Análise fatorial e aplicação da teoria de resposta ao item na escala do desconforto do trato vocal em pacientes com disfonia

Abstract

Investigating and knowing the vocal symptoms of the patient allows the speech-language pathologist a more complete view of their needs, since not always the vocal alteration identified by the clinician is perceived with the same intensity by the patient. In this context, the Vocal Tract Discomfort Scale (EDTV) was developed with the objective of measuring the intensity and frequency of vocal tract discomfort symptoms, filled by the individual himself, through the use of qualitative descriptors. The Item Response Theory (TRI) can be used in the validation of the EDTV in a satisfactory way, considering each item individually, without prioritizing the total scores to characterize the attribute studied. Thus, the objective of this study was to develop a decision model based on the application of the Item Response Theory (TRI) in the Vocal Tract Discomfort Scale (EDTV) protocol. It is a documentary, quantitative, descriptive and cross-sectional research. The database (BD) of this research was formed from the insertion of the demographic variables (sex, age, profession), laryngeal diagnosis, auditory perceptive evaluation performed by a specialist voice judge and EDTV item item. All these variables were extracted from the charts of 310 patients belonging to the Integrated Laboratory of Voice Studies (LIEV). Descriptive and inferential statistical analysis was carried out, from the Exploratory Factor Analysis (AFE) and Confirmatory (AFC) and the TRI. The sample of participants was composed of patients of both sexes, with a mean age of 39.36 (SD = 15.80), being the majority female and non-professional voice. It was observed, through the AFE, that the items of the EDTV were organized in four factors, according to inter-agency correlations. Regarding the AFC results, the EDTV items were organized into three factors, since the last two factors found in the AFE correlated and constituted a single factor. Thus, as a final solution of the AFC, the EDTV was organized in: Factor 1 called "Hybrid" formed by the items corresponding to the frequency and intensity of the symptoms of burning (F1 and I1), sore throat (F4 and I4) and irritated throat (F7 and I7); Factor 2, called "Inflammatory process / tissue injury", formed by the frequency and intensity of symptoms of itching (F5 and I5) and sensitive throat and Factor 3, "Musculoskeletal discomfort", formed by the items of frequency and intensity of the tightening symptom (F2 and I2), dryness (F3 and I3) and ball in the throat (F8 and I8). Regarding the TRI data, it is possible to affirm that the EDTV items, from the result of their parameters of difficulty and discrimination, were adequate, even taking into account that some items had to be excluded from the analysis, this may have influenced in the ability of some factors to cover the entire extent of the latent trait. In general, the research was completed with two factors that presented the best psychometric parameters, "Vocal hyperfunction with phonotraumatic lesion" and "Vocal hyperfunction without phonotraumatic lesion", and the items excluded from the analysis - itch and sensitive throat - were the factor Inflammatory process / tissue injury, which were not included in the TRI analysis. It is concluded that the EDTV presents itself as an instrument that brings the understanding of discomfort and can be considered valid and accurate for the vocal selfevaluation, being able to be used for research purposes and complement in the diagnosis with items with good psychometric qualities.