1 Harvard Graduate School of Education, Cambridge, Massachusetts, USA.
2 Department of Otolaryngology, Harvard Medical School, Boston, Massachusetts, USA.
Otolaryngol Head Neck Surg. 2018 Jun;158(6):1093-1100. doi: 10.1177/0194599818760528. Epub 2018 Mar 7.
Objectives (1) To assess the 11-item Inner Effectiveness of Auditory Rehabilitation (Inner EAR) instrument with item response theory (IRT). (2) To determine whether the underlying latent ability could also be accurately represented by a subset of the items for use in high-volume clinical scenarios. (3) To determine whether the Inner EAR instrument correlates with pure tone thresholds and word recognition scores. Design IRT evaluation of prospective cohort data. Setting Tertiary care academic ambulatory otolaryngology clinic. Subjects and Methods Modern psychometric methods, including factor analysis and IRT, were used to assess unidimensionality and item properties. Regression methods were used to assess prediction of word recognition and pure tone audiometry scores. Results The Inner EAR scale is unidimensional, and items varied in their location and information. Information parameter estimates ranged from 1.63 to 4.52, with higher values indicating more useful items. The IRT model provided a basis for identifying 2 sets of items with relatively lower information parameters. Item information functions demonstrated which items added insubstantial value over and above other items and were removed in stages, creating a 8- and 3-item Inner EAR scale for more efficient assessment. The 8-item version accurately reflected the underlying construct. All versions correlated moderately with word recognition scores and pure tone averages. Conclusion The 11-, 8-, and 3-item versions of the Inner EAR scale have strong psychometric properties, and there is correlational validity evidence for the observed scores. Modern psychometric methods can help streamline care delivery by maximizing relevant information per item administered.
（1）使用项目反应理论（IRT）评估 11 项听觉康复内在有效性（InnerEAR）量表。（2）确定是否可以通过项目子集准确表示潜在的潜在能力，以便在高容量临床情况下使用。（3）确定 InnerEAR 量表是否与纯音阈值和言语识别分数相关。设计：前瞻性队列数据的 IRT 评估。设置：三级保健学术门诊耳鼻喉科。参与者和方法：使用现代心理测量方法，包括因子分析和 IRT，评估单维性和项目特性。回归方法用于评估言语识别和纯音测听评分的预测。结果：InnerEAR 量表是单维的，项目在位置和信息上有所不同。信息参数估计值范围为 1.63 至 4.52，较高的值表示更有用的项目。IRT 模型为识别具有相对较低信息参数的 2 组项目提供了基础。项目信息功能表明哪些项目提供了微不足道的额外价值，并逐步去除，创建了一个 8 项和 3 项 InnerEAR 量表，以更有效地进行评估。8 项版本准确反映了潜在结构。所有版本与言语识别分数和纯音平均值中度相关。结论：InnerEAR 量表的 11 项、8 项和 3 项版本具有较强的心理测量特性，观察到的分数具有相关性效度证据。现代心理测量方法可以通过最大化每个项目的相关信息量来帮助简化护理服务。
