OBJECTIVE: It remains unclear whether the "need" for care at a trauma center should be based on anatomic injury (the current standard) or specialized resource use. We investigated whether anatomic injury severity scores adequately explain hospital resource use. METHODS: This was a retrospective cohort study including children and adults meeting statewide trauma criteria and transported to 48 hospitals from 1998 to 2003. The injury severity score (ISS) was considered as both continuous (range 0-75) and categorical (0-8, 9-15, and >or= 16) terms. Specialized resource use was defined as: major surgery (with and without orthopedic intervention), mechanical ventilation > 96 hours, blood transfusion, intensive care unit (ICU) stay >or= 2 days, or in-hospital mortality. Resource use was assessed as both a binary variable and a continuous term. Descriptive statistics and simple and multivariable linear regressions were used to compare ISS and resource use. RESULTS: 33,699 injured persons were included in the analysis. Within mild, moderate, and serious anatomic injury categories, 8%, 26%, and 69%, respectively, had specialized resource use. When the resource use definition included orthopedic surgery, 12%, 49%, and 76%, respectively, had specialized resource use. Whereas there was fair correlation between ISS and additive resource use (rho = 0.61), ISS explained only 37% of the variability in resource use (adjusted R-squared = 0.37). Resource use within anatomic injury categories differed by age group. CONCLUSIONS: The standard anatomic injury criterion for trauma center "need" (i.e., ISS >or= 16) misclassifies a substantial number of injured persons requiring critical trauma resources. Out-of-hospital trauma triage guidelines based on anatomic injury may need revision to account for patients with resource need.