Table of Contents Author Guidelines Submit a Manuscript
Journal of Probability and Statistics
Volume 2011 (2011), Article ID 523937, 19 pages
http://dx.doi.org/10.1155/2011/523937
Research Article

Classification and Regression Trees on Aggregate Data Modeling: An Application in Acute Myocardial Infarction

1INSERM EMI 0106, 21000 Dijon, France
2Université de Bourgogne, Service de Biostatistique et Informatique Médicale, CHU, 21000 Dijon, boulevard Jeanne d'Arc BP 77908, 21079 Dijon Cedex, France
3Department of Statistics, University of Georgia, Athens, GA 30602-1952, USA
4CEREMADE CNRS UMR 7534, Université de Paris, Dauphine 75775 Paris Cedex 16, France
5Service de Cardiologie, Centre Hospitalier du Bocage, BP 77908, 21079 Dijon Cedex, France

Received 22 October 2010; Revised 24 March 2011; Accepted 25 May 2011

Academic Editor: Peter van der Heijden

Copyright © 2011 C. Quantin et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

Cardiologists are interested in determining whether the type of hospital pathway followed by a patient is predictive of survival. The study objective was to determine whether accounting for hospital pathways in the selection of prognostic factors of one-year survival after acute myocardial infarction (AMI) provided a more informative analysis than that obtained by the use of a standard regression tree analysis (CART method). Information on AMI was collected for 1095 hospitalized patients over an 18-month period. The construction of pathways followed by patients produced symbolic-valued observations requiring a symbolic regression tree analysis. This analysis was compared with the standard CART analysis using patients as statistical units described by standard data selected TIMI score as the primary predictor variable. For the 1011 (84, resp.) patients with a lower (higher) TIMI score, the pathway variable did not appear as a diagnostic variable until the third (second) stage of the tree construction. For an ecological analysis, again TIMI score was the first predictor variable. However, in a symbolic regression tree analysis using hospital pathways as statistical units, the type of pathway followed was the key predictor variable, showing in particular that pathways involving early admission to cardiology units produced high one-year survival rates.