Upon doing:

kmsurvival <<- survfit(Surv(time, event) ~ Risk)

I get a KM plot similar to :

KM Plot

Now, for a particular patient in the data; say patient number 15, can I find out his/her probability of survival at a given time, using the kaplan meier estimate?


1 Answer 1


You have to see in which groups this patient is, and then do

kmsurvival <- survfit(Surv(time, event) ~ Risk)
summary(kmsurvival, times = 40)

to get the survival probability at 40 (days?). If you want to condition on additional information for that patient (e.g., age, sex, BMI, etc.), you would need to fit a Cox model and get the estimated survival probability from it (though under the proportional hazards assumption).

  • $\begingroup$ Say patient 2,5,3 fall in group1. How does the above solution specifically let me see probability of survival for patient number 2 after 40 months? $\endgroup$
    – rj dj
    Commented Sep 13, 2018 at 12:39
  • 1
    $\begingroup$ For patients in the same group, the survival probabilities will be the same. You don't use any other information for them. $\endgroup$ Commented Sep 13, 2018 at 13:16

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.