Comparing the Performance of Three Severity Scoring Systems for ICU Patients: APACHE III, SAPS II, MPM II

Kim, Eun Kyung; Kwon, Young Dae; Hwang, Jeong Hae

Abstract

OBJECTIVE
To evaluate the predictive validity of three scoring systems; the acute physiology and chronic health evaluation (APACHE) III, simplified acute physiology score (SAPS) II, and mortality probability model (MPM) II systems in critically ill patients. METHODS: A concurrent and retrospective study conducted by collecting data on consecutive patients admitted to the intensive care unit (ICU) including surgical, medical and coronary care unit between January 1, 2004, and March 31, 2004. Data were collected on 348 patients consecutively admitted to the ICU (aged 16 years or older, no transfer, ICU stay at least 8 hours). Three models were analyzed using logistic regression. Discrimination was assessed using receiver operating characteristic (ROC) curves, sensitivity, specificity, and correct classification rate. Calibration was assessed using the Lemeshow-Hosmer goodness of fit H-statistic. RESULTS: For the APACHE III, SAPS II and MPM II systems, the area under the receiver operating characterist ic (ROC) curves were 0.981, 0.978, and 0.941 respectively. With a predicted risk of 0.5, the sensitivities for the APACHE III, SAPS II, and MPM II systems were 81.1, 79.2 and 71.7%, the specificities 98.3, 98.6, and 98.3%, and the correct classification rates 95.7, 95.7, and 94.3%, respectively. The SAPS II and APACHE III systems showed good calibrations (chi-squared H=2.5838 p=0.9577 for SAPS II, and chi-squared H=4.3761 p=0.8217 for APACHE III). CONCLUSIONS: The APACHE III and SAPS II systems have excellent powers of mortality prediction, and calibration, and can be useful tools for the quality assessment of intensive care units (ICUs).