Effects of remote ear-nose-and-throat specialist assessment screening on self-reported hearing aid benefit and satisfaction

Abstract

Objective

To explore the impact of remote versus in-person ear-nose-and-throat (ENT) specialist screening before hearing treatment on self-reported hearing aid (HA) benefit and satisfaction among adult first-time HA users.

Design

Participants were randomised to either remote or in-person ENT assessment before treatment initiation. Hearing ability, hearing quality, and treatment satisfaction were assessed pre- and post-HA treatment using the SSQ12, IOI-HA, and selected items from the 2021 Danish national Patient-Reported Experience Measures. Average daily HA usage was also recorded.

Study sample

751 adult potential first-time HA users with self-reported hearing impairment were included; 501 participants were remotely assessed in private or public audiological clinics, and 250 control group participants were assessed in-person by private ENT specialists. Of the 658 participants who completed the entire trial, 454 received HAs.

Results

No significant post-treatment HA benefit differences were found between groups. Remotely assessed HA recipients in private clinics expressed slightly higher staff and waiting time satisfaction. Participants with normal hearing and mild/moderate hearing loss reported higher pre-treatment hearing ability and quality. No significant difference in average daily HA usage was observed between groups.

Conclusions

Findings suggest that remote screening does not compromise patient-reported HA benefit and satisfaction when compared to in-person screening.

Keywords:

• Remote assessment
• ear-nose-and-throat assessment
• hearing aid (MesH Term)
• patient-reports
• hearing aid benefit
• hearing ability
• hearing quality
• patient satisfaction

Introduction

Currently, approximately 161 private ear-nose-and-throat (ENT) specialists in Denmark can prescribe hearing aid (HA) treatment for adults with hearing impairment following physical assessments, as stipulated by existing guidelines (The Danish Health Authority 2015). During these in-person consultations, ENT specialists evaluate an audiological examination (audiometry and tympanometry), inspect the external ear canal and the tympanic membranes bilaterally by oto-microscopy, and gauge the patient’s need and motivation for treatment before recommending an appropriate treatment course.

Upon recommendation for HA treatment, patients can seek help from any of the 21 public or 359 private audiological clinics in Denmark. Although private clinics are bound by technical standards and existing regulations (The Danish Health Authority 2015, 2019), they can only serve adults with mild (21–40 dB hearing level air conduction (AC) thresholds) and moderate (41–60 dB hearing level AC thresholds) hearing loss (HL) and cannot treat adults with severe hearing issues or children. The cost at private clinics is partially covered by a public subsidy, with patients also contributing a personal payment in some cases. In contrast, public clinics offer HA services at no cost to the patient. In both settings, technical audiologists offer detailed counselling on HA usage before treatment begins. Adults with severe HL (exceeding 60 dB hearing level AC thresholds), severely asymmetrical HL (asymmetry exceeding a pure-tone-average (PTA) of 30 dB hearing level AC thresholds and/or a difference of 20% or more in speech discriminations score (DS) between the two ears), or serious ear disorders, and children with HL (regardless the severity), however, necessitate expert assessment and treatment at a public, hospital-based audiology or otology department, where ENT specialists with audiology or otology expertise ensure accurate examination and audiological or surgical treatment for this patient group.

According to a 2021 FORCE Technology Lab report, the number of HAs administered in Denmark increased by 20% in 3 years, from 140,570 in 2018 to 168,402 in 2021, with about half served by private clinics (Ravn and Jørgensen 2021). However, organisational inefficiencies, bottlenecks, and capacity issues within the Danish hearing healthcare system often disadvantage many patients. Prolonged waiting lists for private ENT assessments and public HA treatment result in delayed diagnostics and treatments (The Danish Ministry of Health 2018). With demographic changes projecting a significant increase in the population aged 60 and above by 2050 (WHO 2021), the number of individuals suffering from age-related HL in Denmark and globally is expected to rise. Age-related HL, if untreated, can elevate the risk of dementia, anxiety, and social isolation (Dalton et al. 2003), underscoring the need for innovative measures to equip Danish hearing healthcare for future challenges.

As a response, a digital and remote ENT specialist assessment (RESA) method was developed in 2020 by a group of ENT specialists in the North Denmark Region. This method enables ENT specialists to assess patients digitally, remotely, and without direct patient-physician contact, identifying potential first-time HA users with complicated HL or serious ear disorders requiring specialised assessments and treatment while providing efficient diagnosis and treatment pathway for patients with mild and moderate HL—here referred to as “simple” HL. (Siggaard et al. 2023) RESA screening accuracy was tested and compared to the existing physical ENT specialist assessment (PESA) routine in a randomised controlled trial in 2021–2022, named the Innovation of Hearing Rehabilitation and Effects of Reform (InHEAR) trial, including 751 hearing-impaired adults without previous HA experience. Findings showed that RESA screening sensitivity for cases of complicated HL and serious ear disorders among participants were 85% (95% confidence interval (CI): 71 − 94%), which was significantly higher than PESA screening sensitivity of 20% (95% CI: 3 − 56%). (Siggaard et al. 2023)

Nevertheless, the digitisation of the ENT specialist assessment process may have profound implications for patients and their treatment decisions, especially if serious ear disorders such as cholesteatomas go undetected. While this shift introduces new challenges, the elimination of direct patient-physician interactions in the RESA screening might particularly impact shared comprehension and decision-making, potentially reducing perceived treatment benefit and overall patient satisfaction. Variations in the extent of hearing rehabilitation counselling performed by audiology assistants or hearing therapists across private and public audiological clinics could also affect patient’s perception of their HA benefit. Consequently, the aim of this study was to investigate the effects of RESA screening on self-reported HA benefit and satisfaction compared to PESA screening.

