A challenge facing all guideline developers is how to keep their guidelines up-to-date. The practice of using a fixed interval to update guidelines, such as every 3 or 5 years, may not be efficient in terms of resources or keep fast-moving guidelines sufficiently up-to-date. I will present a conceptual framework for considering when guidelines need updating, that includes periodic surveillance. Next I will present the results of a programme of surveillance of a large number of systematic reviews, a conceptually similar process. I will then relate our experience with implementing a programme of surveillance for guidelines produced by the American College of Physicians. I will conclude with thoughts and speculations about the future directions for surveillance and updating of guidelines