A. Introduction: Injury continues to be a leading cause of death and disability in theUnited States. Although trauma death and injury rates are generally higher in rural areas, the historical focus of traumasystem development has centered on designation of urban based trauma centers and few patient outcome studies have beenconducted in the rural setting. North Coast EMS, which covers a three-county area with a population of 215,000 in therural/remote northwestern corner of California, identified high preventable injury death rates in the mid-1980’s.Between 1988 and 1991, a California EMS Authority special project funded development of a rural trauma model whichdemonstrated statistically significant major trauma patient survival relative to the national Major Trauma PatientOutcome (MTOS) expected norms (J Trauma 1994; 36: 395-400). This model included a prehospital major trauma patientprewarning system, transport to the closest hospital, use of emergency department trauma teams and expeditious movementof injured patients to definitive care. In 1992, an attempt to compare trauma patient outcome between North Coast EMS andHarbor-UCLA Medical Center failed due to incompatible databases. A multivariate logistic regression extension of theTRISS method was therefore developed to compare trauma system type. Special project funding was requested to provide auniform data collection process.

B. Project Description: Between December, 1995 and April, 1997, NorthCoast EMS, in partnership with Los Angeles County EMS and Harbor-UCLA Research and Education Institute (REI), receivedCalifornia EMS Authority Prevention 2000 Block Grant support for a rural/urban major trauma patient outcome study. Twoconsecutive special project contracts funded a single research project designed to contribute to our understanding oftrauma patient care and trauma system effectiveness in rural and urban settings. The TRISS method and the new logisticregression coefficient were utilized to determine the effect of trauma system type on patient survival in the two studypopulations, rural and urban. The new extension is described as follows: After controlling for known TRISS regressioncoefficients (i.e., age, mechanism of injury, vital signs and Injury Severity Score), the effect of additional factors,such as type of trauma system, can be measured. This unknown coefficient allows determination of the relativeefficacy of the two types of trauma systems. If the new coefficient is found to be statistically significantly differentfrom zero, this would imply that the type of trauma system influences the probability of survival, afteraccounting for all other patient characteristics known to influence outcome. If the new logistic regression coefficientis found not to be statistically significant from zero, this would imply that the type of trauma system has littleor no effect on the probability of survival.

C. Tasks/Methodology: North Coast EMS and REI trained datacollection specialists at both sites to uniformly retrieve data at the ten study hospitals, county coroners offices andeach of the approximately thirty transfer receiving hospitals. Data was collected on all trauma patients with an InjurySeverity Score (ISS) score of 10 or greater who arrived at the eight community hospitals and the two Level I TraumaCenters between September 1, 1995 and August 31, 1996. The null hypothesis was established as follows: "In a populationof moderately and severely injured trauma patients, the type of system, rural vs. urban, in which patients are treateddoes not effect their survival, after adjusting for their underlying risk of death using TRISS methodology." All dataobtained from this project was collected on laptop computers utilizing the same relational database program (Paradox 4.5,Borland Inc.) and statistical analysis was performed using the PROC LOGISTIC function of the SAS statistical analysissystem (on a Digital Equipment Corporation VAX Minicomputer). Statistical tests included Chi square, Wilcoxon rank sumand multivariate logistic regression extension of the TRISS method.

D. Outcomes - A total of 1,123 traumapatients were entered into the study. 337 (30%) of the trauma patients were enrolled at the eight rural hospitals and 786(70%) were enrolled in the two trauma centers. A combined total of 345 (31%) patients underwent surgery within six hoursof presentation and a total of 157 (14%) died. The subjects at the urban trauma centers were younger (median age 31 urbanvs. 41 rural, P < 0.0001), more ethnically diverse (Caucasian 38% urban and 86% rural, P < 0.0001), and had ahigher proportion of males than females (75% urban and 62% rural, P < 0.001). The urban population was also moreseriously injured with a higher median Injury Severity Score (17 urban and 14 rural, P < 0.0001), a lower Glasgow ComaScale (P < 0.0001), a lower systolic blood pressure (P < 0.0001), and a lower Revised Trauma Score (P < 0.0001).More patients in the urban trauma centers suffered penetrating injuries (25% urban vs. 9% rural, P < 0.0001). Aftercorrecting for differences in patient population, there was no increase in mortality associated with being treated in arural hospital (Odds Ratio 0.77; 95% confidence interval 0.419 to 1.16). Survival in the eight rural hospitals in theNorth Coast EMS region was not statistically different than the two urban trauma centers in Los Angeles as compared toMajor Trauma Patient Outcome norms.

E. Conclusion - Populations of trauma patients treated in urban andrural areas are fundamentally different. The urban patients are younger, more ethnically diverse, have greater injuryseverity and a higher proportion of penetrating injuries. Results suggest that trauma care systems, such as the ruralsystem in this study, with low volume and high blunt trauma rates, can effectively care for its population of traumapatients with an enhanced, committed trauma system which allows for expeditious movement of patients toward definitivecare, without hospital bypass of major trauma patients to designated trauma centers. Study results also verify that highvolume, high penetrating trauma populations can be effectively managed by trauma centers.