I am working with interval censored data and I am fitting an AFT model with the survreg() function from the survival package.

model = survreg(Surv(dat$MIN, dat$MAX, "interval2") ~ factor1 + factor2 + factor3,
                    dist = "loglogistic",
                    data = dat)

The summary() of the model looks fine and the values are realistic.

My question of a more general kind is, how can I check the model assumptions of my AFT model? Are there any diagnostic plots which I could use, or something alike? All I have found so far only works for proportional hazard models (e.g. in the icenReg package) or you need to use the Weibull distribution (e.g. SurvRegCensCov package). Can anyone help?


1 Answer 1


You can extract the residuals from it and evaluate whether their distribution matches the assumed one. However, you will need to account for the fact that these residuals are censored because they are calculated based on the observed event times.

For an example on how to do this in R, see Section 3.4 in my Survival Analysis in R Companion.

  • 1
    $\begingroup$ Great, that worked, thank you! Is there also a way to check the accelerated failure time assumption? $\endgroup$
    – mela
    Commented Dec 7, 2018 at 13:35
  • 1
    $\begingroup$ Also see https:hbiostat.org/rmsc chapter with case study in parametric survival modeling $\endgroup$ Commented Aug 10, 2023 at 14:31

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.