Search Results

The Bose Hearing Aid is intended to amplify sound for individuals 18 years of age or older with perceived mild to moderate hearing impairment. It is adjusted by the user to meet the user's hearing needs. No pre-programming or hearing test is necessary. The device is intended for direct-to-consumer sale and use without the assistance of a hearing care professional.

The Bose® Hearing Aid is a user-fitted wireless air-conduction hearing aid intended for use by individuals 18 years of age and older with perceived hearing impairment. It incorporates microphones on the earbuds for audio input into the ear, and it can be controlled wirelessly via Bluetooth® using a handheld device (iOS or Android) through the Bose Hear mobile app. There is an on-device user control (in-line remote) on the right cable (attached to neckband) that allows separate control of hearing aid ("world") volume and streaming audio volume, as well as control of directional hearing aid mode. In addition to hearing aid functionality for environmental listening, the Bose Hearing Aid can be used for placing and receiving telephone calls and for streaming audio from a Bluetooth compliant mobile device that has been paired with the Bose Hearing Aid. The controls accessible through the Bose Hear mobile application and on the hearing aid are used by the user to configure parameters, settings, and listening modes.

Here's a breakdown of the acceptance criteria and the study that proves the Bose Hearing Aid meets them, based on the provided text.

Acceptance Criteria and Device Performance

The Bose Hearing Aid underwent various tests to demonstrate its safety and effectiveness. These tests included electrical and battery safety, electromagnetic compatibility, mechanical performance, electroacoustic performance, usability engineering, and biocompatibility.

Here's a table summarizing the acceptance criteria and the reported device performance for several key areas:

Study Details for Effectiveness (Clinical Validation)

2. Sample sizes used for the test set and the data provenance (e.g., country of origin of the data, retrospective or prospective)

• Test Set Sample Size:
  • Phase I Study: 50 adult subjects (49 with mild to severe hearing loss, 1 with normal hearing).
  • Phase II Study: 75 adult subjects with mild to moderately severe hearing loss. These were divided into two groups for the field trial:
    • Pro-Fit Group: 37 subjects
    • Self-Fit Group: 38 subjects
• Data Provenance: The text does not explicitly state the country of origin.
• Study Design:
  • Phase I: Laboratory study, prospective.
  • Phase II: Prospective, two-arm, pre-market study.

3. Number of experts used to establish the ground truth for the test set and the qualifications of those experts

• Phase I & II (Ground Truth for Professional Fitting): "Several participating licensed audiologists" were used for professional fitting.
  • Qualifications: "Licensed audiologists." No further details on years of experience are provided.

4. Adjudication method (e.g., 2+1, 3+1, none) for the test set

• The text describes a comparison between "professionally-selected settings" and "DRC user-selected settings." For establishing the "professionally-selected settings" (which serves as a reference or "ground truth" for comparison to self-fitting), there is no mention of a formal adjudication process (like 2+1 or 3+1) among audiologists who performed the fitting. The implication is that each audiologist applied standard professional fitting practices.
• For the sound quality comparison in the Phase II study, it involved "blind sound quality comparisons conducted in their everyday lives," where subjects compared their own settings to the clinician-selected settings. This is a direct patient preference assessment rather than an expert adjudication of ground truth.

5. If a multi-reader multi-case (MRMC) comparative effectiveness study was done, If so, what was the effect size of how much human readers improve with AI vs without AI assistance

