I assume that when you say parameters that you mean independent variables. Thus you may find our contribution of use to you it's shown as an attachment and is available on Researchgate. Best wishes David Booth

Binning them into 2 categories is helpful if you want to compare above/below a set threshold, otherwise you can simply include continuous variables in a cox regression model. The hazard ratio you get out will then correspond to one unit change in your continuous variable.

only caveat is that you are assuming the relationship increases linearly which may not be the case e.g. if age is your continuous variable and increase from age 20 to 21 has a very different effect from increasing age from 47 to 48 then this is a nonlinear relationship and would not be modelled well as a continuous variable.

alternatively you can catergorise it into multiple bins and then use the binned variable giving a hazard ratio for each bin compared to a reference bin e.g. HR for age 40-50 compared to group on 20-30 if 20-30 is your reference bin

Jamie Wallis Thank you for the insightful reply.

If we have all patients group and two subgroups within all patients group, should we use the same cut off values for all three groups or determine cut off values for each group separately?

Using the same cut off values makes for much easier comparisons if you plan to make comparisons