中文版视功能损害生命质量量表信效度检验的Rasch模型分析

摘要

目的：对中文版视功能损害生命质量量表（CfaVIQ）进行Rasch模型分析，并检验其信度和效度。方法经翻译、回译和文化调适后形成27个条目的CfaVIQ，对215例视功能损害患者进行问卷调查。运用Winsteps统计软件进行Rasch模型项目拟合分析，测评条目靶向和确定分离指数。应用Rasch模型测评量表的内部一致性信度、重测信度、结构效度和区分效度，应用ROC曲线和曲线下面积评价效标效度。结果 CfaVIQ的Rasch模型项目拟合分析表明：个人分离指数和项目分离指数分别为3.55和8.01，均大于标准2；个人和项目的加权拟合统计量的残差均方和未加权拟合统计量的残差均方均在0.5～1.5，证明CfaVIQ符合Rasch模型。Wright map分析显示：样本能力和条目难度呈正态分布。内部一致性的分离指数为3.55，分离信度为0.93；重测信度的Pearson相关系数为0.92；结构效度分析CfaVIQ与视力、视敏度、视功能相关生命质量量表-25的Pearson积差相关分别是-0.67、-0.14和0.79；区分效度显示：条目10和条目23的未加权拟合统计量的残差均方值＞1.5，其余条目的残差均方值都介于0.5～1.5；效标效度指数为0.922（95%的置信区间为0.858～0.987）。结论 CfaVIQ具有良好的信度和效度，Rasch模型为CfaVIQ的分析提供了更全面和客观的信息，具有良好的应用前景。%Objective To explore the psychometric properties of the Chinese version of the Functional Ability of Visually Impaired Questionnaire (CfaVIQ) by using Rasch Analysis. Methods The faVIQ was translated and adapted into the Chinese-version faVIQ (CfaVIQ). The CfaVIQ was completed by 215 visionlly impaired patients. The data were Rasch-analyzed with the Rating Scale Model of Andrich for calculating item fit statistics, determining the separation index by using Winsteps Rasch measurement program. Results Rasch model analysis indicated individual separation index (3.55) and project separation index (8.01) were both arithmetic greater than 2; the inlier-sensitive fit Mean Square (infitMnSq) and outlier-sensitive fit Mean Square (outfitMnSq) of Items, Persons were both within the range of 0.5-1.5, indicating great coincidence to the model;Wright map analysis showed the distribution of the sample ability and item difficulty was ideally normal. Coefficient of internal consistency was 3.55, separation reliability was 0.93. Pearson correlation coefficient of test-retest reliability was 0.92. Pearson correlation coefficient between CfaVIQ and vision (-0.67), visual acuity (-0.14). The National Eye Institute 25 Item Visual Function Questionnarie (NEI-VFQ25) (0.79) indicated a good construct validity. All the items′outfitMnSq, except outfitMnSq>1.5 of item 10 and item 23,were within the range of 0.5-1.5. Criterion Validity was 0.922 (95%CI:0.858-0.987). Conclusions The CfaVIQ is a culturally specific vision quality of life measure instrument. It can sensitively assess the quality-of-life in the visually impaired group. Rasch measurement model can provide comprehensive and positive information and has good application prospects for analysis of CfaVIQ.