Construction Of A Clinical Prognosis Prediction Model For Lower Extremity Arteriosclerosis Obliterans Treated With Both Traditional Chinese And Western Medicin

Purpose:Through the analysis and mining of real-world electronic case data,the machine learning algorithm is used to build a prognostic prediction model of ASO combining Chinese and Western conservative treatment,which provides a reference scheme for the individualized and precise treatment of this disease in clinic,and the constructed model is evaluated and verified.Material and method:1.Information on ASO patients who were hospitalized from January 2014 to January 2022 was retrospectively collected through the HIS Medical record Management system.According to the TCM outcome indicators of patients at discharge,they were divided into two groups:significant group and non-significant group.2.Statistical analysis: Excel software is used to clean the original data,which mainly includes demographic information,clinical characteristics,laboratory test results and imaging examination reports.Lasso regression and Logistic regression analysis using R software,thus in a prognostic prediction model.3.Model evaluation and validation: the subject operating characteristic(ROC)curve and the ROC curve area(AUC)are used to evaluate the predictive performance of the constructed model,that is,the differentiation degree of the model(Discrimination).The calibration curve(Calibration curve)was used to evaluate the accuracy of the model prediction.Decision curve(DCA)was used to verify the availability and benefit of the model,namely the clinical utility.External validation performed through the validation set.Results:1.A total of 157 patients meeting the criteria were collected,including 115 in the significant group and 42 in the non-significant group.2.High-dimensional variables were selected for age,BMI,disease duration,gender,dorsal foot artery pulsatile abnormality,positive limb position test,resting pain,intermittent claudication,limb ulcer,smoking,dyslipidemia,cilostazol,prostadil,vascular Fukang capsule,acupuncture,TCM rubbing,and TCM fumigation.3.Univariate analysis yielded 8 variables with a p-value of <0.2: age,disease duration,sex,resting pain,dyslipidemia,cilostazol,acupuncture,and septal moxibustion.4.Multivariate analysis yielded five variables with a p-value <0.05: disease duration(OR:2.73,95%CI:0.10-0.28,p=0.006),Age(OR: 2.26,95%CI:0.03-0.09,For p=0.023),resting pain(OR: 2.44,95%CI:2.65-6.49,For p = 0.0.14),dyslipidemia(OR: 2.74,95%CI:1.38-3.80,For p=0.006),cilostazol(OR: 2.32,95%CI:0.99-2.31,p=0.019).5.The ROC curve,the calibration curve and the decision curve of the modeling group and the validation group are drawn: the AUC of the modeling group is AUC(0.94)> validation group(0.88)> 0.75,the modeling group calibration curve is very close to the diagonal line(ideal)and better than the validation group,the threshold probability is 0.75-0.98,and the threshold probability of the validation group is 0.78-0.98.Conclusion:regulating qi-flowing method1.The combination of Chinese and Western therapy can improve the symptoms of ASO patients and improve their quality of life.2.The constructed prognostic prediction model has good differentiation and high calibration,which can more accurately predict the prognosis of patients after receiving the intervention.