Accommodating more than 3 longitudinal biomarkers

Description:

In our problem, we have 21 longitudinal biomarkers with one survival outcome. Our goal is to find out which of the biomarkers are significantly associated with the survival outcome. I have used this package with one longitudinal biomarker at a time and found out that not only the mean trajectory but also its slope is significantly affecting the survival outcome for some of the biomarkers. My next plan is to use all the longitudinal biomarkers together to use their association. For variable selection, we can use some standard shrinkage prior for the 'assoc' and the 'assoc_d' parameters.

In this package, I have noticed that the user is not allowed to use more than 3 longitudinal variables. I want to know what can be the issue with using more than 3 longitudinal biomarkers. You have shared the main part of the stan code in a conference paper. My plan is to use that to implement my idea. Do you have any suggestions regarding this?

Start by reviewing the package's current restriction to three longitudinal variables and the Stan code referenced in the issue. Trace how the assoc and assoc_d parameters represent biomarker associations, then determine what would need to change to support more than three biomarkers and how the requested variable-selection approach could be evaluated.