Strata Academy
Risk-of-Bias Tables & Figures for Journal Club
Traffic-light plots, ROB 2 summary figures, AMSTAR 2 presentation, and slide-ready layouts for student critical appraisal sessions
Quick answer
Risk-of-bias visualisations translate domain judgements into tables readers can scan in seconds. ROB 2 uses traffic-light plots (green/yellow/red) per domain per study; systematic reviews also present weighted bar charts showing proportion of low-risk studies. For journal club, show one slide per tool: ROB 2 for trials, AMSTAR 2 item checklist for reviews — and always narrate domains, not just colours.
1. Why visualise risk of bias?
Domain-level judgements are tedious in prose. A traffic-light plot lets journal club attendees see in three seconds that four of five trials have 'some concerns' in Domain 4 (measurement) while randomisation looks clean.
Visualisations also discipline the facilitator — you cannot skip domains if the slide has five columns per study.
Examiners and reviewers expect Cochrane-style figures in dissertations that meta-analyse RCTs. A narrative paragraph saying 'mostly low risk' without a figure is insufficient for dissertation-grade synthesis.
Separate study-level RoB (ROB 2) from review-level quality (AMSTAR 2). Use different figure types — confusing them is a common journal club error.
2. ROB 2 traffic-light plots
Each row is an included study; each column is a ROB 2 domain (D1–D5) plus optional overall judgement. Cells are green (low), yellow (some concerns), or red (high risk).
Overall judgements are algorithm-derived — not an average of colours. A single high domain may drive overall high risk depending on the ROB 2 algorithm.
Present outcome-specific plots when domains differ by endpoint (e.g. mortality objective vs pain subjective). Journal club can use the primary outcome plot only to save time.
Narration script: 'Notice Trial 3 — red in Domain 5. They registered HbA1c primary but published only change from baseline at 6 weeks. How might that affect the pooled estimate?'
- D1 — Randomisation
- D2 — Deviations from intended interventions
- D3 — Missing outcome data
- D4 — Measurement of the outcome
- D5 — Selection of the reported result
3. ROB 2 summary bar charts
Summary plots show the percentage of included studies rated low, some concerns, or high risk within each domain — weighted by sample size in meta-analysis contexts.
Useful when you have twelve trials and the traffic-light plot is too wide for one slide. Journal club: show summary first, then one traffic-light slide for the most influential trial.
Cochrane RevMan exports both study-level and domain-summary figures. If building manually, double-check weighting — examiners notice arithmetic errors.
Link the summary figure to GRADE: if Domain 4 is largely 'some concerns' for a subjective primary outcome, expect GRADE to downgrade for risk of bias.
4. Presenting AMSTAR 2 in journal club
AMSTAR 2 is naturally a table: 16 items with yes/partial yes/no ratings. For slides, highlight the seven critical items — comprehensive search, duplicate screening, excluded studies listed, RoB integration, etc.
Use one slide with critical items colour-coded. A single 'No' on comprehensive search may downgrade overall confidence to critically low regardless of other items.
Do not convert AMSTAR 2 to a traffic-light plot unless your audience knows the tool — the checklist format is clearer for teaching.
Pair with PRISMA flow on the previous slide: search quality and screening transparency support AMSTAR 2 interpretation.
5. ROBINS-I and QUADAS-2 figures
ROBINS-I uses domain judgements for non-randomised intervention studies — similar traffic-light conventions with different domain labels (confounding, selection, etc.).
QUADAS-2 for diagnostic accuracy uses signalling questions across four domains with applicability concerns — often shown as paired 'risk of bias' and 'applicability' plots.
Journal club on a diagnostic paper: one QUADAS-2 slide beats ten minutes on sensitivity/specificity algebra if the group is clinically focused.
Match the tool to the included study design in the review you are appraising — mixed-design reviews may need three figure types.
6. Slide layout tips for facilitators
Slide 1: PICO + which RoB tool. Slide 2: PRISMA flow (if review). Slide 3: Traffic-light or AMSTAR 2 critical items. Slide 4: Forest plot with RoB narrative tied to domains.
Use large fonts — traffic-light cells must be legible on a hospital teaching room projector.
Animate row-by-row only if it helps — over-animation annoys busy firms.
Include citation and outcome label in the figure footer ('ROB 2 — primary outcome: mortality at 90 days').
Provide a handout PDF with full traffic-light plot; slide shows summary only.
- Legible colours with colour-blind-safe palette where possible
- Footer: tool name, outcome, citation
- Never present overall risk without showing domains
- Handout with full plot; slide with summary
7. Dissertation and poster standards
Dissertations meta-analysing RCTs should include ROB 2 traffic-light plot and summary figure in results — exported at publication resolution.
Caption every figure: tool version, judgement date, number of studies, outcome specification.
Posters: one half-column for traffic-light summary; do not shrink until illegible.
Software: RevMan, robvis, Covidence exports, or R robvis package. Record software version in methods for reproducibility.
If conducting your own review, complete ROB 2 before building forest plots — RoB should inform GRADE, not be an afterthought figure.
Interactive version (quizzes, walkthroughs) loads when JavaScript is enabled.