One of the district and universal aspects of emergency medical service (EMS) is the belief that before its implementation many people were dying or being killed by ill-equipped, poorly trained "hearse drivers" and that this tragic state of affairs has been rectified by the advances in the prehospital phase of care. Except for cases of nontraumatic, out-of-hospital cardiac arrest there is almost no convincing scientific evidence to prove that prehospital care has had an impact on morbidity or mortality. At the very foundation of this problem is the lack of a set of broad-based, well-conceived, accurate, reliable, uniform EMS data. Many attempts have been made to develop a uniform EMS data set, but without a national consensus these have not achieved wide distribution. In 1992, with the assistance of the National Highway Traffic Safety Administration, the national consensus process began with a series of meetings involving many EMS agencies and organizations. This culminated in August 1994 with the development of an 81-item uniform EMS data set. We detail the prior attempts at data set development and outline the process leading to the this uniform, national EMS data set.