Generalized Linear Model Based On The Statistical Analysis Of China’s Major Disease Mortality

Under the situation of the economic development and social progress with a quite high speed, people are concerning about their health day by day. Recently, mortality rate of Chinese urban citizen has been increasing mainly because of the chronic diseases, such as cancer, heart disease, brain vessel disease and breathe system disease, consisted the first death reason in China and surpassing the average level of the world. Meanwhile, the death rate of this two kinds of diseases are increasing in rural area in China. Otherwise, heart disease, breath system disease, poisoning and physical damage are also part of the death factor, according to Chinese Statistic Bureau. Both in theory and practice, the death rate of these diseases is usually mentioned, such that the research upon the death rate of diseases is practically significant. The death indicator is the basic indicator reflecting the health of citizen. In addition, the death level and death factor also reflect the economy and development in this society. Monitoring the death is the key method to prevent and control the diseases, which also reflects the medical situations, and the efficiency of the medical prevention. Through the summary and research on the morbidity of kinds of common diseases, it is quite useful and unique to prevent the diseases, promote the development of curing the diseases, and improve the medical level.This paper will research the death rate of urban and rural areas in China, based on the data from Chinese Statistic Yearbook including year of1990,2000, and2010, in order to figure out the influences of the years, regions, and diseases toward the death rate.In addition, this paper will implement the methodology that logistic general regression to analyze the three dimensional contingency table. The structure of this paper will be briefly introduced as following:Chapter1Introduction. This chapter will introduce the research significance, research contents, research methodology, and creation and limitation of this paper.Chapter2Literature review. This part reviews and summarizes the researching results of the literature researching the diseases death rate, through the time span, region span, disease kind and origin of diseases, pointing out the disadvantages of those methodologies and describing the research methodology of this paper.Chapter3Theoretical basis of the model. Introduce the theoretical basis of generalized linear model and logistic linear model, pointing out the reason that this paper chooses the model of three dimensional contingency table.Chapter4Data description and construction of the model. This part mainly designs the practical research, introduce the resources of the data and descriptive analysis of the data.Chapter5Empirical analysis. Choose the simplest model with a comparatively high R-square, according to the likelihood statistic test upon the data.Chapter6Conclusion. This paper empirically analyzes the main effects and interactive effects the data. The main effect shows the influences of residential situation, year and diseases type. The interactive effect shows the crossed influences of two explanatory variables toward the explained variable.Chapter7Discussion and recommendation. Based on the conclusion, this part figures out the political advices for improving the situation of the serious diseases.The conclusion including:Firstly, only the death rate of breathe system diseases has decreased in recent10years among these main diseases discussed above; the significance of differences between urban and rural areas includes the heart disease and cancer, respectively; within dynamic comparison, the significant difference is the heart disease.Secondly, the death rate of cancer, heart disease, and brain vessel disease among urban citizens is significantly higher than the rural citizens; however, the death rate breathe system disease is significantly lower. This shows the medical care and medical condition are much better among urban citizens, but the medical level could not solve the problems of serious diseases, such as cancer, heart disease and brain vessel disease.Although this paper is also limited by the data and other aspects, the research methodologies is improved as following: Firstly, the model of this paper is constructed upon the generalized linear regression and logistic linear regression, which is able to explain the internal connection among variables.Secondly, early empirical researches only consider the single factor, like the time, region or disease type, rather than the interactive influences toward the death rate. Therefore, this paper construct the three dimensional contingency table, analyzing the single effect and interactive effect toward the death rate among time, region and main disease.The limitation of this paper as following:Based on the data of the very year, this paper gets the much more comprehensive data from general population survey, keeping the quality of the initial data. However, the research data is comparatively few, lacking of continuity. At the same, this paper does not include the influences of age structure because that the age structure of China tends to be old, in order to simplify the research.