Unsupervised learning and discriminant analysis applied to identification of high risk postoperative cardiac patients.

A set of 200 patients in the 6 hours immediately following cardiac surgery was analysed within a multidimensional space of 13 commonly monitored physiological variables in order to identify high risk patterns. The application of an unsupervised learning (clustering) method to these data clearly showed the existence of two well-separated classes of low and high risk patients. A stepwise discriminant analysis was then applied to patients representative of the two classes in order to find those variables which, over time, possessed the greatest separation power. The latter always included the oxygen delivery (DO2), an index related to the oxygen content in the blood (Pv(-)O2 or avO2D) and a myocardial contractility index (VF or LAP).

[1]  W. Shoemaker,et al.  The severity index score as a method for data reduction in the surgical ICU. , 1983, Computers and biomedical research, an international journal.

[2]  Keinosuke Fukunaga,et al.  Introduction to Statistical Pattern Recognition , 1972 .

[3]  W. Knaus,et al.  A COMPARISON OF INTENSIVE CARE IN THE U.S.A. AND FRANCE , 1982, The Lancet.

[4]  E. Forgy,et al.  Cluster analysis of multivariate data : efficiency versus interpretability of classifications , 1965 .

[5]  W. Shoemaker,et al.  Evaluation of the biologic importance of various hemodynamic and oxygen transport variables: Which variables should be monitored in postoperative shock? , 1979, Critical care medicine.

[6]  O. Chiara,et al.  Hypermetabolic response after hypothermic cardiopulmonary bypass. , 1987, Critical care medicine.

[7]  W. Shoemaker,et al.  Therapy of critically ill postoperative patients based on outcome prediction and prospective clinical trials. , 1985, The Surgical clinics of North America.

[8]  David J. Hand,et al.  Discrimination and Classification , 1982 .

[9]  D. F. Morrison,et al.  Multivariate Statistical Methods , 1968 .

[10]  John,et al.  Blood gas calculator. , 1966, Journal of applied physiology.

[11]  J. Siegel,et al.  Abnormal Vascular Tone, Defective Oxygen Transport And Myocardial Failure In Human Septic Shock , 1967, Annals of surgery.

[12]  T. Abe.  Influence of cardiac surgery using cardio-pulmonary bypass on metabolic regulation. , 1974, Japanese circulation journal.

[13]  F. Gerbode,et al.  Oxygen Consumption after Open Heart Surgery Measured By a Digital Computer System , 1970, Annals of surgery.

[14]  Michael R. Anderberg,et al.  Cluster Analysis for Applications , 1973 .

[15]  W. Knaus,et al.  APACHE II: a severity of disease classification system. , 1985 .

[16]  Brian Everitt,et al.  Cluster analysis , 1974 .

[17]  W. Shoemaker,et al.  Cardiorespiratory monitoring in postoperative patients: I. Prediction of outcome and severity of illness. , 1979, Critical care medicine.

[18]  W. Shoemaker,et al.  Probability of survival as a prognostic and severity of illness score in critically ill surgical patients , 1985, Critical care medicine.

[19]  O. Chiara,et al.  Hemodynamic and Metabolic Response After Cardiopulmonary Bypass: Continuous Monitoring of CO2 Production * , 1988 .

[20]  W. Shoemaker,et al.  Cardiorespiratory monitoring in postoperative patients: II. Quantitative therapeutic indices as guides to therapy. , 1979, Critical care medicine.

[21]  Keinosuke Fukunaga,et al.  Application of the Karhunen-Loève Expansion to Feature Selection and Ordering , 1970, IEEE Trans. Computers.

[22]  W. Knaus,et al.  A COMPARISON OF INTENSIVE CARE IN THE U.S.A. AND FRANCE , 1982, The Lancet.

[23]  Thomas Marill,et al.  On the effectiveness of receptors in recognition systems , 1963, IEEE Trans. Inf. Theory.

[24]  W. Shoemaker.  Circulatory mechanisms of shock and their mediators. , 1987, Critical care medicine.

[25]  Brian Everitt,et al.  Graphical Techniques for Multivariate Data. , 1978 .