rmean in survfit

classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

rmean in survfit

Dinesh W
This post has NOT been accepted by the mailing list yet.
I am using survfit to generate a survival curve. My population is such that my x axis is in days and i have a starting population of say 10,000 of which say only 2000 are left as of day 365. When I try to print rmean it does not print and in any case I am not interested in that. I actually want to sum up all the daily survival values starting at 1.0 (S_0) for day 1 through 0.2 (S_365) through day 365. I then want to assume "r% retained each day" and compute the remaining integral as sum of a geometric series through infinity ( (S_365 * r) * 1/(1-r) ) or through a specific future time period (730 for example in which case the summation portion 1/ (1-r) would change).  Is there an easy way to do this?
Reply | Threaded
Open this post in threaded view
|

Re: rmean in survfit

Frank Harrell
One approach is to use the rms package's cph and Mean.cph functions.  Mean.cph (cph calls coxph and can compute Kaplan-Meier and other survival estimates) can compute mean restricted life.
Frank
Dinesh W wrote
I am using survfit to generate a survival curve. My population is such that my x axis is in days and i have a starting population of say 10,000 of which say only 2000 are left as of day 365. When I try to print rmean it does not print and in any case I am not interested in that. I actually want to sum up all the daily survival values starting at 1.0 (S_0) for day 1 through 0.2 (S_365) through day 365. I then want to assume "r% retained each day" and compute the remaining integral as sum of a geometric series through infinity ( (S_365 * r) * 1/(1-r) ) or through a specific future time period (730 for example in which case the summation portion 1/ (1-r) would change).  Is there an easy way to do this?
Frank Harrell
Department of Biostatistics, Vanderbilt University