Research methods guide
How to assess certainty of evidence with GRADE
Assess certainty for each outcome across its body of evidence. Document risk of bias, inconsistency, indirectness, imprecision, and publication bias.
A worked review addresses 30-day readmission. Define population and comparison, then rate certainty for that outcome. Example RR 0.82 (95% CI 0.58–1.17), with a prespecified important benefit threshold of RR 0.80.
Documenting domain judgments
Record relative and absolute effects and baseline risk assumptions. These rows illustrate reasoning.
| Item | Worked example | Verification |
|---|---|---|
| Bias and inconsistency | Three trials, one with substantial missingness and influence on the pooled estimate. | Examine direction, intervals and clinical differences alongside I². |
| Imprecision | The interval crosses important benefit and no effect. | At assumed control risk 100/1,000, intervention is about 82/1,000 (58–117); cite the baseline risk source. |
| Indirectness and publication bias | Trials mainly involve tertiary hospitals, and a relevant completed trial has no posted results. | Assess applicability and potential missing evidence separately. |
Apply the method to your materials
Maintain an outcome-specific record
Keep included studies, risk-of-bias judgments, estimates, thresholds and reasons for any downgrading or upgrading.
Check the Summary of Findings
Recompute absolute risks and footnote judgments. Explain missing evidence and limitations.
Questions that arise in practice
Can all outcomes share one certainty rating?
Rate them separately because evidence and precision can differ by outcome.
Verified sources
The official sources below were checked on 2026-11-12. Access conditions and update dates are rechecked before publication.
RigorMed workspace
Start with your research materials
Bring your research question, available materials, and the steps you want to check.
