Test-Retest Reliability of Audiometric Assessment in Individuals With Mild Dementia

Accurate assessment of hearing is critically important regardless of a person's cognitive ability. The degree to which hearing can be reliably measured in adults with mild dementia has not been determined.

To obtain quantitative measures of reliability to evaluate the degree to which audiologic testing can be accurately conducted in older adults with mild dementia.

This repeated-measures cross-sectional study consisted of a comprehensive audiologic assessment on 2 occasions separated by 1 to 2 weeks performed in the department of otolaryngology at the Washington University School of Medicine from December 3, 2018, to March 4, 2020. Participants were 15 older adults with a verified diagnosis of mild dementia and 32 older adults without a verified diagnosis of mild dementia who were recruited from the Knight Alzheimer Disease Research Center at Washington University in St Louis.

Test-retest reliability was assessed for tympanometry, acoustic reflex thresholds, otoacoustic emissions, hearing sensitivity, speech reception threshold, speech perception in noise, and hearing handicap, using standard clinical audiology measures.

A total of 47 older adults (26 women; mean [SD] age, 74.8 [6.0] years [range, 53-87 years]), including 32 with normal cognitive function and 15 with very mild or mild dementia, completed the study protocol. For participants with mild dementia, high test-retest reliability (Spearman ρ > 0.80) was found for most measures typically included in a comprehensive audiometric evaluation. For acoustic reflex thresholds, agreement was moderate to high, averaging approximately 83% across frequencies for both groups. Scores for the screening Hearing Handicap Inventory for the Elderly at time 1 and time 2 were highly correlated for the group with normal cognitive function (r = 0.84 [95% CI, 0.70-0.93]) and for the group with mild dementia (r = 0.96 [95% CI, 0.88-0.99]). For hearing thresholds, all rank-order correlations were above 0.80 with 95% CIs at or below 15% in width, with the exception of a moderate correlation of bone conduction thresholds at 500 Hz for the group with normal cognitive function (r = 0.69 [95% CI, 0.50-0.84]) and slightly wider 95% CIs for low-frequency bone conduction thresholds for both groups. For speech reception thresholds, correlations were high for groups with normal cognitive function (r = 0.91 [95% CI, 0.84-0.95]) and mild dementia (r = 0.83 [95% CI, 0.63-0.94]).

Test-retest reliability for hearing measures obtained from participants with mild dementia was comparable to that obtained from cognitively normal participants. These findings suggest that mild cognitive impairment does not preclude accurate audiologic assessment.