基于知识图谱和社交网络的在线健康社区医患服务匹配研究

In China, the demand for better and diverse medical services has been increasing in recent years. With the advance of the Internet and the popularity of the Web and mobile devices, online health communities (OHCs) have provided new channels for the search of medical and healthcare services, which facilitate not only healthcare seeking of patients for better services, but also healthcare granting of physicians to a wide range of patients. In this project, we intend to investigate both the service seeking and granting behaviors in online health communities and focus on the automatic patient-physician matching problem in these communities. To explain it in more detail, we intend to solve three closely related research questions, in order to meet the demands of medical treatment, user preference and social trust, respectively. Firstly, to meet the basic demand of medical treatment, we employ the “symptom-treatment” knowledge graph to match the symptom entities from the patient content and treatment entities from the physician expertise. In this way, we build the initial connections from one patient to some specific physicians. Secondly, we utilize the initial connections, and integrate the features of patient preference and physician preference (e.g. education, income and attitude) into the modeling of the quality of service. We think that, the patient-physician pairs with the highest quality of service can be treated as the cross-layer connections between the patients network and the physicians network, thereby forming the multi-layer social network. Finally, both the factors of social trust and user preference are comprehensively considered into the patient-physician matching with the aim of shortest social distance and maximal social trust in the multi-layer network. By employing the wide adopted technologies of big data, knowledge graph and social network in solving the above three progressive problems, we provide a series of feasible solutions to patient-physician matching problem and contribute to the industry with improved quality of medical service and reduced cost of medical treatment.

在中国，大众对健康服务的需求日益增加且呈多元化。随着互联网等技术的发展，在线健康社区成为满足该需求的新方式：其不仅为患者寻求健康服务提供便利，而且为医生开拓服务范围提供新途径。本项目以在线健康社区为研究对象，聚焦患者端和医生端社区的融合以及医患匹配，拟解决三个相关的科学问题。首先，以满足基本治疗需求为目标，研究基于知识图谱的疾病症状与医生专业擅长之间的匹配；接着，在满足治疗需求的基础上，以提高服务满意度为目标，研究综合考虑医生和患者偏好（教育、经济、情感等）的匹配模型；最后，以促进医患沟通和缩短社交距离为目标，进一步考虑医患组成的多层社交网络，研究基于多层网络的社交匹配问题。通过综合利用大数据、知识图谱、社交网络方法解决以上三个科学问题，本项目的研究成果，包括提出的新模型、解决方案及算法等，一方面可对服务模型和匹配方法有所贡献，另一方面可应用于医疗实践中提升就医效率，并节省就医成本。

医患服务匹配问题，是关系个人健康需求和全国医疗资源分配的基础问题。基于知识图谱的医患服务匹配问题，旨在借助知识图谱这一工具，搭建医生的专业背景与专业特长与患者的非医学专业背景与非医学专业特长之间的桥梁，完成全国范围内寻找与匹配医生。基于社交网络的医患服务匹配问题，旨在借助医患的在线社交痕迹，进一步提升匹配过程中的服务满意度。为此，本项目以心理与精神健康相关疾病为案例，提出并验证了自动化的匹配模型和方法，并搭建了配套的服务支持和照护系统。本项目取得了一系列理论成果和工程进展。其中理论成果发表论文4篇，培养研究生2名。相关理论成果已在应用到医疗服务照护系统的实现，并在临床推广使用中。