Materials and methods

The Health Research Ethics Committee for Northern Denmark was notified about the trial before patient recruitment. Upon receipt of the notification, the Committee decided that it was not necessary with a formal application before patient recruitment could be initiated (case no. 2020-000992).

Data presented in this manuscript are collected as part of a larger trial and part of the results are already published elsewhere (Siggaard et al. 2023).

Between March 2021 and May 2022, 751 adult, potential first-time HA users with subjective hearing-impairment were recruited via an advertisement on the North Denmark Region’s official Facebook page. The sample size was determined based on an estimated prevalence of complicated HL and serious ear disorders in the intended population of 5%, a minimum power of 80%, and a p-value of 0.05 as specified in greater detail elsewhere (Siggaard et al. 2023). The recruitment process was completed without medical intervention or pre-assessment, similarly to how hearing-impaired adults acquire non-prescribed over-the-counter HAs in the U.S. as established by the U.S. Food and Drug Administration (FDA) (Food and Drug Administration Department of Health and Human Services (HHS) 2022). However, after responding to the Facebook advertisement, prospective participants were contacted by phone by trial administrators and informed first orally and followingly in writing about all aspects of trial participation before recruitment was completed. The exclusion criteria were: (1) previous HA usage, (2) Danish language deficiency, (3) dementia and other mental illnesses rendering consent for participation unattainable, and (4) serious physical comorbidity that would hinder physical attendance.

Preceded by a written consent to participate, participants were randomly appointed one of two arms by an automatic 2:1 allocation ratio; a test group (TG) of 501 participants and a control group (CG) of 250 participants. The test group was further divided randomly into 251 test group 1 (TG1) participants and 250 test group 2 (TG2) participants; TG1 participants underwent RESA screening in one of 12 participating private audiological clinics in the North Denmark Region; and TG2 participants in one of the region’s 5 public audiological clinics. The subgroup division of TG participants aimed to explore, whether RESA screening affected the participants’ perception of HA benefit and satisfaction differently in private versus public audiological clinics. The CG participants underwent PESA screening in private ENT specialist clinics. The trial was divided into three stages: (1) intervention, (2) treatment and (3) “gold standard” ENT specialist assessment. The latter was a 30-minute assessment conducted in-person by ENT specialists with expertise in medical audiology or otology.

Stage 1—intervention

After recruitment and the random group allocation, TG participants underwent RESA screening, a two-step-procedure: The first step involved a standardised examination conducted in-person in either private (for TG1 participants) or public (for TG2 participants) audiological clinics by certified technical audiologists with a minimum of two years clinical experience. The standardised examination comprised three elements; (1) a medical history on ears and hearing was obtained by a self-administered multi complaint risk assessment tool called the Consumer Ear Disease Risk Assessment (CEDRA) designed to screen adult, potential first-time HA users for 104 targeted ear diseases related to HL (Klyn et al. 2019; Kleindienst et al. 2017). For participants who self-reported presence of tinnitus in CEDRA, the Tinnitus Handicap Inventory (THI) was employed to quantify self-reported tinnitus-related disability. Patients received links by email to complete CEDRA and THI digitally at home before the physical audiological examination was conducted; (2) audiological tests matching the required standards as described in the Danish Executive Order on HA treatments in private audiological clinics (The Danish Health Authority 2019) were performed consisting of pure-tone AC and bone conduction (BC) thresholds, a standard speech discrimination test, acoustic reflex tests, and a standard 226 Hz tympanometry; and (3) visual still images of the tympanic membranes were obtained digitally by video-otoscopy. The second step of RESA screening involved four individual ENT specialists, two private ENT specialists and two ENT specialists with audiological expertise, who assessed the test results digitally, remotely and without consulting the participants directly. Irrespective of subgroup allocation, all TG participants were assessed by all four digital assessors. However, in cases where the four assessors diverged in classifying TG participants into uniform diagnostic subgroups, such variation could pose complexities in proffering consistent treatment recommendations due to differing referral decisions. To address this issue, an automated algorithm was employed to assign each TG participant to a specific assessor for the treatment referral process, both systematically and randomly. The allocation was blinded for both participants and assessors. The ‘gold standard’ ENT specialist assessment at Stage 3 ensured that potentially misdiagnosed patients were accurately identified and prescribed the appropriate treatment regimen. In total, six TG and 8 CG participants were incorrectly diagnosed as “simple” cases compared to the Stage 3 “gold standard” assessment. Furthermore, 12 TG participants and one CG participant were erroneously identified as “complicated” at Stage 1. These cases are described in detail elsewhere (Siggaard et al. 2023). Detailed data on the variability in RESA screening sensitivity and specificity across the four assessors will be discussed in a subsequent report.

Conversely, participants allocated to the CG group were examined and assessed physically and in-person by private ENT specialists in accordance with current Danish PESA guidelines (The Danish Health Authority 2015). This involved an in-person consultation, an audiological examination (audiometry and tympanometry), and an objective examination of the tympanic membranes obtained by oto-microscopy.

