Construction Of Cardiovascular Disease Risk Prediction Tool For Population In Southern Xinjiang

Objective: Based on the Framingham prediction model,the cardiovascular disease(CVD)risk prediction model for the population in southern Xinjiang was established through calibration and reestimation.Calculate the risk stratification cutoff value of newly included variables in this population.Based on the newly established model,analyze the CVD risk stratification threshold of the population,draw a risk matrix,and build an online risk calculator for the southern Xinjiang population.Methods: The COX risk regression model was used to analyze the correlation between various in-dicators to be included and the risk of CVD in the southern Xinjiang population.The selected predictive indicators will be included in the Framingham model and the model coefficients will be re-estimated.Restricted cubic spline plot was drawn to analyze the association between newly included indicators and the risk of CVD.The ROC curve was used to obtain the risk stratification cuto-ff value.Survival curve was used to determine the optimal risk stratification threshold.Finally,based on the coefficients of the optimal model,use Microsoft Excel to draw a CVD risk matrix diagram,and use JAVA Script to build an online CVD risk calculator.Results:(1)Multivariate COX risk regression analysis showed that waist circumference was a risk factor for CVD in the male population(HR waist circumference=1.013,95% CI: 1.004,1.021),while e CRF and exercise were independent protective factors for CVD risk(HR e CRF=0.852,95%CI: 0.787,0.921;HR exercise=0.592,95% CI: 0.405,0.867).The inclusion of waist circumference,estimated cardiopulmonary function,and exercise in the male model significantly improved the discrimination by 13.8%(c NRI=0.138,P=0.020)and achieved the best calibration(pseudo R2=0.345).(2)In the female population,waist circumference is a risk factor for CVD(HR waist circumference=1.010,95%CI: 1.004,1.017),while e CRF and exercise status are independent protective factors for CVD(HR e CRF=0.845,95% CI: 0.770,0.928;HR exercise=0.669,95% CI: 0.503,0.890).The inclusion of e CRF and exercise in the female model significantly improved the discrimination by 14.8%(c NRI=0.148,P=0.001)and achieved the best calibration(pseudo R2=0.324).(3)In the male population,the cutoff value of male waist circumference is 96.95 cm,and the e CRF is 11.06.The stratified cutoff value of the e CRF variable in the female population is 8.17.After evaluating the original threshold,10% and 20% were selected as the new risk stratification thresholds.(4)The male risk prediction matrix is composed of four variables: age,blood pressure,physical exercise,and waist circumference.The female risk prediction matrix replaces waist circumference with e CRF.According to the new risk threshold,the risk matrix is divided into low-risk areas(≤ 10%),medium-risk areas(10%~20%),and high-risk areas(>20%),represented by green,yellow,and red,respectively.Building an online calculator based on the best risk prediction model.The risk prediction calculator includes: age,total cholesterol,high-density lipoprotein cholesterol,systolic blood pressure,smoking,diabetes,and other traditional variables.Three variables,waist circumference,estimated cardiopulmonary function,and physical exercise,were added to the male model.Two variables,estimated cardiopulmonary function,and exercise status,were added to the female population.Conclusion: Based on the Framingham model,the CVD prediction model achieved the best discrimination and calibration when the male population included waist circumference(optimal cutoff value: 96.95cm),physical exercise,and e CRF(optimal cutoff value: 11.06);The inclusion of e CRF(optimal cutoff value:8.17)and physical exercise among female populations achieved the best discrimination and calibration.Based on the best prediction model,a CVD risk prediction matrix diagram(high,medium,and low risk)was drawn,and an online risk calculator was built and applied in practical work.