Triage of mass casualties in situations in which patients must remain on-scene for prolonged periods of time, such as after a catastrophic earthquake, differs from traditional triage. Often there are multiple scenes (sectors), and the infrastructure is damaged. Available medical resources are limited, and the time to definitive care is uncertain. Early evacuation is not possible, and local initial responders cannot expect significant outside assistance for at least 49-72 hours. Current triage systems are based either on a shorter time to definitive care or on a longer time to initial triage. The Medical Disaster Response (MDR) project deals with the scenario in which specially trained, local health-care providers evaluate patients immediately after the event, but cannot evacuate patients to definitive care. For this type of scenario, a dynamic triage methodology was developed that permits the triage process to evolve over hours or even days, thereby maximizing patient survival and resulting in a more efficient use of resources. This MDR system incorporates a modified version of "Simple Triage and Rapid Treatment" (START) that substitutes radial pulse for capillary refill, coupled with a system of secondary triage termed, "Secondary Assessment of Victim Endpoint" (SAVE). The SAVE triage was developed to direct limited resources to the subgroup of patients expected to benefit most from their use. The SAVE assesses survivability of patients with various injuries and, on the basis of trauma statistics, uses this information to describe the relationship between expected benefits and resources consumed. Because early transport to an intact medical system is unavailable, this information guides treatment priorities in the field to a level beyond the scope of the START methodology. Pre-existing disease and age are factored into the triage decisions. An elderly patient with burns to 70% of body surface area is unsalvageable under austere field conditions and would require the use of significant medical resources-both personnel and equipment-and would be triaged to an "expectant area." Conversely, a young adult with a Glasgow Coma Scale score of 12 who requires only airway maintenance would use few resources and would have a reasonable chance for survival with the interventions available in the field, and would be triaged to a "treatment" area. The START and SAVE triage techniques are used in situations in which triage is dynamic, occurs over many hours to days, and only limited, austere, field, advanced life support equipment is readily available. The MDR-SAVE methodology is the first systematic attempt to use triage as a tool to maximize patient benefit in the immediate aftermath of a catastrophic disaster.