|
|
Cox Proportional Hazards model |
|
Message/Author |
|
|
Hi, I am running a monte carlo for a survival analysis that uses the Cox Proportional Hazards model. I obtained some betas from the monte carlo that show adequate power. The best beta is .2, which I can convert into an odds ratio. However, I have been advised to report relative risks for a grant application. Does Mplus give parameter values that can be used in the Relative Risk formula? Thanks, Mary |
|
|
I think you will need to use output:BASEHAZARD; That will give you the baseline hazard function and from there you can obtain the comulative baseline hazard at a particular time point as well as the survival function and the relative risk. See formulas (6), (7) and (8) in http://www.statmodel.com/download/Survival.pdf You can then put all that into model constraint (see User's Guide example 5.20 on how to add new quantities to be computed as well as their standard errors). You can probably treat the comulative baseline hazard as a constant, but if you want to account for the uncertainty of the comulative baseline hazard you can use ANALYSIS: BASEHAZARD = ON; in combination with something like SURVIVAL = t(20*1); see User's Guide example 6.21. |
|
|
Dear Dr. Asparouhov, Thank you for the help. I didn't know there was a way to request the hazard function. Mary |
|
|
Hi, I am wondering whether a Cox proportional Hazard model is also appropriate when only the Independent variable is time varying (e.g., internationalization Speed on firm Performance)? Would you model it the same way in Mplus and what would be the timcensoring command if we do not have considered, e.g., MNCs that did not enter a specific market? Thank you! Carolina |
|
|
I don't know of a reason why it would not be. You might find this useful https://www.statmodel.com/download/lilyFinalReportV6.pdf In particular Section 4, equation (10). http://statmodel.com/examples/penn.shtml#lily It might be easier to start with discrete time survival though, see User's Guide example 6.19. |
|
Back to top |
|
|