Assess risk of bias in non-randomized studies of exposures using the ROBINS-E framework with 7 domains, 5 judgment levels, and autosave.
Add your studies and enter their names. Click each colored circle to cycle through judgments: + Low risk, ! Some concerns, − High risk, × Very high risk, ? No information. The Overall column should reflect the most severe domain judgment. When done, export the table as a high-resolution PNG for your manuscript.
Load sample data to see how the tool works, or clear all fields to start fresh.
| Study | D1 Confounding | D2 Measurement of exposure | D3 Selection of participants | D4 Post-exposure interventions | D5 Missing data | D6 Measurement of outcome | D7 Selection of reported result | Overall Overall | |
|---|---|---|---|---|---|---|---|---|---|
ROBINS-E (ROBINS-E Development Group, 2023) • Generated with Research Gold
Click Add Study to create a row for each non-randomized exposure study in your systematic review. Enter descriptive study labels so reviewers can identify each assessment at a glance. Progress is autosaved to your browser.
Evaluate whether the study adequately controlled for confounding variables that could distort the exposure-outcome association. Consider directed acyclic graphs and the target trial framework to identify critical confounders.
Assess how the exposure was measured and whether misclassification could bias results. Consider whether biomarkers, questionnaires, or environmental monitoring provide accurate and consistent classification across exposed and unexposed groups.
Work through selection of participants, post-exposure interventions, missing data, outcome measurement, and selection of the reported result. Each domain uses five judgment levels from Low risk to Very high risk.
Examine the traffic-light table showing domain-level judgments for each study and the robvis-style summary bar chart displaying the percentage distribution of judgments across all included studies.
Download high-resolution PNG files of both the traffic-light table and the summary bar chart. These figures are formatted for direct inclusion in manuscripts, supplementary appendices, or GRADE evidence profiles.
Need this done professionally? Get a complete systematic review or meta-analysis handled end-to-end.
Get a Free QuoteROBINS-E evaluates confounding, exposure measurement, participant selection, post-exposure interventions, missing data, outcome measurement, and selection of the reported result. Each domain targets a distinct mechanism through which bias can distort the exposure-outcome association.
ROBINS-I assesses studies of deliberate interventions, while ROBINS-E is designed for studies where the exposure is not assigned by investigators. The domain structure, signaling questions, and judgment framework differ to reflect the unique causal challenges of exposure research.
ROBINS-E assessments are anchored to a hypothetical target trial that the observational study attempts to emulate. Defining the target trial clarifies what constitutes adequate confounding control, appropriate exposure timing, and relevant outcome measurement windows.
In exposure studies lacking randomization, confounding is almost always the most difficult domain to satisfy. Reviewers must consider whether directed acyclic graphs, propensity scores, or restriction adequately address measured and unmeasured confounders.
ROBINS-E provides a detailed, domain-based judgment framework with explicit signaling questions. The Newcastle-Ottawa Scale uses a simpler star-based scoring system. Use ROBINS-E for rigorous Cochrane-style reviews and Newcastle-Ottawa for rapid quality assessments.
ROBINS-E domain judgments feed directly into the GRADE framework. Studies rated High or Very high risk of bias typically warrant downgrading the certainty of evidence by one or two levels for the risk of bias domain in your GRADE evidence profile.
The overall risk of bias for each study should reflect the most severe domain-level rating. A single domain judged as Very high risk means the study overall receives a Very high risk judgment, regardless of how other domains are rated.
Published by the ROBINS-E Development Group in 2023 in the BMJ, ROBINS-E provides a structured, domain-based framework for evaluating risk of bias in observational studies that estimate the health effects of exposures rather than deliberate interventions. Air pollution and lung cancer, lead exposure and neurodevelopmental outcomes, dietary patterns and cardiovascular disease, and workplace chemical contact and respiratory illness are all questions where randomized experiments would be unethical or infeasible. Unlike ROBINS-I, which was designed for intervention studies, ROBINS-E explicitly addresses the causal and methodological challenges unique to exposure research, where the exposure is not deliberately assigned and its measurement is often subject to substantial error.
The seven domains reflect distinct bias mechanisms in exposure-outcome associations. Confounding (Domain 1) is typically the most challenging because exposure studies lack randomization and often lack a well-defined comparator group. Measurement of the exposure (Domain 2) addresses whether biomarkers, questionnaires, or environmental monitoring provide accurate and consistent classification. Post-exposure interventions (Domain 4) considers whether clinical or behavioral changes triggered by knowledge of exposure status may distort the association. The overall judgment reflects the most severe domain-level rating: a single domain at "very high risk" means the study overall is rated "very high risk." When the evidence base contains substantial risk of bias, consider downgrading certainty using the GRADE certainty framework.
The target trial framework is central to ROBINS-E assessments (Morgan et al., 2024). Before rating each domain, reviewers should specify the hypothetical randomized experiment that the observational study attempts to emulate. This involves defining the eligible population, the exposure contrast of interest, the timing of exposure assignment, the follow-up period, and the outcome measurement strategy. Anchoring domain-level judgments to a clearly articulated target trial prevents vague or inconsistent assessments across review team members and ensures that bias judgments are clinically meaningful rather than purely methodological.
