文献速递耳科与耳神经

中文版言语空间听觉质量量表信度与效度的评估

[Evaluation of reliability and validity of the Chinese version of the Speech, Spatial and Qualities of Hearing Scale].
PubMed 文献 · Zhonghua Er Bi Yan Hou Tou Jing Wai Ke Za Zhi · 2025-01-01

研究对言语空间听觉质量量表(SSQ)进行汉化,基于Rasch模型和经典测量理论评估中文版SSQ的信度与效度;共纳入400份有效问卷,其中听力损失受试者281例、听力正常受试者119例。结果显示中文版SSQ具有良好的信度和效度,可作为我国听力损失人群听觉能力自我评估工具在临床应用。

要点

  • 中文版SSQ由言语感知(14题)、空间听觉(17题)和听觉质量(18题)三个维度构成,各维度均具单维特性。
  • 三个维度信度指数均大于0.8,分离指数均大于2.0;Cronbach′s α系数为0.991。
  • 中文版SSQ对听力损失预测的AUC为0.789(P<0.001),平均得分<8.5分时敏感度61.3%、特异度80.1%。

核心结论

  • 中文版SSQ具有良好的信度和效度,可以作为我国听力损失人群听觉能力自我评估的工具在临床应用。

研究设计与样本

  • 采用Brislin模式对SSQ进行系统性翻译形成中文版SSQ。
  • 采用随机抽样法选取2022年10月至2023年10月在首都医科大学附属北京朝阳医院耳鼻咽喉头颈外科就诊的416例患者进行调查,获得400份有效问卷。
  • 听力损失受试者281例(年龄16~92岁、平均59.4岁,女性164例、男性117例),听力正常受试者119例(年龄18~72岁、平均39.9岁,女性82例、男性37例)。
  • 基于Rasch模型和经典测量理论,通过SPSS 24.0及Winstep 3.72.3统计学软件评估信度和效度。

主要发现

  • Rasch模型分析显示中文版SSQ由言语感知维度(14个题目)、空间听觉维度(17个题目)和听觉质量维度(18个题目)构成,每个维度均具有单维特性,量表整体所测量的数据可以适配Rasch模型分析。
  • 三个维度信度指数均大于0.8,分离指数均大于2.0;三个维度的题目难度分布相对集中(其均值均为:Logit 0),个体能力相对较广(其均值分别为:Logit 0.67、Logit-1.71和Logit 0.83)。
  • CTT分析显示中文版SSQ的Cronbach′s α系数为0.991,其三个维度子量表Cronbach′s α系数均大于0.95,仅第28题与第30题的项总计相关性系数低于0.7,其余题目项总计相关性系数均大于0.7。
  • 区别效度良好(t值范围9.604~12.268,P值均<0.01);探索性因子分析显示3个主成分因子的贡献率为78.49%,所有题目在相应公因子上的载荷值均大于0.4,旋转后3个公因子与原版SSQ的三个维度的题目基本保持一致。
  • 中文版SSQ对听力损失的预测价值较高(AUC为0.789,P<0.001),其平均得分<8.5分时预测听力损失的敏感度为61.3%,特异度为80.1%。

关键数值与推荐

  • 中文版SSQ三个维度信度指数均大于0.8,分离指数均大于2.0。
  • 中文版SSQ的Cronbach′s α系数为0.991,三个维度子量表Cronbach′s α系数均大于0.95。
  • 3个主成分因子的贡献率为78.49%,所有题目在相应公因子上的载荷值均大于0.4。
  • 中文版SSQ对听力损失的AUC为0.789(P<0.001),平均得分<8.5分时预测听力损失的敏感度为61.3%,特异度为80.1%。

