Methodologic Issues in the Evaluation of Early Detection Programs