Article Text
PDF
Abstract
Purpose The Netherlands Longitudinal Study on Hearing (NL-SH) was set up to examine associations of hearing ability with psychosocial, work and health outcomes in working age adults.
Participants Inclusion started in 2006 and is ongoing. Currently the sample comprises 2800 adults with normal and impaired hearing, aged 18–70 years at inclusion. Five-year follow-up started in 2011, 10-year follow-up in 2016 and 15-year follow-up in 2021. All measurements are web-based. Participants perform a speech-in-noise recognition test to measure hearing ability and fill out questionnaires about their hearing status, hearing aid use, self-reported hearing disability and coping, work status and work-related outcomes (work performance, need for recovery), physical and psychosocial health (depression, anxiety, distress, somatisation, loneliness), healthcare usage, lifestyle (smoking, alcohol), and technology use.
Findings to date The NL-SH has shown the vast implications of reduced hearing ability for the quality of life and health of working-age adults. A selection of results published in 27 papers is presented. Age-related deterioration of hearing ability accelerates after the age of 50 years. Having a history of smoking is associated with a faster decline in hearing ability, but this relationship is not found for other cardiovascular risk factors. Poorer hearing ability is associated with increased distress, somatisation, depression and loneliness. Adults with impaired hearing ability are more likely to be unemployed or unfit for work, and need more time to recuperate from work effort.
Future plans Participant data will be linked to a national database to enable research on the association between hearing ability and mortality. Linking to environmental exposure data will facilitate insight in relations between environmental factors, hearing ability and psychosocial outcomes. The unique breadth of the NL-SH data will also allow for further research on other functional problems, for instance, hearing ability and fall risk.
Trial registration number NL12015.029.06.
- mental health
- audiology
- occupational & industrial medicine
Data availability statement
Data are available upon reasonable request. Access to NL-SH data is available through collaborative agreements. Please contact Professor Sophia E Kramer (se.kramer@amsterdamumc.nl) for further information.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
STRENGTHS AND LIMITATIONS OF THIS STUDY
This cohort study has up to 15 years of follow-up on psychosocial health and work-related outcomes among young and middle-aged adults with and without hearing impairment. The latter group has been under-represented in earlier research.
Studies with Netherlands Longitudinal Study on Hearing (NL-SH) data have added to the body of knowledge on associations between hearing ability and work outcomes, which few studies addressed before.
The fully web-based design allows participants to provide measures at a time and place convenient to them. Compared with traditional methods, internet sampling increases geographic reach and is more effective for enrolling individuals who are hard to reach or relatively healthy at baseline.
The study uses a convenience sample of about 2800 individuals who actively sought out to test their hearing online and, after being presented with their results, volunteered to participate. Because of participants’ pre-established interest in hearing impairment, results may not be completely generalisable to the general population.
Response rate at 10-year follow-up can be considered acceptable for a fully web-based study and is sufficient to detect meaningful differences in longitudinal outcomes. However, as the NL-SH is not a large-scale population study, the sample may not be large enough to detect small differences in continuous outcomes or differences in proportions of less common outcomes.
Introduction
Over 1.5 billion people live with some degree of hearing loss worldwide. Its prevalence will almost double by 2050 due to increased life expectancy.1 Although the prevalence of hearing loss is highest in people aged 65 years and older, a substantial number of adults <65 years also experience hearing difficulties.1 2 Because of differences in the demands of life between these younger adults and adults of retirement age, the impact of hearing loss may differ between them. Two decades ago, only a few quantitative studies in the international literature had focused on mental health of adults with hearing loss below the age of 70. Most of these studies measured self-reported hearing status and all were cross-sectional.3 Moreover, few studies had addressed associations between hearing status and work outcomes.4 These studies showed lower participation in the workforce, but the implications among working people with hearing impairment were lesser known. Productivity losses due to reduced performance while at work, sickness absence, being unfit for work and early retirement have substantial costs for society. This may become a larger problem as deferral of state pension age results in a higher proportion of workers with hearing loss. In the same vein, healthcare use of working-age adults with hearing loss was under researched at that time.4 Better knowledge of the impact of hearing loss on working-age adults could lead to better support, and prevention of the negative consequences of hearing loss. The Netherlands Longitudinal Study on Hearing (NL-SH) was therefore initiated in 2006, to gain more insight into the cross-sectional and longitudinal relationship between hearing ability and various domains of life including psychosocial health, physical health, work-related outcomes and healthcare use among individuals aged 18–70 years. Insight in the role that hearing solutions can play in mitigating these outcomes was also included in the research goals. As data collection progressed, the breadth of the database allowed researchers to study research questions beyond those initially identified when the cohort was set-up, for example, the association between age and decline in recognition of speech-in-noise and determinants and outcomes of tinnitus. Additionally, between 20 November 2020 and 12 January 2021 data were collected to study effects of the COVID-19 restrictions on adults with hearing loss.
This paper outlines the design of the NL-SH, describes baseline characteristics of the study population included up to 1 February 2022, and presents key findings thus far.
