Abstract
Purpose
Hearing loss is highly prevalent among older adults in China, yet hearing aid uptake remains substantially low. Understanding preferences for updated hearing aid attributes is essential for improving accessibility and promoting adoption, particularly amid rapid technological advancement such as over-the-counter (OTC) devices. This study aimed to elicit preferences for hearing aid attributes and examine preference heterogeneity across demographic and socioeconomic subgroups among older adults with hearing loss living in six cities in China.
Patients and Methods
A discrete choice experiment (DCE) was conducted among 346 community-dwelling adults aged over 60 years with hearing loss in six cities of China. Twenty-one choice sets were generated and divided into three versions. An opt-out option was included in each choice set. Six attributes were included: out-of-pocket (OOP) cost, accessibility (OTC vs audiologist-fitted), effectiveness in noisy settings, style, battery type, and connectivity. Mixed logit and conditional logit models were estimated, with model fit compared using log likelihood, AIC, and BIC. Preference heterogeneity was examined through interaction terms between participant characteristics and attributes with significant random parameters.
Results
Effectiveness in noisy settings was the most influential attribute (Coefficient=3.01, 95% CI: 2.41, 3.62), followed by accessibility. Participants preferred OTC hearing aids (Coefficient=0.89, 95% CI: 0.60, 1.18), receiver-in-canal style (Coefficient=0.40, 95% CI: 0.05, 0.76), and rechargeable batteries (Coefficient=0.38, 95% CI: 0.17, 0.59), and showed significant aversion to higher OOP costs (Coefficient=−0.03, 95% CI: −0.04, −0.02). Significant heterogeneity was observed. Cost sensitivity was lower among individuals with higher education, married status, and prior hearing aids use, but greater among those with dementia risk. OTC preference was stronger in the higher education group and weaker among participants with dementia risk.
Conclusion
Older adults from six cities in China prioritize performance in noisy environments and improved accessibility when selecting hearing aids. OTC accessibility appears to be valued by respondents and may warrant further evaluation as part of broader hearing-care strategies.
Keywords: hearing rehabilitation, preference, over-the-counter, hearing aid attributes
Introduction
Hearing loss (HL) is a common and undertreated cause of disability worldwide, affecting more than 1.5 billion people and imposing a substantial economic burden.1,2 Its prevalence increases markedly with age, with moderate or greater HL rising from 15.4% among adults in their 60s to 58.2% among those aged over 90 years.3 In China, around 11% of people aged 60 years suffered from disabling HL in 2006,4 and the proportion rapidly increased to 58.58% in 2015.5 Hearing aids can improve communication and related health outcomes, including fall prevention;6,7 however, their uptake remains below 10% in many developing countries and is approximately 6.5% in China.8,9 Understanding the attributes valued by older adults when selecting hearing aids is therefore important for improving hearing aid accessibility and adoption.
Previous studies have investigated the determinants for hearing aids use among older adults with HL, especially in developing countries. These determinants included demographic, socioeconomic status, and cognitive status. A population-based study found that hearing aids use was associated with higher socioeconomic status, including longer years of schooling, urban residence, and higher income, as well as increased age among Chilean older adults.10 Compared with cognitively healthy older adults, those with dementia risk have a 2.07-fold higher risk of discontinuing hearingaid use within one year of fitting.11 Participants with dementia risk might experience declines in working memory, impaired executive function, and emotional difficulties, which may hinder their ability to learn device operation and adapt to amplified sound,12 resulting in different preferences for hearing aids attributes. Higher self-perception of hearing loss was also found to be related with hearing aids adoption.13 Comparatively fewer studies concluded that some product determinants, such as the type of hearing aids, effectiveness in noisy background settings, and the insertion gain of hearing aids, were also exerted significant effects on hearing aids adoption.14 Among Chinese older adults, several sociodemographic features and product factors like acceptance on hearing aids function were related to elevated prevalence of hearing aid acquisition.9
However, evidence on the preferences on hearing aids use among older adults with HL was limited. Most studies were conducted in developed countries and revealed that participants preferred better performance of hearing aids use, including effectiveness in noisy environments and quiet environments, along with shorter duration of hearing aids provision when selecting their hearing aids.15,16 Only one previous study conducted a discrete choice experiment (DCE) among Chinese older adults and concluded that effectiveness in noisy settings was the most important attributes when selecting hearing aids.17 Nevertheless, samples of this study were from rural areas in China. Moreover, in China, public financial support for adult hearing aids remains limited, and hearing aids are commonly purchased out of pocket through hospitals or commercial hearing-aid shops and clinics.18 The acquisition of hearing aids among older adults remains low, with affordability, limited knowledge of hearing aids, and unequal access to hearing-care services constituting important barriers.9,19 In this context, more accessible delivery pathways, including direct online purchase and remote fitting or adjustment, may complement conventional facility-based services and reduce some of the financial and geographic barriers to hearing aid access. Therefore, the present DCE included OTC access as an attribute to examine whether older adults valued this more accessible delivery pathway.
To address these knowledge gaps, using discrete choice experiment (DCE), which is widely used in consumer preference studies, we aimed to identify the preferences on hearing aids attributes among older adults with hearing loss from six cities in China. We also aimed to investigate how the participants’ preferences when selecting hearing aids varied across demographic and socioeconomic factors, as well as dementia risk status. Because the etiology of hearing loss was not determined, this study addressed hearing loss among older adults generally rather than specifically diagnosed age-related hearing loss.
Methods
Data Collection and Study Participants
Data were collected between August and December 2025 from selected communities in Tangshan, Baoding, Langfang, Zhangjiakou, and Cangzhou in Hebei Province and Xicheng District in Beijing Municipality. These study areas were selected to include communities from different geographic locations and socioeconomic contexts,20,21 thereby enhancing the geographic and socioeconomic diversity of the study sample. One sub-district was selected from each of the study city chosen previously. Then, one community was selected from each of the sub-district previously. Thus, six communities from the selected study cities were chosen. Local communities were approached with the assistance of community health service centres, community administrators, and audiologists. After permission to conduct the survey had been obtained, potentially eligible older residents were informed about the study through face-to-face invitations during routine community health services or in public areas of the communities. Interested residents were screened for eligibility by trained research staff, and those who met the inclusion criteria were invited to participate after providing informed consent.
Participants who are living in the above six communities were recruited using convenience sampling. To improve the representativeness of the sample, quota sampling was additionally conducted considering the sex and age distribution of the older adults with hearing loss in a population-based survey in four provinces of China on ear and hearing disorders.5 Participants were classified by sex and predefined age groups, and recruitment within each stratum continued until the target number for that stratum was reached. Recruitment progress was monitored regularly at each site and the quota targets were applied across each study area separately. When a quota for a particular stratum had been filled, further participants from that stratum were no longer enrolled, whereas recruitment continued for strata that had not yet reached their targets. The same eligibility criteria, recruitment materials, screening procedures, and quota requirements were applied across all six study areas. Medical ethics approval of this study was approved by the Institutional Review Board of Peking University (#IRB00001052-24123). Informed consent for the data collection and the use of the data was obtained from all subjects.
