Turn on suggestions

Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.

Showing results for

- JMP User Community
- :
- Discussions
- :
- discrete time cox

- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page

Highlighted

- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Get Direct Link
- Email to a Friend
- Report Inappropriate Content

Dec 13, 2018 1:56 AM
(2637 views)

Hi

Is that possible to perform a Cox Proportional Hazard Regression using discrete time with JMP Pro 12 ?

How to do it with the Fit proportionnal hazard platform ?

In my database, there is one patient per row and i know that the event occured between two visits of the study.

Thanks

Florent

1 ACCEPTED SOLUTION

Accepted Solutions

Highlighted

- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Get Direct Link
- Email to a Friend
- Report Inappropriate Content

Discrete survival time data is not supported in the Fit Model/Proportional Hazard platform. However, you can run a logistic model after the original survival data is transformed, as shown below.

In this example, the response variable for logisitic regression will be "quit", a 0 and 1 discrete variable, where there are multiple records for those whose survival time (represented by "dur") is more than one year.

Please refer to 7.5.2 in https://data.princeton.edu/wws509/notes/c7s6 for details about the data transformation as well as the analogy between the logistic regression and the discrete time proportional hazards model.

3 REPLIES 3

Highlighted

- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Get Direct Link
- Email to a Friend
- Report Inappropriate Content

Discrete survival time data is not supported in the Fit Model/Proportional Hazard platform. However, you can run a logistic model after the original survival data is transformed, as shown below.

In this example, the response variable for logisitic regression will be "quit", a 0 and 1 discrete variable, where there are multiple records for those whose survival time (represented by "dur") is more than one year.

Please refer to 7.5.2 in https://data.princeton.edu/wws509/notes/c7s6 for details about the data transformation as well as the analogy between the logistic regression and the discrete time proportional hazards model.

Highlighted
##

- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Get Direct Link
- Email to a Friend
- Report Inappropriate Content

Re: discrete time cox

Thank you very much, Jiancao.

In your example, how do you use the "censor" data in the logistic regression model ?

Florent.

Highlighted
##

- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Get Direct Link
- Email to a Friend
- Report Inappropriate Content

Re: discrete time cox

You’re welcome. You no longer need that since the values of the binary response variable indicate the censoring status.