RESUMO
OBJECTIVES: The purpose of this study was to investigate group differences between singing teachers and speech-language pathologists when rating dysphonic speaking voices and whether training using reference samples and the Consensus Auditory-Perceptual Evaluation of Voice (CAPE-V) improves inter-rater reliability between and among the two groups. Differences in use of terminology and ratings could reveal potential for miscommunication in the team treatment of singers with voice disorders. STUDY DESIGN: This is a prospective pre test post test cohort study with between and within group comparisons. METHODS: Recorded samples of dysphonic speaking voices were rated by 18 experienced singing teachers with free written descriptions and an Overall Severity (OS) rating of 0-100. Participants were then trained in application of the CAPE-V with verbal definitions and reference samples exemplifying characteristics of disordered voice. Participants rated the samples a second time using the CAPE-V. The pre and post training ratings of participants were compared to composite ratings of six speech-language pathologists. RESULTS: Descriptive statistics indicated the mean aggregate Overall Severity rating of speech-language pathologist (SLP) raters as 25.79 (SD = 6.10, SE = 2.49), as compared to 35.05 (SD = 12.72, SE = 3.00) for singing teachers. Differences in ratings were more pronounced in samples rated by SLPs as "mild" (OS 6-20) or "mild-moderate" (OS 21-35). ANOVA revealed statistically significant group differences between SLPs and singing teachers for the parameters Overall Severity (P = 0.0109, F = 7.8) and Strain (P = 0.0085, F = 8.35). While CAPE-V training did not significantly change the OS ratings of singing teachers, it did improve their inter-rater reliability from 0.67 pre training to 0.83 post training, with agreement similar to that of SLP raters (0.86). After training, participants responded "yes" to the presence of dysphonia in disordered samples more frequently. CONCLUSIONS: The results support the recommendation of training singing teachers in perceptual evaluation of speaking voice to increase sensitivity to the presence of organic voice disorders and to encourage compatibility in terminology used among SLPs and singing teachers.
RESUMO
The purpose of this nonrandomized prospective study was to quantify the inter- and intrarater reliability of experienced speech-language pathologist's perceptual ratings of voice in pediatric patients post-laryngotracheal reconstruction (LTR). Moderate to severe dysphonia is common in this population. Using the sentence portion of the Consensus Auditory Perceptual Evaluation-Voice (CAPE-V) rating scale, three experienced speech-language pathologists independently rated randomized voice samples of 50 participants ages 4-20 years, who had acquired or congenital airway conditions requiring at least one LTR on the six salient perceptual vocal attributes. Data collection and listening conditions were carefully controlled. Seventeen (34%) of the samples were randomly selected for rerating at a later time. Estimates of interrater reliability were strongest for perceptual ratings of breathiness (intraclass correlation coefficient [ICC]=71%), roughness (ICC=68%), pitch (ICC=68%), and overall severity (ICC=67%). Reliability was lower for ratings of loudness (ICC=57%) and strain (ICC=35%). For each rater, the intrarater reliability on all but one parameter (strain) was moderate to strong (ICC=63-93%). There was a strong interrater eliability for four of six vocal parameters rated using the CAPE-V in a population of children and adolescents with marked dysphonia. The parameter of strain, when rated by auditory sample alone and apart from the clinical context, was difficult to rate.