Improving the predictability of ICU illness severity scales

This study demonstrates how to improve the predictability of one of the commonly used ICUs severity of illness scales, namely APACHE II, by using the consistency-driven pairwise comparisons (CDPC) method. From a conceptual view, there is little doubt that all items have exactly equal importance or contribution to predicting mortality risk of patients admitted to ICUs. Computing new weights for all individual items is a considerable step forward since it is based on reasonable to assume that not all individual items have equal contribution in predicting mortality risk. The received predictability improvement is 1.6% (from 70.9% to 72.5%) and the standard error decreased from 0.046 to 0.045. This must be taken as an indication of the right way to go.

[1]  B. Jennett,et al.  Assessment of coma and impaired consciousness. A practical scale. , 1974, Lancet.

[2]  R. Tadeusiewicz,et al.  Pairwise Comparisons and Visual Perceptions of Equal Area Polygons , 2009, Perceptual and motor skills.

[3]  W. W. Koczkodaj.  A new definition of consistency of pairwise comparisons , 1993 .

[4]  W. W. Koczkodaj.  Testing the accuracy enhancement of pairwise comparisons by a Monte Carlo experiment , 1998 .

[5]  J. Fischer,et al.  The history of the Glasgow Coma Scale: implications for practice. , 2001, Critical care nursing quarterly.

[6]  Waldemar W. Koczkodaj,et al.  About Business Decision Making by A Consistency-Driven Pairwise Comparisons Method , 2009 .

[7]  T. L. Saaty.  A Scaling Method for Priorities in Hierarchical Structures , 1977 .

[8]  Tom Fawcett,et al.  An introduction to ROC analysis , 2006, Pattern Recognit. Lett..

[9]  Yaseen Arabi,et al.  Assessment of performance of four mortality prediction systems in a Saudi Arabian intensive care unit , 2002, Critical care.

[10]  Yaseen Arabi,et al.  External validation of a modified model of Acute Physiology and Chronic Health Evaluation (APACHE) II for orthotopic liver transplant patients , 2002, Critical care.

[11]  J. L. Gall,et al.  APACHE II--a severity of disease classification system. , 1986, Critical care medicine.

[12]  Ryszard Janicki,et al.  On Consistency in Pairwise Comparisons Based Numerical and Non-Numerical Ranking , 2010, FCS.

[13]  R M Poses,et al.  Prediction of survival of critically ill patients by admission comorbidity. , 1996, Journal of clinical epidemiology.

[14]  Waldemar W. Koczkodaj,et al.  On distance-based inconsistency reduction algorithms for pairwise comparisons , 2010, Log. J. IGPL.

[15]  W. W. Koczkodaj.  Statistically Accurate Evidence of Improved Error Rate by Pairwise Comparisons , 1996 .