Tab. 1:

Preliminary mapping of the concepts of the modified Bradford Hill considerations (OECD, 2018) to the concepts of certainty in the evidence in the GRADE approach

Assessing certainty in AOPs using the modified Bradford Hill considerations		Relationship to GRADE
Modified Bradford Hill considerations	Criteria for high confidence
Biological plausibility (Is there a mechanistic relationship between an upstream and downstream KE which is consistent with established biological knowledge?)	Extensive understanding of the KER based on extensive previous documentation and broad acceptance (e.g., mutation leading to tumors), i.e., an established mechanistic basis	Biological plausibility is not explicitly part of GRADE, being operationalized under the indirectness domain. Further development of GRADE to respond to considerations around biological plausibility is currently being considered (Whaley et al., 2022).
Essentiality (Are downstream KE and/or the adverse outcome prevented if an upstream KE is blocked?)	Direct evidence from specifically designed experimental studies illustrating essentiality for at least one of the important KE.	Testing of counterfactuals is not part of GRADE. It might become a necessary part of the evidence for developing knowledge of mechanisms. “Essentiality” might be a question that is addressed via a direct SR question rather than as a certainty criterion.
Empirical support	Dependent change in related events following exposure to stressors, accompanied by evidence for temporal, dose-response, and incidence concordance. The more evidence of each, the greater the confidence in the empirical support.	Dose-response is a distinct domain within GRADE. Temporality is accommodated under risk of bias assessment (though risk of bias is not currently part of AOP development). The possibility of increased certainty from seeing more of an upstream event than a downstream event may be additional considerations for the GRADE domains.