Usage Guide
ROBINS-I
How to assess risk of bias for non-randomised interventions
1
Set up the assessment
- 1Create a new ROBINS-I assessment
- 2Define your research question and target trial
- 3Specify pre-intervention confounders
- 4Upload the study manuscript
2
Two-phase AI analysis
- 1Phase 1: AI analyses the document for context, confounders, and preliminary settings
- 2Review and adjust the AI suggestions
- 3Phase 2: AI evaluates all 7 bias domains using the context from Phase 1
- 4Review domain-level judgments with rationales
Tip: The two-phase approach gives more accurate results than single-pass assessment.
3
Review and export
- 1Review each domain: Confounding, Selection, Classification, Deviations, Missing Data, Measurement, Reporting
- 2Override any judgment
- 3Generate the overall risk of bias rating