Virtual audiometric testing using smartphone mobile applications to detect hearing loss

Abstract

Objective: The COVID-19 pandemic drove the need for remote audiometric testing in the form of mobile applications for hearing assessment. This study sought to determine the accuracy of two smartphone-based hearing assessment applications, Mimi and uHear, against the gold standard of in-clinic audiometric testing.

Methods: One hundred patients that presented to clinic for hearing assessment were randomly assigned to take either the Mimi or uHear hearing test alongside standard audiometric testing. Hearing thresholds measured using mobile applications were compared to those from audiometric testing to assess validity. Patient satisfaction was measured using a questionnaire that queried if the app met the user's need, if they would recommend the app to others, and how likely they were to use the app again.

Results: Using Mimi, there were no differences in average hearing levels measured at any frequency when compared to standard audiometric testing. uHear overestimated hearing loss at 500 and 1000 Hz (p < .001 for both) by 5-10 Hz, and underestimated hearing loss at 6000 Hz (p < .001) by 5-10 Hz compared to standard audiometric testing. When stratified by level of hearing impairment, uHear overestimated impairment in those with normal hearing (p < .001). Mimi had higher sensitivity (0.971) and specificity (0.912) for hearing loss (defined as a pure tone average for 500, 1000, 2000, and 4000 Hz greater than 25 dB) than uHear (0.914 and 0.780, respectively). However, uHear outranked Mimi on all three questions in the satisfaction questionnaire (p = .01, p = .03, and p = .02, respectively).

Conclusion: Mimi appears to be a reasonable substitute for standard audiometric testing when individuals cannot present to clinic for gold standard testing; however, the Mimi user experience can be improved.

Level of evidence: Level II.

Keywords: audiometry; hearing loss; hearing test; mobile applications; validation.

FIGURE 1

Bland–Altman plots comparing audiometric results for standard audiometric testing versus the Mimi hearing test app, for both right and left ears. There is slightly more disagreement between audiometric testing and the Mimi app at lower frequencies (250–1000 Hz). At higher frequencies, the Mimi app appears to estimate hearing thresholds well

FIGURE 2

Bland–Altman plots comparing audiometric results for standard audiometric testing versus uHear hearing test app, for both right and left ears. At 250–1000 Hz, uHear appears to systematically overestimate hearing loss by 5–10 Hz. At 2000 Hz, uHear appears to estimate hearing thresholds well. Above 2000 Hz, there is more disagreement between audiometric testing and the uHear app, and uHear appears to generally underestimate hearing loss at these higher frequencies