nav[aria-label="Primary Navigation"] { padding: 0; & ul { list-style: none; width: 100%; display: flex; flex-direction: row; justify-content: start; align-items: start; gap: 30px; padding: 0; & li { margin: 0; } & ul li { list-style: none; } } }

Survival Analysis when "time to event" is category/interval

vincem

Hi there

Does JMP have a module/script to run a Survival analysis when we group the event times into intervals (as en example, time to event split into a category, 0 - 5 days, 6 -10 days, 11-15 days etc...).

Thanks for your input, Vince

Accepted answers

All comments

vincem

i think waht i need is how to fit a "piecewise Cox regression" when tiem to event is in categories

Mark_Bailey

How do you observe time to event, are you observing each day? Why are you grouping the observations this way? Is it for interpretation?

You can use interval censoring this way if appropriate. See the description of assigning your time data columns to the Y role.

Mark_Bailey

Based on the limited description you provided, it is difficult to answer that question, but you intervals are in Y, not X, so segmenting the model does not seem reasonable to me.

vincem

Apologies in the limited info i provided...

I am using the proportional hazards module...

I have as time to event variable (continuous variable which is observed each day) and a censor variable pre-defined. Curious to know if i categories the time to event variable how does oen fit a COX PH?

gregbowers

Hello!

Yes, JMP does offer capabilities for conducting Survival analysis, including scenarios where event times are grouped into intervals. You can achieve this using the "Interval Censoring" option in the Fit Life by X platform. This allows you to specify intervals for the event times and perform Survival analysis accordingly. Thanks for sharing https://community.jmp.com/t5/Discussions/Survival-Analysis-when-quot-time-to-event-quot-is-category/td-p/733575 view

Thank you

vincem

Thank you

Mark_Bailey

The time to event role uses a continuous data column.

vincem

Given that I’m interested in estimating a hazard rate per time ‘interval ‘is this possible?

So assume I have pre-specified time to event intervals, as an example 0 to 5 days, for instance and also have a continuous time to event variable - my main interest is estimating a hazard ratio for the interval (0 to 5 days) for instance.

prishasingh

Thanks