I'm interested in calculating a time to event model (Cox Proportional Hazards Model) and one of my independent variables has multiple measurements. I'm planning on using a time varying covariate for the independent variable with multiple measurements.
I want to see if a change in this independent variable is associated with our event and I'm not sure how to model this. The only idea I have is to enter the baseline measurement in the model along with an indicator if the measurement has crossed a threshold that we determine ahead of time.
Would it be appropriate to enter the baseline measurement and 0/1 indicator of change of that measurement in the same model? I think that would let us see the effect of change of the measurement while also taking the baseline measurement into account.