Graphing restricted splined, trying to graph it to original model (rcspline)

[Martin Imelda Borg]

Hello all. I'm trying to graph SURGICAL EXPERIENCE vs CANCER RECURRENCE
The hypothesis: The more the number of cases (surgical experience) the less likely the recurrence.

data:

* Example generated by -dataex-. For more info, type help dataex
clear
input float experience byte bcr float(dead ttlastfollowbcr modelpgrade5 modelpgrade6 modelpgrade7 modelpgrade8 modelpgrade9 mkexperience1 mkexperience2 mkexperience3 mkexperience4) byte(_est_full_spline _est_no_spline)
0 0 0 33.379875 0 0 1 0 0 0 0 0 0 1 1
0 1 0 2.1355236 0 0 1 0 0 0 0 0 0 1 1
0 0 0 1.4455853 0 1 0 0 0 0 0 0 0 1 1
0 0 1  1.708419 0 1 0 0 0 0 0 0 0 1 1
0 0 0  80.91991 1 0 0 0 0 0 0 0 0 1 1
0 1 0  14.98152 0 1 0 0 0 0 0 0 0 1 1
0 1 0  12.02464 0 0 1 0 0 0 0 0 0 1 1
0 0 0  101.0924 0 1 0 0 0 0 0 0 0 1 1
0 1 0  1.379877 0 0 1 0 0 0 0 0 0 1 1
0 0 0 102.07803 0 1 0 0 0 0 0 0 0 1 1
end

Apart from using LR /AIC/BIC I would like to graph model to compare THE SPLINE MODEL vs NON SPLINE MODEL.
I don't need to do this step, but just curious to see how it would look like

I came across this graph which I though looked really neat. I can't figure out how to plot it though
This is the closest I got to:

//model with full spline 
mkspline mkexperience = experience, knot(`k1' `k2' `k3') cubic

stcox modelpgrade6 modelpgrade7 modelpgrade8 modelpgrade9 ///
    experience spexperience1 spexperience2 spexperience3 quart, dist(llog) cluster(anonsurgeonid) nolog
estat ic 
estimates store full_spline

//model with linear term - no spline
stcox modelpgrade6 modelpgrade7 modelpgrade8 modelpgrade9 ///
    spexperience1 quart, dist(llog) cluster(anonsurgeonid) nolog
estat ic 
estimates store no_spline

lrtest full_spline no_spline 

rcspline bcr experience, knots(`k1' `k2' `k3') ci showknots

This is the resulting graph:



Questions:
1. How do I insert my LINEAR model into the graph to compare model 1 vs model 2 just like the author does in the picture above, with the HR on the y x axis?

I've trying the help for rcspline and the regress options doesn't specify how I can add on the first model

Are there any thoughts on this?

[Martin Imelda Borg]

Clyde Schechter I'm sure you're busy, just wondering if you can help here?