Kaplan Meier Filter censored right data

[jmarchand1]

Hello,

I don't see in the code or when i use the KaplanMeier Fillter function, the impact of censored right data after the next death event?
S(ti/ti-1) = (Ni - Di) / Ni = 1 - Di / Ni
Ni = (number live ti-1) - (number death à ti) - (number censored between ti-1 et ti)

Could you please explain me.

[CamDavidsonPilon]

The Di and Ni are functions of time. A censored death will change Ni, and a death even will change both Ni and Di.

[jmarchand1]

Thank you for your answer.
I am surprise to not see on the plot, an impact on the next time step after several censure compare to a time interval where there is no censure.
Several censure should decrease more sharply the next time interval probability ?

[CamDavidsonPilon]

Several censure should decrease more sharply the next time interval probability ?

Not necessarily. If deaths are nil that interval, nothing changes