Predicting abnormal auditory brainstem response outcomes using risk factors

Area of Science:

• Pediatric Audiology
• Medical Informatics

Background:

• Auditory brainstem response (ABR) testing is crucial for diagnosing hearing loss in non-cooperative children.
• Current ABR scheduling uses fixed 60-minute operating room (OR) blocks, leading to inefficiency and long wait times (average 82 days).
• A significant portion of ABR tests (63%) show normal hearing and complete quickly, indicating potential for optimized scheduling.

Purpose of the Study:

• To develop and validate a predictive model for identifying children at higher risk of abnormal ABR results.
• To improve operating room (OR) block utilization and patient access to ABR testing.
• To reduce wait times for pediatric hearing loss evaluations.

Main Methods:

• A retrospective study of 239 children undergoing sedated ABR testing was conducted.
• Clinical risk factors (e.g., autism, NICU stay, syndrome diagnosis, UNHS referral) were extracted from electronic health records.
• Logistic regression was employed, with models evaluated for discrimination, calibration, and clinical utility.

Main Results:

• The predictive model showed moderate discrimination (AUC ≈ 0.68).
• Universal newborn hearing screening (UNHS) referral and syndrome diagnosis were associated with increased ABR risk; autism diagnosis with decreased risk.
• At a 35% risk threshold, the model identified 53% of abnormal ABR cases with 75% specificity.

Conclusions:

• A pre-test risk stratification model can effectively identify children at higher risk for abnormal ABR.
• Implementing such a model has the potential to enhance OR block utilization, improve patient access, and streamline workflow efficiency.