Non-randomized studies of exposures present unique measurement challenges that ROBINS-E addresses through Domain 2 (measurement of the exposure). Unlike intervention studies where treatment assignment is typically well-documented, exposure assessment often relies on self-reported questionnaires, environmental monitoring with spatial or temporal gaps, or biomarker measurements that may reflect acute rather than chronic exposure levels. Differential exposure misclassification, where measurement accuracy varies by outcome status, can create spurious associations or attenuate true effects. The Cochrane Methods Group recommends that reviewers document what exposure assessment method was used, its validated sensitivity and specificity, and whether assessment was blinded to outcome status.
Domain 4, post-exposure interventions, is unique to ROBINS-E and has no direct equivalent in ROBINS-I. In exposure studies, participants or their clinicians may initiate treatments or behavioral changes after learning about exposure status. For example, workers informed of high lead levels may receive chelation therapy, or patients told they carry a genetic risk variant may alter their diet. These post-exposure interventions can modify the outcome independently of the original exposure, introducing bias if they differ between exposed and unexposed groups. Reviewers should assess whether the study design accounts for these pathways or whether results may be distorted by downstream interventions.
Choosing the correct bias assessment tool depends on study design. For randomized controlled trials, apply the Cochrane RoB 2 tool. For non-randomized studies of interventions, use the ROBINS-I tool. For simpler quality scoring of cohort or case-control studies, the Newcastle-Ottawa Scale provides a star-based system. Once all studies have been appraised, bias assessments should inform interpretation of pooled estimates in your forest plot and guide sensitivity analyses that exclude high-risk studies.
When integrating ROBINS-E results into your systematic review, present domain-level judgments in a traffic-light table and summarize the distribution of ratings using the robvis-style bar chart exported from this tool. The Cochrane Handbook recommends narrative synthesis of bias concerns alongside quantitative pooling, noting which domains are most frequently rated at high risk across the included evidence base. If most studies score poorly on confounding, this should be explicitly stated in the discussion and reflected in the corresponding GRADE domain when assessing certainty of evidence. Sensitivity analyses excluding studies rated at high or very high risk provide evidence of whether pooled estimates are robust to potential bias.
ROBINS-E (Risk Of Bias In Non-randomised Studies of Exposures) is a tool published in the BMJ by the ROBINS-E Development Group (2023). It assesses risk of bias in studies estimating the effect of an exposure rather than an intervention. Use ROBINS-E when your systematic review includes observational studies examining environmental, occupational, nutritional, or other non-interventional exposures. For studies of interventions, use ROBINS-I instead.
ROBINS-E evaluates bias across seven domains: (1) confounding, (2) measurement of the exposure, (3) selection of participants into the study or into the analysis, (4) post-exposure interventions, (5) missing data, (6) measurement of the outcome, and (7) selection of the reported result. Each domain targets a specific mechanism through which bias can affect the estimated exposure-outcome association.
ROBINS-I evaluates studies of interventions (deliberate treatments or policies), while ROBINS-E evaluates studies of exposures (environmental pollutants, dietary factors, occupational hazards). ROBINS-E uses 'some concerns' as its second judgment level instead of ROBINS-I's 'moderate.' The domain structure also differs: ROBINS-E includes 'measurement of the exposure' and 'post-exposure interventions' in place of ROBINS-I's 'classification of interventions' and 'deviations from intended interventions,' reflecting the distinct causal structure of exposure studies.
ROBINS-E uses five judgment levels: Low risk of bias, Some concerns, High risk of bias, Very high risk of bias, and No information. The overall judgment for each study should reflect the most severe domain-level rating. A single domain rated 'very high risk' means the overall is 'very high risk' regardless of other domains.
ROBINS-E is designed for cohort studies, case-control studies, cross-sectional studies, and other non-randomized designs that estimate the effect of an exposure on a health outcome. Common applications include environmental epidemiology (air pollution, water contaminants), occupational health (workplace chemical exposures), nutritional epidemiology (dietary patterns, nutrient intake), and social epidemiology (socioeconomic factors, neighborhood characteristics).
Yes. This tool uses autosave to preserve your study names, domain-level judgments, and all assessment data in your browser's local storage. If you close the tab or refresh the page, your work will be restored automatically the next time you open the tool. No account or login is required.
Assessing non-randomized studies of interventions instead? Use the ROBINS-I assessment tool with 7 domains and traffic-light visualization. For randomized controlled trials, apply the Cochrane RoB 2 risk of bias tool. For cohort and case-control study quality, score methodological rigor with the Newcastle-Ottawa Scale scoring tool. When you are ready to visualize your pooled results, generate publication-ready figures with the forest plot generator.
Reviewed by
Dr. Sarah Mitchell holds a PhD in Biostatistics from Johns Hopkins Bloomberg School of Public Health and has over 15 years of experience in systematic review methodology and meta-analysis. She has authored or co-authored 40+ peer-reviewed publications in journals including the Journal of Clinical Epidemiology, BMC Medical Research Methodology, and Research Synthesis Methods. A former Cochrane Review Group statistician and current editorial board member of Systematic Reviews, Dr. Mitchell has supervised 200+ evidence synthesis projects across clinical medicine, public health, and social sciences. She reviews all Research Gold tools to ensure statistical accuracy and compliance with Cochrane Handbook and PRISMA 2020 standards.
We conduct full risk of bias assessments, GRADE evaluations, and complete systematic reviews with rigorous methodology that satisfies peer reviewers. Average turnaround: 2-4 weeks.