[R] Checking for different hazard distributions in interval censored data

Asa Johannesen bsaj at leeds.ac.uk
Tue Oct 30 11:43:16 CET 2012

Hi all!

I have two survival data sets looking at similar effects in different settings. One data set is only right censored, but the other is interval censored. In the right censored data set, there is an effect of one factor that causes very different shapes in survival curves (and non-proportional hazards) and I'd really like to say that the same factor has a similar effect in the other data set. However, as the data should really go in an "interval2" Surv object, I can't use cox.zph or a plot of that to say anything statistical about the curve shapes or hazards. I can of course just plot the survival curves and say that they look like they have different shapes, but I was hoping you lot would have a better idea.

Is there a manual way to plot the curves you get from plot(cox.zph()), perhaps using an ictest object (from the package interval)? Are there any plotting methods or tests out there in a package that would help me comment on shapes of interval censored survival curves? Can I cheat and pretend the data are not interval censored for the purposes of getting this information? (some way to say which distribution the curves follow would also be awesome as I've done something like that using survreg with the right censored data)

This is probably really obscure, so I'm not too hopeful. Any ideas would be very welcome.



More information about the R-help mailing list