Test-retest reliability of wideband reflectance measures in infants under screening and diagnostic test conditions.

Objective: The main goal of this study was to examine the test-retest reliability of wideband reflectance (WBR) measures collected from infants in screening and diagnostic hearing test environments. In addition, the results of WBR testing for infants who passed and failed otoacoustic emission (OAE) screening were examined to determine whether these measures distinguished between the two groups.

Design: Repeated WBR measures were collected from two groups of infants, one group tested in an outpatient hearing screening setting and the other group in a diagnostic test setting. For a total of 127 infants and a control group of 10 adults, repeated WBR measurements were made with the probe left in place between the two tests (T1 and T2) and after reinsertion of the probe (T3) for a total of 3 measurements. Test-retest differences were calculated for each individual across one-third octave frequency bands, and the mean and 90th percentile were calculated by subject group and OAE results. WBR patterns were also compared between infants who passed versus failed OAE screening.

Results: Mean test-retest differences were smaller for the diagnostic group than for the screening group. Test-retest differences were largest for the reinsertion condition and for the frequencies below 500 Hz. While the low frequencies were variable, the test-retest differences were smallest in the mid-frequency range which is thought to be the frequency range most sensitive to middle ear dysfunction. Test-retest performance did not differ between infants who passed or failed OAE screening. However, infants who failed OAE screening had significantly higher WBR in the range from 630 to 2000 Hz than infants who passed OAE screening.

Conclusions: Test-retest performance was poor for frequencies below 500 Hz, but in general test-retest differences were small across the important mid-frequency range. Reinsertion of the probe between repeated tests yielded larger and more variable test-retest differences. Careful monitoring of probe fit and testing while infants are in a quiet state appears to be critical for obtaining reliable WBR results. Analysis of WBR results indicated significantly higher reflected energy in the mid-frequency range for infants who failed OAE screening than for those who passed OAE screening. Although conclusions are limited by the fact that the true status of the middle ear and cochlea were not known for the infants in this study, this result may indicate that a number of these infants failed OAE screening due to transient or permanent middle ear dysfunction which was detected by WBR.