Comparison of Logistic Regression and Discriminant Analyses for Predicting Survival in Patients With Severe Trauma

The objective of this study was to evaluate the precision of survival predictions for severely injured patients by comparing logistic regression and discriminant analyses. Survival was defined as in-hospital survival, from the initiation of treatment to discharge from the hospital. The study utilized group classification accuracy as a measure to evaluate the efficacy of each method. Data analysis was conducted using severely traumatized patients recruited from the trauma center of Maharaj Nakorn Chiang Mai Hospital. The findings indicated that logistic regression analysis revealed a significant link between survival in severely injured patients and hospital arrival duration, respiratory rate, and blood lactate levels. This model achieved 95.0% accuracy in overall group classification. In contrast, discriminant analysis shows that blood pressure, respiratory rate, Glasgow Coma Scale, and blood lactate levels were significantly associated with trauma patient survival. The discriminant analysis demonstrated an overall classification accuracy of 93.0%.