BACKGROUND
In-situ audiometry is a hearing aid feature that enables the measurement of hearing threshold levels through the hearing instrument using the built-in sound generator and the hearing aid receiver. This feature can be used in hearing aid fittings instead of conventional pure-tone audiometry (PTA), particularly in places where no standard audiometric equipment is available. Differences between conventional and in-situ thresholds are described and discussed for some particular hearing aids. No previous investigation has measured and compared these differences for a number of current hearing aid models by various manufacturers across a wide range of hearing losses.
PURPOSE
The purpose of this study was to perform a model-based comparison of conventionally and in-situ measured hearing thresholds. Data were collected for a range of hearing aid devices to study and generalize the effects that may occur under clinical conditions.
RESEARCH DESIGN
Research design was an experimental and regression study.
STUDY SAMPLE
A total of 30 adults with sensorineural hearing loss served as test persons. They were assigned to three subgroups of 10 subjects with mild (M), moderate to severe (MS), and severe (S) sensorineural hearing loss.
INTERVENTION
All 30 test persons underwent both conventional PTA and in-situ audiometry with four hearing aid models by various manufacturers.
DATA COLLECTION AND ANALYSIS
The differences between conventionally and in-situ measured hearing threshold levels were calculated and evaluated by an exploratory data analysis followed by a sophisticated statistical modeling process.
RESULTS
At 500 and 1500 Hz, almost all threshold differences (conventional PTA minus in-situ data) were negative, i.e., in the low to mid frequencies, hearing loss was overestimated by most devices relative to PTA. At 4000 Hz, the majority of differences (7 of 12) were positive, i.e., in the frequency range above 1500 Hz, hearing loss was frequently underestimated. As hearing loss increased (M→MS→S), the effect of the underestimation decreased. At 500 and 1500 Hz, Resound devices showed the smallest threshold deviations, followed by Phonak, Starkey, and Oticon instruments. At 4000 Hz, this observed pattern partly disappeared and Starkey and Oticon devices showed a reversed effect with increasing hearing loss (M→MS→S). Because of high standard errors for the estimates, only a few explicit rankings of the devices could be established based on significant threshold differences (5% level).
CONCLUSIONS
Differences between conventional PTA and in-situ threshold levels may be attributed to (1) frequency, (2) device/hearing loss, and (3) calibration/manufacturer effects. Frequency effects primarily resulting in an overestimation of hearing loss by in-situ audiometry in the low and mid frequencies are mainly due to sound drain-off through vents and leaks. Device/hearing loss effects may be due to leakage as well as boundary effects because in-situ audiometry is confined to a limited measurement range. Finally, different calibration approaches may result in different offset levels between PTA and in-situ audiometry calibration. In some cases, the observed threshold differences of up to 10-15 dB may translate to varied hearing aid fittings for the same user depending on how hearing threshold levels were measured.