While a case-control design is not bad, it isn't the best choice for a study of prognosis because it is retrospective.  It is much easier, faster, and cheaper (in general) than a cohort study.