The Research On Risk Assessment Method Of Elderly Depression

Depression is a common mental disorder among the older adults and it will be beneficial to build a practical depression classifier. In this study, the prevalence of depression in the elderly is statistical analysis and data mining methods are used to identify potential risk factors, we also seeks to augment simple objective screening questions to subjective screening questions and to build a practical risk score model of later life depression for screening high risk individuals.The main contributions of this dissertation are listed below:Currently the elderly depression risk assessment primarily focus on problems subjective variables, which would cause low sensitivity and specificity, and the data collection is difficulty, which lead to the inconvenience for implementation. Aiming at the above problem, a practical elderly depression risk assessment method with a high sensitivity and specificity was presented. This method is based on analysis of risk factors, and select 5 properties as a risk factor, elderly depression risk assessment model to build simple method based on multivariate logistic regression. This method is not only convenient data acquisition, and higher sensitivity and specificity of the test at the Beijing Hospital elderly health data, respectively, 89.4% and 79.5%. The method is simple and good to easy complete.Univariate analysis method for the elderly depression risk factors, ignoring the impact of the interaction between factors, can not reflect multiple factors working together on the characteristics of depression. To address the problem, multivariate analysis methods are combined with data mining. Here we proposed a multi-factor quantitative analysis of highrisk group in the degree. Firstly, the various risk factors and prevalence of depression in the elderly high-risk is computed with linear correlation analysis, and getting the correlation and ordination. Then the properties importance are analyzed by attribute importance Estimate coefficient in logistic regression methods and IncMSE index in random forest method based on the previous major N variables of linear correlation analysis. Finally, selecting risk factors for elderly depression with cross-union fusion method using univariate and multivariate analysis result. As research results shown, factors whose IncMSE values greater than 30 or more including cognitive function, stand up sit, emotional state and other factors, factors including score rating, cook, walk half a mile and other factors. Meanwhile, the study found that risk factors for olderly depression can reflect mood, loss of energy, the physical activity, the attitude of their own health, the level of cognitive factors. The results show that this method can be quantitatively analyzed risk of depression under the action of multiple factors; and depression in the elderly is combined effect by many factors, including not only the emotional factors, but also physical health and other factors.Analysis prevalence of depression caused by the characteristics of not careful interplay between the factors and effects of the problem and the law less clear, detailed analysis of the prevalence of depression in the elderly in China is characterized by high-risk groups. In this study, statistical methods, gender, age, educational level, self-rated health, cognitive function, place of residence under a single factor in the prevalence of depression in high-risk characteristics, as well as a number of factors intersect at high risk of illness is characterized by depression, for detailed statistical and analytical results. Statistics show that men, high cultural level, self-rated health, high cognitive ability, living in suburban group performance for lower detection rate of depression in patients with high-risk disease. There is interaction between the various factors to varying degrees, including age, health self-assessment, cognitive function does not affect women at high risk of depression than men characteristics; improve sex education level does not affect the characteristics of depression and lower; age, sex, culture With the cognitive level does not affect the improvement of the detection rate of decline is characterized by depression; junior high school or graduate level and cultural level, the rate of depression in high-risk populations disobedience peri-urban areas of gender characteristics of women than men overall.