INTRODUCTION

Most US adults will experience hearing loss (HL) at some point in their lifetime, with prevalence reaching 50% at age 70 years and 80% by 80.1,2,3 Despite this high prevalence, hearing screening is not commonly performed in adults, leaving most affected Americans without a diagnosis. As a result, 80% of persons with HL in the USA do not receive treatment.1,4 Furthermore, untreated HL is associated with significant impacts on quality of life, cognitive impairment and other morbidities, fall risk, and patient-provider communication.5,6,7,8,9,10 HL incurs yearly societal costs of upwards of $194B due to associated and independent risk of poor general health and dementia.7,9,11 Stakeholders are increasingly calling for identification of cost-effective interventions to diagnose, link, and treat persons with HL.12,13,14,15

Screening for HL in primary care settings increases diagnosis of HL and downstream treatment uptake.16,17,18 However, current US screening practices are varied and include single questions, surveys, and sound tests, with some professional societies recommending regular hearing screening of older adults.19,20 A recent United States Preventative Services Taskforce (USPSTF) update found that while several screening modalities successfully identify persons with HL, and that HL treatment is beneficial, there was insufficient evidence to make a recommendation on screening asymptomatic adults over age 50 years.21,22,23 A key limitation of the evidence cited by the USPSTF was the absence of a randomized trial that linked hearing screening to quality-of-life outcomes in the general population. In this study, we used a model to estimate the long-term clinical and economic effects of different adult hearing screening paradigms in the USA and identify key sources of uncertainty to guide future research and policy.

METHODS

Analytic Overview

We used Decision model of the Burden of Hearing Loss Across the Lifespan (DeciBHAL-US) to simulate Current Detection (CD), e.g., current rates of symptomatic presentation and uptake of hearing healthcare, compared to different hearing screening schedules that varied by age at first screen and screening frequency.24 In previously published validation exercises, DeciBHAL accurately modeled observed HL natural history and treatment.24 We simulated a cohort of 40-year-old persons without HL throughout their remaining lifetime, with yearly probabilities of acquiring HL and subsequent hearing aid (HA) uptake and discontinuation. Screening effectiveness was simulated as an increase in baseline HA uptake. We measured undiscounted and discounted (3%/year) effectiveness, using quality-adjusted life-years (QALYs), and estimated health system costs (2020 USD). We considered incremental cost-effectiveness ratios (ICERs) of <$100,000/QALY to be cost-effective.25,26 This study followed the Second Panel on Cost-Effectiveness in Health and Medicine and Consolidated Health Economic Reporting Standards.27,28

Model Overview

DeciBHAL-US is a microsimulation model of HL natural history, detection, diagnosis, and treatment (see Appendix 1 for health state diagrams, and the published validation for further details).24 Simulated persons experience yearly age- and sex-specific probabilities of acquiring bilateral sensorineural HL, conductive HL, or both, based on published National Health and Nutrition Examination Survey (NHANES) estimates.1 Persons with sensorineural HL experience age-related decline in pure tone average (PTA) hearing level (Table 1).15 After acquiring HL, patients experience age- and severity-specific yearly probabilities of acquiring HAs and, once using HAs, rates of discontinuation. Adults with severe-to-profound HL may receive cochlear implantation after use of HAs for at least 1 year.29

Table 1 Selected Model Inputs
Full size table

Simulated Screening Schedules

We simulated screening schedules starting at ages 45, 55, 65, and 75 years and lasting throughout the lifetime with frequencies of screening every 1 and 5 years. We simulated a sound test screening modality, however, to account for imperfect implementation, incorporating a lower specificity to capture costs from falsely positive tests acquiring audiological services (described below).21

Model Input Data

Natural History of Hearing Loss

Yearly probabilities of bilateral sensorineural HL ranged from 0.8 to 10.4% for males and 0.1 to 9.2% for females.1,30,31,32,33 Age-specific decline in HL is modeled as a yearly decibel (dB) increase in PTA (mean=1.05dB/year; SD=0.4) that was derived from a longitudinal study of older adults and validated to the Baltimore Longitudinal Study on Aging.34,35 Simulated persons may also acquire conductive HL.36,37,38,39,40,41,42

Hearing Aid Uptake

Simulated persons with HL experience yearly probabilities of HA acquisition that vary with age and severity (Table 1).4,43 After acquiring HAs, simulated persons have yearly probabilities of discontinuation that range from 13% in year 1 after acquisition and decline to 4% in year 10.44,45 These combined inputs were previously calibrated to NHANES estimates of HA use prevalence.4 DeciBHAL-US projects average person-time of HA use by aggregating each year a simulated person uses a HA and dividing by the total cohort size.

Cochlear Implant Uptake

Patients with severe or profound HL experienced a 1.3% annual probability of cochlear implantation after at least 1 year of HA use, and a 1% probability of CI discontinuation thereafter.46,47

Screening Effectiveness

We simulated the effectiveness of HL screening schedules as a multiplier on calibrated baseline rates of HA uptake among simulated persons with HL. We based this multiplier on one randomized control trial and one non-randomized before-and-after study that demonstrated increased rates of hearing healthcare uptake after screening implementation that ranged from 1.50 to 2.0 (using a tone-emitting otoscope or four-frequency screening device).18,48 For our base-case screening effectiveness parameter (1.62), we combined the risk ratios of hearing healthcare uptake after screening versus no screening using inverse variance weighting. We assumed that among persons who had screen-suggested HL, there would be a similar risk ratio for HA uptake in veterans and non-veterans (though the absolute rate is different). We also assumed that the risk ratio for increased hearing healthcare uptake was equivalent to the risk ratio of HA uptake in one study that measured only hearing healthcare uptake. Given this uncertainty, we varied screening effectiveness in sensitivity analysis from 1.05 to 2.25.

Screening Test Characteristics

While we simulated a sound testing screening strategy, we incorporated test characteristics from other screening test modalities to capture imperfect implementation. To estimate the number of persons without HL who might be referred for a hearing diagnostic test based on results of hearing screening, we used a false positive probability of 24% (1-specificity, 74%), based on pooled estimates of the Hearing Handicap Inventory for the Elderly from the USPSTF Evidence Review.21 We similarly incorporated a pooled sensitivity of 80%. These screening test characteristics affected model-projected costs, but not screening effectiveness because test characteristics were already incorporated in downstream HA uptake estimates.

Health State Utilities

We assumed a population average utility of 0.84 for persons without HL, and mild and moderate HL utility values derived from the published literature of 0.71 and 0.65.49,50,51 We incorporated data from a recent systematic review on the utility benefits of HAs (+0.11) and cochlear implants (+0.17) and vary these in sensitivity analysis (Appendix 2).50,51,52,53

Screening Costs

Screening test costs were applied to persons without and with HL (schedule-dependent, either yearly or every 5 years). For persons without HL, we included the cost of a screening test ($2; device cost and personnel time amortized across all persons screened) and the proportion of persons receiving a false positive test (24%) who would seek an audiology diagnostic test (43%; $295), coming to a net cost of $33.18,21,54,55 For persons with HL but without hearing treatment, we included the proportion of persons that receive a true positive screen (80%) and seek an audiology diagnostic test but do not acquire HAs (51%). Along with the screening test cost ($2), the total was $120.18,54,55

Hearing Healthcare Costs

We included costs of an audiology diagnostic test ($295), HA one-time purchase ($3690 for the device) and recurring operational and replacement costs ($910), and costs of cochlear implantation ($54,380 one-time, $1260–1400 yearly recurring).12,55,56,57,58,59,60 For HAs, we assumed 84% of fittings were binaural and accounted for the proportion of reduced cost HA fittings done through the US Department of Veteran’s Affairs.12 Recurring HA costs included batteries and replacement of the HA device every 5 years.59,61 In sensitivity analyses, we considered lower HA device costs that may be seen with over-the-counter HAs.

