Do trauma safety-net hospitals deliver truly safe trauma care? A multilevel analysis of the national trauma data bank

Background: Patients treated at safety-net hospitals, facilities that care for a high percentage of uninsured patients, are known to have worse outcomes. This study seeks to analyze whether care at trauma safety-net hospitals (TSNH) accounts for the known mortality disparity between uninsured and insured trauma patients. Methods: A retrospective analysis of trauma patients (age, 18-64 years) in the National Trauma Data Bank (6.2; 2001-2005) with moderate to severe injury (Injury Severity Score ≥9) was performed. TSNH were defined as facilities treating ≥47% uninsured trauma patients. The main outcome measure was adjusted mortality of patients treated at TSNH versus non-TSNH. A multilevel analysis using multiple logistic regression and generalized estimating equations was performed to control for both hospital and patient-level characteristics (age, gender, insurance, injury severity, shock, and type and mechanism of injury). Subset analyses by hospital trauma level designation and patient injury severity and type were also performed. Results: Collectively 343,053 trauma patients were treated at 46 TSNH and 413 non-TSNH. TSNH patients (n = 36,774) were more likely to be minorities (55% vs. 27%; p \u3c 0.05) compared with non-TSNH patients (n = 306,279). Unadjusted mortality was greater in TSNH versus non-TSNH patients (6.8% vs. 4.6%; *p \u3c 0.05). After controlling for patient- and hospital-level factors, patients at TSNH and non-TSNH facilities had equivalent odds ratio of death = 0.93 (95% confidence interval = 0.65-1.32). Similar results were obtained in all subset analyses. Conclusion: Patients treated at TSNH have equivalent mortality compared with those treated at non-TSNH. Disparate trauma outcomes due to insurance status are not explained by differences between trauma treating institution