Objective: To compare the AVQI and ABI indices results in different sizes of continuous speech tasks.
Method: This is a cross-sectional analytical observational study with a convenience sample composed of 86 teachers from public education (48 women and 38 men, age range 27-55 years), speakers of Brazilian Portuguese. All teachers underwent a videolaryngoscopy and auditory-perceptual assessment and presented laryngeal examination and voice quality without disorders. The voice and speech samples used for the AVQI and ABI (version 03.01) scores were obtained from the mid-three seconds of a sustained vowel [a] and a spontaneous count from 1 to 20, at the usual frequency and intensity. Before running the AVQI and ABI scripts, the continuous speech sample was edited using Praat software (version 6.1.47), in eleven stimuli of different sizes: counting from 1 to 10 (15 syllables) until 1 to 20 (42 syllables). Statistical analysis was performed using the MINITAB 17. A descriptive analysis of the data was carried out with measures of central tendency and dispersion. It was considered that the count from 1 to 10 represents the moment 1, the count from 1 to 11 represents the moment 2, and so on until the count from 1 to 20, which relates the moment 11. Friedman and Tukey Pairwise test was used with a confidence interval of 95%.
Results: The speech task increase generated a statistically significant difference in AVQI scores, from the count of 1 to 17 (33 syllables) (p-valor _ 0.000) for the men’s group and from 1 to 16 (29 syllables) (p-valor _ 0.000) for women’s group. ABI scores didn’t show significant differences with the speech task increase for the women’s group (p=0.081) and the men’s group (p=0.089).
Conclusion: In Brazilian Portuguese, the speech task size influenced the AVQI results, reinforcing the importance of using standardized speaking tasks, to increase the voice analysis reliability. The ABI was stable, regardless of the task size, in both sexes.