Based On The Meta-Analysis Of The Risk Of Stroke Studyby Logistic Regression Model

Cardiovascular disease has become one of the most important factors of human death; the incidence of cerebral apoplexy is as people raising the level of economic and into a rising trend year by year. In addition, stroke is the first factor leading to human disability; it is considered to be safety a major threat to human health life, the refractory disease. It happened not only bring patients physical and mental influence, but also created a heavy financial burden of the family and society. Now, temporarily without reasonable treatment measures could be cured completely in patients with stroke, so it is particularly important to reasonably prevent the happening of the disease. According to the survey, the risk factors of stroke are various, including abnormal physical indicators, bad living habits, and genetic factors, etc. Due to the different kinds of risk factors and the intensity of various factors could cause the risk of disease may also in difference, it hasn’t found a perfect prevention and control measures temporarily. Therefore, to establish the incidence of individual stroke health risk assessment model, which important to reduce the risk of incidence and provide the reasonable health management measures for prevention.There are three parts of this study: First of all, by making the exclusion criteria and retrieval strategy, searching the academic literature database and conference papers in the third military medical university library materials, obtaining the studies of stroke risk factors, extracting the data, and taking the system review and meta-analysis; In the second part is calculating the risk factors of stroke risk value in the meta-analysis and take them into the optimized Logistic regression model, established the individual stroke incidence of health risk assessment model of Asian and western. The third part is the model validation, extracting the 2,000 cases of Chongqing, Sha Pingba district investigation data and 2,200 cases of chronic diseases Framingham cohort data, to Validate two models, and the effect is used evaluated by ROC curve.The meta-analysis results of Asian: there are 41 studies have been included. In the Asian studies, there are 11 factors into the model, including TC≥4.84 mmol/L, SBP≥140 mm Hg, Smoking and drinking, long-term exercise, take salty, family history of stroke, hypertension, diabetes, hyperlipidemia and coronary heart disease, and their pooled OR are 1.45, 2.34, 1.82, 1.50, 0.55, 1.93, 3.28, 4.32, 1.96, 1.82 and 2.73.The health risk assessment model of Asian as follows:11109876543211.19X+0.60X+0.67X+1.46 X Pgit L=α+1.19X+0.66X+0.60X-0.40X+0.60X+0.85X+0.37X+)(oThe meta-analysis results of Western: there are 14 studies have been included. In the Western studies, there are 6 factors into the model, including SBP≥140 mm Hg, smoking, hypertension, diabetes, hyperlipidemia and coronary heart disease, and their pooled OR are1.29, 1.73, 1.73, 1.94, 1.09 and1.47. The health risk assessment model as follows:The health risk assessment model of Western as follows:654321α++++++=39.009.066.055.055.025.0)( XXXXXXPLogitThe results of Validation: the true incidence of AUC in Asian is 0.631(0.543-0.720),and the Predicting Outcome is 0.635(0.544-0.726), the result of difference test is Z=0.380, P=0.704, the Asian model prediction results and verify the true incidence of crowd are consistent; the true incidence of AUC in Western is 0.684(0.625-0.743),and the Predicting Outcome is 0.638(0.578-0.697), the result of difference test is Z=0.0694,P=0.945, the Western model prediction results and verify the true incidence of crowd are consistent;This study used the system evaluation and Meta-analysis, to comprehensive evaluation of the related research about stroke risk factors. On this basis, the Asian and the western health risk assessment model of individual stroke incidence is established, and obtained a good prediction result. The establishment of the model, gives the help to individual stroke incidence forecast, and provides the evidence based medical science to disease control and prevention evidence of evidence-based medicine, also provides a new train of thought and exploration direction for other chronic diseases health risk assessment model.