The primary goal of triage is to identify the majority of field trauma victims at risk for life-threatening injuries. When triage criteria are made sufficiently sensitive to accomplish this goal, high rates of overtriage occur. Orange County's original physiologic criteria were associated with an overtriage rate of 18-40% depending on the definition of a major trauma victim. During the first year's experience with the original physiologic criteria, 21% of non-CNS motor vehicle trauma deaths occurred in nondesignated hospitals. When the criteria were made more sensitive by adding non-time dependent triage criteria such as anatomic and mechanism of injury criteria, deaths in nondesignated hospitals dropped to 4.4%, but the rate of overtriage doubled. Despite this apparent high rate of overtriage, only 5.5% of all paramedic transports were for patients judged to have met field triage criteria. Based on this experience, an approach is suggested for evaluating the balance between over- and undertriage that occurs for a given set of triage criteria. Once this balance has been defined, triage guidelines can be modified to meet regional triage objectives.