Construction And Validation Of A Logistic Regression-based Model For Predicting The Risk Of Cardiac Arrest In Emergency Rooms

Objective:Sudden cardiac arrest is an important issue threatening human life and health worldwide and is the leading cause of death.Therefore develop and validate clinical prediction models for emergency room cardiac arrest patients.Method:Patients who were in the emergency department of the First Affiliated Hospital of Xinjiang Medical University from January 1,2020,to July 31,2021,and suffered cardiac arrest were retrospectively included in this study.General information,clinical signs,clinical symptoms,and laboratory tests were collected from the patients,and the outcoming patients were observed to have cardiac arrest within 24 hours.Patients were divided into modeling and validation groups according to a 7:3 ratio.To avoid the problem of collinearity,LASSO regression was used to screen variables within the in-progress model and incorporated into a multifactorial logistic regression analysis to construct a predictive model for the occurrence of cardiac arrest in emergency patients.The area under the Receiver Operating Characteristic(ROC)curve(AUC),calibration curve,and decision curve(DCA)were used to assess the value of the predictive model.Results:A total of784 emergency room patients were included in the study.Ten variables were finally screened using LASSO regression and Logistic regression,and a prediction model for the risk of developing cardiac arrest was constructed(Nomogram plot):-4.503+2.159×Mews score+2.095×chest pain+1.670×abdominal pain+2.021×hematemesis+2.015×extremities wet cooling+5.521×tracheal intubation+0.388×lactate-0.100×albumin+0.768×K~++0.001×D-dimer.The model showed good calibration and discrimination,with an area under the ROC curve(AUC)of 0.984(95%CI,0.976～0.993)in the modeling group and an AUC of 0.972(95%CI,0.951～0.993)in the validation group.The clinical prediction model has good calibration and higher clinical benefit.Conclusion:To construct a predictive model for the risk of cardiac arrest occurring in the emergency room based on Mews score,chest pain,abdominal pain,hematemesis,extremities wet cooling,tracheal intubation,lactate(venous blood),albumin,K~+,and D-dimer.It is used for early warning and timely adjustment of treatment strategies for individualized treatment of patients with cardiac arrest in the emergency room.