Stage 2—treatment

Irrespective of group affiliation and screening method, all participants were divided roughly into three diagnostic subcategories: (1) those with objectively normal hearing equivalent to a PTA hearing level of 20 dB or better, where PTA referred to the average AC hearing thresholds at 500, 1000, 2000, and 4000 Hz (AC-PTA-4), (2) those with simple HL including cases of mild (21–40 dB hearing level AC thresholds) and moderate (41–60 dB hearing level AC thresholds) HL where concomitant serious ear disorders were neither suspected nor diagnosed, and (3) those with complicated HL including symmetric and asymmetric HL exceeding 60 dB hearing level AC thresholds with differing stapedial reflex patterns, mild or moderate HL with severe AC-PTA-4 asymmetry exceeding 30 dB and/or a difference of 20% or more in speech DS between the two ears, and/or serious ear disorders such as external auditory canal (EAC) pathology (e.g. atresia, infection, cholesteatoma), middle ear pathology (e.g. tympanic membrane perforation or retraction, infection, otosclerosis, cholesteatoma), severe or catastrophic tinnitus-related disability assessed by an ENT specialist (for CG participants) or measured by a THI score of 58 or higher (for TG participants), or retro-cochlear or cerebral pathology (e.g. otogenic vertigo, vestibular schwannoma, tumour, vascular or neurological disorders). TG participants diagnosed with objectively normal hearing received an additional physical ENT specialist assessment prior to trial termination to ensure that no subjective symptoms or objective findings that would require treatment or further examination were overlooked. Participants with simple HL were offered HAs in either private (for TG1 participants) or public (for TG2 participants) audiological clinics. To imitate the existing PESA routine, CG participants with simple HL were able to choose, which participating private or public clinic would be administering them HAs. Irrespective of group affiliation, participants with complicated HL and/or suspected or diagnosed ear disorders were referred to the Department of Audiology at Aalborg University Hospital in Aalborg, for an audiometric re-examination and a physical assessment performed in-person by ENT subspecialists in medical audiology or otology in accordance with national guidelines (The Danish Health Authority 2015).

Stage 3—“gold standard” ENT specialist assessment

Two to four months after treatment initiation, all participants irrespective of group affiliation received a physical “gold standard” assessment by ENT specialist within medical audiology or otology. It was defined by an in-person, 30-minute consultation with the patients, an assessment of the audiological test results (audiometry and tympanometry), and the performance of an objective oto-microscopy of the EAC and the tympanic membrane on both ears. Based on their assessment, the ENT specialists then re-categorised the participants into the same three diagnostic categories.

Study data were collected and managed using REDCap^® electronic data capture tools hosted at Aalborg University (Harris et al. 2009, 2019).

Questionnaires

At specific trial points, participants automatically and digitally received disease-specific questionnaires regarding perception of hearing ability, quality in different listening environments, HA benefit, and satisfaction, along with generic questionnaires regarding patient satisfaction. An e-mail link directed participants to the specific questionnaire where all items needed to be completed before responses were saved and stored. Up to three reminders, including links to the questionnaire, were re-sent automatically by e-mail at two-day intervals if the questionnaire was not completed. To minimise bias in the data analysis related to HA quality variance, a selection of four different behind-the-ears (BTE) HA models comparable in quality and technical characteristics were applied in all participants (see Supplementary Material). Other or higher-quality HA models were administered only in cases of severe or complicated HL. The administered HA model and the average daily HA use time were registered 2–4 months after treatment initiation at Stage 3.

The International Outcome Inventory for Hearing Aids (IOI-HA)

An electronic version of the 2014 revised seven-item IOI-HA was applied to quantify the patient-perceived HA benefit at Stage 3 (Thunberg Jespersen, Bille, and Legarth 2014; Thorén, Andersson, and Lunner 2012). This tool has been applied widely and internationally owing to its simplicity, interpretability, and ease of use for both clinicians and patients (Cox, Alexander, and Beyer 2003). The IOI-HA addresses seven outcome domains of HA treatment effectiveness: (1) The daily use of HAs, (2) the perceived benefit of HAs, (3) residual activity limitation, (4) satisfaction, (5) residual participation restrictions, (6) impact on others, and (7) quality of life changes after HAs. Each item has a five-point response scale ranging from poorest to highest outcome or level of quality. The definitions of these five response options vary across different items. Despite this variance, previous studies have assigned the individual item scores ranging from 1 to 5, with higher scores indicating better outcomes for each respective item (Liu et al. 2011; Houmøller et al. 2022). Additionally, some studies have employed a subscale analysis within the IOI-HA framework, categorising it into two distinct groups: items oriented internally and those oriented externally, in relation to the effectiveness of HAs (Kramer et al. 2002; Brännström and Wennerström 2010). However, in this study, individual rating responses were juxtaposed rather than aggregated into a total score. This approach was adopted because the rating scales varied between items, rendering the summation of individual responses inappropriate. Although several HA treatment effect parameters are covered by the IOI-HA, they are addressed only by one item each. Consequently, the tool does not provide detailed information regarding patient-perceived hearing ability and quality of hearing in different listening environments or patient-reported treatment satisfaction. Therefore, the IOI-HA was applied as a supplement to the other questionnaires in the present study.

