Z. G. Hashmi1,2, M. P. Jarman1, T. Uribe-Leitz1, J. W. Scott1, N. R. Udyavar1, J. Havens1, A. Salim1, A. H. Haider1 1Brigham And Women’s Hospital,Boston, MA, USA 2Sinai Hospital Of Baltimore,Department Of Surgery,Baltimore, MD, USA
Introduction: The recent National Academies of Sciences, Engineering and Medicine (NASEM) report states that 20,000-30,000 trauma deaths could be prevented each year if all patients were to receive the highest quality of trauma care. While this burden has been quantified, the nationwide geographic distribution of these preventable trauma deaths remains unknown. Knowing where these deaths occur in each state is important to appropriately allocate resources for the optimal care of the injured. The objective of this study is to identify the geographic distribution of preventable trauma deaths for the state of Florida.
Methods: Adult trauma patients(age≥16) with blunt/penetrating injury in the Healthcare Cost and Utilization Project(HCUP) Florida State Inpatient Database(SID) 2010-2014 were included. Hospitals were linked to the United States Office of Management and Budget-defined Core Based Statistical Areas(CBSAs) using the American Hospital Association supplemental file. CBSAs are distinct geographic units with a high level of socioeconomic integration allowing appropriate population-level comparisons. Preventable deaths were defined as lives which could have been saved if treated at the best-performing CBSA-quintile for in-hospital mortality. We performed hierarchical logistic regression using an empiric Bayes approach to generate Reliability-Adjusted in-hospital mortality rates for each CBSA. These rates were then used to benchmark each CBSA into a performance quintile. Next, generalized linear modeling was used to calculate the relative-risk(RR) of mortality at each quintile, relative to the best-performing CBSA-quintile. This RR was then used to calculate the number of preventable deaths at each CBSA-quintile compared to the best-performers.
Results: A total of 405,126 patients representing 33 CBSAs were included. Overall, 15.8%(1319/8344) of all trauma deaths were deemed preventable. Most of these deaths[78.6% (1037/1319)] occurred at the two worst performing CBSA-quintiles. Figure1 demonstrates that while most of the burden appeared to be concentrated in south/central Florida, isolated areas outside of this cluster were also identified. Separate benchmarking for older trauma patients(≥65 years) demonstrated a higher proportion of preventable deaths(27.6%) versus younger patients(8.3%).
Conclusion: Preventable trauma deaths have a heterogeneous geographic distribution and disproportionately affect certain patient populations. This study shows the feasibility of mapping these deaths using state-specific data. Similar nationwide mapping can offer a unique insight for regional prioritization of quality improvement and resource allocation to achieve the NASEM goal of “Zero Preventable Deaths After Injury.”
