Survival analysis of two time to even... PreviousNext
Mplus Discussion > Growth Modeling of Longitudinal Data >
Message/Author
 Tim Stump posted on Saturday, May 09, 2009 - 9:37 pm
I have two time to event variables which I'm interested in modeling simultaneously: time to first hospitalization and then time to second
hospitalization (or rehospitalization). The data I have are from a clinical trial and there are two treatment groups which I would like to compare. I'd like to know if anyone has created a model for this scenario. Any suggestions or examples you might have would be much
appreciated. Thanks.
 Amir Sariaslan posted on Sunday, May 10, 2009 - 3:03 am
Tim,

Katherine Masyn covers this topic (multiple event-models in the DTSM context) in her lectures available here: http://www.ats.ucla.edu/stat/mplus/seminars/DiscreteTimeSurvival/default.htm

You may also want to take a look at her dissertation, which is available under Papers->Survival Analysis on the Mplus website.

Sincerely,
Amir
 Melanie Wall posted on Thursday, December 15, 2011 - 9:37 am
I am looking for example Mplus code to fit a model like the one in Section 3.1 "Frailty models" of the paper entitled...
Continuous Time Survival in Latent Variable Models by
Tihomir Asparouhov, Katherine Masyn, Bengt Muthen.

I looked at Katherine's course notes and I don't think I saw an example like the leukemia time to relapse and time to death example covered.
 Linda K. Muthen posted on Thursday, December 15, 2011 - 11:43 am
I will email these to you.
 Tim Stump posted on Thursday, April 12, 2012 - 11:34 am
Is it possible to model two parallel time to event processes with mplus? For example, for a particular cohort of subjects I'd like to model time from a screening date to 1) first evidence of depression and 2) first evidence of stroke. I've looked at Tihomir Asparouhov's paper-Continuous Time Survival in Latent variable models, but I thought this was for sequential processes. I'm specifically interested in the parallel case and if my two time-to-event processes are correlated. If this is possible with mplus, would you have mplus code to estimate this model? Thanks.
 Linda K. Muthen posted on Thursday, April 12, 2012 - 1:39 pm
Yes, this is possible. See examples under discrete-time and continuous-time survival in the user's guide. Just extend them to include two processes. I would fit each process separately as a first step.
 Tim Stump posted on Friday, April 13, 2012 - 5:38 am
Thank you. I see now that I can estimate both parallel survival process simultaneously in a single mplus run, but is it possible to estimate a correlation (or covariance) between the two parallel survival processes or does the model assume they are uncorrelated?
 Linda K. Muthen posted on Friday, April 13, 2012 - 5:45 am
You can estimate a covariance. Once you get your model going, send the output and your license number to support@statmodel.com and I will help you set this up.
 David Kerr posted on Monday, April 22, 2013 - 10:20 am
My question is similar to Tim Stump's: I want to test whether one survival process (age at 1st abuse experience right censored at baseline of an intervention trial) predicts concurrent or future survival processes, such as age at first suicide attempt (not censored at baseline) or time to first attempt after baseline. I have output for the latter (simultaneous cox regressions of time to first abuse and time to post-baseline attempt). I would appreciate any suggestions.
 Wen-Hsu Lin posted on Wednesday, October 21, 2015 - 1:31 am
Is it possible to use my discrete survival function to predict a distal outcome? I used 6.17 example but included w8sex on f, along with several controls. However, Mplus results showed that "to avoid singularity, some parameters were fixed." I assumed this means no go. Can it be fixed? Thank you.
 Bengt O. Muthen posted on Wednesday, October 21, 2015 - 3:44 pm
We need to see the output. Please send to Support along with your license number.
Back to top
Add Your Message Here
Post:
Username: Posting Information:
This is a private posting area. Only registered users and moderators may post messages here.
Password:
Options: Enable HTML code in message
Automatically activate URLs in message
Action: