Exploring The Validity of Acoustic Measurements and Other Voice Assessments
Exploring The Validity of Acoustic Measurements and Other Voice Assessments
复制标题
DOI：
10.1016/j.jvoice.2021.12.014
复制
发表时间：
2024-04-26
期刊：
影响因子：
2.2
通讯作者：
Behlau, Mara
中科院分区：
文献类型：
--
作者：
Englert, Marina;Latoszek, Ben Barsties, V;Behlau, Mara
目的：在一系列语音评估中，有不同的部分来评估语音功能和语音质量。建议使用声学对语音质量进行客观评估，声学语音质量指数03.01(AVQI)和声学呼吸指数(ABI)是这一领域的领先指标。这项研究的目的是验证这两种声学测量与嗓音评估的其他维度(语音残疾指数-10(VHI-10)和喉部和声带的喉部成像评估)之间的有效性。方法：我们使用了150名参与者(37名发声健康参与者和113名发声障碍患者)的回顾数据库。为了检验声学测量与其他语音评估之间的有效性，采用皮尔逊相关系数和基于似然比的接收器工作特性(ROC)统计量分析了同时效度和诊断精度。结果：VHI-10评分与AVQI、ABI呈中-显著相关，r=0.477，P<；0.001(r(2)=0.228)；r=0.426，P<；0.001(r(2)=0.181)。喉部病变分别与AVQI(r=0.362，P<；0.001，r(2)=0.131)和ABI(r=0.371，P<；0.001，r(2)=0.138)呈低相关。ROC曲线下面积(AUC)在0.701~0.737之间，除ABI和VHI-10之间(AUC=0.689)外，其余基本合理。AVQI和VHI-10之间的AUC最高，阈值为2.10；ABI和VHI-10之间的AUC最高，阈值为3.77。但敏感性最低的是ABI和VHI-10，特异性最低的是喉部诊断和ABI。结论：AVQI和ABI在参照物不是对嗓音质量的听觉-知觉判断时，其同时效度和诊断精度显著降低。然而，在某些效度方面，AVQI似乎是一种比ABI更合理的声障和喉部病变的分类方法。AVQI和ABI值都可以扩大其对临床用户的评估范围，从而显示出个体嗓音检查方法之间的更多关系。
Objective: There are different parts of a battery of voice assessments to assess voice functions and voice quality. The objective assessment of voice quality using acoustics is recommended and two measurements such as the Acoustic Voice Quality index, 03.01 (AVQI) and the Acoustic Breathiness Index (ABI) are leading in this domain. The purpose of this study was to verify the validity between these two acoustic measurements and other dimensions of the voice assessments (the Voice Handicap Index-10 (VHI-10) and the laryngeal imaging evaluation of the larynx and vocal folds). Methods: We used a retrospective database of 150 participants (37 vocally healthy participants and 113 dysphonic patients). To test the validity between the acoustic measurements and the other voice assessments the concurrent validity and diagnostic precision were analyzed using Pearson correlation coefficient and the receiver operating characteristic (ROC) statistics with likelihood ratios. Results: The VHI-10 score presented moderate-significant correlations with the AVQI and the ABI, r= 0.477, P < 0.001 (r(2)= 0.228) and r= 0.426, P < 0.001 (r(2)=0.181), respectively. The larynx alteration presented low-significant correlations with the AVQI (r= 0.362, P < 0.001, r(2)= 0.131), and ABI (r= 0.371, P < 0.001, r(2)= 0.138), respectively. The area under the curve (AUC) of ROC was almost reasonable ranging from 0.701 to 0.737; except between ABI and VHI-10 (AUC= 0.689). The highest AUC was between the AVQI and VHI-10 at a threshold of 2.10; the highest specificity was between the ABI and VHI-10 at a threshold of 3.77. However, the lowest sensitivity was between ABI and VHI-10 while the lowest specificity was between the laryngeal diagnosis and the ABI. Conclusion: AVQI and ABI presented significantly lower concurrent validity and diagnostic precision when the reference is not the auditory-perceptual judgment of voice quality. However, AVQI seems to be for some validity aspects a more reasonable classifier of vocal handicap and laryngeal alteration than ABI. Both AVQI and ABI values can broaden their range of assessment for the clinical user, thus showing more relationships between the individual voice examination methods.