Good test–retest reliability of high frequency (≥8 kHz) pure-tone audiometry (HFPTA) is essential to detect significant changes in hearing threshold caused by ototoxicity. While the test–retest reliability of HFPTA in adults has been extensively studied, such investigations in children are scant. This study aimed to evaluate the test–retest reliability of the HFPTA in normal-hearing children with particular reference to the criteria for ototoxicity.
Participants were 125 children aged between 4 and 13 yr, with normal hearing in the 0.25 to 4 kHz range and normal tympanometric findings. The participants were divided into three age groups, 4 to 6 yr (16M; 16F); 7 to 9 yr (22M; 30F); and 10 to 13 yr (24M; 17F), for investigating possible age effects in the test–retest reliability of HFPTA. The instrumentation for the HFPTA procedure was an Interacoustics AC40 audiometer with Koss R/80 high-frequency headphones, calibrated to meet Australian standards. Hearing thresholds at 8, 9, 10, 11.2, 12.5, 14, and 16 kHz were measured in a sound-treated chamber using a modified Hughson-Westlake procedure with a 5 dB step size. Testing began with an ear and test frequency selected at random; the subsequent test frequencies were randomly selected. After acquisition of hearing threshold data at all frequencies, the other ear was tested using the same procedure. After the first HFPTA test, the headphones were removed and carefully replaced. A second HFPTA test was performed with the ear order reversed. The order of testing the ear for the next participant was reversed.
Good test–retest reliability of HFPTA was achieved with no significant difference in mean HFPTA thresholds across test and retest conditions for all age groups. An age effect in the test–retest reliability of HFPTA was evident with the 4- to 6-yr-old, 7- to 9-yr-old, and 10- to 13-yr-old children demonstrating normal variability of thresholds (within ±10 dB) in 89.9%, 93.0%, and 97% of ears tested, respectively. When the variability of test–retest thresholds was assessed at each frequency, the 4 to 6 yr group showed significantly lower percentage of normal variability at 14 kHz. In identifying significant deterioration of hearing thresholds across test–retest conditions in relation to the ASHA (1994) ototoxicity criteria, the three age groups (youngest to oldest) demonstrated false-positive rates of 24.6%, 11%, and 7.6%, respectively.
Overall, this study demonstrated high test–retest reliability of HFPTA in all but the 4 to 6 yr group. With a false-positive rate of 24.6% for ototoxicity for the youngest group, it is recommended that the HFPTA should not be used alone in assessing the possibility of a genuine threshold shift for this age group. If possible, the HFPTA should be supplemented with an objective test of auditory function to confirm the diagnosis. For children aged 7 yr or older, the HFPTA test is promising as a useful tool to identify hearing impairment in the extended high-frequency range (>8 kHz). However, interpretation of HFPTA findings in serial testing for monitoring hearing in a child should be made with due attention being given to the frequency of the stimulus, age of the child, and the associated nonzero false-positive rates.