Hazard ratio with confindence interval for dummy variable I have covariate Length which is 1 or 0, I'd like to get an estimate with confidence interval for hazard ratio of having an event when Length=0 vs Length=1.
I have came up with this:
proc phreg data=halibutB;
model Time*Censor(0)=Length;
run;

And here the output:

But I am not really sure how to read it. And how can I make the distincion (0 and 1) and add a confidence interval?
 A: The primary calculations in proportional hazards analysis are done with the Cox regression coefficients, which are exponentiated to get the corresponding hazard ratios (HR).
The Parameter Estimate of -0.2897 is the Cox coefficient for the Length=1 versus the Length=0 cases. (I assume here that your variable coding and SAS are using Length=0 as the reference level for that categorical variable.)
The Standard Error is that of the Cox coefficient. The Chi-Square test evaluates whether value of the Cox coefficient is significantly different from 0 (or a HR of 1), based on that Standard Error (Wald Test).
I expect that SAS has a way to display confidence intervals directly, but I don't use SAS. The 95% confidence intervals (CI) for the Cox coefficient can be obtained by adding or subtracting 1.96 times the Standard Error from the estimated Cox Coefficient, as explained for example here, based on a normal distribution of the estimated Cox coefficient. You then exponentiate those CI to get the CI for the HR. The CI for the HR will not be symmetrical about the estimate of the HR.
