ABSTRACT
BACKGROUND: The ability to accurately adjust for the severity of illness in outcome studies of critically ill patients is essential for clinical audits and trials. Several studies have been carried out to validate the acute physiology and chronic health evaluation (APACHE II) score in Korean ICUs. However, few reports have been presented that compare the performance of the APACHE II score and diagnostic category weighted APACHE II models in the surgical ICU population of Korea. The aim of this study was to validate APACHE II and compare the performance of the APACHE II and adjusted APACHE II models for emergency admission in a surgical intensive care unit (SICU) population. METHODS: A retrospective analysis of the prospective ICU registry was conducted in the SICU between October 2007 and February 2011. Calibration and discrimination were determined by the Hosmer-Lemeshow test and the area under the receiver operating characteristic (AUC) curve from patients. RESULTS: This study included 854 patients. SICU mortality was 9.4%. For APACHE II and adjusted APACHE II, AUCs were 0.791 and 0.757, respectively. Hosmer and Lemeshow C statistics showed good calibration for APAHCE II and for adjusted APACHE II (p > 0.05). CONCLUSIONS: The ability of the APACHE II system in predicting group outcome is validated in a surgical ICU population by a receiver operating characteristic curve and logistic regression analysis. Mortality rates predicted using APACHE II exhibited good calibration and moderate discrimination. Diagnostic category weighted adjusted APACHE II did not improve the mortality prediction.