Introduction

In the United Kingdom (UK), one in 5 adults experiences hearing loss, with approximately 1·2 million people suffering from severe–to–profound loss [1]. By 2030, hearing loss is expected to be a top 10 health burden, costing the UK £38·6 billion per year [2]. A cochlear implant (CI), which bypasses inner ear damage providing a clearer auditory signal to the brain (compared to hearing aids), is an effective treatment for this degree of hearing loss [3]. Implantation has consistently been demonstrated to improve patient quality of life [4].

Patients receiving treatment from the National Health Service (NHS) in the UK would typically require their primary care practitioner to make a referral to Otolaryngology or Audiology secondary care service for their hearing or other otological symptoms. In the secondary care setting, a pure tone audiogram can be arranged. If the patient meets the pure tone audiometric threshold criteria for cochlear implantation, they should have a discussion about their potential eligibility for cochlear implantation. If the patient agrees, they can be referred to a cochlear implant centre where a multidisciplinary team will conduct further detailed audiological testing including speech audiometry and clinical evaluation before proceeding with surgery and rehabilitation. Speech audiometry is not routinely performed for standard audiological assessment for hearing loss. There are only 20 cochlear implant centres across the UK. Many patients are routinely under audiology services to fit and maintain their conventional hearing aids and can be referred for cochlear implantation if their hearing deteriorates.

In the UK, the National Institute for Health and Care Excellence (NICE) has recommended that unilateral cochlear implantation is considered for adults who have severe-to-profound hearing loss. In March 2019, the definition of severe-to-profound hearing loss was updated by NICE to be greater than or equal to 80 dBHL at 2 or more frequencies. Prior to this update, the definition was greater than or equal to 90 dBHL at 2 or more frequencies.

Despite the reported benefits of implantation, it has been recognised that many adults who could benefit are never implanted [5]. The implantation rate is thought to be as low as 5% to 10% of those potentially eligible [6], although there has been no recent assessment of this across the UK. Although disparities have been documented in hearing healthcare and implant uptake, the majority of studies evaluating access to implants focus on what happens once patients are assessed at implant centres, rather than the stages leading up to and including referral for assessment [7,8]. Analysing patterns at these gateway stages in a patient’s cochlear implant journey, may lead to a deeper understanding of the referral pathway, enabling identification of the main drivers of inequalities in care, as well as the impact of key healthcare disparities, such as ethnicity and socioeconomic status.

The first aim of the study was to examine what proportion of patients who had met the pure tone audiometric threshold criteria for cochlear implantation as per the NICE guidelines had discussion about their eligibility and referred for further assessment by a cochlear implant team. The second aim was to identify if any determinants of socioeconomic and ethnic inequalities were associated with referral for these patients. This was done via a multicentre national assessment of eligible patients meeting the pure tone audiometric threshold criteria. Our working hypothesis posited that not all eligible patients are consistently referred for assessment and that referral rates may differ across ethnic or socioeconomic cohorts. Assessing disparities between patients who had a symptom of hearing loss and sought medical care to those who did not, and those who were referred from primary care to secondary care and those who were not was outside the scope of this study.

Methods

This study is reported as per the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guideline (S1 Checklist). The study protocol was published in advance (https://entintegrate.co.uk).

Results

Baseline patient characteristics

Data from 36 hospitals across England, Wales, and Scotland were submitted, representing 6,760 adults (Fig 1). No hospitals in Northern Ireland participated. The largest proportion of data submitted was from centres in the Midlands and East of England region (30·4%). In total, 80 hospitals registered their interest; 24 were not eligible for involvement due to noncompatible software and 20 withdrew following registration.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Map of the number of adults submitted to the study according to the region of residence.

