Dear all

I have the following data set titled "EyeSurgery"

The data contains 3000 IDs of patients. The variables are listed below with an example of the first 5 rows.

ID LeftEyeangle RightEyeAngle LeftEyeSurgery RightEyeSurgery LeftTimeToOp RightTimeToOp

1 acute Neutral 1 0 8 -

2 acute acute 1 0 24 -

3 obtuse acute 1 1 15 38

4 obtuse neutral 0 1 18 19

5 neutral neutral 0 0 - -



I am trying to calculate/draw a patient-matched survival analysis curve of "Acute" vs "Obtuse" vs "Neutral" using Cox Proportional Hazards Model. As it is patient-matched(each patient has 2 eye angles), there will be multiple survival curves analysis as the analysis will have to be somehow split into 4 groups "Acute vs Neutral", "Obtuse vs Neutral" and "Acute vs Obtuse" and "Neutral vs Neutral".

My code is :

stset LeftTimeToOp RightTimeToOp, failure(LeftEyeSurgery==1, RightEyeSurgery==1)

stcox, estimate

But this does not seem to work

Would be incredibly grateful if anybody can assist with this issue? It may be a case of reorgansing the data but not sure how to run a mutiple variable survival analysis.


Many thanks in advance