A Decision Analytic Approach to Coronary Care Patient Classification.
Abstract
The purpose of this project is to categorize cardiac patients using an expected utility criteria. Bayesian updating using symptomatic information has been a successful tool for disease diagnosis. However, in classification of cardiac patients, discriminant analysis using the BMD05M computer package achieves comparable results with far less effort. Discriminant scores can be evaluated and normalized as probabilistic estimates of the patient's cardiac condition. By incorporating these probabilities with subjective value estimates for possible misclassifications, the expected utility of assigning patients to each cardiac category can be calculated. Patients are then classified into cardiac categories based upon highest expected utility. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Oct 01, 1976
- Accession Number
- ADA050910
Entities
People
- Terry A.. Bresnick
Organizations
- United States Military Academy