Black shading indicates data was not submitted from this region. The image of the UK was adapted from mapSVG under the creative commons license (https://mapsvg.com/maps/united-kingdom).

https://doi.org/10.1371/journal.pmed.1004296.g001

Following implementation of inclusion criteria and de-duplication of adults, 6,482 adults met pure tone audiological threshold criteria (S1 Flowchart). The baseline characteristics of the included adults are shown in Table 1. The mean age was 72 ± 19 years (mean ± standard deviation); 53 9% of adults were male (Fig 2A), and 75·3% of participants were white, 6·3% were Asian, 1·5% were black, 0·05% were mixed, and 4·6% were self-defined as a different ethnicity. Ethnicity data were missing in 11·8% of participants. As compared to the other ethnic groups (Asian, black, mixed, other), white adults were older (62, 50, 58, 64 versus white 74 years, respectively) (Fig 2B). The age distributions of white and Asian adults were skewed to the left. Age (continuous variable) was moderately positively associated with presence of multimorbidity (binary variable) (Pearson point-biserial correlation coefficient r = 0.3; p < 0.001), and 311 adults who already had a cochlear implant were submitted, leaving 6,171 eligible adults for inclusion in the regression analysis (S1 Flowchart).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2.

(a) Population pyramid by age and sex assigned at birth. (b–d) Population age distribution of Asian (b), black (c), and white (d) adults.

https://doi.org/10.1371/journal.pmed.1004296.g002

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Baseline characteristics of included patients, stratified by ethnicity.

Data are presented as n (column %) or mean [95% CI], other than age (mean ± SD). dBHL: decibel hearing level. Values are given prior to removal of implanted patients.

https://doi.org/10.1371/journal.pmed.1004296.t001

Primary outcome

Association of ethnicity and socioeconomic predictors on likelihood of referral for CI.

Overall, 9.7% (n = 546/6,171) of eligible adults meeting pure tone audiological threshold criteria were referred for implantation assessment. The results of the logistic regression models for the association of predictor variables are shown in Table 2. In the fully adjusted model, the deprivation index of adults was significantly associated with likelihood of referral after completing audiometric evaluation, with adults from the least deprived locations being more likely to be referred than those from the most deprived. This association was strongest for adults from the 4th (OR: 2·19; 95% CI: [1·31, 3·66]; p = 0.002), 5th (OR: 2·32; [1·41, 3·83]; p = 0.005), 6th (OR: 2·32; [1·41, 3·83]; p < 0.001), and 8th deciles (OR: 2·07; [1·25, 3·42]; p = 0.004), where 10th is the least deprived region. There were significant regional differences, with adults living in London (OR: 0·40; [0·29, 0·57]; p < 0.001) being less likely to be referred after completing audiometric evaluation compared to adults from the Midlands and East. Ethnicity was not significantly associated with likelihood for referral.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Results of simple and multivariable logistic regression analysis for the primary outcome: whether patients were referred for surgical management of their hearing or not.

https://doi.org/10.1371/journal.pmed.1004296.t002

Secondary outcome

Association of ethnicity and socioeconomic predictors on likelihood of discussion about referral after completing audiometric evaluation.

Overall, only 35.7% (n = 2,204/6,171) of eligible adults had a documented discussion about their eligibility of being considered for surgical management of their hearing loss. The results of the logistic regression models for the association of socioeconomic and ethnicity factors are shown in Table 3. In the fully adjusted model, the deprivation index of adults was significantly associated with likelihood of being informed of their eligibility for referral (noted as “discussion”) (p < 0·001). Adults from the most deprived location were less likely to be discussed with than those from less deprived areas (deciles 4th–8th inclusive, where 10th is least deprived, Table 3). There were regional differences (p < 0·001), with London and the Northern regions having a lower OR for discussion compared to adults from the Midlands and East: North of England (OR: 0·74; [0·62, 0·89]; p = 0.001), London (OR: 0·44; [0·35, 0·56]; p < 0.001). In the fully adjusted model, ethnicity was significantly associated with likelihood of discussion (p < 0·001), whereby Asian (OR: 0·58; [0·43, 0·79]; p < 0.001) and black adults (OR: 0·56; [0·34, 0·92]; p = 0.021) were less likely to have a discussion compared to white counterparts.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Results of simple and multivariable logistic regression analysis for the secondary outcome: whether patients had a discussion informing them that they were eligible for surgical management of their hearing loss, or not.

