Screening for Voice Disorders in Older Adults (RAVI)—Part III: Cutoff Score and Clinical Consistency

Abstract

The aim of the study was to determine the cutoff score and clinical consistency of "Screening for Voice Disorders in Older Adults" (RAVI-Rastreamento de Alterações Vocais em Idosos). This is a prospective, nonrandomized, cross-sectional diagnostic study. A sample of 301 subjects, including both sexes, aged 60 and more, and all of whom were living in either a community or an institution, was studied. To determine which subjects had or did not have voice problems, we used a composite reference standard (auditory-perceptual analysis of sustained vowel phonation, auditory-perceptual analysis of connected speech, and vocal self-assessment). The best cutoff score was identified using the receiver operating characteristic (ROC) curve analysis. The clinical consistency indicators were co-positivity, co-negativity, positive and negative predictive values, positive and negative likelihood ratio, and test efficiency. The significance level was 5%. The area under the ROC curve was 0.763 (95% confidence interval: 0.706-0.821), and the best cutoff score for determining which older adults had or did not have a voice disorder was 2. All clinical consistency indicators were satisfactory: co-positivity (79%), co-negativity (60%), predictive positive value (51%), negative predictive value (84%), positive likelihood ratio (2.01), negative likelihood ratio (0.34), and test efficiency (69%). RAVI has satisfactory indicators of clinical consistency and is able to determine which older adults have voice disorders by a cutoff score of 2. The use of RAVI as a screening tool is recommended to help determine the prevalence of voice disorders in older adults.