2022 NHCA Annual Virtual Conference

There is a long history of using audiometric criteria to identify people likely to experience hearing trouble in daily life. Some audiometric criteria favor sensitivity (i.e., maximizing correct identification of people reporting hearing trouble), while others favor specificity (i.e., minimizing misidentification of people without reported hearing trouble as hearing impaired). Positive predictive value (PPV, i.e., likelihood of criterion accurately predicting hearing trouble) depends on high specificity. The difference between the population prevalence of hearing trouble (16% in those aged 20-49 years) and the PPV is an indication of a criterion’s usefulness. In this presentation, we report sensitivity, specificity, and predictive values for various audiometric criteria, ranging from highly sensitive (e.g., any audiometric threshold worse than 15 dB HL) to highly specific (e.g., pure tone average thresholds worse than 40 dB), using data generalizing to the U.S. population (National Health and Nutrition Examination Survey - NHANES). The most sensitive criteria yield PPVs approximating the population prevalence of self-reported hearing trouble, meaning that positive decisions based on those criteria are as accurate as using no criteria. The preponderance of people identified using these criteria do not report hearing trouble, so these criteria cannot be used to predict hearing trouble with reasonable confidence.