https://doi.org/10.1371/journal.pmed.1004296.t003

Discussion

Only 9.7% (n = 546) of adults meeting pure tone audiometric threshold criteria for cochlear implantation assessment in a 6-month period in 2021 were referred for further evaluation, and 35.7% (n = 2,204) had a documented discussion about their eligibility for referral. Our hypothesis was upheld as there were disparities in the management of adults based on ethnic and socioeconomic factors, although the finding that males were less likely than females to be referred for assessment after audiometric evaluation for higher-level care was unexpected. Individuals from more deprived areas, specific regions, and older age groups were less likely to be referred. Moreover, adults from Asian and black backgrounds were less likely to have discussions informing them of their eligibility for cochlear implant assessment.

This study adds to the field of cochlear implant candidacy research by highlighting potential targets for increasing referral discussions for implants and suggests the need to explore the reasons behind the demographic, socioeconomic, and ethnic disparities to improve outcomes. According to the principles of shared decision-making and “no decision about me, without me” [10], all potentially eligible adults of cochlear implantation should be informed and given the option for assessment, and all members of the clinical team including audiologists and physicians should play a role in cochlear implant discussion and referral.

There was a trend of lower likelihood for discussion and referral after audiometric evaluation for implants among adults from socioeconomically deprived backgrounds. Similar findings have been observed in studies conducted worldwide. In Australia, recipients were more likely to live in the least 2 deprived deciles [7]. Similarly, in the US, adults with private insurance were 13 times more likely to receive cochlear implants compared to those under Medicare cover [11]. In an implant unit in the US, higher socioeconomic status patients were more likely to be implanted than those with lower socioeconomic status [8]. In countries with private healthcare systems, health-related financial inequalities may contribute to varied access across different sociodemographic groups. In nationalized healthcare systems like the UK, other factors such as patient and caregiver education, health literacy, social acceptance of cochlear implants, and unconscious bias may also play a role [12].

Geographic variations in management were observed, potentially influenced by factors such as clinician beliefs or service capacity. Adults from London were found to be the least likely to have discussions or referrals after audiometric evaluation for assessment compared to the East/Midlands. While population heterogeneity exists between regions, the fact that all identified adults were eligible for referral suggests differential management based on geographic location. Appreciating the underlying reasons can help inform loco-regional policies to establish coherent referral pathways. This in turn can enable monitoring of their effectiveness and highlight which aspects of the pathway may need additional targeted support and education.

Despite cochlear implant surgery being free at the point of access for adults in the UK, additional expenditures such as transportation, travel time or distance, parking for multiple appointments, as well as the potential need to take unpaid time away from employment, may deter some individuals. This study indicated that adults who had their referral candidacy confirmed with audiometric assessment at an implant centre were more likely to be referred, possibly due to geographical proximity between the patient’s home and the implant centre, but also possibly due to a greater intra-departmental appreciation of cochlear implantation as an option for these patients, and reduced administrative burden in referring these patients to colleagues in the same centre. In Australia, the majority of adult recipients lived near major urban centres [7]. Among American veterans, rural adults waited longer to obtain both hearing aids and implants, due to longer commutes and lower incomes [13]. This suggests the importance of centre location, and innovative approaches such as virtual consultations or remotely located audiologists to help redistribute implantation delivery and improve patient access [14].

