Related Experiment Video
Updated: Aug 20, 2025
06:04
Systematic Hearing Performance Evaluation Process for Adolescents with Cochlear Implantation at Early Ages
Published on: March 24, 2023
441
Development of a Predictive Model for Individualized Hearing Aid Benefit
Krish Suresh, Kevin Franck1, Julie G Arenberg2
1Frequency Therapeutics, Lexington.
Summary
A new random forest model accurately predicts hearing aid benefit using patient data. Key predictors include hearing aid use duration, hearing loss severity, and age, aiding in patient candidacy assessment.
Area of Science:
- Audiology
- Biomedical Engineering
- Data Science
Background:
- Hearing aid benefit is highly individualized and difficult to predict.
- Patient-reported outcome measures like the Client Oriented Scale of Improvement (COSI) are crucial for assessing hearing aid success.
Purpose of the Study:
- To develop a predictive model for individualized hearing aid benefit.
- To interpret model predictions for global and individual insights.
Main Methods:
- A dataset of 1,286 patients with hearing loss was compiled, including demographic, medical, and audiological data.
- The Client Oriented Scale of Improvement (COSI) questionnaire measured hearing aid benefit.
- A random forest model was developed and validated using fivefold cross-validation, with hyperparameter tuning.
- Shapley Additive Explanations (SHAP) were used for model interpretation.
Main Results:
- The random forest model achieved a root mean squared error (RMSE) of 0.80, outperforming eXtreme gradient boosting (RMSE of 0.85).
- Significant predictors of hearing aid benefit included shorter duration of hearing aid use, higher pure-tone average in the better ear, and younger age.
Conclusions:
- A predictive model for hearing aid benefit has been successfully developed, offering individualized explanations.
- Predictive modeling shows promise as a tool for assessing hearing aid candidacy and expected outcomes.
