慢性病药品采购价格指数预测研究——基于山东省药品集中采购数据的分析  

Study on purchasing price index predicting of chronic diseases drugs——Analysis based on the data of centralized drug purchasing in Shandong Province

作　　者：蔡文鼎(翻译) 魏新江 CAI Wending;WEI Xinjiang

机构地区：[1]鲁东大学数学与统计科学学院

出　　处：《价格理论与实践》2023年第3期108-111,204,共5页Price:Theory & Practice

基　　金：国家自然科学基金(61973149);山东省自然科学基金重点项目(ZR2020KF029);2021年度省社科规划研究项目(21CSDJ20)。

摘　　要：药品集中采购对于稳定药品价格和降低患者用药负担具有重要意义。本文基于山东省药品集中采购数据,编制慢性病药品链式拉氏价格指数,评估山东省药品集中采购政策的实施效果。构建LSTM、双向LSTM和ConvLSTM模型,预测价格指数波动趋势,为政府及管理者提供科学的决策依据。研究结果表明:慢性病药品价格大幅下降,大大减轻了慢性病患者的用药负担,未来慢性病药品价格将呈现稳中有降的趋势,双向LSTM和ConvLSTM模型的预测精度优于传统LSTM模型,两模型泛化能力更好,对山东省慢性病药品价格指数预测具有一定的应用价值。据此,提出健全药品集中采购价格监测体系、创新药品价格监测技术手段和深化药品生产流通体制改革等建议,推动药品采购与监测技术研究一体化建设。Centralized drug purchasing is of great significance for stabilizing drug prices and reducing Patient's drug burden.Based on the data of centralized drug purchasing in Shandong Province,the chain Laspeyres price index of chronic disease drugs is compiled to evalu-ate the implementation effect of centralized drug purchasing policy in Shandong Province.By establishing LSTM,bidirectional LSTM and ConvLSTM models,it can predict the fluctuation trend of price index and provide scientific decision-making basis for the government and managers.The research results show that the price of chronic disease drugs in Shandong Province has dropped significantly and it will show a trend of stability and decline in the future,which has reduced the drug burden of chronic disease patients.The prediction accuracy and gen-eralization ability of the bidirectional LSTM and Conv LSTM model is better than that of the traditional LSTM model,which has certain application value for the prediction of the price index of chronic diseases in Shandong Province.In view of this,suggestions were put for-ward to improve the monitoring system of centralized drug purchase prices,innovate the technical means of drug price monitoring and deepen the reform of drug production and distribution system.Promoting the integration of drug purchasing and monitoring technology re-search.

关 键 词：慢性病药品 价格指数预测 LSTM模型 双向LSTM模型 ConvLSTM模型 

分 类 号：R95[医药卫生—药学]