P. P. Parikh1, P. J. Parikh4, B. Guthrie4, L. Mamer4, M. Whitmill1, T. Erskine2, R. Woods1, J. Saxe3 1Wright State University,Department Of Surgery,Dayton, OH, USA 2Ohio Department Of Public Safety,Emergency Medical Services,Columbus, OH, USA 3Marian University,Surgery,Indianapolis, IN, USA 4Wright State University,Department Of Biomedical Industrial, And Human Factors Engineering,Dayton, OH, USA
Introduction:
The American College of Surgeons (ACS) developed the National Field Triage Decision Scheme (NFTDS) that has been adapted by many trauma centers in the nation. Some states have modified the NFTDS in an effort to make it more relevant to their population. However, quantitative evidence of the efficacy of these guidelines for a regional trauma is unclear. The objective of this study is to evaluate how the NFTDS and state guidelines compared against observed rates, and against a statistically derived-model, on triage errors.
Methods:
This study included the state trauma data (EMS data merged with Trauma Registry) from Ohio Department of Public Safety (ODPS). The EMS Incident Reporting System (EMSIRS) 2 data were used for the study that ranged from 2008-2012, which accounted for a total of 4914 patient records. Out of these, for 817 patient records “triage protocol” was used for trauma triage decisions. Both the Ohio state and NFTDS were simulated using these data. A predictive model based on multivariate logistic regression was developed and validated using a train-test approach (AUC=0.76). We used the ISS method to identify triage errors for these three approaches.
Results:
As shown in Table 1, the model performed better than NFTDS and state level guidelines by significantly reducing over-triage (OT) rates for the same under-triage (UT) rates, and vice versa. Some factors that were significant and thus added in the statistical model, but not present in the NFTDS were; blunt injury, severe pain, complaint in head, chest, abdomen, and whole body, and change in responsiveness. The observed OT and UT rates were 38.31% and 4.04%, respectively for the state of Ohio. The statistical model performed similar to observed rates. However, since every region has their own version of “protocol,” many such versions existed in the state of Ohio. The model, therefore, may help standardize the triage decision scheme in the state.
Conclusion:
Use of national and state guidelines would have resulted in significantly higher OT rates. The statistical model has many other factors that such guidelines do not have, which might have resulted in better performance. It might be helpful revising these guidelines to include those factors. Further, standardizing the “triage protocol” in the state would certainly help identify factors that affect OT and UT rates, and help improve the performance statewide.