A short form of the speech, spatial and qualities of hearing scale (SSQ12)

The 12-item SSQ12 was applied to measure a range of hearing disabilities across three domains: speech comprehension in silent and noisy listening environments (five items), spatial hearing and sound localisation ability (three items), and sound segregation and quality (four items) (Noble et al. 2013) The tool has typically been used complementarily with IOI-HA or other behavioural or experimental measures of hearing ability (Pedersen et al. 2023). Each item is scored on a scale ranging from 0 to 10. Average scores within this scale quantify the range of hearing disability within each domain and overall, lower scores reflect inability or absence of quality, and higher scores reflect ability and presence of quality. The questionnaire was applied at Stage 1 prior to HA treatment initiation to quantify the averaged hearing disability severity at baseline and to compare between groups.

For a retrospective self-report on hearing ability and quality after HA treatment when compared to before HA treatment initiation, a “benefit” version (SSQ12-B) was applied at Stage 3 (Jensen et al. 2009). In the SSQ12-B version, items are scored on a scale ranging from “-5” to “5.” Negative scores signify a decline in hearing ability and quality following HA treatment. Scores that are zero or near zero indicate no noticeable change in hearing ability and quality after the treatment. On the other hand, positive scores represent an improvement in hearing ability and quality after HA treatment.

Comparative analyses were conducted to determine significant differences in response distributions within the three domains of the SSQ12 and SSQ12-B surveys (speech comprehension, spatial hearing, and quality of hearing) as well as the overall responses.

Satisfaction questionnaires

To achieve a more in-depth analysis of patient satisfaction throughout the trial course, generic satisfaction questionnaires were applied at all trial stages. From the 35-item validated 2021 Danish national Patient Reported Experience Measures (PREM) for planned outpatient visits to somatic hospital departments, 18 items were selected that were in alignment with the present study (The Center for Patient Involvement 2021). The 35-item PREM, and additional versions covering somatic and psychiatric in- and outpatient visits to the hospital, is applied annually and is mandatory in all Danish hospitals as a tool for quantifying patient-perceived treatment quality and satisfaction. The 18 selected PREM items (PREM18) were divided into five themed subgroups and applied at Stage 1 (PREM18-1) and Stage 2 (PREM18-2). The first four subgroups addressed satisfaction with: (1) the clinical staff during the stage-specific visit (four items), (2) the waiting time in the clinic (one item), (3) the information given prior to, during and after the stage-specific visit (six items) and (4) overall treatment satisfaction (five items). In documenting participants’ overall satisfaction with the assessment and treatment arc of the project, rather than the specificities of the Stage 3 visit, only the five items of the fourth themed subgroup explicitly targeting overall treatment satisfaction were utilised upon concluding Stage 3 (PREM4-3). Consequently, items related to the clinical staff, wait times, and information pertinent to the Stage 3 visit were intentionally omitted. The 16 items in the first four subgroups were scored on a numbered five-point Likert scale ranging from “1” (poorest level of satisfaction) to “5” (highest level of satisfaction). Response options such as “unsure” and “irrelevant for me” were available for all 16 items. The fifth subgroup addressed occurrence of clinical errors or malpractice and consisted of one item with a dichotomous “Yes” or “No” response option. Conditioned by a “Yes”-answer in the first item, a second item presented a list of four predefined response options allowing the patient to elaborate on the consequences of the clinical error or occurrence of malpractice. Participants were able to provide one additional free-text comment for the three subgroups regarding the clinical staff during the stage-specific visit; the information given prior to, during and after the stage-specific visit; and overall treatment satisfaction.

Statistical methods

All statistical analysis were conducted using the statistical Rv4.1.2 software (R Foundation for Statistical Computing 2021).

For the SSQ12 and the SSQ12-B questionnaires, data were normally distributed. Therefore, the ANOVA test was employed to compare mean scores among the three groups (TG1, TG2, and the CG) within the surveys’ three domains (speech comprehension, spatial hearing, and quality of hearing) as well as the overall survey scores. Furthermore, linear regression analyses were performed on overall as well as domain SSQ12, and SSQ12-B scores when adjusting for gender, age, randomisation group, type of HA applied, and severity of HL (normal hearing, simple HL, and complicated HL). For the IOI-HA, the variables to be compared were ordinal. The non-parametric Kruskal Wallis Rank Sum test was applied for comparing the distribution of IOI-HA responses among the three groups, given the variation in IOI-HA response definitions between items rendered them not directly comparable. The Kruskal Wallis Rank Sum test was also employed to compare distributions of scores within the five themed subgroups and overall scores for PREM18-1 and PREM18-2, and PREM4-3 between the three groups, as data were highly skewed. Although the Kruskal Wallis test does not directly compare parameters like means or medians, in practice, medians are often inspected when interpreting the results, as they are robust measures of central tendency for non-normally distributed data. Post-hoc tests (Dunn test with Bonferroni adjustment) were conducted for pairwise comparisons of response distributions within specific survey subgroups of the PREM18-1 and the PREM18-2 where significant differences were detected. Pearson’s Chi-square test was employed to compare the distribution of positive/neutral vs. negative free-text comments between the three groups. The Kruskal Wallis Rank Sum test was employed to compare the distribution of usage time of HAs across the three groups, since data were not normally distributed. A p-value below 0.05 was deemed statistically significant for all comparisons.