The presence of a cochlear implant champion increased the likelihood of discussion of cochlear implant assessment referral eligibility based on pure tone audiometric criteria, but it did not translate into more referrals (in fact, marginally fewer patients were referred although p = 0.04). This suggests that not all eligible adults are suitable for various reasons, including audiological and healthcare factors, or patient choice. For a patient, the decision to undertake assessment and surgical management of their hearing loss is complex [5]. The champion program may need review to ensure clinicians have sufficient time to perform their role to aid identifying eligible adults and providing further information. Auditbase software has the capability to enable far more extensive coding of when referral discussions take place. This would allow clearer insight of where and when these discussions happen, although may require additional training locally, and additional time to undertake this extra task. Alternative diagnostic testing, such as speech testing, could aid the clinical decision-making process [15] and explain regional variations in management.

Adults from Asian and black backgrounds were less likely to have a discussion informing them of their eligibility for referral after audiometric evaluation. Variations in clinician referral rates by ethnicity have also been shown to affect referrals for a range of hearing and non-hearing services [16,17]. A study in the US also found that African-American and Asian patients were less likely to accept proceeding with cochlear implantation [18]. There may be factors such as language barriers or clinician biases which have not been explored by our work. It is noted that while our data suggested that not having English as a first language did not significantly impact the discussion of cochlear implants, it did correlate with reduced referral rates. This is consistent with a study from the USA which revealed that non-English speaking patients received fewer referrals compared to English-speaking patients [19]. This may suggest difficulties in adequately explaining cochlear implantation as an option or cultural differences in approaching hearing loss treatment. Compared to white counterparts, minority ethnic groups are less likely to use hearing aids [16] and report lower familiarity with cochlear implants [20].

It is important to note that this study focused on individuals who had already sought hearing healthcare, potentially underrepresenting Asian and black communities. Reasons for underrepresentation may include cultural norms of elder care that mitigate the impact of hearing loss, systemic healthcare access barriers in minority ethnic groups, or the overrepresentation of younger adults who migrated to the UK. Encouraging older Asian or black communities to seek hearing tests is crucial moving forward and an area for future research.

Older adults were less likely to be referred or informed of their eligibility for referral for assessment. Older patients declining cochlear implantation has also been found by a study in the US [18]. Age discrimination may play a role in perceiving hearing loss as an inevitable consequence of ageing [21]. However, evidence suggests improved speech scores and quality of life following implantation in adults aged 65 to 80 years, with few complications when managed appropriately [22]. Age was positively correlated with multimorbidity (i.e., older patients were more likely to suffer from multimorbidity) and may play a role in patients’ likelihood to accept a referral for cochlear implant assessment. A recent population cohort study examining multimorbidity in surgery indicated that 11.2% of patients undergoing elective surgery had multimorbidity yet multimorbid patients accounted for 50.2% of deaths after elective spells [23]. Recognising they are at higher risk of poor outcomes may influence the decision-making process in declining assessment even at this early stage in the patient pathway.

Men were also found to be less likely to be referred or informed of their eligibility for referral for assessment. Interestingly, a study in the US showed that men were not less likely to proceed with surgery compared to women [18]. The influence and magnitude of our effect size for sex was an unanticipated finding. This may be explained by the divergent health-seeking behaviours that exist between men and women. Previous studies on gender differences in the patterns of use of conventional hearing aids have found that hearing-impaired men were less likely to be acceptant of hearing loss, had lower motivation to seek hearing rehabilitation, and were more concerned about their appearance [24,25]. Despite this, few sex-specific initiatives for the management of hearing loss exist.

The degree of hearing loss was associated with likelihood for referral after audiological evaluation; adults with hearing thresholds worse than 90 dBHL (i.e., profound loss) were more likely to be referred than those with severe hearing loss, despite the latter adults still meeting pure tone threshold criteria for referral. This suggests that eligible adults with severe hearing loss, and/or their clinicians, may perceive that they are managing their hearing loss or that they lack awareness of the latest referral guidelines [3]. There may be a concern that implants may sacrifice their residual hearing; however, appropriate counselling about hearing preservation surgery can be done by a cochlear implant team [26].

The strengths of this study include the large sample size, focus on adults (often work focuses on the paediatric population), and the inclusion of all eligible candidates meeting pure tone audiological threshold criteria, rather than those who have already been referred for assessment. There are a number of limitations to be discussed.

