Background: The auditory brainstem response (ABR) is an essential tool for assessing the hearing sensitivity of newborns and other patients who are unable to provide behavioral responses. During a typical ABR session, clinicians rely on extensive training and expertise to rapidly identify the hearing threshold of the patient by adaptively optimizing the acoustic stimuli presented to the patient. Nevertheless, ABR sessions are notoriously long, which may lead to an incomplete diagnosis or necessitate additional appointments, which may impose a significant burden on the caregiver. Furthermore, clinical best practice requires clinicians to identify ABRs based on visual inspection of the recordings, which may introduce errors that lead to inaccurate diagnoses. Purpose: The long-term goal of this study was to develop a clinical decision support system (CDSS) to enable faster and objective ABR threshold estimation. This study sought to validate an algorithm that suggests optimal stimuli based on accruing data during an ABR session. Research Design: Two estimates of ABR thresholds at 500, 1000, 2000, and 4000 Hz were measured. The first estimate followed the algorithm and the second estimate followed clinical best practice. Behavioral audiograms were also measured. Study Sample: Fifty adult ears with typical hearing and hearing loss up to 80 dB HL. Data Collection and Analysis: ABRs were recorded using the Interacoustics Eclipse platform in a sound-attenuating booth. The two ABR thresholds were compared via the mean absolute difference and mean bias. The intraclass correlation coefficient (ICC) between the two ABR thresholds and behavioral thresholds were computed. Results: We found that 72 percent of the ears had a mean absolute difference of approximately 10 dB or less and 96 percent had a mean absolute difference of approximately 20 dB or less. The ICC for the algorithm and clinician threshold estimates was 0.78, indicating good agreement. Conclusions: This algorithm can estimate ABR thresholds with clinically acceptable accuracy for a majority of ears. Additional development and validation are necessary before the algorithm is clinically viable. Clinical Relevance Statement: A CDSS that suggests optimal stimuli for rapid and objective ABR threshold estimation would permit shorter clinical sessions, allow clinicians to spend more time on other aspects of clinical care, and expand access to rural areas that do not have dedicated audiology clinics.
Petersen et al. (Thu,) studied this question.