Sensitivity Analysis

We performed a one-way sensitivity analysis, varying model parameters across their plausible ranges. We selected parameters with the most underlying uncertainty. We then varied multiple parameters simultaneously across their plausible ranges. We included a sensitivity analysis incorporating World Health Organization (WHO) assumptions (HL treatment improves utility by 1 severity stratum) for HA utility benefits. We additionally performed probabilistic uncertainty analysis (PUA), assigning distributions to the most influential parameters identified in deterministic analysis and assessing the impact of their joint uncertainties on our cost-effectiveness outcomes (Appendix 3).62,63

Budget Impact Analysis

We projected 5-year undiscounted incremental costs of a screening schedule for the current US adult population compared to CD, simulating 59 cohorts of persons (ages 40–99), scaled to the US population, over 5 years to estimate 5-year costs for each schedule (Appendix 4).

RESULTS

Clinical Results

HA use increased with earlier age of onset of screening and increased frequency of screening, ranging from 1.20 average person-years under CD to 1.68 for yearly screening starting at age 45 years. Mean age at first HA with CD was 79.2 years, whereas yearly screening schedules initiated before age 75 reduced this age by 0.2 to 0.8 years depending on the age at first screen. Per-person lifetime undiscounted QALYs for CD were 32.107 (Table 2). Compared to CD, 5-yearly schedules increased undiscounted lifetime QALYs by 0.01–0.02, and yearly screening schedules imparted increases of 0.04–0.07 QALYs.

Table 2 Cost-Effectiveness Results of Alternative US Hearing Screening Schedules
Full size table

Costs and Cost-Effectiveness

CD had lifetime undiscounted per-person hearing healthcare costs of $3300, 5-yearly screening schedules had per-person costs of $3630–3960 (varying with age at first screen, 45–75 years), and yearly schedules $4780–6490. Using discounted costs and QALYs, annual screening beginning at ages 75, 65, and 55 years were all considered cost-effective, with ICERs of $39,100/QALY, $48,900/QALY, and $96,900/QALY, respectively. Annual screening beginning at age 45 years had an ICER of $234,600/QALY. Compared to annual screening strategies, 5-year screening strategies beginning at 65, 55, and 45 were eliminated by weak (or extended) dominance because they were less effective and less efficient (i.e., had a higher ICER) than yearly screening beginning at age 75 (Appendix 5).

One-Way Sensitivity Analysis

The ICER for yearly screening beginning at age 55 years was most sensitive to single parameter variations in audiology diagnostic test cost, screening effectiveness, HA device cost, screening test false positive probability, and HA utility benefit (Fig. 1). Varying these parameters across their plausible ranges, the yearly screening schedule beginning at age 55 exceeded the $100,000/QALY threshold (Fig. 1, dashed line).

Fig. 1
figure 1

One-way sensitivity analysis on the cost-effectiveness of yearly screening beginning at age 55 years. This is a tornado diagram illustrating the effects of variation of single parameters (row) across their plausible range on the incremental cost-effectiveness ratio (ICER) of yearly screening beginning at age 55. The y-axis crosses the x-axis at the base-case ICER of $96,900/QALY. Each sensitivity analysis is presented with a base-case value, range explored, and the resulting ICERs from that range are plotted. The long-dashed line represents the US willingness-to-pay threshold of $100,000/QALY, and the short-dashed line indicates an ICER of $50,000/QALY.

Full size image

Multi-way Sensitivity Analysis

With base-case HA utility benefit inputs, yearly screening beginning at age 55 years remained under $100,000/QALY as long as effectiveness was >1.55 and device cost was <$6580 (cut points; Fig. 2A). With screening effectiveness ranging between 1.25 and 1.45, yearly screening beginning at age 65 years was <$100,000/QALY. When screening effectiveness was lowered to 1.05–1.15, none of the simulated screening schedules was cost-effective.

