I am currently conducting a study that follows patients over 30 days to see if they have an event.
Patients may have multiple events, meaning they may enter the study and be followed for 30 days then exit and then re-enter
In order to account for this I used multilevel survival analysis with panel data
I started with set up of the survival set:
stset failuretime, failure(failure==1) origin(origintime)
Then I setup panel data:
xtset idvariable
Then I ran an empty model;
xtstreg , d(weibull)
I got the results in picture 1, so I assumed a mixed effects model would be needed
When I run a weibull distribution using mestreg and a lognormal distribution I get completely opposite coeffecients so I realized the distribution really matters. (picture 2)
I followed the instructions in the stata manual (picture 3) to try to determine which distribution to use
In picture 4 are the results, the p-value of kappa is greater than 0.05 but the p-value for the Wald test is less than 0.05
I assumed the lognormal distribution would be the correct thing to use
Am I correct in my assessment and procedures I used?
How else can I determine which distribution to use?
Is mestreg appropriate here?
If I use the lognormal distribution can I get it to display hazards ratios?
Related Posts with Help with multilevel survival analysis
Fixed effects (in Accelerated Failure Time Survival Model)Hi, For my research I am examining private equity strategies and their subsequent exit-types. My da…
After regression, how can I do calculation using r2 and N?After regression, I want to calculate the scalar, that is r2*N How should I do that? I know my bad …
When Stata if command is false, calling on Mata causes errorsHi everyone, Using Stata 15.1 MP8, I've encountered a variety of errors that seem to be caused by ca…
how can I turn year of birth into age?Hello! I'm trying to recode a variable named V241 in the World Values Survey data set (which can be …
STataSta …
Subscribe to:
Post Comments (Atom)
0 Response to Help with multilevel survival analysis
Post a Comment