Validity and Reliability of Whispered Voice Test as a Screening Test for Hearing in the Elderly People

Abstract

Background and Objectives :The purpose of this study is to evaluate the validity and reliability of the whispered voice test, which has been suggested as a screening test tool for hearing in the elderly group.

Subjects and Method :A prospective multicenter trial has been designed. One hundred and seventy two elderly subjects from four university hospitals were included in this study. A whispered voice test using a number list and number/word list was performed by two testers for each elderly subject. To evaluate their test-retest reliability, a second whispered voice test was done 1-2 weeks later. The result of the whispered voice test was analyzed with the criteria of 26 dB and 40 dB hearing level, checked by pure tone audiometry. Sensitivity, specificity, and cut-off point of whispered voice test at each gold standard, inter-observer reliability and test-retest reliability have been investigated.

Results:There were significant differences with respect to correctly answered numbers in the whispered voice test between the normal and impaired hearing group at both criteria of 26 dBHL and 40 dBHL. With the 26 dB criteria of hearing level which showed higher inter-observer reliability of the number/word list, the whispered voice test has 86.0% sensitivity and 58.9% specificity with the cut-off point of 3 out of six on the number set. The test-retest reliability of whispered voice test using the number list was as high as 0.9 of correlation coefficient. However, the inter-observer reliability of whispered voice test was less than 0.6 (Cohen’s kappa value).

Conclusions:As a screening test for hearing in the elderly group, the whispered voice test has quite high sensitivity and test-retest reliability. The environmental sound level and test method for each tester, however, should be monitored to get better specificity and inter-observer reliability.