When to use the stabilize within trials feature?

This is a more broad question, and may be a worthwhile webinar. I'm curious what situations come to mind in which one would choose to use any of the "stabilize" features. For example in microsimulation, "Stabilize between strategies",  or within a microsimulation sensitivity analysis, "Stabilize trials within Sensitivity Analysis". I've always opted not to stabilize as using a large number of trials should stabilize the results anyways. Just curious if there are any "go to" situations in which one would opt to stabilize. 



1 comment
  • That's a good question. And yes, running a sufficient number of iterations in the model makes the stabilization relatively unimportant.

    One situation where trial-level stabilization is helpful is when running 1-way sensitivity analysis. It provides less noise in the graphs, so the lines are not jagged because the same set of patients is pushed through for each variable value.

    I think there could be debate on the value of stabilization at each level. That's why we chose to provide options for the modeler to choose when and where to stabilize simulation elements.


    Comment actions Permalink

Please sign in to leave a comment.

Didn't find what you were looking for?

New post