基于数据挖掘技术的一次性医用材料费用控制研究

Since the excessive growth in medical expense is a common problem for most of countries all over the world, there has been increasing attention to the expense control nowadays. However, quantitative study on mechanism of growth in medical expense, strategy of solutions, as well as appropriate model under comprehensive and dynamical circumstances has not been well stated previously. In addition, recent research suggested that both usage and expense of medical consumables accounted to the most part in all of the medical items; it has been an urgent and key issue to control the rapid growth in medical consumables expense. Therefore, based on previous work, we are going to start with expense control of medical consumables in order to find solutions to solve the tough problem. In the current study, data will be collected through massive data sampling, cross-sectional observation, as well as survey follow-up. And then the technique of data mining will be conducted to clarify 2 questions. On the one hand, mechanism of growth in medical consumables expense will be elucidated comprehensively and systematically, using Structural Equation Model which concerns the suppliers, customers, service side and health insurance institutions. On the other hand, several data mining techniques will be proceeded to explore possible optimization strategies on expense control of medical consumables. The cost-effectiveness will be estimated by Decision tree-Markov mixed models. And the simulation of the policy changes and evaluation of the intervention effect will be processed using Monte Carlo models. Moreover, Association Rules, Decision Tree and Neural Network model will be integrated to develop a forecast and early warning system for expenditure of medical consumables. In conclusion, the project will provide not only some potential tools for quantitative analysis and simulation in medical expense control, but also the scientific references to prevent the excessive growth in medical expenditure.

医疗费用的过快增长是当前世界各国面临的共同难题，费用控制已成为研究热点，但在整体视阈、动态环境下，费用增长机制与控制策略的定量研究及其适用方法鲜有报道。最新研究显示，一次性医用材料的使用量和费用在所有医疗项目中增长速度最快。因此，我们在前期工作基础上，以一次性医用材料为切入点，抛砖引玉，通过海量医疗卫生数据抽样、现场调查、人群随访等方式搜集数据，结合数据挖掘技术开展两方面研究：其一，从医用材料的供方、需方、服务方、医保方等多环节入手，采用结构方程模型全面、系统的阐释医用材料费用增长机制；其二，构建决策树-Markov混合模型分析高值医用材料的成本-效果，通过蒙特卡洛技术模拟政策变化并评价其仿真干预效果，基于关联规则、决策树和神经网络技术开发医用材料费用预警系统，探寻一次性医用材料费用控制的优化策略。本研究可为医疗费用控制研究提供量化分析模拟工具，也为遏制医疗费用过快增长提供科学依据。

目前我国正处于医疗保障制度改革的关键时期，而医改的核心问题都可以归结为医疗费用的控制和约束问题。最新研究显示，一次性医用材料的使用量和费用在所有医疗项目中增长速度最快。.本项目以一次性医用材料为切入点，通过医疗卫生数据抽样、现场调查等方式搜集数据，结合数据挖掘技术开展两方面研究：（1）一次性医用材料费用过快增长的形成机制研究；（2）一次性医用材料费用控制的优化策略研究。.从天津银海2003-2011年住院参保人群资料库中随机抽取30%比例的住院参保人群开展研究。结果显示：医用材料费在医疗费用中的比例及其位次在逐年上升。医院级别、病种、年度对一次性医用材料的发生费用影响相对较大。三级医院中高值一次性医用材料的使用增长迅速，冠心病等循环系统疾病、骨科疾病等病种一次性医用材料费用比重较大。材料费的结构变动贡献率最大，是引起住院医疗费用结构变动的主要费用项目。应从住院医疗费用的内、外部影响因素入手，积极探索控制医疗费用的有效方法，提高医护人员医疗服务价格和技术附加值，缩短无效住院日，积极探索适合我国国情的DRGs等。.依据天津市2010-2015年天津市骨科疾病住院病人数据库，使用蒙特卡洛模拟，在起付标准、报销比例以及起付线与报销比例组合变化下进行研讨。不仅要加大对三级医院的报销比例，更要在政策上，对于病人去社区医院进行积极引导。.对天津市心脏支架适应症患者中进行为期1年的随访调查。结果显示：高值/一般支架两组的心血管事件发生情况的差异无统计学意义，两组在躯体活动受限程度、心绞痛稳定状态、心绞痛发作情况、疾病认知程度四个维度的差别均无统计学意义，而在治疗满意程度方面的差别有统计学意义，且支架费占比高组SAQ评分较高。.基于数据挖掘技术开发医用材料费用预警系统，探寻一次性医用材料费用控制的优化策略，为医疗费用控制研究提供量化分析模拟工具，也为遏制医疗费用过快增长提供科学依据。.