Determining Medical Urgency of Voice Disorders Using Auditory-Perceptual Voice Assessments Performed by Speech-Language Pathologists

Abstract

Objective: This study examined whether speech-language pathologist auditory-perceptual voice assessments can predict the medical urgency of voice disorders. Methods: Twenty speech-language pathologists (SLPs) evaluated 25 voice samples recorded during initial voice evaluations. Voice samples represented a range of dysphonia severity (mild-severe) balanced across patient diagnoses. Diagnoses included: benign lesions, laryngeal cancer, non-organic voice disorders, laryngeal edema (associated with LPR), and laryngeal paralysis or paresis. Laryngeal cancer and severe unilateral laryngeal paralysis were considered urgent disorders. While blinded to patient information, SLPs rated severity of voice quality, predicted patient diagnosis, and determined whether the patient should be seen urgently by a laryngologist. SLPs were then given basic medical history information and rated medical urgency of voice disorder a second time. Results: On average, SLPs correctly identified 65% of urgent voices and 87% of nonurgent voices when blinded to patient information. Accuracy improved significantly to 86% for urgent voices with medical history information ( P < .001) and decreased to 77% for nonurgent voices. Accuracy was better when severity of voice quality was severe for urgent voices and mild for nonurgent voices ( P < .001). SLPs indicated that patient smoking history and severity of dysphonia were most influential in their decision making. Diagnostic accuracy of auditory-perceptual assessments was poor. Conclusions: SLPs identified 86% of medically urgent voice disorders when auditory perceptual assessments were combined with medical history information. Further work is needed to determine what medical history information is most crucial to rating accuracy and what speech tasks might best separate urgent and nonurgent patients.