1

Survival times are often right censored and left truncated. From my experience, it does not seem like OpenBUGS allows for both. Truncation is denoted as T( , ) and censoring as C( ,). For instance, a censored observation known to lie above 10 days, but cannot have been observed below 2 days (left truncation) seem like it should be coded as thus:

y[i]~dexp(lambda)T(2 , )C(10, )

where the absent second values denote no upper limits. But it seems OpenBUGS can only do one or the other; using both in the code invariably results in a compilation error. Since this is a typical and reasonable modeling scenario, is there any way to do this in OpenBUGS?

ATJ
  • 1,711
  • 1
  • 15
  • 20

0 Answers0