Trained investigators conducted face-to-face interview using electronic questionnaires with participants from communities of the above areas with the help of audiologists. Before each interview after an introduction of the background and purpose of the study, the participants were asked whether they had time and were willing to participate in the survey. The inclusion criteria included older adults aged 60 years or above, with hearing loss, and were willing to participate in our current study with informed consent. According to the rule of thumb by Orme et al,22 the following formula about the sample size calculation was used in this study.
 |
In the above formula, c represented the maximum number of levels among all attributes, t represented the number of choice sets, and a represented the number of options in one choice set. In our DCE model, c=3, t=7, a=3, and the non-response rate was set at 25%. Thus, a minimum of 96 samples would be needed for the research. Initially, 624 participants were conveniently selected and 482 of them responded to our survey, with the response rate of 77.24%. Then, 69 participants were excluded because the absent of hearing loss. After consistency check, a total of 346 participants were finally included into analysis. Detailed information could be found in Figure S1.
Measurements
Study Design of Discrete Choice Experiment (DCE)
The design of DCE was in agreement with the key criteria of good research practices for conjoint analysis raised by the International Society for Pharmacoeconomics and Outcomes Research (ISPOR).23
First, we preliminary selected nine hearing aid attributes and their levels according to previous studies17,24 and focus group discussion, including out-of-pocket (OOP) cost, accessibility, appearance, effectiveness in quiet settings, effectiveness in noisy settings, feedback, water/sweat resistant, battery life, and place of origin. Among the nine attributes, seven of them (OOP cost, appearance, effectiveness in quiet settings, effectiveness in noisy settings, feedback, water/sweat resistant, and battery life) were from previous studies.17,24 Two of them (accessibility and place of origin) were from focus group discussion of eight members in which two members majored in geriatric nursing, two members majored in geriatric psychology, three members majored in public health, and one member majored in audiology. In addition, one level was added for the attribute of feedback according to the focus group discussion. The preliminary attributes and levels were shown in Table S1.
Then, experts including audiologists, hearing aids research and development (R&D) personnel, and otolaryngologists were consulted to revise the attributes and their levels of hearing aids. Five experts were invited via Email for one round consultation. Among them, two audiologists (expert A and B) come from hospital and engaged in hearing aids fitting and adjustment for over six years. Two experts (expert C and D) come from hearing aids R&D companies and had over five years of work experience. One otolaryngologists (expert E) come from hospital and had over 10 years of first-line clinical experience in the treatment of hearing loss. Both expert A, B, C, and D advised to delete the level of “Simulated sound amplifier” for accessibility attribute because the sound amplifier could harm older adults with hearing loss and did not belong to hearing aids. In addition, expert A suggested that the level of OOP cost should be 1000 RMB, 1001–5000 RMB, 5001–8000 RMB, and over 8000 RMB. Expert B also suggested that aural fullness exerted a substantial effect on hearing aids use preferences and should be added as an attribute. Expert C advised to delete the place of origin as well as water/sweat resistant attribute because older adults with hearing loss might not understand the differences between these levels and most of the hearing aids are water/sweat resistant. Besides, expert C suggested revising the level of OOP cost to 1000 RMB, 5000 RMB, and 10000 RMB. Expert D advised deleting the attributes of effectiveness in quiet settings, feedback, and place of origin. In addition, expert D suggested adding the attributes of rechargeable batteries and connection to a smartphone app of hearing aids. Expert E suggested that a medium level between “somewhat effective” and “more effective” of effectiveness in quiet settings and noisy settings should be added. Following their suggestions, the authors deleted the attributes of effectiveness in quiet settings, feedback, water/sweat‑resistant, and place of origin, added the attribute of connectivity, and revised the attributes of appearance, battery life, and effectiveness in noisy settings. In detail, appearance was substituted by style, in which the appearance and aural fullness were described for each style.
Finally, six hearing aids attributes and their levels were determined and included in DCE model, which is presented in Table 1. The six most important attributes/levels comprised: (1) out-of-pocket (OOP) cost; (2) accessibility of hearing aids; (3) effectiveness in noisy settings; (4) style of hearing aids; (5) battery type; (6) connectivity. For the attribute of accessibility, contents presented to participants were as follows. For the audiologist-fitted hearing aids, the authors told participants that these hearing aids can be purchased from public hospitals or private hearing care center, and required face-to-face fitting and adjustments for many times by an audiologist. Thus, the accessibility of these hearing aids was relatively lower. For the OTC hearing aids, the authors told participants that these hearing aids can be purchased from retailers through online shopping apps such as Taobao, and the fitting and adjustments were also conducted by specialized audiologists for many times, despite the fact that it happened online through dedicated smartphone apps by the manufacturer. Thus, the accessibility of these hearing aids was relatively higher.
Table 1.
Final Attributes and Levels of Hearing Aids
| Attributes | Levels | Description for Participants |
|---|---|---|
| Out-of-pocket cost | 1000 RMB | The amount of money you spent to buy hearing aids. |
| 5000 RMB | ||
| 10000 RMB | ||
| Accessibility | Audiologist-fitted hearing aids | These hearing aids require multiple adjustments by an audiologist, with relatively lower accessibility. |
| Over-the-count hearing aids | These hearing aids can be purchased and fitted directly from retailers online, and do not require offline adjustments, with relatively higher accessibility. | |
| Effectiveness in noisy settings | Not very effective | Situations where there are multiple sounds coming from multiple sources, such as in restaurants. |
| Effective | ||
| Very effective | ||
| Style | Behind-the-ear (BTE) | The appearance of hearing aids is fully visible, with moderate aural fullness. |
| Receiver-in-canal (RIC) | The appearance of hearing aids is partially visible, with low aural fullness | |
| In-the-canal (ITC) | The appearance of hearing aids is fully invisible, with severe aural fullness. | |
| Battery type | Non-rechargeable batteries | Whether the hearing aid requires battery replacement or is rechargeable. |
| Rechargeable batteries | ||
| Connectivity | With a dedicated mobile app | Whether the hearing aid can be connected to a dedicated smartphone app provided by the manufacturer. |
| Without a dedicated mobile app |
After the determination of attributes and their levels, several choice sets with different combinations of attributes and levels should be constructed through experimental design. The DCE design included a full factorial design and fractional factorial design (such as orthogonal design and D-efficiency design).25 The former design would result in 216 hypothetical hearing aids alternatives combining the six attributes with each level and was unacceptable for participants. We thus generated 21 choice sets by the D-efficiency experimental design using SAS 9.2 software. Each choice set contained two alternatives and an opt-out option. To further reduce response reliability caused by presenting a single individual with a large amount of choice sets, we also used a blocked design which divided the 21 choice sets into three different versions containing seven choice sets, each according to their month of birth.26 In addition, to verify the logical consistency of the respondents, one choice set was repeatedly set up within each block and participants who did not select the same answer for the same choice set were deemed as failing the consistency check.27 Therefore, each DCE questionnaire contained eight choice sets, with each question describing two alternative hearing aid scenarios in terms of the six attributes and a no‑hearing‑aid option. The interview concerning DCE would be started with a detailed description of the hearing‑aid attributes and their levels, followed by an example of DCE question. Then, participants were shown corresponding block of questionnaire and asked, “If you were actually offered the two hearing aids above, which would you prefer to choose?”, with the answers including alternative A, alternative B, and neither of them.
After the determination of DCE questionnaires, a pilot study, completed by six participants, was administered face-to-face by a trained interviewer at Department of Otolaryngology, Head and Neck Surgery, Peking University First Hospital in August 2025. Participants could understand the attributes and levels of hearing aids and were willing to trade between the attributes. Thus, no further revision was conducted on the attributes and levels.
