why do I get a horizontal line for hazard ratio - using P.Lambert data

[Denise Vella]

Hello , I'm trying to test, modelling effect of patients on hormone treatment (yes or no) on time using p.lambert data instead of p.dickman's data
(I find trying out with other datasets helps me understand the concept better

1. Why do I get a horizontal line in hazard ratio ?
My tVC is pr_1
2. Does it matter how many degrees of freedom to pick? I picked 5 because they used 5.

clear all

use https://www.pclambert.net/data/rott2b, clear
stset rf, f(rfi==1) scale(12) exit(time 60)

//Model effect of age on survival 

***** WHY DID THEY CHOOSE 60  ???
rcsgen age, df(3) gen(agercs) center(60)


//Model effect over 5 year survival  

//Create spline for year of diagnosis 
rcsgen year, df(3) gen(yearspl) orthog


stpm2 hormon yearspl* age pr_1, scale(h) df(5) eform 
estimates store m1

stpm2 pr_1 hormon yearspl* age, scale(h) df(5) eform tvc(pr_1) dftvc(3)
estimates store m2

//the same time varying vs non time varying 
estimates table m1 m2, eform equations(1) ///
b(%9.6f) modelwidth(12) keep(pr_1 hormon yearspl* age)


//model used
stpm2 pr_1 hormon yearspl* age, scale(h) df(5) eform tvc(pr_1) dftvc(3)


//Prediction hazard ratio as hormone vs non hormon since years of diagnosis 
range temptime 0 10 51

predict hr, hrnumerator(hormon 1) ci timevar(temptime)


twoway (rarea hr_lci hr_uci temptime, color(red%25)) ///
(line hr temptime, sort lcolor(red)) ///
, legend(off) ysize(8) xsize(11) ///
ytitle("Hazard ratio (hormon/not on hormon)") name("hr", replace) ///
xtitle("Years since diagnosis")

[Denise Vella]

Hi I tried again, and still get a horizontal line...

//Trying time varying effect of hormonal treatment since years of diagnosis

clear all

use https://www.pclambert.net/data/rott2b, clear
stset rf, f(rfi==1) scale(12) exit(time 60)
generate hormonYES=(hormon==1)

// spline variables for year of diagnosis
rcsgen year, df(3) gen(yearspl) orthog

stpm2 hormonYES year* age enodes pr_1, scale(hazard) df(4) eform nolog tvc(hormon) dftvc(3)

range temptime 0 10 100

predict hr, hrnumerator(hormonYES 1) ci timevar(temptime)
twoway (rarea hr_lci hr_uci temptime, color(red%25)) ///
(line hr temptime, sort lcolor(red)) ///
, legend(off) ysize(8) xsize(11) ///
ytitle("Hazard ratio (hormonY/hormonNo)") name("hr", replace) ///
xtitle("Years since surgery")