Background
The New Injury Severity Score (NISS) was introduced to improve outcome prediction based on anatomical severity scoring in trauma victims. This study was conducted to evaluate whether the NISS can give a better mortality prediction than the Injury Severity Score (ISS) in blunt abdominal trauma patients. Methods: Individuals who visited Chonbuk University Hospital from January of 1996 to October of 2000 were included in this study if a laparotomy was done due to blunt abdominal trauma. The data on the patients were retrospectively collected to identify the injuries and the outcomes of treatment. Both the NISS and the ISS were calculated. The abbreviated injury scale (AIS-90) protocol was used for the severity calculation. The power of the two scoring systems to predict mortality was gauged through a comparison of the misclassification rates, receiver operating characteristic (ROC)curves, and Hosmer-Lemeshow goodness-of-fit statistics. Results: The mean ISS and the NISS of the nonsurvivors were 25.9±7.7 and 29.2±7 . 8 , respectively. The misclassification rates for the ISS and the NISS were not significantly different. Although there was a significant (p<0.05) difference between the area under the curve (AUC) of the ISS (0.813; 95%CI=0.756~0.869) and the AUC of the NISS (0.830; 95%CI=0.778~0.882), they only had fair a predictive abilities. The Hosmer-Lemeshow statistics showed a slightly better fit for the NISS (x2=10.33, p=0.171) than for the ISS (x2= 1 5 . 1 6 , p<0.05). However, these statistics for the ISS and the NISS indicate a poor predictive performance. Conclusions : Both the ISS and the NISS performed poorly in predicting outcome when the data were limited to patients with blunt abdominal trauma.
Comments on this article
DB Error: no such table