Question

Suppose you are to simulate a survival data with event time T following an exponential distribution with a mean of 42, and cnesoring time C following an exponential distribution with a mean of 18. For the simulated data, what is the expected percent of censored cases?

Answer 1

Run the code below in R Studio:

 set.seed(1001); n = 100000; T = rexp(n , rate = 1/42); C = rexp(n , rate = 1/18); prop = C/T; prop = prop[prop <= 1]; mean(prop)

Output:

> set.seed(1001) > n = 100000 > T = rexp(n , rate = 1/42) > C = rexpcn , rate = 1/18) > prop - C/T > prop = prop[prop <= 1] > mean(prop) [1] 0.3090527

Here we used to simulate the percent of censcored cases. Since censoring time should be less than event time, we exclude the cases where T > C and the expected value is estimated as 0.3090527