Results

Out of the 751 participants included, 658 (88%) completed the entire trial course. In total, 40 (5%) were lost to follow-up, 52 (7%) withdrew from the study due to illness or other non-specified reason, and 1 (0.1%) participant died during the trial. Of the 658 participants who completed the trial, 201 (31%) had normal hearing, 406 (61%) had simple HL, and 51 (8%) had complicated HL or serious ear disorders. Three participants in the latter group received other intervention than HAs. Thus 454 (69%) received HAs during the trial. Age and gender were evenly distributed between TG1, TG2 and the CG as presented elsewhere (Siggaard et al. 2023). All 751 included participants were invited to complete questionnaires at Stage 1 (SSQ12 and PREM18-1), and the 454 HA recipients were invited to complete PREM18-2 at Stage 2, and SSQ12-B and IOI-HA at Stage 3. Both HA and non-HA recipients who were not lost to follow-up after Stage 2 were invited to complete PREM4-3 at Stage 3. Response rates to all questionnaires applied are presented in Table 1.

Table 1. Questionnaire response counts and rates.

	Responders
Eligible responders (randomised participants)	Test group1 (n = 251)	Test group 2 (n = 250)	Control group (n = 250)	Total (n = 751)
Stage 1, responders, number (%)
SSQ12	247 (98)	240 (96)	243 (97)	243 (97)
PREM18-1	242 (96)	226 (90)	214 (86)	682 (91)
Eligible responders (HA recipients)	Test group 1 (n = 166)	Test group 2 (n = 151)	Control group (n = 137)	Total (n = 454)
Stage 2, responders, number (%)
PREM18-2	146 (88)	140 (93)	114 (83)	400 (88)
Stage 3, responders, number (%)
SSQ12-B	149 (90)	147 (97)	136 (99)	432 (95)
IOI-HA	155 (93)	151 (100)	137 (100)	443 (98)
Eligible responders (HA and non-HA recipients), responders, number (%)	Test group 1 (n = 231)	Test group 2 (n = 231)	Control group (N = 137)	Total (N = 581)
PREM4-3	211 (91)	191 (90)	134 (98)	443 (98)

Explanatory text.

The table displays the count of questionnaire responses and corresponding response rates for all questionnaires utilised during the trial. The quantity of eligible responders is presented for each stage of the trial.

HA: Hearing aid; IOI-HA: The International Outcome Inventory for Hearing Aids; SSQ12: The 12-item short version of the Speech; Spatial and Qualities of Hearing Scale; SSQ12-B: The 12-item short HA benefit version of SSQ12; PREM18-1: The 18-item Patient Reported Experience Measures (PREM) questionnaire applied at Stage 1; PREM18-2: The 18-item PREM questionnaire applied at Stage 2; PREM4-3: The four-item PREM subgroup questionnaire regarding overall patient satisfaction applied at Stage 3

Figure 1 offers a comparison of the mean domain and total scores, and interquartile ranges (IQRs), a measure of statistical dispersion calculated as the difference between the upper and the lower quartile in data set, for both the SSQ12 (completed at baseline) and SSQ12-B (completed 2-4 months after HA treatment) among the three groups. No statistically significant differences were found between the groups. After adjusting for gender, age, and randomisation group on the total SSQ12 score, it was found that participants with normal hearing and simple HL significantly scored 1.1 scale points (95% CI: 0.6–1.6), p = 0.001, and 0.6 scale points (95% CI: 0.1–1.1), p = 0.02, respectively, higher than those with complicated HL. Similar outcomes emerged from the regression analyses conducted for both speech and spatial domain scores. Notably, individuals with normal hearing displayed significantly higher scores in the qualities of hearing domain when compared to those with either simple or complicated HL (see Supplementary Material for detailed covariate information). This suggests superior hearing ability and quality in the first two diagnostic subgroups when compared to the latter at baseline. There were no significant differences in domain or total SSQ12-B scores between the three randomisation groups when adjusting for age, gender, type of HA model used, and diagnostic subgroup.

Figure 1. Distributions of responses to the speech, spatial, and qualities of hearing scale prior to and following hearing aid treatment. Explanatory text: Each boxplot corresponds to the domain scores of speech comprehension, spatial hearing, and qualities of hearing, as well as an overall score. The y-axes denote the SSQ12 (0–10) and SSQ12-B (−5 to 5) score ranges, respectively. The vertical whiskers extend from the minimum to the lower quartile (base of the box) and from the upper quartile (top of the box) to the maximum, capturing data outside the IQR. Dots represents outliers. ANOVA was used for statistical analysis. Please note, no statistically significant differences were detected between the groups either pre -or post-hearing aid treatment, with all comparisons yielding p-values exceeding 0.05. F-statistics for the SSQ12 speech, spatial, and qualities of hearing domain scores, and total score were 0.15, 0.68, 0.16, and 0.09, respectively. Similar F-statistics for the SSQ12-B domain and total scores were 0.19, 0.45, 0.41, and 0.18, respectively.

Figure 2 juxtaposes the distributions of IOI-HA response distributions between the three groups The inter-group comparisons of response distributions revealed no significant differences between groups, and consequently, post-hoc tests were not performed on these data.

