Hello,

I am using Stata 16 and want to run a joint model analysis to simultaneously model a longitudinal outcome and time-to-event. I am investigating the impact of pneumonia on cognitive function, my longitudinal outcome therefore is cognitive function and my time-to-event outcome is death (survival). I have longitudinal data with 15 years of observation period and 712 participants. Next to my two outcomes of interest, I want to include covariates of age, sex, education, and number of hospitalisations in my model. I have stset my data prior to running the stjm command.

My question is now from a rather theoretical perspective, I managed to run the stjm command but do not understand on how to specify the following options that are required by the command:

ffp(numlist) specifies power transformations of the time variable to be included in the longitudinal submodel as fixed effects. t0 is used as the time of measurements. Values must be in {-5, -4, -3, -2, -1, -0.5, 0, 0.5, 1, 2, 3, 4, 5}.

rfp(numlist) specifies power transformations of the time variable to be included in the longitudinal submodel as fixed and random effects. t0 is used as the time of measurements. Values must be in {-5, -4, -3, -2, -1, -0.5, 0, 0.5, 1, 2, 3, 4, 5}.

gh(#)

I specified my model as the following:

stjm MMSE age sex education number_hospitalizations pneumonia, panel(ID) survmodel(weibull) ffp(1) gh(10) survcov(age gender education number_hospitalizations pneumonia) timeinteraction(pneumonia)

in which MMSE is the cognitive function and pneumonia the hospitalisation for pneumonia. I understand that either ffp() or rfp() have to be specified and I now just tried around on when my code would compile, however I feel like I do not have an understanding on how the specification of the values between -5 and 5 change what is being done in the computation. When I do not specify the gh() option, I get the following error message:

" BLUP calculation failed Variance of BLUPs for a specific panel could not be estimated Try increasing the number of quadrature nodes"

Has anyone else experienced this error and can help me to understand what exactly I have to specify with ffp/rfp and gh()?

Any advice is greatly appreciated!