Evaluation of APACHE II system among intensive care patients at a teaching hospital.

CONTEXT The high-complexity features of intensive care unit services and the clinical situation of patients themselves render correct prognosis fundamentally important not only for patients, their families and physicians, but also for hospital administrators, fund-providers and controllers. Prognostic indices have been developed for estimating hospital mortality rates for hospitalized patients, based on demographic, physiological and clinical data. OBJECTIVE The APACHE II system was applied within an intensive care unit to evaluate its ability to predict patient outcome; to compare illness severity with outcomes for clinical and surgical patients; and to compare the recorded result with the predicted death rate. DESIGN Diagnostic test. SETTING Clinical and surgical intensive care unit in a tertiary-care teaching hospital. PARTICIPANTS The study involved 521 consecutive patients admitted to the intensive care unit from July 1998 to June 1999. MAIN MEASUREMENTS APACHE II score, in-hospital mortality, receiver operating characteristic curve, decision matrices and linear regression analysis. RESULTS The patients' mean age was 50 +/- 19 years and the APACHE II score was 16.7 +/- 7.3. There were 166 clinical patients (32%), 173 (33%) post-elective surgery patients (33%), and 182 post-emergency surgery patients (35%), thus producing statistically similar proportions. The APACHE II scores for clinical patients (18.5 +/- 7.8) were similar to those for non-elective surgery patients (18.6 +/- 6.5) and both were greater than for elective surgery patients (13.0 +/- 6.3) (p < 0.05). The higher this score was, the higher the mortality rate was (p < 0.05). The predicted death rate was 25.6% and the recorded death rate was 35.5%. Through the use of receiver operating curve analysis, good discrimination was found (area under the curve = 0.80). From the 2 x 2 decision matrix, 72.2% of patients were correctly classified (sensitivity = 35.1%; specificity = 92.6%). Linear regression analysis was equivalent to r(2) = 0.92. CONCLUSIONS APACHE II was useful for stratifying these patients. The illness severity and death rate among clinical patients were higher than those recorded for surgical patients. Despite the stratification ability of the APACHE II system, it lacked accuracy in predicting death rates. The recorded death rate was higher than the predicted rate.