Risk Factors Analysis Of Common Chronic Disease And Establishment Of A Risk Factors Appraisal Model

Objective:Based on the investigate of the behavioral lifestyle,health status and chronic disease status of people aged 35 and over in Shanxi Province,and the analysis of the social epidemiological characteristics of the factors influencing hypertension and diabetes;the aim of the study was to explore the relationship between the disease status and the behavioral lifestyle,then to develop risk assessment tools for hypertension and diabetes respectively.And the results could provide reasonable suggestions for the prevention and intervention of the occurrence of chronic diseases,and to provide scientific basis for the formulation and implementation of relevant policies.Method:The 3,800 residents in Jinbei,Jinzhong and Jinnan regions in Shanxi province was investigated.based on multistage sampling method and self-made questionnaire and scale the risk factors of hypertension and diabetes were analyzed by single factor and multivariate statistical analysis.After transforming the regression coefficients in the regression model.Then the risk scores of each influencing factor are determined,the risk scoring system of is established,and the risk score-probability curve is plotted,and the risk level of the actual incidence of the disease is stratified.Finally,the predictive effectiveness and calibration ability of the model were evaluated by using the receiver operating characteristic curve?ROC?and the hosmer-lemshow fit test.Results:1.Demographics of the subjects:The 3,586 inhabitants were included,The female and male participants accounted for 47.99%and 52.01%.The sex ratio was 1.08.The average age was 49.5±12.1,The prevalence rate of hypertension was 20.38%,and the prevalence of diabetes was 8.48%.2.Results of single-factor analysis:risk factors of hypertension:diabetes,hypertension family history,male,age,education,divorce or widowed,monthly income,smoking,drinking,do not take part in physical exercise,lack of sleep time,taste salty,eating meat,vegetables and fruits of residents of high blood pressure prevalence rate.risk factors of diabetes:hypertension,diabetes family history,women,age,education,divorce or widowed,monthly income,overweight,waist-hip ratio,a-type personality,smoking,passive smoking,drinking,eating meat,vegetables and fruits,like to eat milk,sweet food residents of the resident diabetes prevalence rate.3.Results of multivariate Logistic regression analysis:the risk factors of hypertension were:diabetes mellitus?or=2.013?,Family history of Hypertension?or=2.929?,male?or=3.592?,age?or=1.385?,A-type personality?or=1.920?,overweight?or=1.311?,Smoking?or=1.623?,Drinking?or=1.759?,Vegetable and fruit?or=1.469?,food bias?or=1.416?.The risk factors for diabetes were:high blood pressure?or=1.710?,Family history of Diabetes?or=1.991?,age?or=1.220?,overweight?or=1.225?,Smoking?or=1.725?,dietary bias?or=2.468?,and sweet Food?or=1.725?.4.Results of risk scores:The total score of hypertension risk scoring system was 22,and the lowest score was1,and the highest score was 5 points.Among them,diabetes?3 points?,the Family History of Hypertension?5 points?,male?4 points?,age?1 points?,type A personality?2 points?,overweight or obese?1 points?,Smoking?2 points?,drinking?2 points?,little or no vegetables?1 points?,food bias?1 points?.The total score of diabetes risk scoring system was 21,and the lowest was 1,and the highest score was 5 points.Among them,hypertension?3 points?,Diabetes Family History?4 points?,age?1 points?,overweight?1 points?,Smoking?3 points?,Eating meat?4 points?,sweet food?5 points?.5.Predictive results of disease incidence:The formula of the incidence rate prediction model was P_???=1/1+exp[-?-5.768+S_c×0.271?]:The risk prediction model of hypertension was tested.The total score of hypertension was 0²2.The minimum prediction probability value was 2.26%,and the maximum prediction probability value was 89.97%.The risk prediction model of diabetes was P_???=1/1+exp[-?-2.567=S_c×0.199?].The total score of diabetes was 0²1.The minimum prediction probability value was 7.1%,and the maximum prediction probability value was 83.4%.According to the risk score,the probability curve showed that the corresponding prediction probability of hypertension and diabetes also increased with the increasing of risk score.6.Results risk stratification:In the hypertensive prediction model,the risk score of low-risk group was 0-2,and the actual prevalence rate was 7.03%.The risk score of Moderate-risk group was 3-7,and the actual prevalence was 15.74%.The risk score of high-risk group was 8-22,and the actual prevalence was 29.98%.The results of the trend card test show that the prevalence of hypertension increases with the increase of the risk level.In the predictive model of diabetes,the risk score of low-risk group was 0-1,and the actual prevalence was 7.7%.The risk score of Moderate-risk group was 2-10,and the actual prevalence was 13.77%.The risk score of high-risk group was 11-21,and the actual prevalence was 30.88%.The results of the trend-card test showed that the actual prevalence of diabetes increased with the increase of the risk level.7.The predictive efficiency and fitting skills test of the model:To evaluate the ability and the fitting effect of the model to identify risk factors for hypertension and diabetes,the area under the ROC curve and the hosmer-lemeshow fitting test were calculated.The predicted model of hypertension was 0.714 below the ROC curve??²=4.105,P=0.735>0.05?.The area of the ROC curve of the Diabetes prediction model is 0.843??²=1.696,P=0.213>0.05?.The evaluation results show that the predictive model of hypertension and diabetes had a good ability to identify risk factors and a better fitting correction capability.Conclusion:1.Risk factors for hypertension were:history of diabetes,family history of hypertension,male,age,type A personality,overweight,smoking,drinking,little or no vegetables,eating meat.Risk Factors for diabetes were:history of Hypertension,family history of diabetes,age,overweight,smoking,eating meat,sweet food.2.The risk prediction model of hypertension was:P_???=1/1+exp[-?-5.768+S_c×0.271?].The risk prediction model of diabetes was:P_???=1/1+exp[-?-2.567=S_c×0.199?].The evaluation results show that the riskassessment model can accurately predict the disease condition,and the calculation process was simple and convenient,which provides the basis for the accurate and comprehensive treatment of hypertension and diabetes mellitus.