Interpreting HR and use of stcrreg command in stata 15.1

[Sradha Kotwal]

Hi, I am trying to do a competing risk regression trying to analyse the competing risks between differnent types of dialysis modalities at set time points after starting dialysis.
I have used the following set of commands
stset year1, failure(rrt_y1==1) enter(rrtstart) scale 365.35
stcrreg rrt_base, compete (rrt_y1=2 3 4 5 6 7 8 9)

Year1=date at 1 year post start of dialysis
rrtstart=start date of dialysis
rrt_base= dialysis modality at start of dialysis (9 options but only 4 occur)
rrt_y1=dialysis modality at 1 year post start of dialysis (9 options)

Stata then gives me the output in HR (as asked by me).
My problem is that I dont quite understand how to interpret the HR.

rrt_base
FacHD HR 1
PD HR 0.19
HHD HR 0.15
Tx HR 0.02

I dont know if compared to facHD the risk of PD is 79% less, but I dont think this is right.
I am trying to develop some transitional probabilities with standard errors to populate a markov model.
I would be very grateful for some help.

[Clyde Schechter]

I am having a lot of difficulty understanding what you are trying to do here. If your goal is to develop transition probabilities between different modalities of dialysis, this doesn't seem like a suitable approach. I don't see the relevance of using -stcrreg-. And even if it is relevant, I don't understand why the entry point would be the start of dialysis, which would imply that everybody already fails (either as a direct failure or in a competing mode) at the moment of entry--so you will have no observations eligible for the analysis anyway. Maybe I'm completely misunderstanding this?

In any case, if you want transition probabilities, have you looked into -xttrans-?