Hearing Loss
Hearing loss was assessed using two questions derived from the Health and Retirement Study and operationalized following a previous HRS-based study.28,29 These brief questions were selected because they were suitable for community-based screening and minimized respondent burden during completion of the DCE. Participants were asked “Is your hearing excellent, very good, good, fair, or poor (using a hearing aid as usual)?” and “Do you ever wear a hearing aid?”. Those who answered “fair” or “poor” to the former question or responded “yes” to the latter question were deemed as having hearing loss. Combining the two items allowed the inclusion of hearing aids users who might report relatively good aided hearing. This definition was used for study eligibility and did not represent an audiometrically confirmed diagnosis.
Covariates
Covariates included age, sex (male, female), educational level (primary school or below, including elementary school and junior high school; secondary school or above, including senior high school, university or above), marital status (married/partnered, unmarried), dementia risk (yes, no), and hearing aids use (yes, no). Dementia risk was assessed by the Ascertain Dementia 8 questionnaire (AD8).30 This scale contained eight items with the total score ranging from 0 to 8. Participants scored above 1 was considered as with dementia risk.31,32 The Chinese version of the AD8 has demonstrated acceptable internal consistency (Cronbach’s α = 0.78), excellent test–retest reliability (intraclass correlation coefficient = 0.96), and good discriminative validity for dementia (area under the receiver operating characteristic curve = 0.93).33 Hearing aids use was ascertained by the question “Do you ever wear a hearing aid?”. Those who answered yes was considered as hearing aids users.
Statistical Analysis
Basic characteristics of the participants were described as mean (standard deviation) or number (percentage) and compared based on the consistency check by T-test or Chi-square test. Both conditional logit model and mixed logit model were conducted to estimate participants’ preferences for hearing aids attributes.
Choice data were arranged in long format, with each row representing one alternative within a choice set. OOP cost was entered into the models as a continuous variable in RMB. For ease of presentation, the cost coefficient and willingness to pay (WTP) estimates were reported in units of RMB 100. All categorical attribute levels were dummy coded. The reference levels (0) were “audiologist-fitted hearing aids” for accessibility, “not very effective” for effectiveness in noisy settings, “behind-the-ear hearing aids” for style, “non-rechargeable batteries” for battery type, and “no dedicated mobile application” for connectivity. The opt-out alternative was represented by an alternative-specific constant (ASC), coded as 1 for choosing neither hearing aid and 0 for choosing either hearing aid alternative. All hearing aid attribute variables were set to zero for the opt-out alternative.
The coefficient calculated through conditional logit model was interpreted as a preference weight and represented the relative contribution of the attribute level to the utility that respondents assign to an alternative. As for the mixed logit model, the estimated mean coefficients represented average preferences, whereas the estimated standard deviation (SD) reflected preference heterogeneity across participants. A statistically significant SD indicated preference heterogeneity for the corresponding attribute-level coefficient. Interaction analyses were therefore focused on random coefficients with statistically significant SDs to examine whether observed respondent characteristics explained part of this heterogeneity. This interpretation and the use of mixed logit models to examine preference heterogeneity followed established recommendations for the analysis of DCE data.34,35
The ASC, OOP cost, and all attribute-level coefficients were specified as random parameters and were assumed to follow independent normal distributions; therefore, the cost coefficient was specified as random rather than fixed. The mixed logit model was estimated by maximum simulated likelihood using 500 Halton draws, with repeated choices linked at the participant level and standard errors clustered by participant. No user-defined convergence options were specified; therefore, Stata’s default maximum-likelihood convergence criteria were applied, including a relative coefficient tolerance of 1×10−4, a log-likelihood tolerance of 0, and a scaled-gradient tolerance of 1×10−5.
In addition, the goodness of fit between these two models was compared according to the following indices: log likelihood (LL), Akaike information criterion (AIC), and Bayesian information criterion (BIC). The model with larger LL and smaller AIC and BIC was considered superior to the other model. WTP for each non-cost attribute level was calculated as the negative ratio of the corresponding mean attribute coefficient to the mean OOP cost coefficient, using the unrounded coefficients from the mixed logit model. The 95% confidence interval for WTP were derived using the Krinsky–Robb parametric simulation method. Specifically, 10,000 joint draws of the mean attribute coefficient and mean cost coefficient were generated from their estimated bivariate normal distribution using the model variance–covariance matrix. WTP was calculated for each draw, and the 2.5th and 97.5th percentiles of the simulated WTP distribution were used as the lower and upper confidence limits, respectively. A fixed random seed of 202610 was used to ensure reproducibility.36
Heterogeneity of different covariates in the preferences for hearing aids attributes was examined by adding the interaction term between each covariate (age, sex, educational level, marital status, dementia risk, and hearing aids use) and attribute-level coefficients with significant SD into the superior model, with separate interaction models estimated for each covariate. If the superior model was mixed logit model, the interaction terms were introduced as fixed effects while the hearing aids attributes with significant SD were introduced as random effects. Interaction analyses were restricted to coefficients with statistically significant SD because these SDs provided evidence of unexplained preference heterogeneity that could potentially be accounted for by observed participant characteristics. This parsimonious approach also limited model overparameterization and the number of exploratory interaction tests, and has been used in previous health-related DCE studies.37,38
To examine the possible effect of significant differences between those who passed and failed the consistency check on the hearing aid preferences, sensitivity analysis among all eligible participants, including those who failed consistency check was conducted. The conditional logit model, mixed logit model, along with the heterogeneity analysis were conducted in consistent with the primary analysis.
All statistical analyses were conducted using Stata 15.0 and SAS 9.2. Statistical significance was defined as a two-tailed P≤0.05.
Results
Characteristics of Participants
Basic characteristics of the older individuals are presented in Table 2. The final analytic sample included 346 participants. Among them, 54.34% of participants were male and the average age was 72.33 ± 8.80 years. The sex distribution among participants in our study was similar to that among participants in a population-based survey in four provinces of China on ear and hearing disorders, in which 52.16% of participants were male.5 However, the age distribution was inconsistent. Specifically, 40.17% of the analytic sample in our study were aged 60–69 years, 35.84% were aged 70–79 years, and 23.99% were aged 80 years or older, compared with 52.96%, 34.09%, and 12.95%, respectively, in the target population. Moreover, participants with a primary school education or below were less common in our final analytic sample than in the target population (54.05% vs 62.34%). More than 80% of the participants were married or partnered. Nearly half of the participants (159, 45.95%) had the educational level of secondary school or above, and over half of them were at dementia risk (182, 52.60%) and used hearing aids (183, 52.89%) respectively. Considering the comparison between individuals passing or failing the consistency check, significant differences were found concerning age, educational level, and whether with dementia risk among the two groups. Specifically, individuals who passed consistency check were comparatively older, with lower educational level, and with less dementia risk.
Table 2.
