GRADE guidance 36: updates to GRADE's approach to addressing inconsistency

Area of Science:

• Clinical epidemiology
• Health technology assessment
• Evidence-based medicine

Background:

• Previous Grading of Recommendations Assessment, Development and Evaluation (GRADE) guidance requires updates to address inconsistencies.
• Systematic review authors face challenges in interpreting variability in treatment effect estimates across studies.

Purpose of the Study:

• To update existing GRADE guidance on assessing inconsistency and interpreting subgroup analyses.
• To provide clarifications on assessing the credibility of effect modifiers.

Main Methods:

• An iterative process involving the GRADE working group through written feedback and meetings.
• Consultation with GRADE working group members.

Main Results:

• Clarified that inconsistency refers to variability in results, not study characteristics.
• Guidance on assessing inconsistency for binary outcomes using relative and absolute effects.
• A stepwise process for assessing the credibility of effect modification analyses and subgroup-specific estimates.

Conclusions:

• The updated guidance addresses practical issues faced by systematic review authors regarding inconsistency in treatment effect estimates.
• This update enhances the GRADE approach for evaluating evidence certainty.