Prediction of blunt traumatic injuries and hospital admission based on history and physical exam

Background: We evaluated the ability of experienced trauma surgeons to accurately predict specific blunt injuries, as well as patient disposition from the emergency department (ED), based only on the initial clinical evaluation and prior to any imaging studies. It would be hypothesized that experienced trauma surgeons' initial clinical evaluation is accurate for excluding life-threatening blunt injuries and for appropriate admission triage decisions. Methods: Using only their history and physical exam, and prior to any imaging studies, three (3) experienced trauma surgeons, with a combined Level 1 trauma experience of over 50 years, predicted injuries in patients with an initial GCS (Glasgow Coma Score) of 14-15. Additionally, ED disposition (ICU, floor, discharge to home) was also predicted. These predictions were compared to actual patient dispositions and to blunt injuries documented at discharge. Results: A total of 101 patients with 92 blunt injuries were studied. 43/92 (46.7 %) injuries would have been missed by only performing an initial history and physical exam ("Missed injury"). A change in treatment, though often minor, was required in 19/43 (44.2 %) of the missed injuries. Only 1/43 (2.3 %) of these "missed injuries" (blunt aortic injury) required surgery. Sensitivity, specificity, and accuracy for injury prediction were 53.2, 95.9, and 92.3 % respectively. Positive and negative predictive values were 53.8 and 95.8 % respectively. Prediction of disposition from the ED was 77. 8 % accurate. In 7/34 (20.6 %) patients, missed injuries led to changes in disposition. "Undertriage" occurred in 9/99 (9.1 %) patients (Predicted for floor but admitted to ICU). Additionally, 8/84 (9.5 %) patients predicted for floor admission were sent home from the ED; and 5/13 (38.5 %) patients predicted for ICU admission were actually sent to the floor after complete evaluations, giving an "overtriage" rate of 13/99 (13.1 %) patients. Conclusions: In a neurologically-intact group of trauma patients, experienced trauma surgeons would have missed 46. 7 % of the actual injuries, based only on their history and physical exam. Once accurate diagnoses of injuries were completed, usually with the help of CT scans, admission dispositions changed in 20.6 % of patients. Treatment changes occurred in 44.2 % of the missed injuries, though usually minimal. Broad elimination of early imaging studies in alert, blunt trauma patients cannot be advocated.