Abstract: Background　Through the evaluation of the current status and development system of family doctors，it is possible to evaluate and optimize the implementation effect of family doctor policies. Objective　To construct a three-level indicator system to evaluate family doctors in Beijing，in order to provide reference for relevant departments in formulating family doctor contracted service policies. Methods　From March to April 2019，firstly，the content analysis method was used to analyze the relative policy objectives in Notice on Further Strengthening the Work Related to Family Doctor Contracted Service issued by Beijing Municipal Health Commission（Jing〔2018〕9） to extract the first-level indicators. Secondly，using the text mining method and ROST content mining system，the weight of each indicator was formed through word segmentation processing，term frequency statistics，and word cloud output. Based on the results of the term frequency statistics，the weight of the indicators was determined. Finally，the content of the indicators was adjusted according to the opinions of experts. Subsequently，the family doctor service among four community health service centers in Beijing was evaluated by the empirical method. Results　This paper finally built a Beijing Family Doctor Evaluation System consisting of three first-level indicators（service preparation and publicity，service content and process，service supervision and guarantee），seven second-level indicators（service preparation, service publicity，service personnel and process，service content implementation，service improvement，service supervision mechanism and service supporting mechanism） and 23 third-level indicators（with unified service agreement text basic content and format，etc），and the weight ranges of these indicators was formed. This evaluation system was applied to four community health service centers in Beijing，and the scores were all above four points，which was in line with the actual situation. Conclusion　The Beijing Family Doctor Evaluation System contructed by text mining method could be used to evaluate the implementation effect of family doctor policies in Beijing.

Key words: Family doctor, Index system, Text mining method, Beijing

摘要： 背景　通过对家庭医生发展现状与发展制度进行评价研究，能够对家庭医生政策实施效果进行评判和优化改进。目的　构建北京市家庭医生评价的三级指标体系，为有关部门制定家庭医生签约服务政策提供参考依据。方法　于2019年3—4月，运用内容分析法，将《北京市卫生和计划生育委员会关于进一步加强家庭医生签约服务有关工作的通知》（京卫基层〔2018〕9号）的政策目标进行内容分析，形成初级指标。进一步运用文本挖掘法，应用ROST内容挖掘系统，通过分词处理、词频统计、词云输出，形成各指标的权重。以词频结果为基础，确定指标权重，结合课题组专家意见，对指标内容进行调整。研究后期采用实证方式对北京市4家社区卫生服务中心的家庭医生服务进行评价。结果　最终形成由3个一级指标（服务准备与宣传、服务内容与流程、服务监督与保障）、7个二级指标（服务准备情况、服务宣传情况、服务人员及流程、服务内容落实情况、服务完善情况、服务监督机制、服务配套机制体制）及23个三级指标（具有统一的服务协议文本基本内容和格式等）及其权重范围组成的指标体系。将指标体系运用到北京市4家社区卫生服务中心进行评价，得到分数均在4分以上，符合实际情况。结论　基于文本挖掘法建立的北京市家庭医生评价体系能够用于对北京市家庭医生服务政策实施效果的评价。

关键词: 家庭医生, 指标体系, 文本挖掘法, 北京