I am doing a regression analysis with hospitalisation as my dependent variable and polypharmacy as my independent variable confounding for age, sex, comorbidity, previous hospitalisation and previous falls. Polypharmacy is a categorical variable with values of 1 to 4. Is it better for me to put all categories of polypharmacy in one model(by simply putting I. in front of it) or should I create different columns for each category of polypharmacy and model for each seperately