Fig. 2
figure 2

Three-way sensitivity analysis on the cost-effectiveness of adult hearing screening schedules in the USA. This figure shows the impact of varying hearing screening effectiveness, hearing aid device cost, and hearing aid utility benefit on the incremental cost-effectiveness ratios (ICERs) of screening schedules. Screening effectiveness was varied from 1.05 to 2.25 on the x-axis, and hearing aid device cost was varied from $800 to 6580 on the y-axis. Panel A shows the base-case hearing aid utility benefit of +0.11 and panel B shows the results assuming a hearing aid utility benefit of +0.01–0.07 (varying with severity, similar to that assumed by the WHO and the Global Burden of Disease). Each combination of parameters yields an ICER that is color-coded, with red indicating that all screening schedules had ICERs >$100,000/QALY, orange that only yearly screening at age 75 was below $100,000/QALY, yellow that yearly screening at age 65 was below $100,000/QALY, green that yearly screening at age 55 was between $50 and 100,000/QALY, and blue that yearly screening at age 55 was less than $50,000/QALY. The X marks the base-case input combination of screening effectiveness and hearing aid device cost.

Full size image

Lowering the HA utility benefit to match assumptions used by the WHO15 and Global Burden of Disease64 increased the base-case ICER of yearly screening at age 55 to $169,700/QALY (Fig. 2B). With these utility benefit assumptions, yearly screening at age 75 years was cost-effective (ICER=$93,500/QALY).

Probabilistic Uncertainty Analysis

We simultaneously varied the utility benefit of HAs, screening effectiveness and false positive probability, audiology diagnostic test cost, and HA device cost across defined distributions (Appendix 6). At base-case WTP of $100,000/QALY, there was large uncertainty around the optimal screening schedule (most effective non-dominated schedule with an ICER under the WTP), with no single schedule representing the optimal schedule in ≥ 50% of simulations.

Budget Impact Analysis

The average annual undiscounted cost of CD over 5 years and for the US population over age 40 years was $12.8B, and for yearly screening beginning at age 55 it was $21.8B (Fig. 3; Appendix 7).

Fig. 3
figure 3

Budget impact analysis. This figure depicts the model-projected average annual outlays over the first 5 years of the Current Detection and yearly screening beginning at age 55 years (55q1). The height of the bars represents the total annual cost of CD and the 55q1 screening schedule, with 2020 USD on the y-axis. The components of the total cost are chronic suppurative otitis media (CSOM) treatment (blue), hearing aid (HA) uptake (gray), hearing aid maintenance (yellow), cochlear implant (CI) costs, and screening costs (orange).

Full size image

DISCUSSION

Our model predicts that hearing screening schedules for US adults would increase uptake of hearing healthcare and improve quality of life. Yearly screening at ages 55+years increased per-person undiscounted lifetime QALYs by 0.07, equivalent to extending full-health survival by 26 days (averaged across all persons without and with HL). Compared to CD, per-person lifetime undiscounted costs increased by ~$2,900, and the ICER was $96,900/QALY for yearly screening beginning at age 55. Simulated screening schedules beginning later in life, such as 65 and 75 years when HL is more prevalent, were even more cost-effective (ICERs<$50,000/QALY) and initiating screening at age 65 may be more clinically feasible than at age 55 given the onset of Medicare coverage.

In sensitivity analysis, applying less optimistic parameter values resulted in the ICER of yearly screening beginning at 55 years to exceed $100,000/QALY, while screening at older ages remained cost-effective. To improve the cost-effectiveness of screening, policymakers could focus on improving referral processes to hearing healthcare after screening (effectiveness) and lowering patient’s costs for audiology diagnosis or purchase of HA devices. There was large uncertainty around the optimal hearing screening strategy in PUA and future efforts should focus on quantifying the value of reduction in parameter uncertainty ascertained by potential research projects through value of information analysis.65 In particular, clinical studies clarifying the impact of hearing screening on HA uptake would likely reduce this decision uncertainty.

