Evaluating a dichotomized measure of self-reported hearing loss against gold standard audiometry: prevalence estimates and age bias in a pooled national data set

Objectives: To evaluate a harmonized binary measure of self-reported hearing loss against gold standard audiometry in an older adult population. Method: Seven nationally representative population-based studies were harmonized and pooled (n = 23,001). Self-report items were recoded into a dichotomous format. Audiometric hearing loss was defined by averaged pure-tone thresholds greater than 25-decibel hearing level in the better ear. We compared age and sex stratified prevalence rates of hearing loss estimated by self-report and audiometric measures. Results: Overall, 56% of men and 43% of women had audiometric hearing loss. There were moderate associations between self-reported and audiometric hearing loss. However, prevalence based on self-report was overestimated for adults aged below 70 years and underestimated for adults aged above 75. Discussion: Self-report of hearing loss is insensitive to age effects and does not provide a reliable basis for estimating prevalence of age-related hearing loss, although may indicate perceived hearing disability.