A Logistic Regression Model for Survival Prediction in Rats with Fixed-Volume Hemorrhage

Hemorrhagic shock is a common cause of death in emergency rooms. Because the symptoms of hemorrhagic shock occur after shock has considerably progressed, early diagnosis of shock is difficult. The purpose of this study was to achieve early diagnosis of hemorrhagic shock using a survival prediction model in rats. We allocated 60 Sprague-Dawley rats into four groups: a no hemorrhage group, and three groups with hemorrhage at a fixed volume of 2, 2.5, and 3 mL/100 g. We analyzed heart rate, systolic and diastolic blood pressures, respiration and body temperature, and then obtained a logistic regression equation predicting survival rates between 0 for death and 1 for survival. The area under the receiver operating characteristic (ROC) curves was 0.986. The Hosmer–Lemeshow goodness-of-fit chi-square value for the evaluation of the calibration of the outcome prediction model was 2.69 with 8 degrees of freedom (P = 0.952). The optimal boundary value between survival and death was determined to be 0.37. By applying the optimal boundary, the sensitivity, specificity, and accuracy of survival prediction by the regression equation were 96.8, 95.4, and 96.0%, respectively.