OBJECTIVES: The aim of the study was to systematically validate the APACHE III scoring system concerning severity of illness classification and prediction of hospital mortality. Such data have not yet been determined in a large population of critically ill patients in germany. METHODS: 531 patients (ICU stay > 4 hours) were prospectively and consecutively investigated. The day-1-scores and risk-of-death predictions of APACHE III and APACHE II were determined. A comparison was performed between both scoring systems, and the correlation with the observed hospital mortality was examined. RESULTS: For both main validation criteria, as were discrimination (areas under the ROC-curves: APACHE III 0.873; APACHE II 0.859) and calibration (goodness-of-fit testings; p > 0.05), both scoring systems provided satisfying results concerning hospital mortality, no system showing a significantly superior performance. Compared to the observed hospital mortality (13.4%), the prediction of APACHE III (13.2%) was extremely accurate, whereas the prediction of APACHE II was higher (16.8%). The standard (mortality index not significantly < or > 1.0) provided by APACHE III was fulfilled, while the standard given by APACHE II was surpassed. The mean scores and the mean risk-of-death predictions for non-survivors were significantly higher compared to survivors (P < 0.001). The individual score values of both systems were found to have a strong correlation (r = 0.922). CONCLUSIONS: APACHE III (like APACHE II) provides a sufficient severity of disease classification and accurately predicts overall hospital mortality in a representatively large german population of a medical ICU. Therefore APACHE III can be regarded as validated for the use in comparable german ICUs. For use as a standard the more recently introduced APACHE III seems to be superior to the established but older APACHE II. However, each user will--depending on the particular questions to be addressed--carefully have to evaluate, if the improvement of prognostic accuracy really justifies the increased amount of workload necessary for calculating APACHE III score and risk prediction.
OBJECTIVES: The aim of the study was to systematically validate the APACHE III scoring system concerning severity of illness classification and prediction of hospital mortality. Such data have not yet been determined in a large population of critically ill patients in germany. METHODS: 531 patients (ICU stay > 4 hours) were prospectively and consecutively investigated. The day-1-scores and risk-of-death predictions of APACHE III and APACHE II were determined. A comparison was performed between both scoring systems, and the correlation with the observed hospital mortality was examined. RESULTS: For both main validation criteria, as were discrimination (areas under the ROC-curves: APACHE III 0.873; APACHE II 0.859) and calibration (goodness-of-fit testings; p > 0.05), both scoring systems provided satisfying results concerning hospital mortality, no system showing a significantly superior performance. Compared to the observed hospital mortality (13.4%), the prediction of APACHE III (13.2%) was extremely accurate, whereas the prediction of APACHE II was higher (16.8%). The standard (mortality index not significantly < or > 1.0) provided by APACHE III was fulfilled, while the standard given by APACHE II was surpassed. The mean scores and the mean risk-of-death predictions for non-survivors were significantly higher compared to survivors (P < 0.001). The individual score values of both systems were found to have a strong correlation (r = 0.922). CONCLUSIONS: APACHE III (like APACHE II) provides a sufficient severity of disease classification and accurately predicts overall hospital mortality in a representatively large german population of a medical ICU. Therefore APACHE III can be regarded as validated for the use in comparable german ICUs. For use as a standard the more recently introduced APACHE III seems to be superior to the established but older APACHE II. However, each user will--depending on the particular questions to be addressed--carefully have to evaluate, if the improvement of prognostic accuracy really justifies the increased amount of workload necessary for calculating APACHE III score and risk prediction.
Authors: P Kellner; R Prondzinsky; L Pallmann; S Siegmann; S Unverzagt; H Lemm; S Dietz; J Soukup; K Werdan; M Buerke Journal: Med Klin Intensivmed Notfmed Date: 2013-04-06 Impact factor: 0.840
Authors: Ashish Nimgaonkar; Dilip R Karnad; S Sudarshan; Lucila Ohno-Machado; Isaac Kohane Journal: Intensive Care Med Date: 2004-01-15 Impact factor: 17.440