应用Rasch分析评估老年听力障碍量表(HHIE)的测量特性
本研究采用Rasch分析评估老年听力障碍量表(HHIE)的测量特性,数据来自380名听力损失成人。结果显示HHIE靶向性好且个体分离信度高,但与Rasch模型拟合不佳且非单维;多数条目无响应依赖且适合3点量表,但2个条目更适合二分量表,9个条目因拟合差或性别差异项目功能而被建议删除。删除并调整后得到16条目版本,拟合良好且单维。
要点
- HHIE整体靶向性好、个体分离信度高,但拟合Rasch模型不佳且非单维。
- 2个条目更适合二分响应量表,9个条目因拟合差或性别DIF被建议删除。
- 调整后16条目版本拟合Rasch模型良好且达到单维性。
核心结论
- HHIE具有若干强测量特性,但需删除拟合差或存在性别差异项目功能的条目并调整2个条目的响应格式以提升质量。
- 修订后的16条目HHIE拟合Rasch模型良好且为单维。
研究设计与样本
- 对380名听力损失成人的HHIE数据进行Rasch分析,参与者主要来自英国Nottingham生物医学研究中心数据库,另从两个听力学诊所和一家听力损失慈善机构在线论坛招募。
- 评估量表整体(Rasch模型拟合、单维性、靶向性、个体分离信度)和单个条目(响应依赖、拟合、差异项目功能、阈值排序)。
主要发现
- HHIE靶向性好且个体分离信度高。
- HHIE与Rasch模型拟合不佳且非单维。
- 多数条目无响应依赖且适合3点响应量表,但2个条目更适合二分响应量表。
- 9个条目因拟合差和/或与性别相关的差异项目功能而被确定为删除候选。
关键数值与推荐
- 删除9个候选条目并调整2个需二分响应量表的条目评分后,得到16条目版本,拟合Rasch模型良好且单维。
临床怎么用
- 在听力学研究和临床中应确保使用高质量结局测量工具,可考虑使用修订后的16条目HHIE以改善测量质量。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
原文
OBJECTIVES: The aim of this research was to evaluate the measurement properties of the Hearing Handicap Inventory for the Elderly (HHIE). The HHIE is one of the most widely used patient-reported outcome measures in audiology. It was originally developed in the United States in the 1980s as a measure of the social and emotional impact of hearing loss in older adults. It contains 25 items that are accompanied by a 3-point response scale. To date, the measurement properties of the HHIE have primarily been assessed via traditional psychometric analysis techniques (e.g., Cronbach's alpha and Principal Components Analysis). However, traditional techniques are now known to have several limitations in comparison to more modern approaches. Therefore, this research used a modern psychometric analysis technique, namely Rasch analysis, to evaluate the HHIE. DESIGN: Rasch analysis was performed on HHIE data collected from 380 adults with hearing loss. The participants were principally recruited from the participant database of the National Institute for Health Research Nottingham Biomedical Research Centre in the United Kingdom. Additional participants were recruited from two UK audiology clinics and the online forum of a UK hearing loss charity. Rasch analysis was used to assess the measurement properties of the HHIE (i.e., fit to the Rasch model, unidimensionality, targeting, and person separation reliability) and its individual items (i.e., response dependency, fit, Differential Item Functioning, and threshold ordering). RESULTS: The HHIE was found to have several strong measurement properties. Specifically, it was well-targeted and had high person separation reliability. However, it displayed poor fit to the Rasch model and was not unidimensional. The majority of the items were free of response dependency (i.e., redundancy) and were suited to the 3-point response scale. However, two items were found to be better suited to a dichotomous response scale. Furthermore, nine items were identified as being candidates for removal from the questionnaire, as they exhibited poor fit and/or Differential Item Functioning (i.e., item bias) associated with gender. The measurement properties of the HHIE could be improved by removing these items and adjusting the scores of the two items that require a dichotomous response scale. These amendments resulted in a 16-item version of the HHIE that had good fit to the Rasch model and that was unidimensional. CONCLUSIONS: It is vital to ensure that high-quality outcome measures are used in audiology research and practice. This study evaluated one of the foremost outcome measures in this field: the HHIE. The results demonstrated that the HHIE had several strong measurement properties. Amending the HHIE, such as by removing items exhibiting poor fit, could further enhance its quality. A unique aspect of this study was the application of Rasch analysis to the evaluation of the HHIE. It is recommended that future studies use modern techniques to develop and identify high-quality, hearing-specific outcome measures.
引用
- Heffernan E, Weinstein BE, Ferguson MA, et al. Application of Rasch Analysis to the Evaluation of the Measurement Properties of the Hearing Handicap Inventory for the Elderly. Ear Hear. 2020. PMID: 31972771. doi:10.1097/AUD.0000000000000832
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。