临床怎么用

  • 中文版SSQ可作为我国听力损失人群听觉能力自我评估的工具在临床应用。

局限性

  • 原文未报告局限性。

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

Objective: To sinicize the Speech, Spatial and Qualities of Hearing Scale (SSQ) and to assess its reliability and validity based on the Rasch model and Classical Test Theory (CTT). Methods: The acquisition of the Chinese version of the SSQ relies on the systematic translation of the SSQ using the Brislin model. A total of 416 patients who attended the Department of Otorhinolaryngology, Head and Neck Surgery, Chaoyang Hospital, Beijing, between October 2022 and October 2023 were randomly selected using the Chinese version of the SSQ, of which 400 questionnaires were valid, with 281 subjects with hearing impairment [aged 16-92 (59.4±14.4) years, 164 females and 117 males] and 119 subjects with normal hearing [aged 18-72 (39.9±13.6) years, 82 females and 37 males]. The reliability and validity of the Chinese version of the SSQ were assessed by SPSS 24.0 and Winstep 3.72.3 statistical software. Results: Rasch model analysis showed that the Chinese version of the SSQ consisted of a speech perception dimension (14 items), a spatial hearing dimension (17 items), and an auditory quality dimension (18 items), each with unidimensional properties, and the data measured by the scale could be adapted for Rasch model analysis. The Chinese version of the SSQ had reliability indices over 0.8 and separation indices over 2.0 for all three dimensions. The distribution of the difficulty of the questions in the three dimensions was relatively concentrated (the three dimensions of means was: Logit 0.00), and the participants competence was relatively broad (the respective means was: Logit 0.67, Logit-1.71, Logit 0.83). Among them, the item-total correlation coefficients (ITCC) of item 28 and 30 were below 0.7, while ITCC values for the rest were all greater than 0.7. The discriminative validity was good (t=9.604-12.268,P<0.01). Exploratory factor analysis showed that the contribution rate of the three principal components was 78.49%, and the loadings of the items on the corresponding common factors were all greater than 0.4. The three rotated common factors were basically consistent with those of the original SSQ in the three dimensions. The Chinese version of the SSQ had a high predictive value for hearing loss (AUC of 0.789, P<0.001), with a sensitivity of 61.3% and specificity of 80.1% for predicting hearing loss in case of its mean score <8.5. Conclusion: The Chinese version of the SSQ has favorable reliability and validity, which can serve as a self-assessment tool of hearing ability for people with hearing loss in China in clinical application. 目的: 汉化言语空间听觉质量量表(Speech,Spatial and Qualities of Hearing Scale,SSQ),评估其信度和效度。 方法: 采用Brislin模式对SSQ进行系统性翻译,形成中文版SSQ。采用随机抽样法选取2022年10月至2023年10月在首都医科大学附属北京朝阳医院耳鼻咽喉头颈外科就诊的416例患者进行中文版SSQ调查,共获得400份有效问卷,其中听力损失受试者281例(年龄16~92岁、平均59.4岁,女性164例、男性117例),听力正常受试者119例(年龄18~72岁、平均39.9岁,女性82例、男性37例)。基于Rasch模型和经典测量理论(classical test theory,CTT),通过SPSS24.0及Winstep 3.72.3统计学软件对中文版SSQ的信度和效度进行评估。 结果: Rasch模型分析显示,中文版SSQ由言语感知维度(14个题目)、空间听觉维度(17个题目)和听觉质量维度(18个题目)构成,每个维度均具有单维特性,量表整体所测量的数据可以适配Rasch模型分析。中文版SSQ的三个维度信度指数均大于0.8,分离指数均大于2.0。三个维度的题目难度分布相对集中(其均值均为:Logit 0),个体能力相对较广(其均值分别为:Logit 0.67、Logit-1.71和Logit 0.83)。CTT分析显示,中文版SSQ的Cronbach′s α系数为0.991,其三个维度子量表Cronbach′s α系数均大于0.95,仅第28题与第30题的项总计相关性系数(item-to-total correlation coefficients,ITCC)低于0.7,其余题目ITCC均大于0.7。区别效度良好(t值范围9.604~12.268,P值均<0.01)。探索性因子分析显示,3个主成分因子的贡献率为78.49%,所有题目在相应公因子上的载荷值均大于0.4,旋转后3个公因子与原版SSQ的三个维度的题目基本保持一致。中文版SSQ对听力损失的预测价值较高[曲线下面积(AUC)为0.789,P<0.001],其平均得分<8.5分时预测听力损失的敏感度为61.3%,特异度为80.1%。 结论: 中文版SSQ具有良好的信度和效度,可以作为我国听力损失人群听觉能力自我评估的工具在临床应用。.

来源:Zhonghua Er Bi Yan Hou Tou Jing Wai Ke Za Zhi,查看原文。版权归原作者所有。

引用

  1. Wu JZ, Zhang J. [Evaluation of reliability and validity of the Chinese version of the Speech, Spatial and Qualities of Hearing Scale]. Zhonghua Er Bi Yan Hou Tou Jing Wai Ke Za Zhi. 2025. PMID: 40850794. doi:10.3760/cma.j.cn115330-20240630-00390

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。