Comparisons on Basic Characteristics of the Participants According to the Consistency Check
| Variables | Overall (n=413) | Passed Consistency Check (n=346) |
Failed Consistency Check (n=67) |
χ2/ta | P value |
|---|---|---|---|---|---|
| N (%) | N (%) | N (%) | |||
| Age in years, mean (SD) | 71.81 (8.85) | 72.33 (8.80) | 69.13 (8.67) | 2.730 | 0.007 |
| Sex | 0.340 | 0.560 | |||
| Male | 227 (54.96) | 188 (54.34) | 39 (58.21) | ||
| Female | 186 (45.04) | 158 (45.66) | 28 (41.79) | ||
| Educational level | 7.461 | 0.006 | |||
| Primary school or below | 211 (51.09) | 187 (54.05) | 24 (35.82) | ||
| Secondary school or above | 202 (48.91) | 159 (45.95) | 43 (64.18) | ||
| Marital status | 0.260 | 0.610 | |||
| Unmarriedb | 64 (15.50) | 55 (15.90) | 9 (13.43) | ||
| Married/partnered | 349 (84.50) | 291 (84.10) | 58 (86.57) | ||
| With dementia risk | 6.997 | 0.008 | |||
| No | 184 (44.55) | 164 (47.40) | 20 (29.85) | ||
| Yes | 229 (55.45) | 182 (52.60) | 47 (70.15) | ||
| Hearing aids use | 0.016 | 0.900 | |||
| No | 194 (46.97) | 163 (47.11) | 31 (46.27) | ||
| Yes | 219 (53.03) | 183 (52.89) | 36 (53.73) |
Notes: aChi-square test was conducted for categorical variables and T-test was conducted for continuous variables; b: Unmarried included those were unmarried (3, 0.73% for the total sample), divorced (4, 0.97% for the total sample), and widowed (57, 13.80% for the total sample).
Abbreviation: SD, Standard deviation.
Preference for Hearing Aids Attributes
Both conditional and mixed logit regression models were conducted to estimate the preference for hearing aids attributes among older adults with hearing loss. Several indices were used to compare the goodness of fit between the above two models, which is shown in Table 3. The Akaike information criterion (AIC) and Bayesian information criterion (BIC) values of the mixed logit model were both smaller than those of conditional logit model, while the log likelihood was larger than that of the conditional logit model. Thus, the goodness of fit of the mixed logit model was better, and results of the preference for hearing aids attributes were mainly based on mixed logit model. Nevertheless, results based on the conditional logit model are also shown in Table S2.
Table 3.
Comparisons on the Goodness of Fit Between Conditional Logit Model and Mixed Logit Model
| Indices | Conditional Logit Model | Mixed Logit Model |
|---|---|---|
| Log likelihood (LL) | −1866.188 | −1588.665 |
| Akaike information criterion (AIC) | 3750.376 | 3213.330 |
| Bayesian information criterion (BIC) | 4586.176 | 3337.367 |
| Sample | 346 | 346 |
| Observation | 7266 | 7266 |
Table 4 displays the results of preference for hearing aids attributes among older adults with hearing loss. In general, the opt-out option exerted negative effect on the preference selection among participants (coefficient for alternative-specific constant (ASC)=−3.90, 95% CI: −5.09, −2.70), indicating that participants preferred to select one of the hearing aid option in one choice set, rather than choosing the opt-out option. Among the six hearing aids attributes included in this study, five attributes (Out-of-pocket (OOP) cost, accessibility, effectiveness in noisy settings, style of hearing aids, and battery type) exerted significant effect on the preferences for hearing aids use. In detail, subjects showed a negative preference for higher OOP cost (Coefficient =−0.03, 95% CI: −0.04, −0.02). Individuals preferred to select over-the-counter (OTC) hearing aids compared to audiologist-fitted hearing aids (Coefficient = 0.89, 95% CI: 0.60, 1.18). Concerning effectiveness of hearing aids in noisy settings, compared with not very effective, effective (Coefficient = 1.92, 95% CI: 1.48, 2.36) and very effective (Coefficient = 3.01, 95% CI: 2.41, 3.62) exerted positive effects on the selection of hearing aids among the participants. Furthermore, older adults with hearing loss preferred to choose receiver-in-canal (RIC) hearing aids (Coefficient = 0.40, 95% CI: 0.05, 0.76) and hearing aids with rechargeable batteries (Coefficient = 0.38, 95% CI: 0.17, 0.59).
Table 4.
Results of Preference for Hearing Aids Attributes and WTP Among Older Adults with Hearing Loss (Based on Mixed Logit Model)
| Attributes and levels | Mean | SD | WTP in RMB (95% CI) |
||
|---|---|---|---|---|---|
| Coefficient (95% CI) | P | Coefficient (95% CI) | P | ||
| ASC (Opt-out) | −3.90 (−5.09, −2.70) | <0.001 | 4.11 (2.97, 5.25) | <0.001 | - |
| OOP costa | −0.03 (−0.04, −0.02) | <0.001 | 0.04 (0.03, 0.05) | <0.001 | - |
| Accessibility (Ref. Audiologist-fitted hearing aids) | |||||
| OTC hearing aids | 0.89 (0.60, 1.18) | <0.001 | 1.09 (0.66, 1.53) | <0.001 | 3304.63 (2289.03, 4541.81) |
| Effectiveness in noisy settings (Ref. Not very effective) | |||||
| Effective | 1.92 (1.48, 2.36) | <0.001 | 0.06 (−0.15, 0.26) | 0.590 | 7148.83 (5600.65, 9143.93) |
| Very effective | 3.01 (2.41, 3.62) | <0.001 | 1.38 (0.96, 1.79) | <0.001 | 11,190.73 (9025.80, 14,102.61) |
| Style (Ref. BTE) | |||||
| RIC | 0.40 (0.05, 0.76) | 0.027 | 1.67 (1.21, 2.12) | <0.001 | 1492.42 (151.97, 2882.33) |
| ITC | −0.36 (−0.66, −0.07) | 0.015 | 1.05 (0.48, 1.62) | <0.001 | −1355.15 (−2574.49, −279.11) |
| Battery type (Ref. Non-rechargeable) | |||||
| Rechargeable | 0.38 (0.17, 0.59) | <0.001 | −0.19 (−0.69, 0.31) | 0.445 | 1419.41 (673.96, 2214.06) |
| Connectivity (Ref. Without mobile app) | |||||
| With mobile app | 0.10 (−0.09, 0.29) | 0.316 | −0.25 (−0.56, 0.70) | 0.127 | 366.25 (−359.37, 1120.06) |
Notes: aIn the unit of one hundred RMB.
Abbreviations: WTP, Willingness to pay; SD, Standard deviation; CI, Confidence interval; ASC, Alternative-specific constant; OOP, Out-of-pocket; BTE, Behind-the-ear; RIC, Receiver-in-canal; ITC, In-the-canal.
The estimated mean willingness to pay (WTP) for hearing aid attributes is shown in Table 4. On average, individuals were willing to spend more than 10000 RMB (USD 1447) to obtain a hearing aid which was very effective in noisy settings than a not very effective one. WTP for OTC hearing aids was over 3300 RMB (USD 478) instead of audiologist-fitted hearing aids. Participants were willing to spend nearly 1500 RMB (USD 217) to get a RIC hearing aid instead of a behind-the-ear (BTE) hearing aid, while more than 1300 RMB (USD 188) should be compensated to the participants to make them select an in-the-canal (ITC) hearing aid than a BTE hearing aid. WTP for hearing aids with rechargeable batteries was 1419 RMB (USD 205) instead of hearing aids with non-rechargeable batteries.
