Establishment And Validation Of An Early Prediction Model For Severe Acute Pancreatiti

Objective:The purpose of this study is to study the relevant risk factors that affect the severity of acute pancreatitis(AP)and establish a precise and personalized risk predictive model.Methods:The relevant information of 3421 patients with acute pancreatitis admitted to Changsha Central Hospital from January 2015 to August 2022 were collected.According to the inclusion and exclusion criteria,2986 patients who met the requirements were selected and randomly assigned to the modeling group(n=2112)and the validation group(n=874)based on the ratio of 7:3.In the modeling group,the risk factors affecting the severity of AP were screened by single factor and multiple factor binary logistic regression analysis.The risk predictive model was established in R software with these indicators,and then the predictive model was visualized with nomogram and dynamic nomogram was made.The external validation method was used to substitute the validation group data into the model to verify the accuracy of the predictive effect of the nomogram.Conclusion:This study established an intuitive risk predictive model for severe acute pancreatitis,and this model had easily obtained clinical and laboratory parameters,which could be used to predict the incidence rate of severe AP patients.Nomogram transforms complex regression equations into visual graphics,making the results of the predictive model more readable and convenient for the evaluation of patients.