An Evaluation of Patient Outcomes before and after Trauma Center Designation Using Trauma and Injury Severity Score Analysis

In June 1990, the Ministry of Health designated 11 hospitals throughout Ontario to be lead hospitals in trauma care. An integral part of a trauma system is the evaluation of care, in particular, outcome of the trauma patients. The Trauma and Injury Severity Score (TRISS) methodology, which offers a standard approach for evaluating outcomes for different populations of trauma patients, was used to determine if there was an improvement in outcomes after the designation of trauma centers of patients involved in motor vehicle crashes (ICD-9-CM, E810.0-825.9), with an Injury Severity Score > 12 for two 12-month periods: one predesignation (1989/1990) and one postdesignation (1992/1993). The Revised Trauma Score, Injury Severity Score, age, and outcome were calculated or abstracted from the hospital chart of each patient at the trauma center. The probability of survival of each patient, the z- and W-statistics of both years were calculated. A measurable improvement was shown in z-statistics between the 2 years from z = -0.40 predesignation to z = +0.72 postdesignation. When the bias introduced by patients intubated before arrival at the trauma center being excluded from TRISS analysis was removed, using a TRISS-like (as per Offner et al: J. Trauma 32:32, 1992) logistic regression equation that allows analysis of intubated patients, the improvement was even greater, with z = +1.34 predesignation and z = +2.97 postdesignation. Only the statistically significant zscore of the postdesignated year required the W-score to be calculated, W = +5.60. This value of W signifies that there were approximately 6 more survivors per 100 patients treated at the trauma center postdesignation than expected from the norms. This study demonstrates and quantifies the improvements in patient outcomes associated with trauma center designation. It also identifies a limitation of TRISS analysis and a method of eliminating the bias in trauma center populations that are largely referred rather than directly off the street.