Prediction of mortality by using the standard scoring systems in a medical intensive care unit in Thailand

Abstract

In order to evaluate and compare the predictive ability of the APACHE II (Acute Physiology and Chronic Health Evaluation II) and the SAPS (Simplified Acute Physiology Score) scoring systems in relation to outcome in a medical intensive care unit (ICU). The authors reviewed consecutive medical ICU admissions (n = 482) at a tertiary hospital over a 2-year period. For each patient, demographic data, diagnosis, APACHE II score, SAPS score and ICU outcome complied during the first 24 hrs of the ICU stay were obtained. The comparison of predictive ability between APACHE II and SAPS was assessed by forward stepwise logistic regression and area under the receiver operating characteristic (ROC) curves. Overall ICU mortality was 36.93%. Mean APACHE II and SAPS scores were 21.17 ± 9.35 and 14.61 ± 6.47, respectively. APACHE II and SAPS scores of nonsurvivors (26.97 ± 8.27 and 18.01 ± 5.84 respectively) were significantly higher than those of survivors (17.77 ± 8.22 and 12.62 ± 5.99 respectively) (p < 0.001). Correlation between both systems was excellent (Pearson correlation coefficient, r = 0.825: p < 0.001). The predicted risk of death calculated by using the APACHE II risk of death equation was 38.98%. The predictive ability to discriminate between survivors and nonsurvivors of APACHE II was higher than SAPS according to forward stepwise logistic regression and area under the ROC curves (APACHE II was 0.788 while SAPS was 0.746). In conclusion, the APACHE II scoring system is an efficient predictor for monitoring the hospital outcome and has more predictive ability than the SAPS in the medical ICU patients.