Confounding or bias amplification? Clues for the researcher seeking causal inference. (For ISEE)
On this page:
Objective: When examining health effects due to exposure mixtures, associations between one exposure and outcome could be affected by the well-known problem of confounding, but also by amplification of bias due to the presence of unknown or unmeasured confounders and correlation between the exposure of interest and co-occurring exposures. However, determining which of these types of bias—or both—may exist is not straightforward, and requires knowledge of and assumptions about, the underlying causal structure.
Material and Methods: We performed a simulation study in the context of perfluoroalkyl and polyfluoroalkyl substances (PFAS) to explore how point estimates changed in single- and multi-PFAS models under two different causal structures—one with simple co-exposure confounding and another with unmeasured common causes of individual PFAS and the health outcome (i.e., reflecting potentially different sources of PFAS). We based our simulation on a study examining the association between PFAS biomarker levels and vaccine response, where correlations between PFAS were moderate to high (range: 0.22 to 0.78), and point estimates were presented for both single- and multi-PFAS models.
Results: When bias amplification was present, the ‘better’ choice is the single PFAS model. However, if the direction of the amplification bias, strength of associations and/or degree of correlation were changed, the balance could shift such that the multi-PFAS model yielded lower bias. Furthermore, we showed that it may be possible to detect bias amplification if there exists a PFAS in the mixture that is not causally related to the outcome.
Conclusion: We showed that distinguishing co-exposure confounding from bias amplification may be possible and is important for interpreting models (i.e., single- or multi-PFAS models) examining the health effects of PFAS mixtures.