Protocolo de avaliação vocal espectrográfica: desenvolvimento e validação

Abstract

Objective: To develop and search for evidence of validity and reliability of the spectrographic vocal evaluation protocol. Methods: Methodological study for development and validation of an evaluation instrument. Extensive literature research has been conducted to develop and substantiate the Spectrographic Vocal Assessment Protocol (PAVE). The PAVE items were sent to five expert judges, non- research speech therapists, doctors and clinicians, with minimum experience of five years in the field of voice and spectrographic acoustic analysis, who individually analyzed the protocol items. For the analysis of concordance between the judges, the Gwet analysis and the Content Validity Ratio formula (CVR) were used. For reliability analysis, two audiologist judges analyzed with the PAVE the spectrographs of 50 women aged 25 to 62 years, 25 with previous otolaryngological diagnosis of normal larynx or grade I and 25 triangular cleft diagnosed with vocal fold alterations. For spectrographic evaluation, emission of the vowel /a:/ was analyzed by the Real Time Spectrogram. For reliability data reliability, inter and intra-rater analysis was performed using the Kendall coefficient. For criterion validity, the t test for independent samples was used to compare the average performance of the protocol between the two groups of subjects. Results: The PAVE was developed based on the literature, subdivided into Broadband and Narrowband. As for the analysis expert judges twenty-four items of the protocol had CVR=1,0 and 25 items had CVR=0,6. Gwet's AC1 analysis indicated substantial agreement (0,656) for broadband and near perfect (0,888) for narrowband items. The reliability of the protocol for broadband items, concordance among judges ranged from moderate (0,462) to almost perfect (0,888), being significant for all items, and intra-rater concordance was almost perfect (0,865-1,000) and significant for all items. For narrowband items, concordance between judges varied from moderate (0,469) to almost perfect (0,922), being significant for all items, except for the item "presence of sub-harmonics in high frequencies", where agreement was moderate (0,504), but not significant, and the intra-rater agreement was almost perfect (0,874-1,000) and significant for all items. Validity of criterion: for most broadband and narrow items, groups with normal or altered diagnosis differed statistically. Conclusion: PAVE presents evidence of satisfactory content validity, criterion and reliability, being reliable for the spectrographic vocal evaluation.