Clustering Users Based on Hearing Aid Use: An Exploratory Analysis of Real-World Data

Area of Science:

• Audiology
• Digital Health
• Data Science

Background:

• Traditional hearing aid use assessment relies on subjective self-reports, which are often inaccurate.
• Smartphone-connected hearing aids offer objective data logging capabilities, overcoming self-report limitations.
• Objective data allows for higher temporal resolution and longitudinal analysis of hearing aid use.

Purpose of the Study:

• To explore daily and hourly hearing aid usage patterns using objective data.
• To identify distinct clusters of hearing aid users based on their usage patterns.
• To validate the identified user clusters using machine learning techniques.

Main Methods:

• Analysis of objective hearing aid use data from 15,905 real-world users over a 4-month period.
• Clustering of 453,612 logged days to identify typical daily hearing aid use patterns (full day, afternoon, sporadic evening).
• Clustering of users based on the proportion of time spent in each typical day pattern, validated by a supervised classification ensemble.

Main Results:

• Average daily hearing aid use was 10.01 hours, with significant between-user and within-user variability.
• Three typical daily use patterns were identified: full day (44%), afternoon (27%), and sporadic evening (26%).
• Three distinct user groups emerged, with the largest group (49%) predominantly exhibiting full-day hearing aid use. Classification accuracy was ~86%.

Conclusions:

• Objective data from connected hearing aids reveals diverse and quantifiable user engagement patterns.
• Distinct user clusters based on hearing aid usage can be identified and validated, paving the way for personalized interventions.
• This approach offers deeper insights into hearing aid adoption and supports the development of tailored hearing healthcare solutions.