早期重症急性胰腺炎预测模型的构建

打开文本图片集

Construction of early predictive model for severe acute pancreatitis

SU Xiaofang， WANG Qinghua*

School of Nursing of Binzhou Medical University， Shandong 256603 China

*Corresponding Author  WANG Qinghua， E⁃mail： wangqinghua@bzmc.edu.cn

Abstract  Objective：To analyze the early risk factors of severe acute pancreatitis and establish an early predictive model.Methods：Data of patients with acute pancreatitis from July 2016 to July 2022 in a grade A tertiary hospital in Shandong Province were retrospectively collected and randomly divided into a modeling group of 630 cases（483 cases of non-severe acute pancreatitis，147 cases of severe acute pancreatitis） and a validation group of 270 cases.General data，vital signs，and serological indicators within 24 hours of admission were collected by the medical record system.Univariate and binary Logistic regression were used to screen the risk factors of severe acute pancreatitis.Combined with the risk factors，Logistic prediction models and early prediction models were established and verified.Results：In the modeling group，it was found by univariate and bivariate Logistics regression that C⁃reactive protein α⁃hydroxybutyrate dehydrogenase，hematokrit（HCT），abdominal fluid and respiratory rate were independent predictors of severe acute pancreatitis.Logistic prediction model was established for independent predictors.The AUC was 0.938（95%CI 0.910⁃0.967），the sensitivity was 93.75%，and the specificity was 87.13%.The ROC curve was used to calculate the truncation value of the independent predictor，and the unweighted prediction model was constructed.The truncation value of the model was 2.000，the AUC was 0.942，the sensitivity was 90.15%，and the specificity was 87.43%.Z⁃test was performed with BISAP and APACHE Ⅱ scores，and the AUC was statistically significant （P<0.05）.Conclusions：C⁃reactive protein，α⁃hydroxybutyrate dehydrogenase，HCT，abdominal fluid and respiratory rate were independent predictors of SAP.Based on the predictors，Logistic prediction model and early prediction model were constructed.The early prediction model had higher specificity，and had higher prediction effect.

Keywords  acute pancreatitis； the degree of severity； early warning； prediction model； influencing factors

摘要  目的：分析重症急性胰腺炎早期发生的危险因素，构建预测模型。（剩余9053字）