Advance directives theoretically enhance individual autonomy and facilitate treatment decision making at the end of life. There is little empirical evidence to support this, however. Based on a national postal survey of 2172 randomly selected medical practitioners (response rate 73%), this paper examines the effect advance medical directives have on (a) treatment prescribing for terminally ill people and (b) the degree of difficulty practitioners experience in making treatment choices. A hypothetical patient with Alzheimer's disease and an acute life-threatening illness was presented with and without an advance directive. With a directive, respondents were more uniform in their choice of treatment, with 86% choosing as the patient had requested. Difficulty with decision making was also less with the directive, 31% vs 45% with no directive. The data indicate that advance directives do affect practitioners' treatment choices in favor of patient wishes and reduce the difficulty practitioners may experience in making them.