Objective: To evaluate Acute Physiology and Chronic Health Evaluation (APACHE) II and Simplified Acute Physiology Score (SAPS) II scoring systems in a single intensive care unit (ICU), independent from the ICUs of the developmental sample; and to compare the performance of APACHE II and SAPS II by means of statistical analyses in such a clinical setting.
Design: Prospective, cohort study.
Setting: A single ICU in a Greek university hospital.
Patients: In a time interval of 5 yrs, data for 681 patients admitted to our ICU were collected. The original exclusion criteria of both systems were employed. Patients <17 yrs of age were dropped from the study to keep compatibility with both systems. Eventually, a total of 661 patients were included in the analysis.
Interventions: Demographics, clinical parameters essential for the calculation of APACHE II and SAPS II scores, and risk of hospital death were recorded. Patient vital status was followed up to hospital discharge.
Measurements and main results: Both systems showed poor calibration and underestimated mortality but had good discriminative power, with SAPS II performing better than APACHE II. The evaluation of uniformity of fit in various subgroups for both systems confirmed the pattern of underprediction of mortality from both models and the better performance of APACHE II over our data sample.
Conclusions: APACHE II and SAPS II failed to predict mortality in a population sample other than the one used for their development. APACHE II performed better than SAPS II. Validation in such a population is essential. Because there is a great variation in clinical and other patient characteristics among ICUs, it is doubtful that one system can be validated in all types of populations to be used for comparisons among different ICUs.