The effect of speaker-specific information on perceptual voice rating tasks
MetadataShow full item record
The effect of speaker-specific information on perceptual voice rating tasks Cara Sauder Chair of the Supervisory Committee: Professor Tanya Eadie Speech & Hearing Sciences Auditory-perceptual voice assessment and visual-perceptual judgments of videolaryngostroboscopic exams are among the most common and important components of a comprehensive clinical voice assessment. Findings from these perceptual voice assessments are used to determine the presence, severity, and/or nature of a voice disorder. However, perceptual measures are inherently subjective. Error and variability among clinicians are a threat to the validity of these important voice assessment measures. One potential source of variability that has not been systematically investigated relates to what is known about a speaker during perceptual voice assessment. For example, exposure to speaker-specific information (e.g., referring diagnosis, case history, voice quality, etc.) may be controlled in a laboratory setting, but is common when perceptual rating tasks are accomplished in a clinical setting. Understanding the effects of speaker-specific information on perceptual voice assessment measures is important for designing clinical and research protocols and for interpreting results from the existing literature. Because findings from a comprehensive voice assessment are used to form clinical impressions, determine a clinical diagnosis, and/or make treatment recommendations, it is also important to consider the overall effect of speaker-specific information on clinical decisions. This document includes a review of the literature (Chapter 1) on the effects of speakerspecific information on auditory- and visual-perceptual videolaryngostroboscopic rating tasks and clinical judgments about voice disorder diagnosis and treatment. Chapter 1 also describes how an existing conceptual model used to explain how an acoustic signal is mapped to auditoryperceptual voice ratings may be extended to visual-perceptual ratings of videolaryngostroboscopy (VLS). Next, three original studies are described in which speakerspecific information that suggests the presence/absence and nature of a voice disorder are provided during auditory-perceptual rating tasks (Chapter 2) and visual-perceptual VLS rating tasks (Chapter 3 and Chapter 4). In Chapter 2, the effect of knowledge of accurate and inaccurate referring diagnosis on auditory-perceptual ratings was investigated. In Chapter 3, the effect of accurate and inaccurate case histories suggesting a particular voice disorder etiology or the absence of a voice disorder (control) on visual-perceptual VLS ratings and other clinical judgments was evaluated. In Chapter 4, the effect of “accurate” auditory cues suggesting different levels of dysphonia severity on visual-perceptual VLS ratings and other clinical judgments was evaluated. The results suggest that the small observed differences in average auditory-perceptual ratings when an accurate referring medical laryngeal diagnosis was present versus absent were unlikely to be clinically meaningful. However, the overall effect of the same referring medical laryngeal diagnoses when they were inaccurate, or inconsistent with the speaker’s true medical laryngeal diagnoses, was greater. These differences in perceived severity were also potentially clinically meaningful. These findings are important because referring medical laryngeal diagnoses are commonly inaccurate in a clinical setting. In Chapter 3, there were also observed differences in the probability of rating VLS parameters as more severe for two of three outcome measures when a preliminary case history (accurate or inaccurate) suggested its presence. Agreement in clinical impressions was increased when case histories were accurate versus inaccurate, and there was also an effect of speaker-specific information on treatment recommendations. Chapter 4 revealed no clinically meaningful effect of the presence of speaker-specific information about dysphonia severity during clinicians’ evaluation of VLS exams on ratings of VLS parameters. Although some trends in diagnostic codes and treatment recommendations were observed when auditory cues were present versus absent, overall agreement in diagnostic coding and treatment recommendations when auditory cues were absent versus present was high. Findings across these three studies are integrated and summarized in Chapter 5. The results support continued investigations into factors affecting variability in perceptual voice rating tasks and clinical judgments to guide decisions about the best way to perform these important clinical voice assessments in clinical and laboratory settings. Findings are also useful for interpreting and evaluating the external validity of the existing literature.
- Speech