Figure 2. Distributions of responses to the international outcome inventory for hearing aids. Explanatory text: The data, being ordinal, are represented through horizontal histograms for each of the seven items. Each bar corresponds to the count of responses from each group on the x-axis, mapped to the respective response definition on the y-axis. The statistical analysis utilised was the non-parametric Kruskal-Wallis Rank Sum test. Notably, no statistically significant differences were discerned across the groups, with all comparisons yielding p-values greater 0.05.

Figure 3 presents a comparison of the median subgroup scores and IQRs for the PREM18-1 and PREM18-2 across the three groups. The H-statistics for the four themed subgroups in PREM18-1, namely Clinical Staff, Waiting Time, Information, and Overall Treatment Satisfaction were computed as follows: 18.14, 108.92, 5.26, and 4.8, respectively, with degrees of freedom equal to two. Corresponding H-statistics for the subgroups in PREM18-2 were 1.80, 32.87, 5.04, and 1.70, respectively. Correspondingly, effect sizes for the four subgroups in PREM18-1 were 0.02 (95% CI: 0.01–0.05), 0.16 (0.11–0.21), 0.005 (−0.002–0.03), and 0.004 (−0.002–0.03), respectively, and in PREM18-2 0.00 (0.00–0.02), 0.08 (0.03–0.14), 0.01 (0.00–0.04), and 0.00 (0.00–0.3), respectively. For PREM18-1, statistically significant differences in subgroup score distributions were identified within the “Clinical Staff” and “Waiting Time” themed subgroups. Specifically, within the “Clinical Staff” subgroup, the CG scored significantly lower than both TGs (CG vs. TG1, p < 0.001 and CG vs. TG2, p < 0.02), while score distributions between the TGs did not differ significantly (p = 0.41). As for the “Waiting Time” subgroup, all post-hoc comparisons were highly significant (p < 0.001), with TG1 scoring higher than both TG2 and the CG, even though TG2 still scored significantly higher than the CG. For PREM18-2, significant differences in subgroup score distributions were found in the “Waiting Time” themed subgroup. Here, the post-hoc test showed that both the CG and TG1 scored significantly higher than TG2 (p < 0.0001). While box plots as shown in Figure 3 primarily provide a summary of distribution, highlighting the median (represented by the thick line) and the IQR, they may not visually portray the subtle differences within the overall distributions identified by the test applied. However, it is worth noting that these slight variances in response distributions likely have a minimal clinical impact.

Figure 3. Distributions of responses to the 18-item patient-reported experience measures at the intervention and treatment stages. Explanatory text: Responses are represented by boxplots, with the bold horizontal line within the box indicating the median, and the box itself illustrating the interquartile range (IQR). Each boxplot corresponds to specific subgroups: clinical staff, waiting time, information provided before, during, and after each trial stage, as well as the overall treatment satisfaction. The y-axes indicate the score range from 1 to 5. The vertical whiskers extend from the minimum to the lower quartile (base of the box) and from the upper quartile (top of the box) to the maximum, encompassing data outside the IQR. Dots presents outliers. The Kruskal Wallis Rank Sum test was utilised for statistical analysis. Notably, significant differences were observed between groups in the “Clinical Staff” subgroup at intervention and in the “Waiting Time” subgroup at both stages, with p-values for these comparisons being less than 0.05. Post-hoc tests were conducted for these subgroup comparisons.

In the overall treatment satisfaction subgroup within the PREM4-3 survey, there were no statistically significant difference between groups (p = 0.96). Both TG1 and TG2 demonstrated a median score of 4.25 (IQR: 4.00–5.00), while the CG had a median score of 4.00 (IQR: 4.00–5.00).

Free-text comments to the three PREM18 subgroups regarding the clinical staff during the stage-specific visit, the information provided before, during and after the stage-specific visit, and/or the overall treatment satisfaction at Stages 1, 2 and 3 were categorised into two groups of predominantly positive/neutral or negative comments within the three groups. Table 2 shows the distribution of positive/neutral and negative comments between groups and the total median scores of the participants who provided free-text comments within each category. Pearson’s Chi-square test showed that there were no statistically significant differences between distributions of positive/neutral vs. negative comments between the three groups.

Table 2. Count and distribution of free-text responses to the 18-item patient-reported experience measures at the intervention, treatment, and follow-up stages.

	Test group 1 (n = 251)		Test group 2 (n = 250)		Control group (n = 250)
PREM18-1 at stage 1	Count (%)	Total median score	Count (%)	Total median score	Count %	Total median score	p-Value
Positive/neutral comments	30 (12)	5.00	27 (11)	4.75	26 (10)	4.38	0.27
Negative comments	39 (16)	3.75	52 (21)	3.67	29 (12)	3.50
PREM18-2 at Stage 2 (Eligible responders)	(n = 166)		(n = 151)		(n = 137)
Positive/neutral comments	14 (8)	4.75	18 (12)	5.00	18 (13)	5.00	0.48
Negative comments	14 (8)	3.75	27 (18)	3.60	32 (23)	4.25
PREM4-3 at Stage 3 (Eligible responders)	(n = 231)		(n = 213)		(n = 137)
Positive/neutral comments	30 (13)	4.75	39 (18)	4.75	17 (12)	5.00	0.10
Negative comments	33 (14)	3.50)	21 (10)	3.75	19 (14)	3.75

Explanatory text.

