Diagnostic Accuracy of the AzBio Speech Recognition in Noise Test

Purpose The goal of this study was to determine the ability of the AzBio speech recognition in noise (SRN) test to distinguish between groups of participants with and without a self-reported SRN disorder and a self-reported signal-to-noise ratio (SNR) loss. Method Fifty-four native English-speaking young adults with normal pure-tone thresholds (≤ 25 dB HL, 0.25-6.0 kHz) participated. Individuals who reported hearing difficulty in a noisy restaurant (Reference Standard 1) were placed in the SRN disorder group. SNR loss groups were created based on the self-report of the ability to hear Hearing in Noise Test (HINT) sentences in steady-state speech-shaped noise, four-talker babble, and 20-talker babble in a controlled listening environment (Reference Standard 2). Participants with HINT thresholds poorer than or equal to the median were assigned to the SNR loss group. Results The area under the curve from the receiver operating characteristics curves revealed that the AzBio test was not a significant predictor of an SRN disorder, or an SNR loss using the steady-state noise Reference Standard 2 condition. However, the AzBio was a significant predictor of an SNR loss using the four-talker babble and 20-talker babble Reference Standard 2 conditions (p < .05). The AzBio was a significant predictor of an SNR loss when using the average HINT thresholds across the three Reference Standard 2 masker conditions (area under the curve = .79, p = .001). Conclusions The AzBio test was not a significant predictor of a self-reported SRN disorder or a self-reported SNR loss in steady-state noise. However, it was a significant predictor of a self-reported SNR loss in babble noise and the average across all noise conditions. A battery of reference standard tests with a range of maskers in a controlled listening environment is recommended for diagnostic accuracy evaluations of SRN tests.