Toward a more quantitative measure to assess severity of dysphonia: preliminary observations.

Rating scales to determine the severity of dysphonia have shown considerable inter-rater variability. The objective of this study, therefore, was to investigate a more quantitative measure to perceptually rate the severity of dysphonia. The study was conducted on 10 participants with abuse-related dysphonia in two conditions, spontaneous speech, and paragraph reading. Four speech-language pathologists, who had experience with voice disorders, and one trained student counted the frequency of the nondysphonic syllables during the above two conditions. The voice samples were both recorded and transcribed for simultaneous analysis. All the recordings were randomized when presented to the raters. The raters circled the nondysphonic syllables while listening to each participant's recording. The nondysphonic syllables were counted, and a percentage of dysphonic severity was calculated for each participant to obtain a dysphonic severity percentage. Specific characteristics (eg, breathiness, noise, strain) were not specifically addressed, as they were components of the percentage of dysphonia obtained. The Cronbach's alpha revealed very high inter-rater reliability and high correlations among the raters for both spontaneous speech and paragraph reading, indicating reduced variability in raters' perceptions. This method appeared to be a more quantitative measure of perceptual ratings than current scales, which use general gradations of dysphonic severity. In addition, a naïve rater was successfully trained to use this method. This technique has the potential to be used in both pre- and posttherapy analysis, as well as during therapy, to determine progress.