Background:  Epidemiological studies provide a key part of the evidence base for cancer hazard identification. However, evidence of association between exposure and disease in a population may be affected by selection bias, information bias, and confounding. Therefore, judgement is required about whether a causal interpretation is supported.


Objectives: Discuss how developments in causal inference methods for observational studies can be drawn upon by Working Groups to support cancer hazard identification.


Methods: A small set of methodological developments will be discussed and illustrated.


Results: There is evidence of increasing use of methods for identifying and quantifying biases by Working Groups conducting cancer hazard identifications.  


Conclusions/Implications for practice or policy: Recent developments in causal inference methods aim to reduce speculation about the direction and magnitude of biases and clarify underlying assumptions when evaluating potential sources of bias.  Evidence of the increasing use of these methods by IARC Working Groups suggests that experts find these approaches useful to support cancer hazard identification.