The Intra-rater and Inter-rater Consistency of GRBAS and CAPE-V in Chinese Context
10.3969/j.issn.1006-7299.2025.01.005
- VernacularTitle:GRBAS和CAPE-V量表中文听感知评估一致性分析
- Author:
Yang LIU
1
;
Hengxin LIU
;
Ge QU
;
Ruixiang LI
;
Dongyan HUANG
Author Information
1. 中国人民解放军总医院第六医学中心/耳鼻咽喉头颈外科医学部/国家耳鼻咽喉疾病临床医学研究中心(北京 100037)
- Publication Type:Journal Article
- Keywords:
GRBAS;
CAPE-V;
Intra-rater consistency;
Inter-rater consistency;
Chinese context
- From:
Journal of Audiology and Speech Pathology
2025;33(1):23-28
- CountryChina
- Language:Chinese
-
Abstract:
Objective To evaluate the intra-rater and inter-rater consistency of GRBAS and CAPE-V in Chi-nese context.Methods Five experts used the"Voice Assessment Assistant"APP to evaluate voice samples extrac-ted from The Sixth Medical Center of Chinese PLA General Hospital Pathological Voice Database V1.0.Fifty-two cases were used for inter-rater consistency analysis,and 38 cases were used for intra-rater consistency assessment.Intraclass correlation coefficient(ICC)was used for intra-rater and inter-rater consistency analysis of CAPE-V,while Cohen's Kappa method was applied for intra-rater consistency analysis of the five features of GRBAS.Inter-rater consistency analysis for GRBAS used Fleiss Kappa method,and Spearman's correlation analysis was conducted for the same features in both scales.Results The results of intra-rater consistency analysis showed that all six fea-tures of CAPE-V exhibited high consistency,with ICC coefficients of overall severity(OS)=0.80,roughness(R)=0.69,breathiness(B)=0.77,strain(S)=0.75,pitch(P)=0.74,and loudness(L)=0.78.For GRBAS,ex-cept for the features G(0.48)and S(0.45)which showed weaker intra-rater consistency,the rest of the features had poor intra-rater consistency.The results of inter-rater consistency analysis showed that the ICC correlation coef-ficients for all features of CAPE-V were greater than 0.85,indicating a high degree of consistency in the scoring of each feature by different raters.For GRBAS,except for the feature G(correlation coefficient=0.48),the rest of the features had correlation coefficients less than 0.40,indicating that the inter-rater consistency was poor for all features except G.When evaluating the same sample with both GRBAS and CAPE-V,the Spearman correlation co-efficients for the overall voice quality OS/G,roughness R,breathiness B,and strain S were 0.89,0.85,0.91,and 0.91,respectively,suggesting a high degree of correlation between the scoring results of the four features in the two scales.Conclusion In the auditory perceptual assessment of Chinese voice samples,the CAPE-V scale has higher intra-rater and inter-rater consistency compared to the GRBAS scale and is more suitable for clinical practice.
