医患关系的微观大数据分析：患方医暴与药价改革如何影响医生行为及患者福利?

In contemporary societies, the doctor-patient interaction has been increasingly eroded by tension and violence. In China, more than 60% of the doctors have experienced on-the-job violence from patients and their families. The problem has drawn wide attention from the media, which may in turn motivate defensive medicine and as a result further worsen their relationship with the patients. Therefore, breaking this potential vicious circle is crucial for establishing a sustainable healthcare system that provides accessible, affordable, and high-quality health services. In this study, we aim to achieve three goals using a combination of economic methods and ‘big data’ technologies. The first is to examine the supply side response to the tension between the medical professionals and the patients. Particularly, we utilize a unique and detailed administrative dataset consisting of approximately 100,000 inpatients admitted between 2012 and 2016 to examine whether doctors take refuge in defensive medicine and identify the economic loss associated with it. The second goal is to evaluate the effectiveness of Drug Pricing Reform in correcting the distorted incentives of the doctors and aligning their interest with the patients’. Theoretically, we establish a multi-task principal-agent model to predict the effects of the reform on the incentives and the behavior of the doctors and the expenditures and the health outcomes of the patients. Empirically, we estimate the causal effects of the media coverage using a difference-in-differences method and a big data technique called deep learning. Findings from this study would provide rich policy implications to foster safe, affordable, and high-quality health care.

医患关系紧张是我国医疗卫生服务事业发展的严重阻力，给医生、患者和管理者都带来了前所未有的挑战。如何应对医患关系问题及其背后的信任危机是紧迫的时代命题，也是促进医生积极履行职业责任、患者获取高质量服务、政府有效控制费用增长的关键。本项目综合运用经典的经济学方法和前沿的大数据方法，基于中国公立医院独特的微观面板数据，分析和检验以患方医暴为代表的社会环境和药价改革为代表的经济环境对医疗系统的影响，旨在探究医生行为与其所受激励的关系，充分理解医患关系的症结和改善方向的可行性。在精准估计患方医暴对医生诊疗行为及患者福利影响的基础上，从理论和实证两方面探究以药价改革为代表的新一轮医改能在多大程度上提供正确的激励机制。本项目把对政策效果的分析落实到医疗服务质量和患者福利的评估上，为改善我国医患关系和推进大数据在健康医疗行业治理中的应用提供理论支撑和路径选择，为发展适合中国国情的医疗卫生服务体系建言献策。

在本项目基金的支持下，截止2020年12月，课题组已在国家自然科学基金委认定的国际权威期刊和会议上发表论文5篇。发表杂志为International Journal of Industrial Organization，IJIO（SSCI,国际A-类，1篇），会议为International Conference on Medical Image Computing and Computer-Assisted Intervention，MICCAI（人工智能领域顶级会议，1篇）和 International Symposium on Biomedical Imaging (ISBI) (医疗领域顶级会议，3篇)。在学术会议方面，项目组负责人受邀参加国际大型学术会议并做报告3次，在国内外一流高校汇报研究成果并进行深度学术研讨4次。在成果转化及科研获奖方面，负责人基于本项目延伸的政策建言发表在《光明日报内参》上，项目组成员论文获得ISBI最佳论文，并被编入计算机科学讲义丛书，取得了较好的学术和社会影响。综上，项目组按照计划完成了项目组预期目标，并在论文发表、学术会议、科研成果转化等方面取得了较好的成果。.项目的推进始终围绕如下目标：通过产业组织理论的视角，结合医疗机构独特的微观面板数据，引用先进的数据科学方法，分析和检验市场经济激励与政府管制对医疗系统的影响，为实现高质量高效率的医疗服务市场寻求正确的改革方向。具体而言（1）经济理论基础：以博弈模型为理论基础，对原研药和仿制药之间的竞争与创新决策进行动态建模，探究了如何平衡药品价格竞争为患者带来的短期福利与加强知识产权保护以鼓励创新提高患者的长期福利。该研究为回答以药价改革为代表的新一轮医改能在多大程度上提供正确的激励机制这个问题上，提供了必要的前期理论基础。相关成果发表在经管类国际权威期刊IJIO上（2）前沿方法积累：项目组重点开展了面向医疗数据的自适应深度神经网络研究，相关成果4篇分别发表于医疗和人工智能领域的顶级会议上。（3）核心问题探索：精确测量了医患关系不和谐的经济后果，梳理了经济后果产生的底层机制，为构建健康和谐的医疗服务市场生态提供了思路，也为其所能产生的巨大价值提供了科学的估算依据，主持人作为第一作者及通讯作者的相关工作论文已返修回国际权威期刊Health Economics。