The complexity of the relationship between ethnicity and COVID-19 outcomes: author's reply

We thank Dr Diaz-Quijano and Mr Melo for their knowledgeable input into the topic of how bias and adjustment can influence causal relationships. We agree there are inherent biases, some unobserved, in hospital-based research and that even our illustrative example [1] is unlikely to account for every factor. This is partly due to, as raised by our colleagues, the complexity of the relationship between ethnicity as exposure of interest, COVID-19 severity as the outcome, and other influencing factors (colliders, mediators, and confounders).
Source: Journal of Clinical Epidemiology - Category: Epidemiology Authors: Tags: Commentary Source Type: research