中文摘要：

目的应用项目反应理论（IRT）进一步分析慢性病患者生命质量测定量表体系共性模块（QLICD-GM）条目。方法用QLICD-GM测定8种慢性病1 024例患者,采用塞姆吉玛等级反应模型计算每个条目的难度、区分度系数和信息量,绘制项目特征曲线,采用MULTILOG软件进行计算和作图。结果在所分析的30个条目中,其区分度为1.17～1.88,难度（程度）都呈严格的单调递增（-2.88～2.27）。根据平均信息量（0.36～0.97）并结合条目特征选出24个好的条目。结论项目反应理论能筛选出信息量较高的条目,且每个条目都能进行精确的定量刻画,可以弥补经典测量理论（CTT）的不足。

英文摘要：

Objective To analyze the general module of quality of life instruments for chronic diseases（QLICD-GM） using item response theory（IRT）.Methods A total of 1 024 patients each with one of the eight chronic diseases were measured by QLICD-GM.Samejima′s Graded Response Model and software MULTILOG were used to calculate difficulty coefficient,discrimination coefficient,and information power,and to draw item characteristic curve.Results Of all 30 items,the discrimination coefficient fell between 1.17 and 1.88,and difficulty coefficient increased monotonically（ranging from-2.88 to 2.27）.Twenty-four good items were selected by average information power（ranging from 0.36 to 0.97） and item characteristics.Conclusion IRT can select items with higher information power and provide accurate quantitative analysis on each item.