To evaluate the efficiency of four self-assessment questionnaires that rate the impact of a voice problem on the individual's life: Voice-Related Quality of Life (V-RQOL), the original and reduced versions of the Voice Handicap Index (VHI) and VHI-10,
Vocal Performance Questionnaire (VPQ), and Voice Symptom Scale (VoiSS).
Methods
Data from 975 subjects, 486 with a diagnosis of dysphonia and 489 vocally healthy individuals, were submitted to the receiver operating characteristic (ROC) curve analysis to obtain the cutoff values that determine the discriminating power of these instruments (presence of dysphonia vs healthy voice).
Results
The ROC curve analysis showed that the most efficient questionnaires were the VoiSS and the VHI. Results showed that they presented as a perfect classification based on their efficiency, specificity, and sensitivity values (all three of them = 1). The VHI-10 and the V-RQOL showed excellent classification (VHI-10: efficiency = 0.991; specificity = 1; sensitivity = 0.981; V-RQOL: efficiency = 0.914; specificity = 0.860; sensitivity = 0.967). Finally, the VPQ showed a good level of classification (efficiency = 0.828; specificity = 0.824; sensitivity = 0.831). The cutoff values for the instruments are as follows: VoiSS = 16 points, VHI = 19 points, VHI-10 = 7.5 points, V-RQOL = 91.25, and VPQ = 20.5 points. These values are important for screening large populations as well as for helping in the decision-making process of clinical management. The cutoff values for maximum sensitivity and specificity of the instruments that did not produce perfect classification are as follows: VHI-10: sensitivity = 5; specificity = 7.5, V-RQOL: sensitivity = 86.25; specificity = 98.75, and VPQ: sensitivity = 15.5; specificity = 31.5.
Conclusions
Both the VoiSS and the VHI are perfect classifiers. The VHI-10 and the V-RQOL are excellent classifiers, and the VPQ is good at discriminating individuals with dysphonia from the ones without dysphonia.