This table presents the count and distribution of positive/neutral and negative free-text responses to the PREM18 survey subgroups including the clinical staff, the information provided before, during and after the stage-specific visit, as well as the overall treatment satisfaction across Test Group (TG) 1, TG2 and the Control Group. Median total scores are depicted for all groups within the positive/neutral and negative free-text response categories. Comparative analyses on the distribution of response categories across groups were conducted using Pearson’s Chi-squared test. Notably, no statistically significant differences were detected among the groups, with all p-values exceeding 0.05. Chi-square statistics were 2.59, 1.47, and 4.62 for the PREM18-1, PREM18-2, and PREM4-3, respectively.

PREM18: The 18-item Patient Reported Experience Measures (PREM) questionnaire; PREM18-1: The 18-item PREM questionnaire applied at Stage 1; PREM18-2: The 18-item PREM questionnaire applied at Stage 2; PREM4-3: The 4-item PREM subgroup questionnaire regarding overall patient satisfaction applied at Stage 3.

Negative feedback primarily centred around insufficient information about the trial, HAs, and individual HL severity, alongside issues with staff communication, technical delay, rushed visits, appointment cancellations, and parking difficulties. Positive comments highlighted satisfaction with HAs, quick examinations and HA fittings, quality information sources including the InHEAR trial webpage and the personalised electronic letters, and the professionalism and kindness of the staff. The trial period was free of any adverse or serious events.

Comparative analysis for the daily average HA usage between groups revealed a median of 7.0 h a day (interquartile range (IQR): 2.3–11.5) in TG1 participants, 7.5 h a day (IQR: 3.2–12.0) in TG2 participants, and 7.0 h a day (IQR: 2.0–11.8) in CG participants. A p-value of 0.5 indicated that there were no significant differences in the daily averaged HA usage between the three groups.

Discussion

Limited physical examination and lack of communication and personal interaction between patients and healthcare providers are factors that may potentially compromise treatment benefit and patient satisfaction when physical medical assessment routines are replaced by novel telemedical, digital and remote assessment alternatives. Therefore, we aimed to investigate whether RESA screening in adult potential first-time HA users compromised self-reported HA benefit and satisfaction when compared to PESA screening.

The SSQ12 and SSQ12-B survey results revealed insights into participants’ hearing ability and quality before and after HA treatment. SSQ12 responses showed no significant difference in self-reported hearing disability across the three groups pre-treatment. However, after adjusting for age, gender, and randomisation group, the data revealed that individuals with normal hearing and simple HL reported significantly higher speech, spatial, and total SSQ12 scores than those with complicated HL. Qualities of hearing domain scores were significantly higher in individuals with normal hearing than in those with simple or complicated HL. This suggests these individuals experienced a higher level of hearing ability and quality pre-treatment, consistent with the premise that severe hearing-impairment poses greater daily challenges. Post-treatment (2–4 months), there was no significant difference in self-reported hearing ability or treatment benefit across the groups, as assessed by SSQ12-B and IOI-HA. This may firstly be explained by the fact that participants with normal hearing who were not administered HAs did not complete the SSQ12-B. Secondly, it may suggest that RESA screening does not somehow undermine self-reported HA benefits compared to PESA screening. And thirdly, it might simply indicate that the HA treatment trajectory is independent of—and thus not directly impacted by—the ENT specialist’s screening method.

PREM survey results offered insights into patient experiences across the three groups. For PREM18-1 completed at the intervention stage, the CG was less satisfied with the “Clinical Staff” than both the TG1 and TG2, however a small effect size of 0.02 suggest limited practical applications for the difference detected between groups. In terms of “Waiting Time,” TG1 was most satisfied, followed by TG2, with the CG least satisfied, a finding of potential practical significance given the large effect size of 0.16. These results might be attributed to private ENT specialists (who conducted PESA screening in CG participants) potentially having less time with the patients compared to private and public audiological clinics due to busier schedules and a higher daily patient turnover, contributing to shorter patient consultations and extended waiting times. The results from PREM18-2 mirrored this trend in “Waiting Time”. Both CG and TG1 registered higher scores than TG2. However, the effect size of 0.08 suggests a modest practical significance for the difference detected between the groups (see Supplementary Material for further details on covariates). Despite the lack of statistically significant differences in the distribution of positive/neutral and negative feedback across groups, the free-text comments yielded nuanced insights into specific facets of patient satisfaction and dissatisfaction. Issues such as scheduling, parking, and the maintenance of high standards of clinical care and professionalism were highlighted as crucial for a successful treatment journey. In essence, the PREM survey results underscored the importance of efficient logistics and quality interactions with staff in shaping patient experiences within hearing rehabilitation.

The average daily HA usage did not differ significantly among participants in the three groups. That said, the relatively low medians of 7.5 h per day in TG1 participants and 7 h per day in TG2 and CG participants and their large IQRs indicate that there is a considerable amount of variation in HA usage within each group. This implies that some individuals may use their HAs for significantly fewer hours than the median value, while others may use them for significantly more hours. Potential reasons for this variability in usage patterns have not been investigated in this study but may include factors such as demographics, hearing impairment severity or environmental influences.

