Evaluation of the consistency of Acute Physiology and Chronic Health Evaluation (APACHE II) scoring in a surgical intensive care unit.

OBJECTIVES: To determine the applicability of the Acute Physiology and Chronic Health Evaluation (APACHE II) scoring system in a Swiss ICU, and to evaluate its utility in evaluating data from 2 yrs of consecutive admissions to show that the predictability of outcome is similar to that predictability observed by Knaus et al. in 1985 (in 5,815 patients), with the provision that large numbers of patients are studied. DESIGN: Prospective clinical trial over 2 yrs, with statistical analysis of the Swiss patients, and between the Swiss patients and the patients studied by Knaus et al. Receiver operating characteristic curves were calculated. SETTING: Surgical ICU in a Swiss university hospital. PATIENTS: A total of 2,061 consecutive patients admitted to the surgical ICU who were classified as postoperative (elective or emergency) and nonoperative. Hospital mortality rate was considered. RESULTS: Patients were 53 +/- 16 yrs of age. Mean APACHE II score was 10.5 +/- 7.0. The mean APACHE II score was significantly (p < .001) lower in the 1,813 survivors (9.0 +/- 5.2) than in the 248 nonsurvivors (21.5 +/- 8.5). The mortality rate was higher among the Swiss patients when compared with the patients studied by Knaus et al. who had postoperative scores of 20 to 29 and nonoperative scores of > 24. The distribution of the scores and mortality rates were stable during the two study periods, as were the differences in mortality rates between the Swiss population and that population studied by Knaus et al. Sensitivity and specificity of the scores were highest in the emergency surgery group (87% and 78%), and lowest in the elective surgery group (76% and 73%). The APACHE equation underestimated the risk of death. CONCLUSIONS: The APACHE II score, because of its consistency over time and the stability of the mortality rates, can be used in our surgical ICU without modification. The calculated risk of death gives no additional information.