Risk of bias visualiser

Pick an instrument, name the studies, answer each signalling question across all of them at once, and take away a Cochrane-style traffic light plot and a Word table. RoB 2, ROBINS-I and QUADAS-2. Nothing is uploaded.

1. Which instrument?

The instrument follows the design of the studies you are appraising, not the question you are asking of them.

    Questions

    Judgements, not data. The work stays in your browser.

    Which instrument should I pick?

    The design of the included studies decides it. RoB 2 for randomized trials, ROBINS-I for non-randomised studies of an intervention, QUADAS-2 for diagnostic accuracy studies. RoB 2 is the tool people mean when they write ROBINS-II, and QUADAS-2 is what they mean by QUADAS; the cards carry the names the tools themselves use.

    Where does the suggested judgement come from?

    For RoB 2 it is the algorithm published with the tool, which maps the signalling answers onto low risk, some concerns or high risk. ROBINS-I and QUADAS-2 publish no such algorithm, so the suggestion there follows those tools' guidance as coded on this page, and the panel says so where you can see it. Every suggestion can be overridden, and a domain you overrode is marked.

    Can I skip the signalling questions?

    Yes. The switch at the top of step 3 replaces the question tables with one dropdown per domain per study, which is what you want when the judgements already exist in a spreadsheet or in RevMan. Leaving the overall cell blank takes it from the domains. Answers typed in either mode are kept when you switch, so nothing is lost by looking.

    What is in the Word file?

    A landscape page with the traffic light as a shaded table, one row per study and one column per domain, then the legend and the key to the domain codes. It is a real .docx written in the browser, so the cells are Word cells rather than a picture: restyle it, drop a column, paste it into a manuscript. The switch above the figure swaps the symbols for the judgement in words.

    Does my work leave the browser, and will it still be here tomorrow?

    Nothing is uploaded; there is no server to upload to. The page does keep the instrument, the studies and the judgements in this browser's local storage, so closing the tab by accident costs you nothing and the page offers to carry on when you come back. That copy stays on your machine and never reaches me. Start again clears it.

    How does this relate to the robvis R package?

    The table in step 4 is shaped like the input robvis expects, with its vocabulary for the judgements, so the copy button gives you something that goes straight into rob_traffic_light(). Use the R package when the figure belongs in a reproducible pipeline; use this page when you want the judgements made and the figure out in one sitting. The weighted summary barplot is the package's, not this page's.