Construction And Application Of Comprehensive ICU Patient Unplanned Readmisson Risk Prediction Nomogram

PurposeTo investigat the actual re-entry situation and related risk factors characteristics of comprehensive ICU patients in a tertiary first-class hospital after being transferred from the ICU,the independent risk factors were analyzed and the risk prediction model for unplanned readmission of ICU patients was constructed and its predictive effect was evaluated.Finally the risk prediction model was draw into a nomogram that was convenient to clinical use,providing clinical ICU nurses with an effective assessment tool to screen and comprehensive ICU high-risk readmission risk patients.in order to prevent and control the unplanned readmission of comprehensive ICU patients.MethodConvenience sampling was used to retrospectively investigate the relevant data of patients discharged from the comprehensive ICU of a tertiary A hospital in Zhenjiang from January 1,2015 to December 31,2019.The ICU admissions department registration form was registered by the hospital ICU office nurse to screen returning patients and non-returning patients.And then the designed data questionnaire was used to analyze the actual readmission situation of the comprehensive ICU patients.Among the all patient data included in the study,the patient data from January 2015 to December 2018 was used as a modeling data set,and the patient data from January to December 2019 was used as an independent verification data set.Firstly,according to whether unplanned readmission occurs,the comprehensive ICU outpatients were divided into readmission group and non-readmission group.Through univariate analysis,the risk factors for unplanned readmission of comprehensive ICU patients are initially screened.The risk factors with ststistically significant differences were included in the analysis of the binary logistic regression model,and the independent risk factors for unplanned readmission of ICU patients were obtained,and the risk prediction model was constructed to convert the obtained risk prediction formula into a visual nomogram.Finally,the modeling data and independent data were used for internal and external verification to evaluate the prediction effect of the model.Based on the independent risk factors of unplanned return of patients in comprehensive ICU,a reference nursing intervention plan was developed through expert group training and group discussion.ResultA total of 1042 patients who were transferred from the comprehensive ICU department and transferred to the general ward were included.55 patients returned to the ICU during the same hospitalization period.There were 37 cases in the modeling group and 18 cases in the independent verification group.The comprehensive ICU unplanned return rate was5.28%.Respiratory failure was the most common reason for patients returning to the comprehensive ICU(19 cases,34.55%).The results of binary logistic regression showed that the unplanned return of comprehensive ICU patients was related to 4 independent risk factors:age(OR=1.029,95%CI:1.001-1.057),the admission diagnosis of respiratory diseases(OR=5.625,95%CI:2.353-13.450),secondary intubation(OR=12.290,95%CI:2.777-54.389)and SOFA score(OR=1.368,95%CI:1.212-1.544).According to the partial regression coefficient value of each risk factor obtained by binary logistic regression analysis,a comprehensive ICU patient unplanned return risk prediction model is obtained,Z=-8.615+0.028×age+0.313×SOFA score+diagnosis of admission[1.727×Respiratory system(0,1)+0.037×trauma(0,1)-0.983×tumor(0,1)]+2.509×secondary intubation(0,1).The internal verification of the modeling data shows that H.Lc~2=10.493,P=0.232(P>0.05),the area under the ROC curve AUC=0.884,(95%CI:0.845-0.923,P<0.05).The maximum value of the Youden Index is 0.658,and the corresponding risk probability P=0.033.The accuracy of the calculation model for the actual return situation of the patients in the modeling group was95.2%,the sensitivity was 91.89%,and the specificity was 73.93%.Externally verified by independent data,H.Lc~2=0.118,P=0.731,area under the ROC curve AUC=0.923,model prediction accuracy=85.02%,model sensitivity=88.89%,model specificity=84.78%.The nomogram score corresponding to the model cutoff value is 93 points,that is,when the patient’s nomogram score is greater than 93 points,the patient is identified as a high-risk returning patient by the model.ConclusionUnplanned readmission may occur after general ICU patients are transferred to the general ward,and the most common cause of return is respiratory failure.Old age,high SOFA score,secondary intubation,and a diagnosis of respiratory system disease are independent risk factors for unplanned readmission of comprehensive ICU patients.By constructing a predictive model for the risk factors of unplanned return to the comprehensive ICU,and verified by internal and external data,the model has a good predictive effect,has a certain promotion and application type,and can provides scientific basis to prevent and control comprehensive ICU patients from returning after being transferred.The comprehensive ICU unplanned return risk prediction model was transformed into a nomogram,which was more convenient for clinical medical staff to use.