Epidemiological studies are increasingly being used as evidence in the doseresponse step of chemical risk assessment (CRA). The use of epidemiologic studies in this step of CRA is attractive since the data remove the animal to human extrapolation uncertainty that comes from traditionally practiced methods. Analytical epidemiological investigations, such as case-control and cohort studies, can provide invaluable information for public health and CRA decisions. However, despite contributing a human generalizability factor to the risk characterization, concerns exist regarding internal and external validity issues that arise with observational study designs, affecting specific needs of evidence-based toxicology. For example, many epidemiologic studies often have inherent limitations depending on the study design used and frequently do not address modes of action or causal relationships. In addition, the observational nature and indirect exposure measures often constrain data interpretation. While systematic review guidelines often recommend using risk of bias analyses to address the extent to which study results can accurately identify the relationship between exposure and outcome, there is still uncertainty and debate on what to do with the studies that do not fit the study quality needs. Thus, balancing the approach and understanding the limitations of using epidemiological data for the use in dose-response is needed to reduce gaps in knowledge related to exposure science for contaminants. This session will provide a brief introduction and overview of the general needs and concerns related to using epidemiological data in dose response analysis using a case study on the derivation of a toxicological reference value for cadmium.