02-18-2016 12:18 PM
I have a measure "MELD score" that us repeated up to 21 days (up to when a patient is dischaged from the ICU), and want to see if the information contained in the pattern of decline can be used to predict a binary outcome (Dead vs not) (Retransplant vs not). I also have linear outcomes.
It seems easy enough to do a cox regresion with MELD as the dependent variable vs time as the independent variable but I want to use the rate of decline in meld to model for a specific clinica outcome. Also I am concerned that there is significant collinearity between meld scores on the subsequent days as they are repeated measures of the same thing on the same individuals.
And yes I'm a complete SAS novice!