Heterogeneity in Preferences for Hearing Aids Attributes
The standard deviation (SD) of random parameters (means) is also presented in Table 4. The magnitude and significant of SD indicated the heterogeneity of individual preferences for each specific attribute and level. The distribution of SD across OOP cost, OTC hearing aids, very effective in noisy settings, RIC hearing aids, and ITC hearing aids exhibited significant statistical differences, indicating heterogeneity in participants’ preferences for above hearing aids attributes (OOP cost, accessibility, effectiveness in noisy settings, and style). Therefore, six interaction models based on mixed logit model were conducted, and the results are shown in Table 5. Compared to participants with a primary or lower education level, unmarried, and without hearing aids use, OOP cost had a smaller impact on the preference for hearing aids selection among those with a secondary or higher education level, married/partnered, and with hearing aids use, whereas it exerted a greater effect among those with a dementia risk than among those without a dementia risk. Moreover, OTC hearing aids were more important in the higher education group than in the lower education group, while they were less important among the dementia‑risk group than among the non‑dementia risk group. Concerning effectiveness in noisy settings, individuals using hearing aids showed a higher preference for a hearing aid which was very effective in noisy settings, compared to individuals who had never used hearing aids. Lastly, the negative preference for ITC hearing aids has significantly decreased among hearing aids users.
Table 5.
Results of the Interactions Between Covariates and the Preference for Hearing Aids Attributes Among Older Adults with Hearing Loss Based on Mixed Logit Model, Coefficient (95% CI)
| Attributes, Levels and Interactions | Age (Oldest Old=1) | Sex (Female=1) | Educational Level (Secondary School or Above=1) |
Marital Status (Married or Partnered=1) |
Dementia Risk (Yes=1) |
Hearing Aids Use (Yes=1) |
|---|---|---|---|---|---|---|
| Interactions with covariates | ||||||
| ASC (Opt-out) *covariate | −0.04 (−1.95, 1.87) | −1.22 (−2.60, 0.15) | 2.56 (1.13, 4.00)*** | −0.19 (−2.09, 1.70) | −1.78 (−3.10, −0.47)** | 0.14 (−1.28, 1.57) |
| OOP cost*covariate | −0.01 (−0.02, 0.01) | 0.01 (−0.01, 0.02) | 0.02 (0.01, 0.04)*** | 0.02 (0.01, 0.04)** | −0.01 (−0.02, −0.01)* | 0.02 (0.01, 0.03)*** |
| OTC hearing aids*covariate | 0.06 (−0.45, 0.57) | 0.21 (−0.23, 0.66) | 0.52 (0.07, 0.98)* | 0.03 (−0.56, 0.63) | −0.63 (−1.06, −0.20)** | −0.27 (−0.70, 0.16) |
| Very effective*covariate | −0.21 (−0.86, 0.45) | −0.01 (−0.57, 0.56) | 0.08 (−0.47, 0.64) | 0.74 (−0.03, 1.52) | −0.27 (−0.80, 0.26) | 0.92 (0.36, 1.48)*** |
| RIC*covariate | 0.07 (−0.65, 0.80) | −0.10 (−0.75, 0.55) | −0.25 (−0.91, 0.40) | 0.33 (−0.58, 1.23) | −0.55 (−1.22, 0.11) | 0.07 (−0.57, 0.72) |
| ITC*covariate | 0.13 (−0.54, 0.80) | 0.03 (−0.51, 0.58) | −0.18 (−0.72, 0.36) | 0.38 (−0.50, 1.27) | 0.24 (−0.30, 0.77) | 0.57 (0.02, 1.12)** |
| Attributes and levels | ||||||
| ASC (Opt-out) | −4.13 (−5.62, −2.64)*** | −3.37 (−4.59, −2.16)*** | −5.52 (−7.26, −3.77)*** | −3.83 (−5.57, −2.08)*** | −3.27 (−4.50, −2.02)*** | −4.20 (−5.63, −2.77)*** |
| OOP costa | −0.02 (−0.03, −0.02)*** | −0.03 (−0.04, −0.02)*** | −0.04 (−0.05, −0.03)*** | −0.05 (−0.06, −0.03)*** | −0.02 (−0.03, −0.01)*** | −0.04 (−0.05, −0.03)*** |
| Accessibility (Ref. Audiologist-fitted hearing aids) | ||||||
| OTC hearing aids | 0.88 (0.58, 1.17) *** | 0.73 (0.42, 1.04)*** | 0.63 (0.33, 0.94)*** | 0.84 (0.26, 1.42)** | 1.15 (0.78, 1.52)*** | 1.01 (0.65, 1.38)*** |
| Effectiveness in noisy settings (Ref. Not very effective) | ||||||
| Effective | 2.02 (1.52, 2.53)*** | 1.88 (1.47, 2.29)*** | 1.91 (1.48, 2.34)*** | 1.89 (1.43, 2.35)*** | 1.89 (1.49, 2.29)*** | 1.89 (1.50, 2.28)*** |
| Very effective | 3.21 (2.44, 3.98)*** | 3.02 (2.35, 3.68)*** | 2.98 (2.29, 3.67)*** | 2.33 (1.45, 3.20)*** | 3.17 (2.51, 3.83)*** | 2.42 (1.85, 2.98)*** |
| Style (Ref. BTE) | ||||||
| RIC | 0.44 (0.16, 0.87)* | 0.36 (−0.11, 0.83) | 0.54 (0.05, 1.03)* | 0.10 (−0.75, 0.96) | 0.69 (0.17, 1.21)** | 0.34 (−0.15, 0.83) |
| ITC | −0.33 (−0.67, 0.01) | −0.37 (−0.75, 0.02) | −0.27 (−0.64, 0.09) | −0.66 (−1.51, 0.18) | −0.52 (−0.93, −0.11)* | −0.70 (−1.14, −0.27)** |
| Battery type (Ref. Non-rechargeable) | ||||||
| Rechargeable | 0.37 (0.14, 0.60)*** | 0.40 (0.19, 0.61)*** | 0.37 (0.16, 0.59)*** | 0.34 (0.14, 0.55)*** | 0.35 (0.15, 0.55)*** | 0.36 (0.16, 0.57)*** |
| Connectivity (Ref. Without mobile app) | ||||||
| With mobile app | 0.11 (−0.10, 0.32) | 0.13 (−0.07, 0.33) | 0.12 (−0.08, 0.32) | 0.10 (−0.11, 0.32) | 0.16 (−0.03, 0.35) | 0.11 (−0.09, 0.30) |
Notes: aIn the unit of one hundred RMB. ***P≤0.001, **P≤0.01, *P≤0.05.
Abbreviations: CI, Confidence interval; ASC, Alternative-specific constant; OOP, Out-of-pocket; BTE, Behind-the-ear; RIC, Receiver-in-canal; ITC, In-the-canal.
Sensitivity Analysis
Findings for conditional logit model (Table S3), mixed logit model (Table S4), as well as the heterogeneity analysis (Table S5) showed similar magnitude and significance compared with main analysis, indicating the robustness of the overall preference estimates after excluding participants who failed the consistency check.
