Simple way to estimate the median time and compare survival distributions in analgesic trials under informative censoring [letter]