Feature request: Select specific state for competing risks model using tbl_survfit

[BG-55]

Hi,
I am a big fan of this package. It's a very helpful resource. I have been doing competing risks analysis and using tbl_survfit to get the survival times. I was wondering being able to select the specific state from the competing risks/multi state model could be added to tbl_survfit? I have multiple outcomes that I would like to know the survival times for, but tbl_survfit only seems to pick the first outcome in my factor to give me the survival times for. Being able to select a specific state or even a vector of states would be very helpful. I am new to Github, so I hope made this correctly. Thank you so much in advance for any help you can provide!

[ddsjoberg]

Hi @BG-55 , thanks for the post! Delighted the package has been helpful for you

Since most users of tbl_survfit() are probably unaware of the existence of competing risks, I don't know if I want to include an argument that would only apply to these models. But have you seen this function https://mskcc-epi-bio.github.io/tidycmprsk/reference/tbl_cuminc.html

It does pretty much the same thing, but specifically for competing risks

[BG-55]

No problem! That makes sense. I usually fit models using survfit rather than cuminc, but I assume they both use the Aalen-Johansen estimator, so I can give that function a whirl. Thank you for letting me know about that function. I had not used it before.

[ddsjoberg]

The tbl_cuminc() function uses cmprsk package to do the calculations, which from my view, is the standard package for competing risks in R. So that could be a good option