GZA Sint-Augustinus, Department of Otorhinolaryngology and Head & Neck Surgery, European Institute of ORL-HNS, Antwerp, Belgium.
SRH University of Applied Health Sciences, Speech-Language Pathology, Düsseldorf, Germany; University of Münster, University Hospital Münster, Department of Phoniatrics and Pediatric Audiology, Münster, Germany.
J Voice. 2024 Nov;38(6):1527.e1-1527.e19. doi: 10.1016/j.jvoice.2022.04.022. Epub 2022 Jun 23.
Acoustic measurements are useful tools to objectively measure overall voice quality. The Acoustic Voice Quality Index (AVQI) has shown to be a valid multiparametric tool to objectify dysphonia severity. The increasing number of validity studies investigating AVQI's validity demands a comprehensive synthesis of the available outcomes.
The aim of the present meta-analysis is to quantify the evidence for the diagnostic accuracy of the AVQI, including its sensitivity, specificity and likelihood ratio statistics, and its concurrent validity and sensitivity to changes in auditory-perceptual voice quality ratings.
Meta-analysis SEARCH STRATEGY: MEDLINE, EMBASE, the Cochrane library and Web of Science were searched from 2010 till April 2021 with an additional manual search, using keywords related to AVQI and common terminologies of validity outcomes. Studies considering the clinical validity of AVQI (ie, diagnostic accuracy, concurrent validity and sensitivity to change), using auditory-perceptual voice quality evaluation as reference, were included.
The Preferred Reporting Items for Systematic reviews and Meta-Analyses of Diagnostic Test Accuracy Studies (PRISMA-DTA) guidelines were used. Quality assessment of included studies was conducted using the QUADAS-2 tool. For the diagnostic accuracy of AVQI, the pooled sensitivity, specificity and likelihood ratio statistics were determined using a summary receiver operating characteristic approach. Weighted correlation coefficient measures (r¯) were used to assess the concurrent validity and sensitivity to change.
A total of 198 studies were screened and 33 articles were included. In total, voice samples of 11447, 10272, and 367 different subjects were considered for analysis of diagnostic accuracy, concurrent validity and change responsiveness, respectively. Satisfying diagnostic accuracy results were found with a pooled sensitivity of 0.83 (95% CI: 0.82-0.83), a pooled specificity of 0.89 (95% CI: 0.88-0.90), a pooled positive LR of 7.75 (95% CI: 6.04-9.95), a pooled negative LR of 0.20 (95% CI: 0.16-0.23), and a pooled diagnostic odds ratio of 47.13 (95% CI: 34.82-63.79). Summary receiver operating characteristic curve analysis showed an excellent AUC value of 0.937 and Q* index of 0.874. Strong correlations of r¯ = 0.838 for concurrent validity and r¯ = 0.796 for sensitivity to change were found.
Our results confirm the general clinical utility of the AVQI as a robust and valid objective measure for evaluating overall dysphonia severity across languages and study methods.
声学测量是客观评估整体语音质量的有用工具。声学分度指数 (AVQI) 已被证明是一种有效的多参数工具，可客观评估嗓音障碍的严重程度。随着越来越多的有效性研究调查 AVQI 的有效性，因此需要对现有结果进行综合分析。
本荟萃分析的目的是量化 AVQI 诊断准确性的证据，包括其敏感性、特异性和似然比统计数据，以及其同时效度和对听觉感知嗓音质量评分变化的敏感性。
荟萃分析
从 2010 年到 2021 年 4 月，使用与 AVQI 相关的关键词和常见有效性结果术语，在 MEDLINE、EMBASE、Cochrane 图书馆和 Web of Science 上进行了搜索，并进行了额外的手动搜索。包括考虑 AVQI 临床有效性（即诊断准确性、同时效度和对变化的敏感性）的研究，使用听觉感知嗓音质量评估作为参考。
使用系统评价和荟萃分析诊断测试准确性研究的首选报告项目 (PRISMA-DTA) 指南。使用 QUADAS-2 工具对纳入研究进行质量评估。对于 AVQI 的诊断准确性，使用汇总受试者工作特征方法确定汇总敏感性、特异性和似然比统计数据。使用加权相关系数度量 (r¯) 评估同时效度和对变化的敏感性。
共筛选出 198 项研究，纳入 33 项研究。总共分析了诊断准确性、同时效度和变化反应性的 11447、10272 和 367 个不同受试者的语音样本。结果发现，诊断准确性令人满意，汇总敏感性为 0.83（95%CI：0.82-0.83），汇总特异性为 0.89（95%CI：0.88-0.90），汇总阳性似然比为 7.75（95%CI：6.04-9.95），汇总阴性似然比为 0.20（95%CI：0.16-0.23），汇总诊断比值比为 47.13（95%CI：34.82-63.79）。汇总受试者工作特征曲线分析显示 AUC 值为 0.937，Q*指数为 0.874，均为优秀。同时还发现，同时效度的 r¯ = 0.838 和对变化的敏感性的 r¯ = 0.796 相关性较强。
我们的结果证实，AVQI 作为一种可靠且有效的客观测量工具，可用于评估各种语言和研究方法的整体嗓音障碍严重程度，具有广泛的临床应用价值。