Discussion
This study examined preferences for hearing aids attributes using discrete choice experiment (DCE) among 346 community-dwelling older adults with HL from six cities in China. We also shed light on their preferences with different demographic and socioeconomic features. OOP cost, accessibility, effectiveness in noisy settings, style of hearing aids and battery type were all significantly associated with the selection of hearing aids, in which effectiveness in noisy settings served as the most important hearing aids attribute, followed by accessibility. In addition, findings from heterogeneity analysis also indicated that preferences for hearing aids attributes differed across educational level, marital status, dementia risk, and hearing aids use. In detail, OOP cost had a smaller impact on the preference for hearing aids selection among those with secondary or above educational level and married/partnered, whereas exerted a greater effect among those with dementia risk. Accessibility of hearing aids was more important in the higher education group while less important among those with dementia risk group. Participants who had ever worn hearing aids had a higher preference for the selection of a hearing aid which was very effective in noisy settings, whereas they were less influenced by the negative effect of high OOP cost on the preferences for hearing aids. Findings of our study provided updated evidence on the preferences when selecting hearing aids among older adults with HL. Hearing health care should pay more attention to the improvement in the effectiveness of hearing aids in noisy settings and focused more on individuals who have lower education level, are unmarried, have dementia risk, and have never worn hearing aids.
Findings of the DCE indicated that older individuals with HL preferred to select hearing aids with less OOP cost, more effectiveness in noisy settings, receiver-in-canal (RIC) hearing aids, and rechargeable batteries, particularly attaching the greatest importance to higher level of effectiveness in noisy settings. Our findings were similar to previous studies. First, studies conducted DCEs among Chinese and US older individuals with hearing loss consistently found that effectiveness in noisy settings exerted the most crucial effect on the preferences for hearing aids.17,24 Second, longer battery life was found to increase participants’ preferences for the selection of hearing aids,17 which was also consistent with our results identifying rechargeable batteries as a significant factor in the preferences for hearing aids. Third, because of the three styles of hearing aids in our study mainly reflecting appearance and severity of aural fullness, participants’ preference for RIC hearing aids embodied the pursuit of comfort when wearing hearing aids. Thus, this finding was similar to a previous study in which rare physical discomfort when hearing aids were in their ear increased the preferences for hearing aids selection.24
Our study further found that accessibility of hearing aids served as the second most important factor of the preferences for hearing aids, and participants with HL preferred to select OTC hearing aids featured by higher accessibility than audiologist-fitted hearing aids. According to the latest world report on hearing, self-fitted OTC hearing aids served as one of the most important advancement in hearing technology and hearing service delivery recently.3 Low accessibility to hearing healthcare was also deemed as one of the largest challenge for the ear and hearing care system in developing countries, according to WHO.3 Considering that recent intervention studies on the effectiveness of OTC hearing aids have concluded that self-fitted OTC hearing aids yielded comparable effectiveness to audiologist-fitted hearing aids during 6-week trial,39 and that individuals were more satisfied with the effectiveness of self-adjusted OTC hearing aids,40 it is imperative to promote the popularization of OTC hearing aids among Chinese older adults. However, considering that OTC hearing aids were not uniformly applicable to older adults with hearing loss and might be appropriate only for older adults with mild to moderate hearing loss,41,42 the measurement of OTC hearing aids preference and the popularization of OTC hearing aids should be based on audiometric assessment for hearing loss severity.
Moreover, this study revealed individual sociodemographic differences across educational level, marital status, and hearing aids use in the preferences for hearing aids attributes. First, OOP cost had a smaller impact on the preference for hearing aids selection among those with secondary or higher educational level, married/partnered, and with hearing aids use. One possible explanation might be that higher educational level and married status stood for higher socioeconomic status, which might be related to higher ear and hearing care literacy,43 and less sensitivity to OOP cost. Further, income may have accounted for some of the observed educational difference and participants with higher educational level were more likely to afford the regular maintenance on hearing aids.44 Participants who had ever used hearing aids might have experienced the improvement in residual auditory function and communicative function,6 thus attaching less importance to the OOP cost. Second, OTC hearing aids were more important in the higher education group than in the lower education group, which may be due to the fact that online air conduction tests and remote self-fitting are needed for OTC hearing aids,45 which require more health literacy. Third, individuals who had ever used hearing aids showed a higher preference for the selection of a hearing aid which was very effective in noisy settings, which might result from more attention paid to the effectiveness of hearing aids when experiencing hearing aids benefits. Nevertheless, it was noteworthy that these analyses were observational interactions within a stated-preference experiment and are vulnerable to residual confounding, multiple testing, model instability and post‑hoc interpretation. Thus, the potential explanations and reasons require confirmation through causal inference design and should be interpreted as exploratory and hypothesis-generating findings.
Heterogeneity in the preferences for hearing aids attributes was also found among participants with different dementia‑risk status. Previous studies have revealed the specific determinants of hearing aids adoption among those with dementia risk, including greater difficulties in maintaining hearing aids use after fitting, limited executive function, and poor using skills.12 Researchers also indicated that absence of expectancy about the positive consequences of hearing aids was a barrier to their use among community-dwelling older adults with age-related HL and dementia.46 Therefore, older individuals with dementia risk might hardly experience positive outcomes from hearing aids and may care more about OOP cost. In addition, older individuals with dementia risk also faced difficulties in mastering knowledge and skills, and thus valued OTC hearing aids less than those without dementia risk. However, the AD8 scale used in our study was a screening tool for dementia risk and cannot be used for clinical diagnosis of dementia or cognitive impairment. Thus, the subgroup interpretation involving dementia risk should also be framed as exploratory and hypothesis-generating.
It was also noteworthy that participants who failed the consistency check were younger, more educated, and more likely to be at risk of dementia than those retained in the primary analysis. This pattern suggests that difficulties in completing the DCE tasks may be more closely related to cognitive vulnerability than to older age or lower educational attainment alone. Their exclusion may therefore have underrepresented the preferences of cognitively vulnerable older adults, including those who appear demographically well equipped to engage with hearing aid information but may still experience difficulties in comparing complex attributes or using self-fitted OTC devices. These findings highlight that age and education alone may be insufficient to identify individuals requiring decision support, and that cognitive screening, simplified information, and professional or caregiver assistance may be needed when providing OTC hearing aids. Nevertheless, sensitivity analyses including all eligible participants yielded results consistent with the primary analysis, supporting the robustness of the overall preference estimates.
Our study has several limitations. First, although the updated and most relevant attributes for hearing aids were selected in the first step of the DCE based on a literature review and expert consultation, we were unable to include all attributes of hearing aid preferences considering the response reliability of participants. Second, differences in the hearing aids accessibility between cities or provinces in China were much more obvious than those in any other country, which limited the extrapolation of the study results. Third, the DCE method was used to measure preferences for hearing aids selection, which is unable to achieve perfect simulation of subjective choice behavior. Further comparison studies between preferences for hearing aids attributes and selection among older hearing aids users in the real world should be conducted in the future. Fourth, neutrality assessment for the accessibility attribute was absent during piloting, which may bring about framing effects in the OTC attribute. However, according to the latest report on hearing from the WHO,3 as well as recent high-level evidence on the effectiveness between audiologist-fitted hearing aids and OTC hearing aids,39,47 accessibility was considered as one of the most obvious difference between these hearing aids, and the effectiveness between these two hearing aids were comparable. Thus, our description on audiologist-fitted hearing aids and OTC hearing aids was deemed to be neutral. Fifth, although participants did not report any difficulty in interpreting the attributes and levels of hearing aids in the pilot study, this study failed to conduct a quantitative analysis to cognitively test the DCE materials. Sixth, despite the use of age- and sex-based quota sampling, the final sample differed from the target population in age and educational distribution, with an overrepresentation of adults aged 80 years or older and an underrepresentation of those with lower educational attainment. Selection bias may therefore remain, and the generalizability of the findings to the broader population of older adults with hearing loss should be interpreted cautiously. Seventh, our findings applied to self-reported hearing loss, rather than clinically characterized hearing impairment via audiometric assessment, and thus were unable to take hearing loss severity into consideration. Moreover, AD8 is a screening tool for dementia risk, not a diagnostic assessment for dementia or cognitive impairment. Therefore, the interpretation of preferences for hearing aids attribute, especially OTC hearing aids, and the heterogeneity analysis should be very cautious.
