ARIMA模型与GM(1,1)模型下我国公立医院人均住院费用和药占比的预测分析

Prediction and Analysis of Per Capita Hospitalization Expense and Drug Proportion in Public Hospitals in China Based on ARIMA Model and GM(1,1)Model

目的:分析我国2002-2018年公立医院人均住院费用和药占比的变化趋势并进行预测,为公立医院价格综合改革提供依据。方法:采用EVIEWS 10.0和MATLAB 2018a软件分别建立ARIMA模型和GM(1,1)模型,预测我国2019-2025年人均住院费用和药占比的变化趋势。结果:采用ARIMA(1,1,1)模型和GM(1,1)模型对人均住院费用预测的平均绝对误差分别为64.544、145.563,平均相对误差分别为0.010、0.022;采用ARIMA(1,2,0)模型和GM(1,1)模型对人均药占比预测的平均绝对误差分别为0.889、2.894,平均相对误差分别为0.023、0.072,因此本研究中ARIMA模型的预测误差小。2025年患者人均住院费用预计为12315元,药占比预计达到10.6%。结论:2002-2025年我国患者人均住院费用逐年增加,增速减缓;药占比逐年下降。今后应继续推行医药卫生体制改革和药品组合政策,控制不合理用药,减轻患者疾病经济负担。

Objective:To analyze the changing trend of the per capita hospitalization expenses and the proportion of drugs in public hospitals from 2002 to 2018,and to make a prediction,so as to provide a basis for the comprehensive price reform of public hospitals.Methods:EVIEWS 10.0 and MATLAB 2018a software were used to establish ARIMA model and GM(1,1)model to predict the changing trend of the per capita hospitalization expenses and drug proportion in China from 2019 to 2025.Results:The average absolute errors of the ARIMA(1,1,1)model and GM(1,1)model were 64.544 and 145.563,respectively.The average relative errors were 0.010 and 0.022,respectively.The average absolute errors of ARIMA(1,2,0)model and GM(1,1)model were 0.889 and 2.894,respectively.The mean relative errors were 0.023 and 0.072,respectively,and the error of ARIMA model was small.In 2025,the average hospitalization expense of patients is estimated to be 12315 yuan,and the proportion of drugs is estimated to be 10.6%.Conclusion:From 2002 to 2025,the per patient hospitalization expenses in China increased year by year,and the growth rate slowed down.The proportion of drugs decreased year by year.In the future,we should continue to carry out the reform of the medical and health system and the policy of drug combination,control irrational drug use and reduce the economic burden of disease on patients.