Alpha Omega Alpha Honor Medical Society

2010 Research Abstract

Reliable Reduction of Trauma Overtriage Based on Serial Prehospital Data: A Conservation of Scarce Health Care Resources

Investigator: Michelle Scerbo, The University of Texas Medical School at Houston

Mentor: John Holcomb, MD, The University of Texas Medical School at Houston

Background: Only 28% of hemodynamically stable and alert trauma patients transported via helicopter required the resources available at our Level I. Investigation of the dynamic relationship of data from prehospital and Emergency Department (ED) records can discern these trauma patients who necessitate the care at a Level I Trauma Center.

Methods: A retrospective evaluation was conducted of adult hemodynamically stable and alert trauma patients transported by helicopter (n=1653) to our Level I center from 2007-2009. Prehospital/ED data included mechanism and location of injury, vital signs, and management. Vital signs were evaluated as temporal trends and intrasubject data variability. Univariate analysis was conducted using Student's unpaired t-test. Random Forest Modeling was used to interpret the importance of these variables to predict the necessity of Level I care.

Results: Univariate analysis revealed that patients that did not necessitate Level I care had a documented loss of consciousness (20% vs. 10% p<0.01), received less crystalloid (258±9cc vs. 299±19cc p<0.05), were evaluated less often for changes in mentation (6.0±0.5 vs. 6.4±0.08 p<0.0001), and had larger temporal trends and intrasubject data ranges for Glasgow Coma Scale (0.4±0.003 vs. 0.8±0.08 p< 0.001; 0.4±0.03 vs 0.7±0.06 p<0.001). The Random Forest Model predicted lower-level admission for 31% of these patients with a Sensitivity of 89%, Specificity of 42%, Negative Predictive Value of 92% and Positive Predictive Value of 34%. 2.8% were improperly classified (under triaged).

Conclusion: While differences between hemodynamically stable and alert trauma patients are not overtly obvious, they do exist. Using complex computer modeling to guide triage decisions may allow more appropriate use of the trauma system.

Updated on September 8, 2011.


© 2017 Alpha Omega Alpha Honor Medical Society