MULTIVARIATE POPULATION-BASED ANALYSIS OF THE ASSOCIATION OF COUNTY TRAUMA CENTERS WITH PER-CAPITA COUNTY TRAUMA DEATH RATES

The purpose of this study was to utilize a large population-based data base to determine the association of trauma centers with per capita county trauma death rates. Methods: Per capita county trauma death rate, the dependent variable in the model, was obtained from a well-validated state Medical Examiner's data base. Over 200 county demographic, prehospital, and hospital trauma care resource variables were obtained from a variety of sources for multivariate modeling. Bivariate analysis identified candidate variables for multivariate modeling, excluding highly correlated independent variables to avoid problems of collinearity. Multivariate linear regression, logistic regression, and stepwise discriminant analysis were used to determine the relative association of the candidate variables with per capita county trauma death rates. Results: Bivariate analysis identified multiple factors associated with per capita county trauma death rates. These included, among others: county rurality, percentage of unemployment, percentage nonwhite, 911 access, and ALS certified EMS. Per capita trauma death rates were significantly lower in counties with trauma centers compared with counties without trauma centers (4.0 +/- 0.5 and 5.0 +/- 1.1 deaths per 10,000 population, p = 0.0001, respectively). Multivariate analysis demonstrated that the presence of a trauma center in the county and ALS were the best medical system factors predicting decreased per capita county trauma death rates. Conclusions: This study is unique in utilizing a regional population-based data base of all trauma deaths in a large state to analyze the association of trauma centers and trauma death rates. Multivariate modeling controlling for other county variables demonstrated that the presence of a trauma center and Advanced Life Support training were the best predictors of per capita county trauma death rates. These findings are consistent with the hypothesis that trauma centers decrease trauma death rates.