Age was treated as a spline variable in cox model and I got an interesting note...

SeaMoon_168 — Wed, 23 Apr 2025 19:26:42 GMT

I tried to use phreg (code shown below) to calculate the Hazard Ratios of these variables and here age was included as a spline.

proc phreg data=test; 
effect spl_age = spline(age / basis=bspline knotmethod=percentilelist(25 75) );
class sex(ref="0") race(ref="0") Hypertension(ref="0") 
Diabetes(ref="0") 
Region(ref="0"); 
model (start,stop)*Censor(0)= sex race Hypertension Diabetes Region 
spl_age/ risklimits eventcode=1;     
id usubjid; 
run;

The code works but I got a message show below: Hazard ratios that cannot be conveniently calculated or displayed are set to missing in the ParameterEstimates table. Use the HAZARDRAT10 statement to compute the needed hazard ratios. I wonder if anything wrong with the code, especially age was treated as a spline. Many thanks.

Re: Age was treated as a spline variable in cox model and I got an interesting note...

Ksharp — Thu, 24 Apr 2025 01:45:03 GMT

I think that is because you are using spline(non-linear) effect for AGE. sas can not easily calculate HR as other variables by using EXP option of model.
Therefore , you need special statement to get it. Like:

model (start,stop)*Censor(0)= sex race Hypertension Diabetes Region spl_age/ risklimits eventcode=1;
HAZARDRATIO age;

topic Age was treated as a spline variable in cox model and I got an interesting note... in Statistical Procedures

Age was treated as a spline variable in cox model and I got an interesting note...

Re: Age was treated as a spline variable in cox model and I got an interesting note...