Partly because physicians can “never say never,” partly because of the seduction of modern technology, and partly out of misplaced fear of litigation, physicians have increasingly shown a tendency to undertake treatments that have no realistic expectation of success. For this reason, we have articulated common sense criteria for medical futility. If a treatment can be shown not to have worked in the last 100 cases, we propose that it be regarded as medically futile. Also, if the treatment fails to restore consciousness or alleviate total dependence on intensive care, we propose such treatment be judged futile. This definition provides clear end points and encourages the profession to review data from the past and perform, prospective clinical studies that not only report treatments that work but also treatments that do not work. We have also argued that, in a variety of settings, physicians have no ordinary ethical obligation to offer futile interventions (Schneiderman et al. unpublished). Although physicians should inform, and discuss all decisions to withhold or withdraw medical treatments with patients, they need not obtain the patient's permission to desist from futile interventions.