Further weaknesses include the observational retrospective nature, dependence on accurate documentation, and potential underrepresentation of certain regions [9] or demographic groups. The models used assume correlations and do not predict causality. There was a reliance upon accurate documentation of patient discussions and it is conceivable that discussions were had for some adults yet not recorded. Furthermore, without the ability to contact patients directly, gender identity was determined by accuracy of NHS sex categorisation. The study was conducted during the COVID-19 pandemic, which may have affected referral rates (S4 Appendix). A further weakness of this study is that it only captures patient assessments performed in secondary care and was not able to capture those performed in other settings, such as high street audiology assessments.

This study only includes individuals who have successfully accessed secondary healthcare and may exclude those encountering barriers such as socioeconomic, systemic, or cultural challenges in accessing primary healthcare. This approach risks disproportionately representing the perspectives of those with easier access, reinforcing disparities and impeding a comprehensive understanding of diverse healthcare needs. To mitigate this, future research should consider approaches to actively incorporate individuals underrepresented due to barriers in seeking healthcare. We also relied on the electronic health records of the services referring to the cochlear implant team and cannot confirm that patients attended and completed a cochlear implant assessment which could give further information about patient barriers to referral rather than referrer barriers.

Speech audiometry, which forms part of NICE cochlear implant candidacy criteria, is not routinely performed in the UK as part of a standard hearing assessment and was therefore not included in our data collection. Clinicians should consider requesting this to aid their decision-making process regarding discussion and possible referral for cochlear implantation.

The effects of untreated or undertreated hearing loss can be diverse, being associated with depression, social isolation, risk of hospitalisation, and mortality [27–30]. There is a strong link between hearing loss and cognitive decline, including dementia. Although the exact cause of this link is undergoing research, treatment of auditory impairment could delay or prevent neurodegenerative changes if they occur due to altered sensory input [31]. An analysis of the UK Biobank data has shown dementia is more likely to develop in black participants than white which reinforces consideration of targeting potential modifiable risk factors of cognitive decline in black ethnic groups [32]. Thus, the results of this research imply that differences at this early stage in candidacy discussions may have substantial implications for further health issues, above and beyond that of hearing loss.

Future research should adopt a nuanced approach to investigate the reasons behind the ethnic, socioeconomic, and other demographic differences in assessment. The referral pathway is likely to be influenced by various factors related to clinicians, patients, and systems. However, the specific contributions of each stakeholder remain unclear in the existing research. Conducting qualitative assessments involving key stakeholders, such as patients and clinicians, could shed light on these contributions. This approach may not only help investigate the dynamics involved but also identify adaptable strategies to enhance accessibility and pinpoint areas for targeted educational initiatives. Utilising information technology systems in audiology can help identify eligible adults. Efforts are underway to create an automatic pop-up identification tool in the next published version of Auditbase. This should be used alongside running regular audit reports and contacting patients accordingly should be encouraged in all UK hospitals.

In conclusion, our findings indicate that a significant number of eligible adults with severe-to-profound hearing loss in Great Britain are not being appropriately referred after meeting pure tone audiometric threshold criteria for cochlear implant assessment or informed about their eligibility for referral. This study highlights that there are notable disparities in patient management based on socioeconomic status, ethnicity, and sex. With this knowledge, clinicians and policymakers may wish to explicitly target these underrepresented patient groups via support and education to those teams directly caring for them to improve discussion and referral rates for cochlear implants. While it is acknowledged that not all eligible adults may opt for surgery, it is crucial that all patients are made aware of their eligibility for further assessment for this potential life-altering intervention. Further research is necessary to understand and address these disparities, with a focus on developing tools to ensure informed decision-making, educating hearing healthcare providers, and investigating the reasons behind cochlear implant declination and the underrepresentation of older ethnic minority groups in seeking hearing assessments in secondary care. The aim is to enhance equitable access to hearing loss treatment.

Supporting information