【作者】 李长平；

【导师】 马骏；

【作者基本信息】 天津医科大学 ， 流行病与卫生统计学， 2004， 硕士

【摘要】 目的 卫生资源配置问题一直是世界各国卫生界的热门话题，如何做好微观调整与宏观调控、使卫生事业发展与国民经济相协调，实现卫生服务的总供求基本平衡，也是我国卫生界一直致力研究的方向。本次研究旨在通过运用计量经济学模型(包括时间序列模型和经典计量经济学模型)对我国和天津市卫生资源的人、财、物三方面进行预测，并通过结构分析，进一步了解我国卫生资源发展与我国人口、国内生产总值(GDP)发展的关系，来探讨计量经济学模型在卫生资源相关研究中的应用。 方法 第一部分：收集了1980-2000年全国医院、卫生院卫生技术人员数、床位数、卫生事业经费数进行分析，首先采用时间序列模型中应用范围很广泛的求和自回归移动平均模型即ARIMA(p，d，q)模型对该人、财、物三方面进行预测。然后建立起该三方面与人口、GDP发展关系的经典计量经济学方程，利用该方程进行预测、比较静力学分析和弹性分析。 第二部分：收集了天津市1980-2000年全国医院、卫生院卫生技术人员数、床位数、卫生事业经费运用计量经济学模型分别对其进行了预测和结构分析。 并且利用2001、2002年收集的数据分别对全国和天津市卫生资源三方面进行了样本期外预测。 结果 运用ARIMA模型对全国医院、卫生院卫生技术人员数、床位数、卫生事业经费数进行预测，其平均相对误差分别为0.0053、0.0046、0.0412。利用经典计量经济学方程进行预测，平均相对误差分别为0.019、0.012、0.015。ARIMA模型对天津市医院、卫生院卫生技术人员数、床位数、卫生事业经费数进行预测，其平均相对误差分别为0.014、0.014、0.029。经典计量经济学方程预测，平均相对误差分别为0.020、0.022、0.010。天津医科大学硕士学位论文利用计量经济学方程进行卫生事业经费相对于GDP发展的弹性分析，其弹性系数分别为0.24、0.46。 对2001年、2002年全国和天津市的医院、卫生院卫生技术人员数、床位数进行样本期外预测。ARIMA模型预测2001年全国医院、卫生院卫生技术人员数、床位数相对误差分别为0.034、0.012，2002年相对误差分别为0.067、0.031。经典计量经济学模型预测2001年相对误差分别为0.013、0.029。ARIMA模型预测2001年天津市医院、卫生院卫生技术人员数、床位数相对误差分别为0.015、0.024，2002年相对误差分别为0.043、0.004。经典计量经济学模型预测2001年相对误差分别为0.094、0.001。2002年相对误差分别为0.046、0.024。结论在本次研究中利用计量经济学模型中两个有代表性的ARIMA模型和经典计量经济学方程模型对卫生资源的人、财、物三方面进行了预测，均取得了较满意的结果。并利用建立起的计量经济学方程进行了结构分析，由弹性系数表明，天津卫生事业经费相对于GDP的弹性变化高于全国总体水平。 综上所述，计量经济学模型在卫生资源的研究，尤其在卫生资源发展的预测和与社会人口、经济发展的关系研究中发挥了明显的优势。在本次研究中，时间序列方法在卫生资源研究的预测中显示了较好的精度，建立起的经典计量经济方程利用岭回归方法对卫生资源的预测也取得了较满意的结果，且通过该方程可了解我国卫生资源与社会人口、经济发展的关系，可以通过对该关系的了解，指导我国卫生资源的发展方向，为实现卫生资源与社会的协调发展提供方法上的指导。更多还原

【Abstract】 Objective Health resources allocation is the popular topic among the health circles all over the world. How to do well in the micro-adjustment and the macro-adjustment, how to coordinate between the development of health undertakings and the national economy, and how to achieve the balance between the supply and demand are the goals we are always working for. The objective of this paper is to probe into application of the econometrics models (including time series model and classic econometric model) for study of the health resources through forecasting the manpower, material and financial resources and also through analyzing structure to learn the relationship among development of the health resources, population and gross domestic product.Methods Part one: We collected health manpower of hospital, beds, and health expenditure in our country from 1980 to 2000. The steps of analysis were as follows: First, the autoregressive integrated moving average model (ARIMA), which was a very popular model of time series, forecasted the three parts of health resources. Then the classic econometric models was set up to describe the relationship among development of health resources, population and gross domestic product, and to undertake forecast, compared statics and elasticity.Part two: Health manpower of hospital, beds, and health expenditure in Tianjin from 1980 to 2000 were forecasted by econometric models and the structure analysis.Results The mean relative errors of the ARIMA model which forecasted the three parts of the health resources in our country were 0.0053, 0.0046, 0.0412 respectively. The mean relative errors of the classic econometric model were 0.019, 0.012, 0.015 respectively.The mean relative errors, which forecasted the three parts of the healthresources in Tianjin, were 0.0053, 0.0046 and 0.0412 respectively through the ARIMA model forecasting. The mean relative errors were 0.019, 0.012 and 0.015 respectively through the classic econometric model forecasting. In addition, elasticity of public health expenditure and development of GDP was 0.24 in our country and 0.46 in Tianjin.Health manpower of hospital and beds from 2001 to 2002 were forecasted out of sample size. The mean relative errors, which forecasted health manpower of hospital and beds from 2001 to 2002 in our country, were 0. 034, 0. 012 and 0. 067, 0. 031 respectively through the ARIMA model forecasting. And the mean relative errors, which forecasted the two aspects of health resources in 2001, were 0. 013, 0. 029 respectively through the classic econometric model forecasting. The mean relative errors, which forecasted health manpower of hospital and beds from 2001 to 2002 in Tianjin, were 0.015,0.024 and 0.043,0.004 respectively through the ARIMA model forecasting. And the mean relative errors, which forecasted the two aspects of health resources, were 0.094,0.001 and 0. 046, 0. 024 respectively through the classic econometric model forecasting. Conclusion In this study, we chose two representative econometric models-the ARIMA model and the classic econometric model-to forecast the three parts of health resources, and achieved satisfied results. The coefficient of elasticity indicated that the change of elasticity of Tianjin was higher than that of our country.In a word, the application of econometric models for the study of health resources showed good forecasting precision. And the time series model showed good forecasting precision .We can learn the relationship among development of health resources, population and economy through the classic econometric model. The method also directs the development of health resources and benefits thecoordination between the health resources and the development of society.更多还原