In a study I am conducting, I am looking at the time to event (use of opioid analgesic) as a marker of effective analgesia (pain relief) effect for morphine given intrathecally prior to the operation. The longer the time, the more effective the analgesic effect. In some participants, there was no additional opioid required (time to event is undefined, but large) and I am unsure how to conduct a statistical analysis of mean difference between the group which received intrathecal morphine and those who did not.

  • 1
    $\begingroup$ By "time-to-event is undefined" do you mean "censored"? In survival analysis is it more frequent to compare the survival between two groups rather than the mean time. One widely used test is the log-rank which is implemented in many statistical softwares, see en.wikipedia.org/wiki/Logrank_test $\endgroup$
    – periwinkle
    Jun 23 at 8:00
  • $\begingroup$ It is not survival. I want to compare two groups which either have intrathecal morphine (ITM) at the time of their operation vs no ITM. A marker for efficacy is the time for a subsequent opioid analgesic to be required to relieve pain, however in some there in no required second opioid. How coul I analyse the difference between these two groups? $\endgroup$ Jun 23 at 8:09
  • 1
    $\begingroup$ "Survival analysis" is used to describe analyses where the outcome is a time-to-event (and not necessarily related to survival/death). It is quite comon that in this setting the time-to-event is "censored": for some patients you do observe this time, but for others you don't observe it. Rather you only have the information for these patients that this time is larger than the duration of their follow-up (these patients are said to be censored). My feeling is that in your setting the patients who do not need second opioid may be considered as being censored. $\endgroup$
    – periwinkle
    Jun 23 at 8:20

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.