I have some data with various tests over many years for people; I have a starting point for eligibility for each person (a particular Diagnosis); however, except for those who died, I don't have an end of eligibility; is there any standard in epidemiology, or anything else, for picking this (e.g., as of date of last test in the data or ...)? citations particularly welcome
added: my worry with using date of last test is that this will bias the result (rates will be too high) but I don't see any other good end point for exposure/eligible to be tested
second addition: the Dx is Spinal Cord Injury and the test is PSA; I am leaning toward multiple "end points" as a form of sensitivity analysis