To conclude, this study identified preferences for several hearing aids attributes among older adults with HL from six cities in China, including OOP cost, accessibility, effectiveness in noisy settings, styles of hearing aids, and battery type. Furthermore, significant heterogeneity in demographic and socioeconomic features containing educational level, marital status dementia risk, and hearing aids use. OTC accessibility appears to be valued by respondents and may warrant further evaluation as part of broader hearing-care strategies, particularly if accompanied by safeguards, education and clinical referral pathways. Nevertheless, the promotion of OTC hearing aids should be very cautious and take the hearing loss severity among older adults into consideration.
Funding Statement
This study was funded by the National Natural Science Foundation of China (grant number 82404385), China Disabled Person’s Federation Hearing and Speech Program (grant number 2024CDPFHS-08) and Clinical Medicine Plus X - Young Scholars Project of Peking University-the Fundamental Research Funds for the Central Universities (grant number PKU2024LCXQ035; PKU2025PKULCXQ041).
Data Sharing Statement
The datasets used and analyzed in this study are available from the corresponding author (Yanan Luo) on reasonable request.
Ethical Approval
Medical ethics approval of this study was approved by the Institutional Review Board of Peking University (#IRB00001052-24123). Informed consent for the data collection and the use of the data was obtained from all subjects. This study complies with the Declaration of Helsinki.
Author Contributions
All authors made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agreed to be accountable for all aspects of the work.
Disclosure
The authors report no conflicts of interest in this work.
References
- 1.Blazer DG, Tucci DL. Hearing loss and psychiatric disorders: a review. Psychol Med. 2019;49(6):891–14. doi: 10.1017/s0033291718003409 [DOI] [PubMed] [Google Scholar]
- 2.World Health Organization. Deafness and hearing loss. Available from: https://www.who.int/news-room/fact-sheets/detail/deafness-and-hearing-loss. Accessed July 18, 2025.
- 3.World Health Organization. World report on hearing. 2021. [Google Scholar]
- 4.Zheng X, Chen G, Song X, et al. Twenty-year trends in the prevalence of disability in China. Bull World Health Organ. 2011;89(11):788–797. doi: 10.2471/blt.11.089730 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Gong R, Hu X, Gong C, et al. Hearing loss prevalence and risk factors among older adults in China. Int J Audiol. 2018;57(5):354–359. doi: 10.1080/14992027.2017.1423404 [DOI] [PubMed] [Google Scholar]
- 6.Carlson ML. Cochlear implantation in adults. N Engl J Med. 2020;382(16):1531–1542. doi: 10.1056/NEJMra1904407 [DOI] [PubMed] [Google Scholar]
- 7.Goman AM, Tan N, Pike JR, et al. Effects of hearing intervention on falls in older adults: findings from a secondary analysis of the ACHIEVE randomised controlled trial. Lancet Public Health. 2025;10(6):e492–e502. doi: 10.1016/s2468-2667(25)00088-x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Orji A, Kamenov K, Dirac M, Davis A, Chadha S, Vos T. Global and regional needs, unmet needs and access to hearing aids. Int J Audiol. 2020;59(3):166–172. doi: 10.1080/14992027.2020.1721577 [DOI] [PubMed] [Google Scholar]
- 9.He P, Wen X, Hu X, et al. Hearing aid acquisition in Chinese older adults with hearing loss. Am J Public Health. 2018;108(2):241–247. doi: 10.2105/ajph.2017.304165 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Fuentes-López E, Fuente A, Cardemil F, Valdivia G, Albala C. Prevalence and associated factors of hearing aid use among older adults in Chile. Int J Audiol. 2017;56(11):810–818. doi: 10.1080/14992027.2017.1337937 [DOI] [PubMed] [Google Scholar]
- 11.Powell DS, Garcia Morales EE, Oh E, et al. Dementia and hearing aid use and cessation: a national study. Am J Audiol. 2023;32(4):898–907. doi: 10.1044/2023_aja-23-00038 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Gregory S, Billings J, Wilson D, Livingston G, Schilder AG, Costafreda SG. Experiences of hearing aid use among patients with mild cognitive impairment and Alzheimer’s disease dementia: a qualitative study. SAGE Open Med. 2020;8:2050312120904572. doi: 10.1177/2050312120904572 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Chang HP, Ho CY, Chou P. The factors associated with a self-perceived hearing handicap in elderly people with hearing impairment--results from a community-based study. Ear Hear. 2009;30(5):576–583. doi: 10.1097/AUD.0b013e3181ac127a [DOI] [PubMed] [Google Scholar]
- 14.Ng JH, Loke AY. Determinants of hearing-aid adoption and use among the elderly: a systematic review. Int J Audiol. 2015;54(5):291–300. doi: 10.3109/14992027.2014.966922 [DOI] [PubMed] [Google Scholar]
- 15.Grutters JP, Joore MA, Kessels AG, Davis AC, Anteunis LJ. Patient preferences for direct hearing aid provision by a private dispenser. A discrete choice experiment. Ear Hear. 2008;29(4):557–564. doi: 10.1097/AUD.0b013e3181734a19 [DOI] [PubMed] [Google Scholar]
- 16.Meister H, Grugel L, Walger M, von Wedel H, Meis M. Utility and importance of hearing-aid features assessed by hearing-aid acousticians. Trends Amplif. 2010;14(3):155–163. doi: 10.1177/1084713810385713 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Zhu D, Shi X, Nicholas S, Ye X, Chen S, He P. Preferences for hearing aid attributes among people with moderate or greater hearing loss in rural china: a discrete choice experiment. Patient Preference Adherence. 2020;14:643–652. doi: 10.2147/ppa.S248522 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Yong M, Willink A, McMahon C, et al. Access to adults’ hearing aids: policies and technologies used in eight countries. Bull World Health Organ. 2019;97(10):699–710. doi: 10.2471/blt.18.228676 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Zheng H, Wong LLN, Hickson L. Barriers to hearing aid adoption among older adults in mainland China. Int J Audiol. 2023;62(9):814–825. doi: 10.1080/14992027.2022.2105263 [DOI] [PubMed] [Google Scholar]
- 20.Beijing Municipal Bureau of Statistics, Survey Office of the National Bureau of Statistics in Beijing. Beijing Statistical Yearbook 2025. China Statistics Press; 2025. [Google Scholar]
- 21.Hebei Provincial Bureau of Statistics, Survey Office of the National Bureau of Statistics in Hebei. Hebei Statistical Yearbook 2024. China Statistics Press; 2024. [Google Scholar]
