Abstract: Background　The development of general practitioners'（GPs'） competence assessment systems has become mature in foreign countries. However，in China，GPs' competence has been assessed mainly by healthcare institutions themselves. Due to different standards and evaluation methods of various institutions，the authority and objectivity of the assessment results cannot be guaranteed. Therefore，it is necessary to establish a unified standard in line with the current development of general practice. Objective　To develop a tiered GPs' competence assessment system using the Delphi method，and based on this，to develop a theory-skill-humanity（TSH） cube model，to improve the GPs' competence assessment more accurately and objectively via promoting the homogenous progress. Methods　From April to May 2019，using literature review and group discussion，we developed a preliminary framework of a three-tiered GPs' competence assessment system consisting of 3 first-tier indices（general practice theory，skills and humanity），44 second-tier indices and 44 third-tier indices. Then we developed the GPs' competence assessment system after revising the framework according to the results analysis data of a two-round Delphi survey with 28 experts conducted from July 1 to September 10，2019. We analyzed the experts' basic personal information，suggestions for the framework，and their positive，authoritative and Kendall's concordance coefficients as well as consensus level in the survey. Results　The familiarity and judgment coefficients were 0.900 and 0.936，respectively，in the first round of survey，and were 0.948 and 0.952，respectively，in the second round.The authority coefficient in the first and second round expert consultation was 0.918 and 0.952，respectively，which were both above 0.70. The importance and accessibility coefficients were 0.471 and 0.409，respectively，for the first round of survey，and were 0.363 and 0.386，respectively，for the second round. The final GPs' competence assessment system includes three domains（general practice theory，clinical skills and humanity），and 43 assessment items. In accordance with this system and previous patents，the TSH model was developed.Conclusion　 Our GPs lusmpetence assessment system developed based on the results of email-based surveys with high expertsent sytive and authoritative coefficients and consensus，includes three first-tier indices，43 second-tier indices and 43 third-tier indices，which is scientific，reliable and reliable verified statistically.

Key words: General practitioners；Theoretical examination；Indicators system；TSH（theory, skill, humanity）model；Delphi method

摘要： 背景　全科医生能力评价体系在国外发展已趋于成熟。但在国内，仍以医疗机构自行考核为主要方式，由于各个机构标准不一，评价方法不一，其结果的权威性、客观性等无法保证。因此需要构建符合目前全科医学发展的统一标准。目的　通过德尔菲法制定全科医生能力评价分级指标体系，根据指标结果构建“全科医生能力评价分级TSH立方体模型”，推动全科医生能力评价同质化进程，更准确、客观地评价全科医生能力。方法　2019年4—5月，通过文献分析、专题小组讨论，初步确立了全科医生能力评价分级指标体系，包括一级指标3个（全科理论知识、全科临床技能、全科人文素养），二级指标、三级指标各44个。2019-07-01至2019-09-10通过德尔菲法进行两轮专家问询（n=28），将其建议进行归纳整理，最终确定并形成全科医生能力评价分级指标体系。记录专家基本资料、专家积极性、专家修改建议、专家权威程度、专家意见集中程度、专家意见协调程度。结果　第一轮专家咨询的熟悉程度和判断系数为0.900和0.936，第二轮为0.948和0.956；两轮权威系数分别为0.918和0.952，均在0.70以上；第一轮专家咨询重要性、可得性协调系数为0.471、0.409，第二轮为0.363、0.386。最终确定了包含全科理论知识、全科临床技能、全科人文素养3个维度，43条考核指标的全科医生能力评价分级指标体系。根据指标结果及前期所获专利完成“全科医生能力评价分级TSH立方体模型”构建。结论　本研究构建了包括一级指标3个，二级指标、三级指标各43个的全科医生能力评价分级指标体系，经统计分析，专家积极系数、权威程度较高，意见较集中，函询结果科学性、可信度和可靠性高。

关键词: 全科医生, 能力评价, 指标体系, TSH模型, 德尔菲法