Causal inference in observational studies: Emma McCoy, Imperial College London