基于经典测量理论和项目反应理论的胃食管反流病生活质量量表QLICD-GERD的项目分析

文献信息

中文摘要

背景： 胃食管反流病（GERD）是一种常见的消化系统疾病，严重影响患者的身心健康，导致健康相关生活质量（HRQoL）显著下降。慢性病患者生活质量量表-胃食管反流病（QLICD-GERD，V1.0）被开发用于通过普适性模块和疾病特异性模块全面评估GERD的影响。本研究旨在采用经典测量理论（CTT）和项目反应理论（IRT）分析QLICD-GERD（V1.0）的心理测量学特性和项目表现，为其进一步完善提供理论和实证支持。
方法： 于2022年1月至2024年12月期间，从中国云南省一家三级医院的消化内科招募了207例胃食管反流病（GERD）患者。采用自填式QLICD-GERD（V1.0）对来自中国云南省一家三级医院消化内科的207例GERD患者（2022年1月至2024年12月）进行调查。应用CTT方法——包括Cronbach's α系数、分半信度、标准差（SD）分析、相关分析和因子分析——评估信度和效度。采用Samejima等级反应模型进行IRT分析，估计项目区分度（a）、难度参数（b1-b4）和平均信息量（I）。使用SPSS 27.0、AMOS和Multilog 7.03进行数据分析。
结果： 在CTT分析中，48个条目中有42个满足至少三种统计方法的标准，被评为良好，6个条目被评为较差。量表的总体Cronbach's α系数为0.848，分半信度为0.905，表明内部一致性令人满意。KMO值为0.755，累积方差贡献率为68.51%。修订后的验证性因子分析（CFA）模型拟合指标有所改善（χ²/df=2.493，RMSEA=0.084，GFI=0.700，CFI=0.691，TLI=0.666，IFI=0.697）。在IRT分析下，项目区分度范围为0.89至1.44，大多数难度参数随等级水平（B1→B4）单调递增。平均信息量值为0.064至0.529：躯体领域9个条目、社会领域2个条目、精神领域1个条目和疾病特异性领域14个条目被归类为较差，20个为良好，2个为优秀。
结论： QLICD-GERD（V1.0）总体表现出良好的心理测量学特性，具有较好的信度、效度和区分度。心理和社会领域表现最佳，而躯体领域和精神领域的若干条目需要改进以提高测量精度。CTT与IRT的结合为优化量表结构、增强其评估GERD患者生活质量的临床适用性提供了综合框架。

英文摘要

BACKGROUND: Gastroesophageal reflux disease (GERD) is a common digestive disorder that severely affects patients' physical and psychological health, leading to a substantial decline in health-related quality of life (HRQoL). The Quality of Life Instrument for Chronic Diseases-Gastroesophageal Reflux Disease (QLICD-GERD, V1.0) was developed to comprehensively assess the impact of GERD using both generic and disease-specific modules. This study aimed to analyze the psychometric properties and item performance of the QLICD-GERD (V1.0) using both Classical Test Theory (CTT) and Item Response Theory (IRT), in order to provide theoretical and empirical support for its further refinement.
METHODS: A total of 207 patients with gastroesophageal reflux disease (GERD) were recruited from the Department of Gastroenterology of a tertiary hospital in Yunnan Province, China, between January 2022 and December 2024. A total of 207 patients with GERD from the Department of Gastroenterology of a tertiary hospital in Yunnan Province (January 2022-December 2024) were investigated using the self-administered QLICD-GERD (V1.0). CTT methods-including Cronbach's alpha, split-half reliability, standard deviation (SD) analysis, correlation analysis, and factor analysis-were applied to evaluate reliability and validity. IRT was used to estimate item discrimination (a), difficulty parameters (b1-b4), and average information (I), using the Samejima graded response model. SPSS 27.0, AMOS, and Multilog 7.03 were used for data analysis.
RESULTS: In CTT analysis, 42 out of 48 items met the criteria of at least three statistical methods and were rated as good, while six items were rated as poor. The overall Cronbach's alpha of the scale was 0.848, and the split-half reliability was 0.905, indicating satisfactory internal consistency. The KMO value was 0.755, with a cumulative variance contribution of 68.51%. The revised confirmatory factor analysis (CFA) model achieved improved fit indices (χ²/df = 2.493, RMSEA = 0.084, GFI = 0.700, CFI = 0.691, TLI = 0.666, IFI = 0.697). Under IRT analysis, item discrimination ranged from 0.89 to 1.44, and most difficulty parameters increased monotonically with grade levels (B1→B4). Average information values ranged from 0.064 to 0.529: 9 items in the physical domain, 2 in social, 1 in spiritual, and 14 in the disease-specific domain were classified as poor, 20 good, and 2 excellent.
CONCLUSIONS: The QLICD-GERD (V1.0) demonstrates overall good psychometric properties, with strong reliability, validity, and discrimination. The psychological and social domains performed best, while several items in the physical and spiritual domains require refinement to improve measurement precision. The combination of CTT and IRT provides a comprehensive framework for optimizing the instrument's structure and enhancing its clinical applicability for assessing quality of life among GERD patients.