• This was not an MRMC comparative effectiveness study involving AI assistance for human readers.
• Instead, it was a comparative effectiveness study comparing human-professional fitting (Pro-Fit Group) vs. user self-fitting (Self-Fit Group) using AI/device-based algorithms (DRCs) without professional assistance.
• The effect sizes reported are on the effectiveness of the self-fitting method compared to professional fitting:
  • Sound Quality Preference (Phase II): Subjects in the Self-Fit Group preferred their own settings over clinician-selected settings "significantly (p < 0.0001) more." (Figure 5a shows a clear shift in distribution towards preferring "Self-Fit" in the Self-Fit group vs. "No Pref" or slight "Pref Clin" in the Pro-Fit group).
  • Satisfaction with Sound Quality (Phase II): Self-Fit group rated themselves "significantly happier with the sound quality" (p < 0.0001). (Figure 5b shows a higher proportion of 4 and 5-star ratings for the Self-Fit group).
  • Speech-in-Noise Recognition (QuickSIN - Phase II): The QuickSIN benefit scores in the Self-Fit Group were "significantly non-inferior" to those of the Pro-Fit Group (p < 0.05). This means there was no statistically significant reduction in benefit when self-fitting compared to professional fitting.
  • Patient-Reported Outcomes (APHAB & SSQ-12 - Phase II): Benefit experienced by the Self-Fit Group was "significantly non-inferior" to the Pro-Fit Group (p < 0.05) for both global and subscale scores. Again, indicating no significant reduction in benefit.
  • Gain Settings (Phase II): Subject-selected gain was "significantly correlated" with professionally-selected gain (r = 0.65, p < 0.0001) and was on average only 1.9 dB less than professionally-fit gain. This quantifies the similarity of the user-derived settings to professional ones.

6. If a standalone (i.e., algorithm only without human-in-the-loop performance) was done

• The "Self-Fit" group in the Phase II study effectively represents the "algorithm only" performance as used by a lay human without professional assistance. The user manipulates the "Dimension-Reduced Controllers (DRCs)" (Loudness and Fine Tuning) which are implemented by the device's algorithms. The core of the study is validating the effectiveness of this self-fitting strategy.
• The Phase I study specifically evaluated the "validity of the DRC method for user self-fitting of signal processing parameters" with a simulated hearing aid. The "ground truth" here was the NAL-NL2 prescriptive gain. This part of the study could be considered akin to a standalone algorithm evaluation, as it assessed how closely user-selected gains via the DRCs matched a conventional prescriptive target.

7. The type of ground truth used (expert consensus, pathology, outcomes data, etc.)

• For Electroacoustic and Mechanical bench testing: Ground truth was established by adherence to established industry standards (e.g., ANSI S3.22, ANSI/CTA 2051, IEC standards).
• For Biocompatibility: Ground truth was established by adherence to ISO 10993-1 and specific biological testing (cytotoxicity, irritation, skin sensitization) with defined pass/fail criteria.
• For Self-Fitting Effectiveness (Clinical Studies):
  • Reference/Comparator Ground Truth: "Professionally-selected settings" by licensed audiologists using a "custom professional fitting application." This essentially serves as an expert-derived ground truth for optimal fitting.
  • Clinical Effectiveness Measures: Standardized objective tests (QuickSIN for speech-in-noise) and validated patient-reported outcome measures (APHAB, SSQ-12 questionnaires) as well as patient subjective preferences and sound quality ratings.
  • Phase I specific: Comparison to NAL-NL2 prescriptive gain as a target for "validity."

8. The sample size for the training set

• The document does not provide details on a specific "training set" sample size for the development of the self-fitting algorithms (DRCs). This is common for regulatory submissions where the focus is on validation rather than the internal development process. The algorithms (DRCs) themselves are the result of Bose's R&D, likely incorporating internal data and audiometric principles, but a distinct "training set" for the algorithm's development and its sample size is not presented in this regulatory summary.

9. How the ground truth for the training set was established

• As the document does not detail a specific "training set" for the AI/algorithm development, it also does not explain how ground truth for such a set was established.
• However, the Phase I study provides insight into the validation of the self-fitting method, showing that "the average deviation from NAL-NL2 prescribed gain for subject-selected gain was 5 dB, which was in line with published reports of preferred gain settings after post-fitting fine tuning." This suggests that NAL-NL2 prescriptive targets and published reports on preferred gain settings served as a key reference or "ground truth" during the development and validation of the DRCs' effectiveness in achieving appropriate gain settings. The DRCs essentially aim to allow users to approximate these audiologically sound targets without professional intervention.

Ask a specific question about this device