【作者】 于惊涛；

【导师】 陈力；

【作者基本信息】 哈尔滨医科大学 ， 社会医学与卫生事业管理， 2012， 博士

【摘要】 客观结构化临床考试是近年来在医学教育领域中应用较为广泛的临床能力考核形式，通常由标准化病人考站、临床操作考站、临床资料测试考站、临床病例分析考站以及客观题笔试考站等组合而成，医学生依据一定次序先后通过这些考站，各考站测评分数经过加权形成最终的考试成绩。在对客观结构化临床考试的应用和研究中，人们证实其具有较好的测验稳定性，即信度良好，也证实了该测验具有较好的同时效度和预测效度，这为客观结构化临床考试的持续推广奠定了基础。但是，对这种由多种考站组合而成、具有情境模拟特点的医学考试，仅有同时效度和预测效度是不够的，对结构效度的探讨应该成为客观结构化临床考试效度评价中不可或缺的一环。目前，对客观结构化临床考试结构效度的研究较为匮乏，一些研究也仅仅是进行探索性因素分析，其证据效果并不明显。因此，本研究将目标定位于客观结构化临床考试的结构效度，评估它的汇聚效度和区分效度，为厘清其结构框架和评分模型导向作出努力。同时，我们也力求方法学上的突破，将验证性因素分析（CFA）、多质多法分析（MTMM）、结构方程模型分析（SEM）、概化理论分析（GT）等方法引入到客观结构化临床考试结构效度的探讨中。本研究以2006年临床医学专业本科毕业生参加客观结构化临床考试的成绩为数据来源，并进行题目质量评估和筛选。采用专家咨询法，将客观结构化临床考试共四类考站进行归类，确定出若干种测评方法（考站）。采用探索性因素分析和验证性因素分析（CFA），分别对临床理论考试、临床思维与技能考站、临床资料与病例考站进行测评的能力维度分析。采用多质多法分析（MTMM）来评估客观结构化临床考试的汇聚效度和区分效度，采用结构方程模型分析（SEM）探讨客观结构化临床考试不同结构模型的拟合情况。采用概化理论分析（GT）探索客观结构化临床考试评分误差和系统变异的主要来源，对其结构效度进行进一步的验证。研究结果显示，客观结构化临床考试包括若干考站，根据这些考站设置的目的、性质、特点和组织形式等，可以将其归类为三个测评方法（考站），即临床理论考试、临床思维与操作考站、临床资料与病例考站。通过探索性因素分析和验证性因素分析，证明在不同测评方法（考站）内均不同程度的测量了三种临床能力维度，即医学知识、操作能力、临床思维。多质多法分析结果显示，客观结构化临床考试具有较高的汇聚效度，但区分效度并不理想。进一步SEM拟合检验证明，客观结构化临床考试评分模型呈现出测评方法（考站）导向的特点。概化理论分析结果显示，客观结构化临床考试的测评维度效应是造成评分误差、引起系统变异的主要原因，验证了OSCE较高的汇聚效度和较低的区分效度，提示我们要加强对能力维度评估的质量和效果。研究结果除了对客观结构化临床考试结构效度进行了量化评估，还证明了CFA、MTMM、SEM、GT等方法的适用性。更多还原

【Abstract】 Objective structured clinical examination (OSCE) has been widely used as aclinical skills assessment form in medical education area, usually consists test stationfor standardized patients, test station for clinical operation, test station for clinicaldata、test station for clinical case analysis, objective multiple-choice questions etc.,medical students pass these test stations according to a certain order; each test stationevaluation scores are weighted to form the final exam results. In the application andstudy of objective structured clinical examination, researchers have confirmed that ithas good test stability: reliability, concurrent validity and predictive validity are good,which laid the foundation of further promoting the objective structured clinicalexamination.However, for the medical test with multiple test station with situation analogy,concurrent validity and predictive validity is not enough, discuss on construct validityshould be an important process in objective structured clinical examination evaluation.At present, study on the effective of objective structured clinical examination is quiteinadequate, existing studies only concentrate on exploratory factor analysis withnon-obvious proof. Therefore, this study will target on the structure validity of theobjective structured clinical examination, evaluate its convergent validity anddiscriminate validity, clarify its structure and model oriented. At the same time, wealso try to make a breakthrough in methodology, introduce confirmatory factoranalysis（CFA）, Multitrait-multimethod approach (MTMM）, structural equationmodel (SEM) Generalizability theory (GT）etc. into the discussion of objectivestructured clinical examination’s structure validity. This study’s data sources come from clinical medical undergraduates’ scoresparticipating in objective structured clinical examination, evaluated and assed for titlequality. Expert consultation has been used to classify four groups of objectivestructured clinical examination, identified several evaluation methods (test station).Exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) has beenrespectively analyses the test dimensions of clinical theory examination and clinicalthinking and clinical skills examination, clinical data test and clinical treatment abilitytest station. Multitrait-multimethod approach (MTMM) analysis has been used toevaluate the convergent validity and discriminate validity of objective structuredclinical examination, structural equation model analysis (SEM) has been used todiscuss different structure model’s fit indexes of objective structured clinicalexamination. Generalizability theory analysis (GT) has been used to quantitativeanalyses the structure validity factors of objective structured clinical examination, anddiscusses the improvement strategy.The results shows that objective structured clinical examination including severaltest stations, which classified as three test methods(test station) according to teststation’s purpose, properties, characteristics and organizational form: clinical theoryexamination, clinical thinking and operating examination, clinical data test andclinical treatment ability the exam. Exploratory factor analysis and confirmatoryfactor analysis have been used to demonstrate that, different measurement methods(test station) evaluate three clinical assessment dimensions: medical knowledge, skills,clinical thinking. Multitrait-multimethod analysis results show that objectivestructured clinical examination has high convergent validity, though discriminatevalidity is not ideal. Further SEM fitting inspection certificates that structured clinicalexamination score model presents evaluation methods (test station)situational-specific. Generalizability theory analysis result shows that: evaluationmethod (test station) number, assessment dimensions have influence on objective structured clinical examination’s structure validity and should be optimized.The results quantitative assessed the objective structured clinical examination’sstructure validity, and proves the applicability of CFA、MTMM、SEM、GT method.更多还原