More operations, more deaths? Relationship between operative intervention rates and risk-adjusted mortality at trauma centers.

INTRODUCTION The Trauma Quality Improvement Project has demonstrated significant variations in risk-adjusted mortality rates across the designated trauma centers. It is not known whether the outcome differences are related to provider-level clinical decision making. We hypothesized that centers with good outcomes undertake critical operative interventions aggressively, thereby avoiding complications and deaths. METHODS The previously validated Trauma Quality Improvement Project risk-adjustment algorithm was used to measure observed-to-expected mortality rates (O/E with 90% confidence intervals [CI]) for 152 Level I and II trauma centers participating in the National Trauma Data Bank (version 7.0). Adult patients (>or=16 years) with at least one severe injury (Abbreviated Injury Scale score >or=3) were included (N = 135,654). Operative intervention rates for solid organ injuries (spleen, liver, and kidney) were compared between the centers classified as high mortality (O/E with CI > 1, n = 35 centers) versus low mortality (O/E with CI < 1, n = 37 centers) using nonparametric tests. RESULTS Low- and high-mortality trauma centers were similar in designation level, hospital and intensive care unit beds, teaching status, and number of trauma, orthopedic, and neurosurgeons. Despite a similar incidence and severity of solid organ injuries, low-mortality centers were less likely to undertake operative interventions. CONCLUSION Trauma centers with higher risk-adjusted mortality rates are more likely to undertake operative interventions for solid organ injuries. Hence, there is a need to focus quality improvement efforts on medical decision-making and perioperative processes of care.