Patient Classification Model of Appendicitis Severity Based on Bayesian Stepwise Discriminant Analysis

To emphasize distinctions in diagnosis and provide valuable guidance for treatment based on patient heterogeneity, it is useful to classify the various forms of appendicitis into three groups. A bayesian stepwise discriminant model was proposed to predict the patient's severity of Appendicitis. The age, BMI index, diabetes status, perforation status, other complications, WBC and hemoglobin index were discriminant factors for patient classification according to a bayesian stepwise discriminant analysis. The classification accuracies for severe appendicitis and mild appendicitis were 96% and 97% respectively. The results showed that the presented method was verified effective and practically applicable.