Severity assessment by APACHE III system in Spain

BACKGROUND: To assess the performance of the prediction equation of the APACHE (Acute Physiology Age and Chronic Health Evaluation) Ill prognostic scoring system when applied in Spain. PATIENTS AND METHOD: Prospective multicenter cohort study that included 10786 adult patients from 86 Spanish intensive care units (ICU). Data collection during first 24 hours of admission: acute physiology score, age and comorbilties, for calculating APACHE Ill score; treatment location prior to ICU admission and main diagnosis admission category for applying the mortality prediction equation of APACHE Ill system. Main outcome was observed hospital mortality. RESULTS: Age was 57.74 (0.16); 68% males. Non-operative patients represented 76% of sample. APACHE Ill score was 53.75 (0.26);observed and predicted hospital mortality were 21.2% and 19.8% respectively, with a standardized mortality ration of 1.07. The chi (2) HosmerLemershow statistic was (H) 135.6, (C) 133.91: p < 0.001. The area under the Receiver Operating Curve (ROC) was 0.808, and correct classification at mortality risk of 50% was 82%. Uniformity of fit was better for non-operative diagnoses and for patients admitted from the emergency area. Calibration was excellent for risk lower than 60% but slightly underestimated observed risks above this level. CONCLUSIONS: The American APACHE Ill equation fit well when applied to Spanish critical patients but with limitations. Discrepancies could be attributed to differences in case-mix and variations in practice style.