Distúrbio de voz em professores: identificação, avaliação e triagem

Abstract

The voice is the main work tool for teachers. Correct identification of a disorder through eficiente methods of assessment and screening will contribute to planning actions and devising specific public policies for this population. Aim: to analyze the different methods of identification and assessment and to propose a screening instrument for voice disorders in teachers. Method: The subjetcs of the study are 252 female teachers of the public school system of São Paulo, who had their voices recorded and analyzed by three speech-language pathologists using the GRBASI scale to classify their voice qualities (with or without disorder). In the same day of voice recording, subjects underwent vocal fold examinations, and the responsible doctor classified the images (with or without disorder). The teachers also answered questions in the Vocal Aspects Domain of the Teacher Voice Production Conditions questionnaire (CPVP) and the Voice Handicap Index (VHI). The data were submitted to statistical analysis, conducted with two purposes: 1) to verify the agreement between voice perceptive-auditory and vocal fold assessment, and 2) to devise and validate a screening instrument for voice disorders in teachers. In order to study the agreement in between both assessments, Kappa s correlation coefficient was calculated with level of significance at p ≤0.05. To devise the screening instrument, the data of 130 teachers (sample A) with and without voice disorder were used. An exploratory factorial analysis was conducted from the list of 21 symptoms originally in the Vocal Aspects domain of the CPV-P, and the items with correlation coefficient greater than 0.50 were chosen. A ROC curve was plotted in order to identify the best cut-off point to select teachers who may have a voice disorder and Cronbach s alpha coefficient (α) was calculated in order to assess the score- s internal consistency. The external validation of the instrument was conducted using the data of the 122 remaining subjects (Sample B). Each teacher s score was calculated and the internal consistency of the data was analyzed using Cronbach s alpha coefficient (α). Next, the mean scores and 95% confidence intervals were calculated. The association between the scores and assessment using GRBASI scale was calculated using the chi-square test (p≤0,05), and concurrent validity was analyzed calculating Spearman s correlation coefficient (r) between the score of the proposed instrument and VHI results. Results: The gross agreement between voice quality and vocal fold assessment was 77.9, and Kappa agreement index between both assessments was k = 0.40 (p < 0.001). The 12 symptoms selected by the factorial analysis to compose the Screening Index for Voice Disorder (SIVD) were: hoarseness, voice loss, breaking voice, low-pitched voice, phlegm, dry cough, cough with secretion, pain when speaking, pain when swallowing, secretion in throat, dry throat and strained speech. Chronbach s alpha coefficient for Sample A was α = 0,86. The score was defined as the simple sum of the number of present symptoms and the cut-off was five points. Chronbach s alpha coefficient for sample B was α = 0.89. For this sample, the instrument s sensitivity with cut-off=5 was 92% and specificity 39%. There was a statistically significant association (p ≤0.001) between the SIVD score and voice quality assessment using the GRBASI scale. There was also statistically significant correlation (p ≤0.001) between the total VHI score and SIVD score, and the same was true for each VHI domain and SIVD score. Final Remarks: The association of perceptive-auditory vocal assessment and vocal fold evaluation should be considered the gold-pattern in voice disorder diagnosis. The SIVD is a valid instrument for screening with a high sensitivity level. Therefore, its use will aid in mapping voice disorders, and in planning public health actions and devising public policies regarding teachers