Assessing Agreement between Frequency-Specific Chirp Auditory Steady-State Response and Pure Tone Audiometry in Adults by Intraclass Correlation Coefficient.

INTRODUCTION

Chirp auditory steady-state response (ASSR) can be used to assess frequency-specific hearing thresholds. However, its reliability has not been confirmed yet. The purpose of this proposed study is to analyze the agreement of thresholds measured by chirp-ASSR and pure tone audiometry (PTA) to investigate the value of chirp-ASSR in hearing threshold evaluation.

METHODS

Participants with normal hearing (age: 18-66, 108 ears) and patients with sensorineural hearing loss (age: 22-82, 75 ears) were tested using PTA and chirp-ASSR at 0.5, 1, 2, and 4 kHz, respectively. Intraclass correlation coefficient (ICC) and Bland-Altman plot were introduced to analyze the agreement between the 2 methods.

RESULTS

One-hundred eight participants underwent both chirp-ASSR and PTA to estimate their thresholds. The ICCs yielded by these 2 methods are 0.757, 0.893, 0.883, and 0.921 (p < 0.001) at 0.5, 1, 2, and 4 kHz carrier frequency, respectively. However, there is a significant difference between the 2 methods at 2 kHz: the mean value of the ASSR thresholds is 5.27 dB HL higher than the value of PTA thresholds. Additionally, the 95% limits of agreement range from -27.48 to 26.66 dB HL at 0.5 kHz, from -18.19 to 17.87 dB HL at 1 kHz, from -12.01 to 22.55 dB HL at 2 kHz, and from -21.29 to 19.17 dB HL at 4 kHz, which are large enough to affect clinical decision-making.

CONCLUSION

In this study, we have confirmed good to excellent correlation between chirp-ASSR and PTA in threshold estimation for adults with and without hearing loss. The degree of correlations is higher for participants with hearing loss and for measurements at high frequencies. However, significant systematic difference and large limits of agreement between the 2 methods have been found. These findings show that chirp-ASSR can be treated as a supplementary method to PTA when evaluating the hearing level, but the 2 methods are not interchangeable due to their systematic difference and large limits of agreement.