- 22.de Bekker-Grob EW, Donkers B, Jonker MF, Stolk EA. Sample size requirements for discrete-choice experiments in healthcare: a practical guide. Patient. 2015;8(5):373–384. doi: 10.1007/s40271-015-0118-z [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Bridges JF, Hauber AB, Marshall D, et al. Conjoint analysis applications in health--a checklist: a report of the ISPOR good research practices for conjoint analysis task force. Value Health. 2011;14(4):403–413. doi: 10.1016/j.jval.2010.11.013 [DOI] [PubMed] [Google Scholar]
- 24.Bridges JF, Lataille AT, Buttorff C, White S, Niparko JK. Consumer preferences for hearing aid attributes: a comparison of rating and conjoint analysis methods. Trends Amplif. 2012;16(1):40–48. doi: 10.1177/1084713811434617 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.World Health Organization. How to conduct a discrete choice experiment for health workforce recruitment and retention in remote and rural areas: a user guide with case studies. World Health Organization. 2012. [Google Scholar]
- 26.Hensher D, Rose J, Greene W. Applied Choice Analysis: A Primer. Cambridge University Press; 2005. [Google Scholar]
- 27.Brinkmann M, Diedrich L, Krauth C, Robra BP, Stahmeyer JT, Dreier M. General populations’ preferences for colorectal cancer screening: rationale and protocol for the discrete choice experiment in the SIGMO study. BMJ Open. 2021;11(1):e042399. doi: 10.1136/bmjopen-2020-042399 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Sonnega A, Faul JD, Ofstedal MB, Langa KM, Phillips JW, Weir DR. Cohort profile: the health and retirement study (HRS). Int J Epidemiol. 2014;43(2):576–585. doi: 10.1093/ije/dyu067 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Wei J, Li K, Kim Y, et al. Initiation of hearing aids use and incident dementia among mid-to-late life adults: the health and retirement study. J Geriatr Psychiatry Neurol. 2025;38(3):172–179. doi: 10.1177/08919887241302107 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Galvin JE, Roe CM, Powlishta KK, et al. The AD8: a brief informant interview to detect dementia. Neurology. 2005;65(4):559–564. doi: 10.1212/01.wnl.0000172958.95282.2a [DOI] [PubMed] [Google Scholar]
- 31.Chen HH, Sun FJ, Yeh TL, et al. The diagnostic accuracy of the Ascertain Dementia 8 questionnaire for detecting cognitive impairment in primary care in the community, clinics and hospitals: a systematic review and meta-analysis. Fam Pract. 2018;35(3):239–246. doi: 10.1093/fampra/cmx098 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Yang L, Yan J, Jin X, et al. Screening for dementia in older adults: comparison of mini-mental state examination, mini-cog, clock drawing test and AD8. PLoS One. 2016;11(12):e0168949. doi: 10.1371/journal.pone.0168949 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Li T, Wang HL, Yang YH, Galvin JE, Morris JC, Yu X. The reliability and validity of Chinese version of AD8. Zhonghua Nei Ke Za Zhi. 2012;51(10):777–780. [PubMed] [Google Scholar]
- 34.Hauber AB, González JM, Groothuis-Oudshoorn CG, et al. Statistical methods for the analysis of discrete choice experiments: a report of the ISPOR conjoint analysis good research practices task force. Value Health. 2016;19(4):300–315. doi: 10.1016/j.jval.2016.04.004 [DOI] [PubMed] [Google Scholar]
- 35.Lancsar E, Fiebig DG, Hole AR. Discrete choice experiments: a guide to model specification, estimation and software. Pharmacoeconomics. 2017;35(7):697–716. doi: 10.1007/s40273-017-0506-4 [DOI] [PubMed] [Google Scholar]
- 36.Hole AR, Kolstad JR. Mixed logit estimation of willingness to pay distributions: a comparison of models in preference and WTP space using data from a health-related choice experiment. Empirical Econ. 2012;42(2):445–469. doi: 10.1007/s00181-011-0500-1 [DOI] [Google Scholar]
- 37.Muntingh ADT, Hoogendoorn AW, Van Schaik DJF, et al. Patient preferences for a guided self-help programme to prevent relapse in anxiety or depression: a discrete choice experiment. PLoS One. 2019;14(7):e0219588. doi: 10.1371/journal.pone.0219588 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Sturkenboom R, Keszthelyi D, Masclee AAM, Essers BAB. Discrete choice experiment reveals strong preference for dietary treatment among patients with irritable bowel syndrome. Clin Gastroenterol Hepatol. 2022;20(11):2628–2637. doi: 10.1016/j.cgh.2022.02.016 [DOI] [PubMed] [Google Scholar]
- 39.De Sousa KC, Manchaiah V, Moore DR, Graham MA, Swanepoel W. Effectiveness of an over-the-counter self-fitting hearing aid compared with an audiologist-fitted hearing aid: a randomized clinical trial. JAMA Otolaryngol Head Neck Surg. 2023;149(6):522–530. doi: 10.1001/jamaoto.2023.0376 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Nelson PB, Perry TT, Gregan M, VanTasell D. Self-adjusted amplification parameters produce large between-subject variability and preserve speech intelligibility. Trends Hear. 2018;22:2331216518798264. doi: 10.1177/2331216518798264 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Tucci DL, Califf RM. Over-the-counter hearing aids: from research to policy to practice. JAMA. 2022;328(23):2299–2300. doi: 10.1001/jama.2022.22123 [DOI] [PubMed] [Google Scholar]
- 42.Jiang F, Dayimu A, Dong Q, et al. Over-the-counter hearing aids reduce depression symptoms in older adults with hearing loss: a randomized controlled trial. Nature Mental Health. 2025;3(5):498–506. doi: 10.1038/s44220-025-00408-4 [DOI] [Google Scholar]
- 43.Furuya Y, Kondo N, Yamagata Z, Hashimoto H. Health literacy, socioeconomic status and self-rated health in Japan. Health Promot Int. 2015;30(3):505–513. [DOI] [PubMed] [Google Scholar]
- 44.Fischer ME, Cruickshanks KJ, Wiley TL, Klein BE, Klein R, Tweed TS. Determinants of hearing aid acquisition in older adults. Am J Public Health. 2011;101(8):1449–1455. doi: 10.2105/AJPH.2010.300078 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45.US. Food and Drug Adiministration. Medical devices; ear, nose, and throat devices; establishing over-the-counter hearing aids. 2022. Available from: https://www.federalregister.gov/documents/2022/08/17/2022-17230/medical-devices-ear-nose-and-throat-devices-establishing-over-the-counter-hearing-aids. Accessed February 15, 2026.
- 46.Hooper E, Brown LJE, Cross H, Dawes P, Leroi I, Armitage CJ. Systematic review of factors associated with hearing aid use in people living in the community with dementia and age-related hearing loss. J Am Med Dir Assoc. 2022;23(10):1669–1675.e16. doi: 10.1016/j.jamda.2022.07.011 [DOI] [PubMed] [Google Scholar]
- 47.De Sousa KC, Almufarrij I, Kruger M, Manchaiah V, Munro KJ, Swanepoel W. Effectiveness of over-the-counter hearing aids versus professionally fitted devices: a systematic review and meta-analysis. Otolaryngol Head Neck Surg Jun. 2026;17. doi: 10.1002/ohn.70306 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The datasets used and analyzed in this study are available from the corresponding author (Yanan Luo) on reasonable request.