Previous model-based cost-effectiveness analyses of adult hearing screening schedules in high-income settings have supported the economic efficiency of screening beginning at ages 55–65 years.54,66,67,68 One study set in the USA, based on a randomized controlled trial of ≥ 50-year-old, primarily male veterans, projected a $1400 cost per additional HA user at 1 year.54 Our results are not directly comparable to this study because we projected costs/QALY and over a lifetime. Two European studies projected ICERs of €3700/QALY ($4700 in 2020 USD) and £1800/QALY ($2800 in 2020 USD) for 5-yearly screening in the Netherlands and the UK.66,67 These ICERs are lower than our estimates, likely due to our incorporation of audiometric evaluation costs for false positives, consideration of HA discontinuation, and the higher costs of HAs and healthcare in the USA.

A major consideration around adult hearing screening is successful implementation in real-world settings.69,70 Challenges throughout the implementation process include appropriate distribution of the hearing screening test, referral to a hearing professional, and linkage to hearing healthcare. Our sensitivity analyses varying screening effectiveness might approximate reductions in effect due to imperfect implementation. Innovative hearing screening strategies could expand access and help to alleviate the burden on primary care settings.71,72

We simulated a sound testing–based screening modality as objective assessment measures, such as tone-emitting otoscopes or four-frequency screening devices, have shown better results in trials of adult hearing screening. However, single-question or other clinical hearing screening tests may be more feasible in a clinic setting.18,54 To be conservative and account for imperfect implementation, we incorporated a higher false positive probability (associated with a survey). However, novel mobile app–based point of care hearing screening modalities would likely reduce false positives and lower overall test costs.73

Affordability of HAs after a positive screen and diagnosis of HL remains a large barrier to increased adoption of HAs.16 HAs are not covered under Medicare and receive limited coverage by other insurers, so persons with HL bear the majority of the several thousand-dollar cost of HAs. We project Medicare coverage of HAs for adults over age 65 years to be very cost-effective (ICER<$50,000/QALY) and would help with affordability issues. Another potential mitigation of this problem is the Over-the-Counter Hearing Aid Act, which allows persons with perceived mild-to-moderate HL to purchase FDA-regulated over-the-counter HA devices without a medical exam or fitting by a specialist.74 While the introduction of this new category of FDA-regulated hearing devices is projected to increase access and lower HA device cost, its overall impacts on the costs and effectiveness of hearing screening are less well understood. Our sensitivity analysis demonstrated that if over-the-counter HAs increase HA uptake at lower cost, hearing screening becomes even more cost-effective.

Our analysis has several limitations. First, we made simplifying assumptions in our model structure and input data; however, we documented these assumptions and tested them in sensitivity analysis. Second, we assumed that the risk ratio of HA uptake due to screening (hearing screening effectiveness) remained constant throughout the lifetime. The true impact of repeated screening on persons with HL and at different ages is unknown; therefore, we varied screening effectiveness extensively in sensitivity analysis. Third, we excluded possible benefits to screening beyond provision of technology, including acknowledgement of a suspected health issue and treatment approaches not based on use of HAs.75 Additionally, there may be benefits to passing a hearing screen, such as peace of mind, that we did not consider. Fourth, we conservatively excluded potential benefits of hearing healthcare on physical and cognitive health and healthcare costs. To the extent that early detection and treatment of HL improves physical and cognitive health, the benefits of hearing screening would be larger than our current estimates. Lastly, due to data limitations we did not include indirect societal costs such as improved workforce participation due to HL and its treatment. Assuming that early treatment of HL improves workforce participation, the cost-effectiveness of hearing screening would improve.76

To conclude, we project annual hearing screening in US adults beginning at age 55 to be cost-effective. In addition, our analysis supports that annual screening beginning at ages 65 and 75 is even more cost-effective and should likely be performed. While future research might inform more certain parameters such as screening effectiveness, our findings were robust over a wide range of assumptions of screening effectiveness. Delaying hearing screening implementation in the hopes of perfect evidence will only further increase the wide hearing healthcare diagnosis and treatment gap and perpetuate long-standing inequities in hearing healthcare. The inclusion of hearing in health assessments of older adults is imperative to their health and well-being, and we demonstrate here that it too is likely a good use of resources.