Highly variable population-based prevalence rates of unilateral hearing loss after the application of common case definitions.

Objectives: This study shows how population-based estimates of the prevalence of unilateral hearing loss (UHL) in children aged 6 to 19 yrs can differ considerably with various applications of commonly accepted case definitions. It also examines demographic variables and risk factors related to UHL. Design: The Third National Health and Nutrition Examination Survey, conducted from 1988 to 1994, is a national population-based, cross-sectional survey. This study examined results of audiometric testing at 0.5 to 8 kHz and demographic data from in-person examination interviews. Three definitions of UHL were used: (1) 0.5, 1, and 2 kHz >= 15 dB pure-tone average (PTA); (2) 0.5, 1, 2, and 4 kHz >= 15 dB PTA; and (3) 0.5, 1, and 2 kHz >= 20 dB or PTA >25 dB at two or more frequencies above 2 kHz (3, 4, 6, and 8 kHz). Case definitions 2 and 3 are not merely subsets of case definition 1. Some overlap exists between the groups, but each case definition classifies a proportion of children who fall uniquely under that case definition. Inclusion of participants based on tympanometry results (test of middle ear function) was also examined as were demographic characteristics and risk factors associated with UHL. Results: Overall, the weighted proportion of children with UHL using case definition 1 was 6.3% (approximately 3,213,000 children nationally); using case definition 2, it was 5.8% (approximately 2,958,000 nationally); using case definition 3, it was 3.0% (approximately 1,530,000 nationally). For all three case definitions, children who failed tympanometry were at higher risk for UHL than children who passed. For case definition 2, children from rural areas were at higher risk for UHL than were children from urban areas. Conclusions: This study demonstrates that different applications of well-accepted case definitions of UHL can influence population-based prevalence estimates, in this study by as much as a factor of 2. These findings highlight the importance of controlling for tympanometry status as a risk factor in such estimates. Which demographic characteristics and risk factors are significantly associated with hearing loss seem to vary depending on the case definition. These findings have implications for the interpretation of prevalence rates and risk factors in the literature on hearing loss in general. Prevalence rate estimates require careful consideration of the case definition of hearing loss, tympanometry status, and demographic characteristics.