Highlight
– In a retrospective cohort of 284,175 US Veterans, pre‑fitting personal factors (health and sociodemographic) were independent predictors of 2‑year hearing aid (HA) use persistence.
– More severe audiometric hearing loss (PTA) predicted greater persistence, but measures of audiometric asymmetry and complexity predicted lower persistence.
– New HA users, those with dementia or other mental health diagnoses, and patients with greater multimorbidity had reduced persistence; persistence peaked at ages 70–79 and was lower in non‑White, Hispanic, and unmarried patients.
– These nonmodifiable, pre‑fitting indicators may serve as flags prompting intensified or tailored management to improve long‑term HA use.
Background
Hearing loss is highly prevalent in older adults and is associated with adverse outcomes including social isolation, reduced quality of life, and increased risk of cognitive decline. Globally, the World Health Organization estimates that hundreds of millions of people have disabling hearing loss, and use of amplification devices remains well below need. Effective hearing rehabilitation depends not only on device fitting and technical performance, but also on patient uptake and sustained use. Understanding who is at risk for early discontinuation of hearing aids allows clinicians and health systems to deploy targeted support to improve long‑term outcomes.
Study design
The study by Naylor and colleagues (Ear Hear. 2025) used linked electronic health records (EHR) and post‑fitting battery order data from the US Department of Veterans Affairs (VA). The initial cohort included 731,231 patients with audiology HA orders from April 1, 2012 to October 31, 2014. After applying inclusion criteria (age ≥50 years, pure‑tone average [PTA] ≥25 dB HL, complete audiograms, availability of a 2‑year HA use persistence measure, a 5‑year clearance period for selected health conditions) and excluding cochlear implant codes, the final analytic sample was 284,175 patients.
Independent (pre‑fitting) variables covered three domains:
- Audiological: PTA severity, PTA asymmetry, audiogram slope, audiogram complexity, and new versus experienced HA user status.
- Health: diagnostic codes for dementia, mild cognitive impairment (MCI), other mental health conditions, multimorbidity (aggregate comorbidity burden), and recent inpatient episodes.
- Demographic: age, race, ethnicity, partnership status, income, and urban–rural residence.
The primary outcome was hearing aid use persistence at 2 years post‑fitting, operationalized using battery order records in the 18 months prior to the 2‑year mark. Multiple logistic regression was used to estimate adjusted associations. Continuous predictors were discretized and missing data imputed.
Key findings
Sample and outcome: Among 284,175 veterans meeting inclusion criteria, two‑year HA use persistence was assessed objectively via battery orders. The analytic model adjusted for a broad set of audiometric, health, and demographic covariates.
Audiologic predictors
– PTA severity: Greater hearing loss severity (higher PTA) was positively associated with 2‑year persistence. In other words, patients with more substantial measured hearing loss were more likely to continue ordering batteries two years after fitting. This aligns with prior work linking perceived benefit and symptom burden to device use.
– PTA asymmetry, audiogram slope, and audiogram complexity: These features—reflecting unequal hearing between ears, rapid high‑frequency drop‑off, or jagged audiometric configurations—were each negatively associated with persistence. Such audiometric patterns may reflect complex rehabilitation needs, reduced perceived benefit from standard amplification, or greater technical and counseling demands.
User experience and health predictors
– New vs experienced HA users: New HA users had lower persistence than experienced users, indicating that early post‑fitting support and habituation are critical for maintaining long‑term use.
– Cognitive and mental health: Presence of dementia and other mental health diagnoses (e.g., serious mood disorders) were independently associated with reduced persistence. Mild cognitive impairment did not show a significant independent association when dementia and other mental health conditions were included in the model, suggesting that more advanced cognitive and psychiatric morbidity exerts stronger effects on adherence.
– Multimorbidity and inpatient care: Greater overall comorbidity burden was associated with lower persistence; recent hospitalizations likely further disrupt continuity of device use and follow‑up care.
Demographic predictors
– Age: Persistence showed an inverted U‑shape across age groups, peaking between approximately 70 and 79 years and declining at younger and older ages. This suggests highest sustained use in the typical retirement‑aged population with both perceived need and capacity to manage devices.
– Race, ethnicity, partnership: Non‑White race and Hispanic ethnicity were associated with lower persistence, as was not being married/partnered. These findings point to sociodemographic disparities and the potential role of social support in sustained device use.
