Also called the Cochrane’s Review Manager, it is a freely available software, initially built to ease the planning of protocols, the writing of reviews, and the storing of information, within the Cochrane organization. Reviews could be systematic or non-systematic. In case the researcher wishes to include a quantitative analysis in its systematic review such as a meta-analysis , RevMan 5 could readily do the job and could additionally produce some interesting graphical outputs of the data entered.
In this short tutorial about this point and click software, RevMan, we intend to plot the Risk of bias summary and Risk of bias graph, of six(6) published randomized controlled trials, which aggregated data are summarized in the following table:
The methodological quality of each one of the six studies, is judged to be at Low risk of bias, Unclear risk of bias or at High risk of bias, based on seven(7) items carefully described in RevMan 5.4.1 user guide(you can find it online).
The first two are classified as selection bias: an inadequate generation of a randomized sequence or an inadequate concealment of allocations prior to assignment, may bias allocation to interventions. The next two are classified as performance bias and attrition bias: which are both due to knowledge of the allocated interventions after assignment. The fifth one is classified as Attrition bias due to the amount, nature or handling of incomplete outcome data. The sixth item is classified as Reporting bias due to selective outcome reporting. The last class of bias (Other bias), include all the bias due to problems not covered elsewhere in the previous items.
From P. E. Verde et al. 2015, we can imagine the full profile of risk of bias of each one the studies, and then try to figure out how they got their summary and graph of risk of bias plot with the RevMan software.
Download and install RevMan 5.4.1 from here. It looks blank like this when you open it for the first time:
Then click on File>>New. A small window ”New Review Wizard” pops up:
We may continue from there by clicking on the “next” button, but we could also just click on the “Finish” button and another page will open immediately. That’s what we employ. We click on “Finish” and we get this:
The left panel contains a list of modules we could explore, but we are going to look into “Studies and references” only. Click on it and get these details on the middle panel:
Right click on “References to studies”, in that middle panel
We then have an opportunity to register the “included studies” in the software, via “Add Study”. We click on that, and enter the first study “Blume 2008” in short. We do that for the five others with their corresponding short names:
Next, we have to enter information on the methodological quality of each one of the studies. To do that, right click on each one of the study, on the first study for instance, and get these options:
Then click on “Edit Study characteristics”
We are only interested to fill that risk of bias table for “Blume 2008”, as: low low high high low low low respectively
Other studies are filled up similarly with these information:
Duzgun 2008: low high high high low unclear high
kalani 2003: low low low low low unclear high
kastenbauer 2003: unclear unclear low high low unclear low
Londahl 2010: low all the way
Tsang 2003: low low low low low unclear and high
We are all set to generate our plots/graphs. We right click on the “Figures” module on the left panel
We click on “Add Figure”.
“Risk of bias graph” and ”Risk of bias summary” are the only two options we are interested in. Below is the first graph, generated with a caption, a short description of the “Risk of bias graph”. We do not know why, the yellow parts are shown as white.
To save the graph, right click on it, then “Edit Figure”. Save it, as an image(.png) or a pdf, or other desirable formats of your choice:
We go back to generate the other figure of interest:
we get this:
We saved it as an image and it looks like this:
Exercise: Which one of the seven(7) bias domains, can be considered as important factors of bias in the six(6) included studies?.
Verde PE, Ohmann C. Combining randomized and non-randomized evidence in clinical research: a review of methods and applications. Research Synthesis Methods 2015; 6(1):45–62.