Reliability of Auditory-Perceptual Voice Evaluation: An Investigation of GRBAS Scale Reproducibility Among Expert
Naoki Takemoto1, Tetsuji Sanuki2, Akihiro Kozaki3
1Department of Otolaryngology-Head and Neck Surgery, Nagoya City University Graduate School of Medical Sciences.
Objective:
The GRBAS scale is widely used for auditory-perceptual evaluation of voice quality, yet its reliability remains a matter of debate. We aimed to examine the intra-rater and inter-rater reliability of the GRBAS scale when applied by experienced evaluators.
Methods:
Two studies were conducted. In study 1, 74 voice samples from patients with various voice disorders were independently assessed twice, at a 2-month interval, by three speech-language pathologists with over 10 years of clinical experience. Intra- and inter-rater reliabilities were calculated using the AC1 statistic. In study 2, 24 additional voice samples were evaluated by five experienced raters (three speech-language pathologists and two phonosurgeons) to further investigate inter-rater reliability.
Results:
In study 1, all raters demonstrated moderate to almost perfect intra-rater reliability across the GRBAS parameters. Inter-rater reliability was moderate to substantial for grade (G), breathiness (B), and asthenia (A), but slight for roughness (R) and strain (S). In study 2, the inter-rater reliability was again moderate for A but consistently slight for S. Across both studies, the parameter S showed limited reproducibility.
Conclusions:
Among experienced evaluators, the GRBAS scale provided high (moderate or above) intra-rater reliability, supporting its use for longitudinal assessments within the same rater. However, the inter-rater reliability remained limited, particularly for R and S. These findings highlight the need for enhanced rater training and the development of complementary objective assessment methods to improve the reproducibility of auditory-perceptual voice evaluation.
