Difference between survival rates at specific time point with respective confidence intervals

[Arlindo Ferreira]

Dear Statalist,
I am interested in calculating the difference between the survival rates of two groups at a specific time point with respective confidence intervals. E.g., I would like to know at month 36 what is the survival rate diference between intervention A and B. Hypothetically, the result I expect should be something like: %A - %B = 5.6%, 95% CI (3.1% - 7.9%).
I tried to play with sts list [sts list, at(36) by(group)], but it only describes the survival rates at specific time points. Also, the option "compare" of sts list only puts the rates side by side without actually reporting the diference with 95% CI.
Any idea?
Thank you all very much in advance,
Arlindo

[Paul Dickman]

I believe stsurvdiff (by Patrick Royston, available from SSC) will do what you want.

From the help: "stsurvdiff computes the difference in the Kaplan-Meier survival curves, and a pointwise confidence interval, at the two levels of a binary "treatment" variable, varname"

You can also use -predict- following -stpm2- (available from SSC).