Tendency And Prediction Study Of Total Expenditure On Health In China Based On Principal Component Analysis And Neural Network Model

BackgroundTotal expenditure on health is defined as the total resource in money form which is consumed by the health care of a country or region from the whole society in certain period of time. The financing level of total expenditure on health represents the degree of attention and investment which a country puts to health service. The financing proportion of it illustrates the proportion which the government, society and individual share and also the level of sustainable development of health service. And the distributions of total expenditure on health between cities and villages and among different institutions are related with the efficiency and equity of health resource allocation. Recently, the total expenditure on health increases year by year. The reason can be explained as that its variation is influenced by the development of society economy, the change of demographic structure, the application of new techniques of medicine, the transformation of disease spectrum and so on. All of these influencing factors which come from the exterior of health system, the suppliers and demanders of health service interact with each other, which bring huge difficulties to predict the tendency of total expenditure on health accurately. After literature review, we found that the majority of former studies on the tendency and prediction of total expenditure on health mainly chose historical data to perform single variable time series prediction, or proceeded regression prediction by using the specific data of influencing factors of some certain years. Till now, there are still no multiple variables time series prediction researches which combine total expenditure on health and its multiple influencing factors mentioned above. We adopt neural network model combining with principal component analysis to study and predict the tendency of total expenditure on health in China. Not only can the model retain the characteristics of these multiple influencing factors and eliminate the interactions of them, but also may reveal the historical characteristics of total expenditure on health. Therefore, we can achieve the dual goals that predicting the tendency accurately and enhancing the accuracy of prediction.ObjectivesThe research adopted neural network model combining principal component analysis to study and predict the tendency of total expenditure on health in China by reviewing the financing level, financing proportion and distribution of total expenditure on health in China from1978to2011. Thus, we may provide methodology guidance for the study of total expenditure on health and provide suggestions for making macroscopic health economic policies.Materials and MethodsThe research adopted neural network model combining principal component analysis to study and predict the tendency of total expenditure on health in China. First, we conducted description analysis by selecting relevant indexes from the aspects of financing level, financing proportion and distribution of total expenditure on health. Then we sought principal components from the data that revealed society economy, the demographic structure, the lives of residents, and the supply and utilization of health service. And last, we combined the time series data of total expenditure on health with the influencing factor model output to establish the prediction model of total expenditure on health which was based on support vector machine neural network model to predict the total expenditure on health in the next short period of time. The data is from2012China Health Statistical Yearbook, China Statistical Yearbook2012and2012Statistical Communique on the2012National Economic and Social Development, etc. We used Excel2007, SPSS17.0, SAS9.0and Matlab7.0to analyze the data.Main Results1) From1978to2011, the total expenditure on health and per capita of total expenditure on health in China approximately increased exponentially. The peak of annual rate of increase of total expenditure on health showed up in2009, which was21.35%. The annual rate of increase in recent3years was13.61%on average, which was higher than both in recent5years (13.30%) and in recent10years (11.88%). The proportion of total expenditure on health in GDP is5.15%, which is highest from1978but still low when compared with developed countries.2) For the past few years, the health expenditure from government rose up gradually, social health expenditure became stable and the out-of-pocket health expenditure decreased yearly. The proportion of government health expenditure in total expenditure on health is30.4%in2011and the proportion of public financing in total expenditure on health is55.89%and2.88%in GDP. The proportion of social health expenditure in total expenditure on health is34.7%and the annual rate of increase in recent3years was22.16%,9.60%and8.58%. The proportion of out-of-pocket health expenditure in total expenditure on health is34.9%in2011.3) From1990to2011, the proportion of urbanization in China went up by94.1%, and per capita of city health expenditure multiplied16times compared with21times for villages. The per capita of city health expenditure in2011(Y2,695.1) is3.09times of that of village(￥871.6). The per capita of city health expenditure is1.50times of that of the average of China, but the per capita of village is just48%of that of the average of China.4) From2000to2011, the percentage of hospital expenditure of total expenditure on health in China rose from64.90%to its peak67.74%in2003, and then went slowly down to61.98%in2011. At the same time, the percentage of public health institutes expenditure of total expenditure on health rose from5.07%to7.69%in2003, and8.09%in2011. The proportion of city and county level hospital in hospital expenditure are64.12%and21.28%while primary health service institutes expenditure is just14.49%.5) As the result of neural network model showed, the proportion of total expenditure on health in GDP in2012would decrease from5.15%of2011to4.86%, it would rebound up to4.91%in2013. From2012to2016, the proportion of total expenditure on health in GDP will be below5%, presenting the trend of first up and then down, and will not fluctuate greatly.Discussions and Suggestions1) The innovation of combining neural network model with principal component analysis to study and predict the tendency of total expenditure on health achieved the goals of predicting the tendency of total expenditure on health accurately and raising the accuracy of prediction, which is significantly meaningful to methodology of studying total expenditure on health. 2) The total expenditure on health in China increases year by year and this trend is more remarkable since new medical reform. The proportion of total expenditure on health in GDP reaches the peak historically. However, there is still a huge gap compared with developed countries. Chinese government should continue to increase the investment to health service and enhance the level of health financing.3) The financing proportion of total expenditure on health in China improves to be reasonable gradually. The health expenditure of the government increase step by step. Nevertheless, the public financing level is still low. The proportion of social health expenditure turn to be stable but the speed slows down. The proportion of out-of-pocket health expenditure decreases year by year. Next, in order to further improve financing proportion and increase public financing, the government should add more input and improve the social medical insurance level, while continuing to reduce the level of out-of-pocket health expenditure.4) The distribution of total expenditure on health in China loses the balance. Even if the expenditure on health in countryside increases fast, its level remains lower than that of cities. We should invest more health expenditure to rural areas to improve the standard of New Rural Cooperative Medical System and strengthen the health human resource and institution construction in rural China.5) In the total expenditure on health in China, medical fees cost in hospitals remain on the top for a long time, especially in those advanced comprehensive hospitals. The expenditure spent in public health institution remains unsatisfying in spite of noticeable increase. We should continue to increase investment towards public health, and consolidate and improve National Essential Public Health Services. At the same time, increase the basic medical institute investment and guide residents there for medical treatment.6) The China health service will develop stably in the next few years. The proportion of total expenditure on health in GDP would maintain below5%from2012to2016if without abrupt policy adjustment, which means that it will not ascend sharply in a short period of time.