...' figures in diagnostic reviews. If you are creating a Risk of bias and applicability concerns figure, and have more than one Test Group, you can choose which groups to include in the figure. See Figures generated by RevMan.