In this paper, retrospectively calculated New Injury Severity Score (NISS) was compared with the prospectively calculated ISS from data derived from the trauma registry of the Trauma Services of the Auckland Hospital, New Zealand as to which test is a better predictor of patient outcome, which is defined as the likelihood of death. The area under the curve (AUC) for ISS and NISS were computed using the non-parametric approach. AUC for ISS=0.95835, and AUC for NISS=0.97350, p less than 0.012. Misclassification rate for ISS was 2.77% and the value for NISS was 2.43%. (A)
Abstract