In recent years, HA benefit and patient satisfaction outcomes have been applied in studies of digital treatment and communication option technologies, tele-audiology, and tele-rehabilitation services in hearing healthcare. (Maruthurkkara, Case, and Rottier 2021; Forde et al. 2022; Şahin, Yavuz Veizi, and Naharci 2021; Almufarrij et al. 2022) In one study, general telemedicine interventions for older adults were shown to improve health outcomes, QoL indicators and patient satisfaction. (Şahin, Yavuz Veizi, and Naharci 2021) In another study from 2022, patient and clinician effectiveness and satisfaction analyses were conducted by telemedicine in 304 adult new and follow-up otolaryngology referrals. The commonest symptoms in new referrals were tinnitus (19%) and HL (18%), and the most common telemedical diagnosis made was sensorineural HL, mainly presbycusis (15%). Overall, 96% were satisfied with the consultation form (Swaminathan, Mughal, and Phillips 2022).

Although randomisation does not guarantee perfectly balanced groups, a strength of the study was the study design that aimed to reduce the risk of both known confounding factors (e.g., baseline characteristics and type of HL) and unknown confounding factors (e.g., genetic, lifestyle, and psychological factors). Comparability in age, gender, and severity of HL was found between groups (Siggaard et al. 2023). Also, the large sample size and the compelling response rates for all questionnaires applied increased the validity of the results. However, despite high response rates ranging from 83% to 100%, it is important not to completely disregard the possibility of selection bias, as individuals who volunteer for the study may be more motivated to receive HA treatment compared to non-volunteering individuals. Additionally, participants who are less satisfied may be less inclined to complete the follow-up questionnaires, resulting in an overrepresentation of satisfied participants among the responders.

In conclusion, our findings suggest that RESA screening does not undermine patient-experienced HA benefit and satisfaction when compared with PESA screening. Digital solutions are increasingly integrated into healthcare, aiding the management of chronic illnesses, and enhancing patient feedback collection for research and quality analysis. While some of the questionnaires employed in this study are endorsed by the Danish Ministry of Health for public healthcare, their use is not commonplace in the private sector, which delivers around half of the Denmark’s’ HA services. Tools like the IOI-HA are underused in public clinics, and the SSQ12 remains unused in both sectors. This lack of uniform data collection and quality control across sectors could impact future care quality. We advocate for a systematic use of such data across sectors in line with forthcoming legislative and clinical guidelines. These will necessitate clinics to report on HA benefits and patient satisfaction to the Danish Health Data Authority using IOI-HA, SSQ12, and other tools for satisfaction outcomes. This approach aligns well with evolving practices in Danish hearing healthcare, offering a useful model for future quality assurance and data analysis.

Author contributions

Lene Dahl Siggaard (LDS) is first author and acts as guarantor. The InHEAR trial management group, LDS, Henrik Jacobsen (HJ), Dan Dupont Hougaard (DDH) and Morten Hoegsbro (MH), conceived, designed, and conducted the study. LDS drafted the original manuscript with contributions from all authors. LDS and MH conducted all statistical analyses and data interpretations with help from the biostatistical advisory service at Aalborg University, Denmark. All authors reviewed, revised, and approved the final manuscript and agreed to be personally accountable for their contributions and to ensure that questions related to the accuracy or integrity of any part of the work, even ones in which the author was not personally involved, were appropriately investigated and resolved, and the resolution documented in the literature. The corresponding author attests that all listed authors meet the authorship criteria and that no others meeting the criteria have been omitted.

Transparency

LDS, the lead and corresponding author, affirms that the manuscript is an honest, accurate and transparent account of the study being reported; that no important aspects of the study have been omitted; and that any discrepancies from the study as planned (and, if relevant, registered) have been explained.

Ethical approval

The trial was registered retrospectively with ClinicalTrials.gov as trial registration prior to participant recruitment is not mandatory in Denmark. However, the Health Research Ethics Committee for Northern Denmark was notified about the trial prior to patient recruitment. Upon receipt of on the notification, the Committee decided that they did not require a formal application before patient recruitment could be initiated (case no. 2020-000992). https://clinicaltrials.gov/ct2/show/NCT05154539

Acknowledgements

This trial was made possible by the collaborative efforts of doctors, secretaries, audiology assistants and administrators at the Department of Otolaryngology and Audiology at Aalborg University Hospital, Denmark, as well as by the participating private and public hearing rehabilitation clinics and private ENT specialist clinics in the North Denmark Region, Denmark. The authors take this opportunity to express their gratitude to everyone who contributed their time and expertise, in particular the trial participants and those who contributed to the feasibility of the study. Their input and understanding were important in ensuring the success of this important study.

Data sharing statement

Written participant consent for data sharing was obtained prior to trial inclusion and randomisation. In pursuance of Danish legislation, data will be available only after approval by the Danish Data Protection Agency and after signing an access agreement.

Disclosure statement

All authors have completed the ICMJE uniform disclosure form at www.icmje.org/coi_disclosure.pdf and declare as follows: no support was received from any organisation for the submitted work; no financial relationships existed with any organisations that might have an interest in the submitted work in the previous three years; no other relationships or activities exist or have existed that could appear to have influenced the submitted work.

Additional information

Funding

This publication presents independent research funded by the Danish Health Data Authority and the North Denmark Region. The views expressed in this publication are those of the authors and not necessarily those of the Danish Health Data Authority or the North Denmark Region. No commercial support was received for this study.