Evaluation of acute physiology and chronic health evaluation III predictions of hospital mortality in an independent database.

摘要: Objective: To assess the accuracy and validity of Acute Physiology Chronic Health Evaluation (APACHE) III hospital mortality predictions in an independent sample U.S. intensive care unit (ICU) admissions. Design: Nonrandomized, observational, cohort study. Setting: Two hundred eighty-five ICUs 161 hospitals, including 65 members Council Teaching Hospitals 64 nonteaching hospitals. Patients: A consecutive 37,668 ICU admissions during 1993 to 1996; 25,448 at hospitals with >400 beds 1,074 <200 beds. Interventions: None. Measurements Main Results: We used demographic, clinical, physiologic information recorded day 1 APACHE equation predict probability for each patient. compared observed predicted all across patient subgroups assessed predictive using tests discrimination calibration. Aggregate death rate was 12.35% 12.27% (p=.541). The model discriminated between survivors nonsurvivors well (area under receiver operating curve = 0.89). calibration showed that number deaths close by model, but when tested deciles risk, goodness-of-fit (Hosmer-Lemeshow statistic, chi-square 48.71, 8 degrees freedom, p<.0001) not perfect. Observed rates were significantly (p <.01) different 55 (84.6%) III's specific admission diagnoses 11 13 residual organ system-related categories. most frequent significant (p<.01) differences included acute myocardial infarction, drug overdose, nonoperative head trauma, multiple trauma. Conclusions: accurately aggregate Further improvements can be achieved more precise disease labeling, improved acquisition weighting neurologic abnormalities, adjustments reflect changes treatment outcomes over time, a larger national database.