The conventional approach to investigate the modification of personal characteristics and pre-existing comorbidities is by inclusion of interaction terms in the regression model of outcome on the exposure and modifier, or by carrying out separate regressions on strata with different modifier status