Recent methods and advancements in chemical risk assessment science encourage the integration of epidemiological evidence in the derivation of toxicological reference values. The causal inference movement has demonstrated that causality is usually not determined based on a single method or study but rather involves analyzing a wide variety of evidence and using a weight of evidence approach. For example, recent developments through the notion of triangulation of evidence in environmental epidemiology reinforce the weight of evidence concept where each approach has potential sources of bias; however, if the results all point to the same conclusion, the causal inference is strengthened. In addition, there are many advantages to the use and incorporation of epidemiological evidence for dose-response, including (1) confirmation of human relevance; (2) the potential to examine relevant exposure levels, and (3) incorporation of exposureresponse uncertainties attributed to genetics, co-exposures, or other exposome or population-based considerations. Current risk assessment guidance does not, however, provide clear instruction for risk assessors on how to interpret the relevance of epidemiologic evidence for causal inference, nor is guidance provided on how to navigate challenges in interpretation of epidemiologic evidence due to uncertain exposure measurements, co-exposures and confounders, heterogeneity in response, or other study design limitations that may limit assessment of the temporal relationships between exposure and response. Despite these limitations, epidemiologic evidence can be a valuable tool and provide a more robust framework for reference value derivation with the integration of toxicological information. However, there is still a lack of guidance in how to best integrate human and animal evidence for quantitative reference value derivation, especially when considering inclusion of uncertainties that may impact confidence in causality. Interpretation of the evidence, including dose-response, outcome severity, and integration with toxicological evidence, must be considered as part of a robust risk assessment process. Growing methods, such as use of triangulation for evidence integration, address some of these concerns. For many analyses based on epidemiologic evidence, assumptions regarding the exposure and response must be made; uncertainty analyses that test the sensitivity of derived reference doses to deviations in these assessments are a critical step in the process. This session will open the dialogue regarding the use of epidemiological data in the hazard characterization and exposure assessment steps of chemical risk assessment. An overview of the general issues related to the use of observational studies in establishing causality will be provided. In addition, tools on how to address the uncertainty related to the derivation of toxicological reference values when using epidemiological data will be explored, including the application of methods for evidence synthesis such as triangulation to evaluate risk of bias across studies and for modeling the data to estimate effect concentrations. A communication/education tool for bridging the epidemiology/risk assessment gap—known as the “Matrix”—also will be described. The session will close with case studies that demonstrate the impact of uncertainties from study selection, exposure assumptions, and dose-response modeling assumptions and the impacts that these may have on toxicological reference values