A risk-of-bias label belongs to a result: check which one

Locate the assessed result

A trial can report several outcomes at several times. Measurement methods, missing follow-up, and reporting choices may differ between them. A bias table referring to one result should not automatically be copied to every other result simply because they came from the same participants and publication.

Cochrane’s RoB 2 guidance requires identifying the specific results being assessed. This article explains that scope. It does not perform an independent RoB 2 assessment, certify any quitting trial, or replace the detailed judgment and methodological expertise required to evaluate a result.

Cochrane: Cochrane Handbook, Chapter 8: Assessing risk of bias in a randomized trial

Read the reasons behind the label

Find the domains and explanations relevant to the judgment, such as allocation, deviations, missing outcomes, measurement, or selection of the reported result. An unexplained color in a graphic supplies less information than the underlying rationale. A low-risk label does not mean every imaginable limitation is absent.

For an invented example, a trial reports a self-reported routine measure and a separately verified quitting outcome at different follow-ups. The relevant measurement and missing-data issues may differ. The example does not claim that either measure is valid or biased; it shows why the assessed result must be identified before borrowing its rating.

Keep risk of bias, size and certainty separate

An effect can be large or small while its bias assessment concerns how it was estimated. A precise estimate can still be systematically distorted. Conversely, concern about bias is not a calculation proving that the true effect is zero. Read the reported estimate with its uncertainty and the methodological assessment together.

A GRADE certainty rating concerns a body of evidence for an outcome and considers several dimensions, including risk of bias. It is not the same assessment as a result-specific trial judgment. Do not rewrite a low-risk result as “high-certainty evidence” unless the synthesis actually supplies that separate conclusion and its scope.

Cochrane: Cochrane Handbook, Chapter 14: Grading the certainty of evidence

Carry the scope into your reading note

Write the study result, time point, effect question, assessment tool, overall judgment, and important reason. If the review only shows an unexplained label, note that you have not established the basis. A brief introductory note should not pretend to be a full formal appraisal or invent an expert reviewer.

For clinical choices, ask a qualified professional about the whole evidence. You can use this map to ask why different outcomes received different judgments without assuming inconsistency or misconduct. Your own tracker record does not undergo a trial’s formal bias assessment merely because you label gaps honestly. The useful outcome is knowing exactly what the published judgment covers.