Preventable death and interpersonal violence in the United States: Who can be saved?

BACKGROUND Public health initiatives to reduce mortality from penetrating trauma have largely developed from patterns of injury observed in military casualties, with a focus on hemorrhage control and use of tourniquets. Recent efforts show that injury patterns differ between civilian mass casualty events and combat settings, and no studies characterize wounding patterns in all types of civilian homicide. We hypothesize that many homicide deaths are due to nonsurvivable injuries, and that an effective strategy to reduce mortality must focus on both primary prevention as well as improvement in trauma prehospital care. METHODS We analyzed homicides from the National Violent Death Reporting System from 2012 to 2015. We excluded deaths due to poisoning, intentional neglect, or unknown weapon. Deaths were classified as "dead on scene" (DOS), "dead on arrival" (DOA), or "dead at or after hospital" (DAH) if the patient was admitted to a hospital. Injury patterns for penetrating weapons (firearms and sharp instruments) were further categorized. RESULTS We included 18,051 homicides, the vast majority of which were due to firearms (n = 12,901 or 71.5%) or sharp instruments (n = 2,265 or 12.5%). The most common injury patterns included wounds to the chest or head, with isolated extremity injuries representing a minority of both firearms deaths (n = 397 of 12,901, 3.1%) and deaths from sharp instruments (n = 50 of 2,265, 2.2%). Furthermore, over half of all deaths occurred prehospital, with only 13.3% of victims admitted prior to death. CONCLUSION The vast majority of deaths from interpersonal violence are due to firearm injuries. Few deaths appear to be related to extremity hemorrhage alone, and over half of all fatally injured died at the scene. Strategies to decrease mortality from interpersonal violence must go beyond treating injuries that have already occurred, and must address violence prevention directly. LEVEL OF EVIDENCE Epidemiological study, level IV.