– No association: Tinnitus, urban–rural location, and mild cognitive impairment (in adjusted models) were not significantly associated with 2‑year persistence when other covariates were accounted for.
Overall interpretation
Although audiometric severity increased the likelihood of continued use, the cumulative, independent contributions of non‑audiologic pre‑fitting factors (health and sociodemographic) were large. Many of these factors are nonmodifiable but are available before fitting and therefore can serve as clinical flags for tailored interventions aimed at improving long‑term use.
Expert commentary and clinical implications
Strengths of the study include its exceptional sample size, use of objective service data to determine persistence, and comprehensive covariate adjustment. These features increase precision and the ability to detect modest associations. The findings offer practical knowledge: clinicians can identify patients at high risk of early discontinuation at the time of referral or fitting and proactively deploy supports.
Mechanistic plausibility
Greater measured hearing loss likely produces greater perceived benefit and motivation to continue using amplification. Conversely, asymmetric or complex audiograms may reduce speech understanding benefit from conventional fittings, complicate device programming (e.g., need for asymmetric prescriptions or assistive technologies), and increase the likelihood of patient frustration and abandonment. Cognitive impairment and psychiatric comorbidity can impair learning, device management, and follow‑up engagement. Social support (e.g., marriage/partnership) may facilitate device management and encourage persistence.
Clinical actions to consider
- Flag high‑risk patients (new users, dementia/mental health diagnoses, high multimorbidity, non‑White/Hispanic, unmarried) at pre‑fitting and allocate enhanced counseling, caregiver engagement, or more frequent early follow‑up.
- For complex audiograms, plan extended fitting sessions, verification with real‑ear measures, and trials of alternative solutions (CROS/BiCROS, frequency‑lowering technologies, remote mic accessories).
- Integrate behavioral supports—for example, structured hearing aid orientation sessions, written instructions, caregiver training, and technology to simplify maintenance (automatic coupling checks, rechargeable systems).
- Address system‑level barriers, including access to batteries and supplies, transportation, and cost that may differentially affect disadvantaged groups.
Limitations and generalizability
Key limitations should temper interpretation. The cohort was drawn from the VA system and therefore is predominantly male and older; findings may not generalize to civilian populations or younger adults. The outcome—battery orders—provides an objective but imperfect proxy for device use; some patients may obtain batteries outside the VA or receive them for devices they do not use. Diagnostic codes and EHR data can suffer from misclassification. The study used discretization of continuous variables and multiple imputation for missing data; while reasonable, these analytic decisions can influence effect estimates. Residual confounding is possible despite adjustment.
Conclusions and research gaps
Naylor et al. demonstrate that pre‑fitting personal characteristics beyond audiometry substantially influence 2‑year HA use persistence. Because many of these factors are nonmodifiable but recognizable before fitting, they provide an empirical basis for stratified care pathways: patients at higher risk for discontinuation can receive augmented counseling, caregiver integration, and follow‑up intensity. Future research should validate these associations prospectively, test targeted interventions (randomized where feasible) to improve persistence, and refine adherence measures that combine objective device telemetry with patient‑reported outcomes. Work is also needed to understand and remediate the race/ethnicity and social support disparities identified.
Funding and clinicaltrials.gov
Funding: As reported by the authors in the primary publication (Naylor G et al., Ear Hear. 2025). Readers should consult the original article for specific funding sources and conflict of interest declarations.
ClinicalTrials.gov: Not applicable—this was an observational EHR‑based cohort study.
References
1. Naylor G, Dillard LK, Zobay O, Saunders GH. Associations Between Pre‑Fitting Factors and 2‑Year Hearing Aid Use Persistence, Derived From Health Records and Post‑Fitting Battery Order Data of 284,175 US Veterans. Ear Hear. 2025 Nov‑Dec;46(6):1595‑1602. doi: 10.1097/AUD.0000000000001694.
2. World Health Organization. World Report on Hearing. Geneva: WHO; 2021. Available at: https://www.who.int/publications/i/item/world-report-on-hearing
3. Lin FR, Metter EJ, O’Brien RJ, Resnick SM, Zonderman AB, Ferrucci L. Hearing loss and incident dementia. Arch Neurol. 2011;68(2):214‑220.
Additional literature on hearing aid uptake, adherence, and strategies to support sustained use is extensive; clinicians should consider integrating evidence‑based behavioral and technical supports aligned with individual patient risk profiles.