Risk of bias traffic light plot generator

Record a judgement for every domain of RoB 2, ROBINS-I, Newcastle-Ottawa, QUADAS-2 or AXIS, and get both figures a review needs: the per-study traffic light and the stacked summary across domains. Download as SVG for a manuscript or PNG at 300 DPI for a slide. Nothing is uploaded and nothing is stored.

Your assessments

Figure heading and instrument

Changing the instrument keeps your studies and resets the domains, because the domains are different questions.

Studies and judgements

Use the label your manuscript uses, usually first author and year. The overall column is worked out for you with the RoB 2 rule, so it cannot contradict the domains.

Assessing more than a handful of studies?

In a Verflux project each study carries its own assessment, two reviewers can appraise independently, and these figures are built from the saved judgements rather than retyped. The ratings also feed the GRADE risk-of-bias domain.

Start a free project See what else it does

Cite this tool

If this tool helped with a published review, a citation is the only payment we ask for.

View on Google Scholar
Rehman, N. U., Saif Ullah, K., & Tufail, U. (2026). Verflux: A browser-based platform for end-to-end systematic reviews and meta-analysis (Version 1.0) [Software]. https://verflux.com
@software{verflux2026, author = {Rehman, Naeem Ur and Saif Ullah, Khuram and Tufail, Usman}, title = {Verflux: A browser-based platform for end-to-end systematic reviews and meta-analysis}, year = {2026}, version = {1.0}, url = {https://verflux.com} }

Which instrument fits which design

The tool has to match the design. The domains are not interchangeable, because they describe different ways a study can mislead you.

RoB 2 — randomised trials
Five domains, from the randomisation process to selective reporting. Judged per result, not per paper, so a trial can be low risk for one outcome and high risk for another.
ROBINS-I — non-randomised studies of interventions
Seven domains, starting with confounding, which is the one that usually decides the rating. ROBINS-I is judged against a hypothetical target trial, so "low risk" means comparable to a well conducted randomised trial.
Newcastle-Ottawa — cohort and case-control studies
Selection, comparability and outcome. Published as a star system; here it is recorded on the same three-level scale so the figure is readable beside the others.
QUADAS-2 — diagnostic accuracy
Patient selection, index test, reference standard, and flow and timing. Each domain also carries an applicability question in the full tool.
AXIS — cross-sectional studies
Design, ethics, response rate, data collection and participants. Useful where no intervention is being compared.

What each judgement means

Low risk the domain is unlikely to have changed the result
Some concerns there is a plausible problem, but not enough to doubt the result outright
High risk the problem could plausibly have produced the result
No information the paper does not report enough to judge

The overall judgement follows the published RoB 2 rule, and this page applies it for you: high risk if any domain is high, or if more than one domain raises some concerns; low risk only when every domain is low; otherwise some concerns.

"No information" is a judgement, not a gap to fill in later. Reviewers notice when a column is suspiciously complete for papers that are thin on method.

Four things reviewers query

  1. The wrong tool for the design. RoB 2 applied to a cohort study is the most common version of this, and it is visible at a glance from the domain names.
  2. One rating for a whole paper. RoB 2 is assessed per result. A trial that reports a primary and a secondary outcome can earn two different ratings.
  3. An overall column that contradicts the domains. Usually a study marked low overall while carrying two domains of concerns.
  4. No mention of who assessed. Cochrane expects two independent assessors and a stated way of resolving disagreement. The figure does not show that, so your methods text has to.

Questions

Which risk of bias tool should I use?

RoB 2 for randomised trials, ROBINS-I for non-randomised studies of interventions, Newcastle-Ottawa for cohort and case-control studies, QUADAS-2 for diagnostic test accuracy, and AXIS for cross-sectional studies.

What is the difference between a traffic light plot and a summary plot?

A traffic light plot shows every study and every domain, so a reader can see which study carries which problem. A summary plot stacks the judgements per domain across all studies, showing where the body of evidence is weakest. Most reviews publish both.

How is the overall judgement decided in RoB 2?

A study is high risk overall if any domain is high risk, or if it raises some concerns in more than one domain. It is low risk only when every domain is low. Otherwise it raises some concerns.

Can two reviewers assess independently here?

Not on this page, which is a figure maker for one person's judgements. Verflux projects record each assessor separately and keep the disagreements.

Is it free?

Yes, with no sign-up. It runs entirely in your browser, so nothing you type is uploaded or stored.

More free tools

No account needed for any of them, and nothing you type is uploaded.

PRISMA 2020 flow diagram Build the flow diagram from your screening counts, with the arithmetic checked. GRADE summary of findings Rate certainty per outcome and export the Summary of Findings table.