IRT等级反应模型在冠心病PRO量表编制中的应用  

Application of IRT Graded Response Model in Coronary Heart Disease PRO Scale

作　　者：曾宪华[1] 马瑞[1] 郭小玲[1] 罗艳虹[1] 张岩波[1] 

机构地区：[1]山西医科大学卫生统计教研室,030001

出　　处：《中国卫生统计》2012年第5期639-641,共3页Chinese Journal of Health Statistics

基　　金：国家自然科学基金资助项目(81273180)

摘　　要：目的探讨项目反应理论中的等级反应模型在基于冠心病患者报告临床结局量表(CHD-PRO)条目筛选中的应用。方法对CHD-PRO初始量表所有条目拟合等级反应模型,依据区分度参数和难度等级参数对条目进行评价和筛选。结果通过等级反应模型理论对量表条目进行筛选,初始量表共计62个条目,建议删除12个条目。结论与经典测量理论和传统方法相比,项目反应理论与等级反应模型能够克服等级资料的限制,增加了对受试对象能力的估计,更加适合量表条目的筛选。Objective To explore the IRT graded response model and its application in the patient-reported outcomes scale of coronary heart disease. Methods All the items of the initial scale of CHD-PRO fit the graded response model, and the items are evaluated and selected accord- ing to the discriminative power index and difficulty index. Results The number of the initial scale contains 62 items. 12 items are proposed to be deleted according to graded response model theory. Conclusion Com- pared with Classical Test Theory and traditional methods, item response the- ory and graded response model could overcome the limitation of ordered data, and considered the estimation of the subjects ability. IRT is a more suitable method to select items for the scale development.

关 键 词：项目反应理论 等级反应模型 冠心病 患者报告临床结局 条目筛选 

分 类 号：R541.4[医药卫生—心血管疾病]