基层全科医生应诊能力评价指标体系的信效度研究

Validity and Reliability of an Evaluation Index System for Consultation Competency of General Practitioners Practicing in Primary Care Settings

背景目前,我国缺乏在农村基层卫生环境中评估全科医生临床能力的研究,缺少适用于农村基层医疗环境的医疗服务能力评价工具是导致该类研究偏少的重要原因之一。目的本文旨在验证前期构建的基层全科医生应诊能力评价指标体系的内部一致性信度和结构效度,为在农村基层中科学、客观测评全科医生应诊能力提供具备较高信效度的评估工具。方法依据基层全科医生应诊能力评价指标体系设计相应的调查量表,以Likert 5级评分法为各选项赋分。于2022年9—12月,采用目的性和分层抽样方法,抽取正在广西壮族自治区乡镇卫生院工作的全科医生或助理全科医生为研究对象,通过“问卷星”平台将调查量表发送给研究对象。基于量表实测数据,计算量表及各部分的Cronbach'α系数、折半信度系数、临界比值(CR)及相关系数;采用验证性因子分析法(CFA)将量表实测数据和假设理论模型拟合,计算模型基本拟合度(PFC)、整体模型拟合度(OMF)及模型内在结构拟合度(FISM)3类指标验证实测数据拟合程度及结构效度。结果共发放600份调查量表,有效回收366份,有效回收率为61.0%。量表应答者来自南宁市、桂林市、梧州市、百色市、贵港市5个城市的204家乡镇卫生院,86.1%(315/366)为全科医师、13.9%(51/366)为助理全科医师。总量表及除基本人口学信息外的各部分的Cronbach'sα系数均>0.700,总量表折半信度Guttman Split-Half系数为0.931。初始一阶7因子模型的各项PFC指标均达理想;除拟合优度指数(GFI)、调和拟合优度指数(AGFI)、标准拟合指数(NFI)外,修正后的最终一阶模型的OMF指标均已达理想;FISM指标中有9个观察变量的项目信度(R^(2))<0.5,其余均达理想;初始二阶模型的PFC与初始一阶模型相似;修正后的最终二阶模型的OMF与最终一阶模型接近,FISM则逊于最终一阶模型。量表实测数据最终与一阶7因子模型拟合良好。结论基层全科医生应诊能力评价指标体系具备较高的信度和效度,可用于农村基层执业环境中全科医生应诊能力测评研究和实际工作。

Background Currently,there is a lack of clinical competence evaluation tools applicable of general practitioners(GPs)practicing in rural settings in China,resulting in the lack of researches on the clinical competence evaluation of GPs in rural.Objective To explore reliability and validity of an evaluation index system for consultation competency of rural GPs developed previously,and provide an evaluation tool with high reliability and validity for scientific and objective assessment of consultation competence of GPs in rural settings.Methods Based on the evaluation system for consultation competency of rural GPs,a corresponding questionnaire was designed,and points were assigned to each option on a 5-point Likert scale.From September to December 2022,GPs or assistant general practitioners(AGPs)who were working in rural township health centers in Guangxi Province were recruited as the research subjects by using the purposive sampling and stratified sampling methods,the questionnaire was distributed through a national web-based survey platform-"WJX"to them.Cronbach'sαcoefficient,split-half coefficient,critical ratio(CR)and correlation coefficients were calculated based on questionnaire data.Confirmative factor analysis was employed to fit questionnaire data and assumed model,and calculate three categories of indicators,including preliminary fit criteria(PFC),overall model fit(OMF),and fit of internal structural model(FISM),to verify the degree of fit and structural validity of the measured data.Results A total of 600 questionnaires were distributed and 366 were validly collected,with an effective recovery rate of 61.0%,86.1%were registered as GPs and 13.9%as AGPs,who came from five cities,including Nanning,Guilin,Wuzhou,Baise,and Guigang,and 204 township health centers in Guangxi Province.Cronbach'αcoefficient for the whole questionnaire or for every section was higher than 0.700,and Guttman Split-Half coefficient was 0.931.The initial first-order model met the PFC well;except for goodness of fit index(GFI),adjusted goodness of fit index(AGFI),and normed fit index(NFI),other indicators related to the OMF reached for the best levels in the modified first-order model.Apart from R^(2) for 9 observable variables less than 0.5,the FISM for the modified first-order model showed good effects.The initial second-order model indicated as similar effects on the PFC as the initial first-order model;the modified second-order model shared similar OMF with the modified first-order model;the modified second-order model was inferior to the modified first-order model in terms of the FISM.Eventually,questionnaire data fitted the modified first-order model better.Conclusion The evaluation index system for consultation competency of rural GPs shows high reliability and validity,which can be used for research and practical work on the evaluation of consultation competence evaluation of GPs in rural settings.