RESUMO
OBJECTIVE: This study evaluated the levels of intraobserver and interobserver agreement for measurements of visuoperceptual variables in videolaryngostroboscopic examinations and compared the observers' behavior during independent versus consensus panel rating. STUDY DESIGN: This is a retrospective study. SETTING: This study was conducted in a single-center tertiary care facility. PARTICIPANTS: Sixty-four patients with dysphonia of heterogeneous etiology were included. EXPOSURE: All subjects underwent a standardized videolaryngostroboscopic examination. MAIN OUTCOME AND MEASURES: Two experienced and trained observers scored exactly the same examinations, first independently and then on a consensus panel. Specific visuoperceptual variables and the clinical diagnosis (as recommended by the Committee on Phoniatrics and the Phonosurgery Committee of the European Laryngological Society and advised by the American Speech-Language-Hearing Association) were scored. Descriptive and kappa statistics were used. RESULTS: In general, intraobserver agreement was better than agreement between observers for measurements of several variables. The intrapanel observer agreement levels were slightly higher than the intraobserver agreement levels on the independent rating task. When rating on the consensus panel, the observers deviated considerably from the scores they had previously given on the independent rating task. CONCLUSION AND RELEVANCE: Observer agreement in videolaryngostroboscopic assessment has important implications not only for the diagnosis and treatment of dysphonic patients but also for the interpretation of the results of scientific studies using videolaryngostroboscopic outcome parameters. The identification of factors that can influence the levels of observer agreement can provide a better understanding of the rating process and its limitations. The results of this study suggest that future research could achieve better agreement levels by rating the visuoperceptual variables in a panel setting.
