Hi Statalist,

I am running a survival analysis to find the difference in occurence of an event between 4 different groups. I am using a multivariable cox regression analysis, however, there is a problem, when I use stcurve after stcox to have the covariate-adjusted survival graph the follow-up time for the groups gets disrupted (the follow-up time for group 2 and 4 is much shorter than the other 2 groups). The Kaplan-Miere graph below shows the correct follow-up time for groups 2 and 3 but the stcurve for these 2 groups continues to the end of the study. Is there anyway that I can fix this?

Thank you,
Mina