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
- :
- Multivariate analysis for survival

- 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

Multivariate analysis for survival

Jul 30, 2019 11:50 AM
(911 views)

HI there

I am struggling to perform Multivariate analysis for survival in my dataset - can anyone help

4 REPLIES 4

Highlighted
##

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

Re: Multivariate analysis for survival

What is your 'multivariate analysis of survival?' What is your data set? What is the struggle?

Learn it once, use it forever!

Highlighted
##

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

Re: Multivariate analysis for survival

I would like to ascertain which variables in my cancer patient data set is independently associated with overall survival

I have a number of variables some continuous some binary

I know usually the analysis that Is performed is Multivariate analysis but I am unsure how to even begin to start this?

I have a number of variables some continuous some binary

I know usually the analysis that Is performed is Multivariate analysis but I am unsure how to even begin to start this?

Highlighted
##

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

Re: Multivariate analysis for survival

That question is easy to answer! Start by opening JMP and selecting **Help** > **Reliability and Survival**. See the chapters about the **Fit Parametric Survival** and the **Fit Proportional Hazards** platforms. These methods are documented and the examples are shown through out the descriptions.

Learn it once, use it forever!

Highlighted
##

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

Re: Multivariate analysis for survival

To do a survival analysis you need to have an entry date such as the diagnosis date or date of the first treatment. The patients will have been followed and some unfortunately have died, for these the date of the last control will be the date of death. They are called uncensored since they are no longer in the census. It is important that in the follow-up of the patients you do not have lost patients or that these are limited to a low percentage, less than 10%. When you consider doing the analysis you indicate a date on which you have the follow-up of all. The difference between the last control date and the diagnosis date will provide you with survival. In another column you must have for each person the vital state in the last control, for the already deceased JMP by default requires that a 0 and 1 be recorded for the living or censored. If this you already knew sorry but it was important to explain.

Before going on to do the multivariate analysis, I advise first to see what role your prognostic variables have. For this, in the Reliability and Survival submenu select Survival and in Grouping you select the variable you want to analyze, in Y, time to event you select survival and in Censor the column with the vital state in the last control. You do this for each of the forecast variable you have. In this part the numerical variables cannot be analyzed if you have not categorized them before (age over 70 or less, for example).

I suggest doing this before any multivariate analysis because, imagine that you do not find any prognostic variable with statistical significance, in general, no further analysis would be necessary. On the other hand, in case of finding prognostic variables with statistical significance, the univariate analysis helps you to select them in the multivariate model, whose objective is to indicate which of these variables are the really prognostic ones.

Article Labels

There are no labels assigned to this post.