论文部分内容阅读
目的使用项目反应理论(item response theory,IRT)对SF-36量表进行分析,探讨更适合于量表分析的方法。方法采用横断面研究。使用SF-36量表调查526例确诊的鼻咽癌患者,使用部分评分模型对反映生理健康的21个条目进行分析:计算PSI(person separation index),估计阈值参数和生存质量,对性别、年龄、婚姻等进行项目功能差异(DIF)分析。结果 PSI=0.85,整个模型的拟合效果好;除了条目3i、3j、8的阈值参数顺序出现颠倒外,其余条目均符合理论假设;DIF分析结果显示,所有的条目都没有非一致性DIF,条目4a在性别上存在一致性DIF,条目3g、10a在病人来源上存在一致性DIF。结论 SF-36生理健康的21个条目基本满足单维性、局部独立性的要求,可以放在一起进行IRT分析。IRT能够克服等级资料的限制,能力参数与阈值参数定义在同一个标尺上,而且存在参数不变性的优点,更加适合量表的分析评价。
Objective To analyze the SF-36 scale using item response theory (IRT) to explore more suitable methods for scale analysis. Methods Using cross-sectional studies. A total of 526 patients with confirmed nasopharyngeal carcinoma were screened using the SF-36 scale, and 21 partial items reflecting the physical health were analyzed using a partial score model: PSI (person separation index), estimated threshold parameters and quality of life were evaluated, and gender, age , Marriage and other project functional differences (DIF) analysis. The results of PSI = 0.85, the fitting effect of the whole model is good; except for the order of threshold parameters of 3i, 3j, 8 are upside down, the rest of the items are in line with the theoretical assumptions; DIF analysis shows that all the items have no inconsistent DIF, Entry 4a has a gender-consistent DIF, entry 3g, 10a has a consistent DIF on the patient’s source. Conclusion The 21 items of physiological health of SF-36 basically meet the requirements of unidimensional and local independence and can be put together for IRT analysis. IRT can overcome the limitation of level data. The ability parameter and the threshold parameter are defined on the same scale, and there is the advantage of parameter invariance, which is more suitable for the analysis and evaluation of the scale.