Making your data and project the best it can be
When should quality and risk of bias be assessed?
We suggest assessing in two stages:
- Before receiving data: assessments should be based on the original publications
- After receiving data: assessments can be revised/updated based on the participant-level data. In some situations, assessments may differ for different participants from the same primary study.
Additional considerations
- IPD may allow the researchers to address risk of bias concerns
- For example, they can apply additional inclusion/exclusion criteria and exclude or reinstate individual participants based on the revised criteria
- IPD may also allow the researchers to address generalizability concerns
- For example, the researchers can conduct subgroup analyses based on study or participant characteristics
Below you will find existing tools for assessing risk of bias and applicability/generalizability in conventional meta-analyses of various study designs.
Extensions to the IPD context are currently in progress by members of our team.
- Randomized controlled trials:
- RoB 2 tool: A revised Cochrane risk of bias tool for randomized trials
- Prediction model studies:
- PROBAST: Prediction model Risk Of Bias ASsessment Tool
- Prognostic factor studies
- QUIPS: QUality In Prognosis Studies tool
- Diagnostic accuracy studies:
- QUADAS-2: A revised tool for QUality Assessment in Diagnostic Accuracy Studies