# [R] Calculating the probability of an event at time "t" from a Cox model fit

aajit75 aajit75 at yahoo.co.in
Mon Dec 19 09:34:45 CET 2011

```Dear R-users,

I would like to determine the probability of event at specific time using
cox model fit. On the development sample data I am able to get the
probability of a event at time point(t).
I need probability score of a event at specific time, using scoring scoring
dataset which will have only covariates and not the response variables.

Here is the sample code:

n = 1000
beta1 = 2; beta2 = -1;
lambdaT = .02 # baseline hazard
lambdaC = .4  # hazard of censoring
x1 = rnorm(n,0)
x2 = rnorm(n,0)

# true event time
T = rweibull(n, shape=1, scale=lambdaT*exp(-beta1*x1-beta2*x2))
C = rweibull(n, shape=1, scale=lambdaC)   #censoring time
time = pmin(T,C)  #observed time is min of censored and true
event = time==T   # set to 1 if event is observed
dataphr=data.frame(time,event,x1,x2)

library(survival)
fit_coxph <- coxph(Surv(time, event)~ x1 + x2 , method="breslow")

library(peperr)
predictProb.coxph(fit_coxph, Surv(dataphr\$time, dataphr\$event), dataphr,
0.003)

# Using predictProb.coxph function, probability of event at time (t) is
estimated for cox fit models, I want to estimate this probability on scoring
dataset score_data as below with covariate x1 and x2.

Is it possible/ is there any way to get these probabilities? since in
predictProb.coxph function it requires response, which is not preseent on
scoring sample.

n = 10000
set.seed(1)
x1 = rnorm(n,0)
x2 = rnorm(n,0)
score_data <- data.frame(x1,x2)