Cohort description
The NL-SH is a web-based prospective cohort study in which individuals with normal hearing and those with hearing impairment between the ages of 18 and 70 years are invited to participate.
Recruitment and inclusion criteria
The study was initiated in 2006. At that time participants were recruited through various approaches. Information about the study was posted on several websites (both hearing-related and non-hearing-related), flyers were distributed at audiology centres and hearing aid dispensers, articles about the study appeared in magazines aimed at patients with hearing loss and in trade magazines of the hearing loss industry, and advertisements were placed in newspapers in the region of Amsterdam, the Netherlands. However, from its initiation, most participants were recruited through the web-based Dutch National Hearing Test (NHT). This hearing test has been available since 2005 and is found at www.hoortest.nl.5 It offers the general public a fast and convenient self-test to screen for hearing impairment. A score is calculated that, based on validated cut-off points, is considered good or insufficient. Depending on the outcome of the test, the participant receives appropriate advice. They are then asked if they are interested to take part in hearing-related research. If interested, they are directed to the NL-SH website at www.hooronderzoek.nl. The introduction page of this website explains the study and provides visitors with an informational brochure. They can register for the study by filling out an online form, asking for their sex, age and contact details (email address, home address and phone number). Age is checked against the inclusion criterion of being 18–70 years of age; no data are stored for those who do not meet this criterion. After enrolment, eligible adults receive a link to the web-based study questionnaire. Informed consent for the study is asked as part of this questionnaire. Due to the nature of the recruitment process, no predefined group of individuals is approached for participation and therefore information on reach or reasons for refusal is not available.
Data collection
Inclusion started in 2006 (T0) and continues still. The 5-year follow-up measurement (T1) started in 2011, the 10-year follow-up measurement (T2) started in 2016 and the 15-year follow-up measurement (T3) started in 2021. Hearing ability was measured with an online speech-in-noise test. All other data were collected from questionnaires. In 2006 and 2007, a subset of participants filled out five additional monthly questionnaires on healthcare use after their baseline measurement; this is further described by Nachtegaal et al.6 A once only questionnaire on hearing-related difficulties encountered due to the COVID-19 restrictions was distributed to all participants on 20 November 2020.7
Eligible participants received a link to the questionnaire. An email invitation for the T1, T2 and T3 measurements was sent, respectively, 5, 10 and 15 years after the T0 hearing test was performed. In the questionnaires, a link to the speech-in-noise test was provided. In general, the hearing test was performed directly after filling out the questionnaires, but a delay of up to 3 months was allowed. For each measurement round, two email reminders and one postal reminder were sent within 3 months after the first invitation for the questionnaire. To those not concluding the hearing test, three email reminders were sent during those 3 months.
Between measurement rounds, participants are kept involved by a three to six monthly emailed newsletter about the study and its results. The newsletter is sent to each participant, unless they opt-out. Loss to follow-up is mitigated by contacting participants through regular mail if their email address shows up as incorrect.
Participants can withdraw their consent any time they want, using a web-based form. The reason for withdrawal is asked, but they can leave this blank.
Speech-in-noise test
As the first and major complaint of adults with hearing loss is difficulty in understanding speech in background noise,8 understanding speech-in-noise is more representative of limitations experienced in everyday listening than understanding speech-in-quiet or pure-tone thresholds.9 The ability to understand speech-in-noise is measured using the protocol of the first internet version of the NHT. Originally developed for use by phone,10 this internet version of the NHT was launched in 2005.5 The test is diotic (ie, identical signals to the left and right ear), and the results are mainly representative of the better ear.11 Participants are instructed to perform the test in a quiet room and, preferably, use headphones. They may however use speakers and have to indicate which transducer they used. Participants with hearing aids are instructed to remove their aids. First, digit triplets are presented in quiet (without noise) and participants are instructed to adapt the volume of their computer to a level where they can clearly hear the presented digits. Then, a total of 23-digit triplets (eg, 6-2-5) are presented against a background of stationary masking noise in an adaptive manner: the noise level is fixed, and the speech level varies. After an incorrect response, the subsequent triplet is presented at a 2 dB higher level, which increases the signal-to-noise ratio (SNR) by 2 dB. If the participant provides a correct response, the subsequent triplet is presented at a 2 dB lower SNR. The speech-reception threshold in noise (SRTn) is calculated by taking the average SNR of the last 20 presentations, corresponding to a score of 50% of the presented triplets understood correctly. SRTn values can reach a level of 4 dB SNR. Based on their score, a participant is categorised into one of three hearing ability groups: ‘good’ (SRTn<−5.5 dB SNR), ‘insufficient’ (−5.5 dB SNR≤SRTn≤−2.8 dB SNR) or ‘poor’ (SRTn>−2.8 dB SNR). These categorisations are based on the standard speech-in-noise test in the Netherlands, which uses sentences within stationary speech-shaped noise.10 Validation of the original NHT version showed a strong correlation (ρ=0.87) with SRTn’s derived from the standard sentences SRTn test.10 The average measurement error (SE of measurement) is estimated to be 0.95 dB.12 Compared with the standard sentence-in-noise test, the NHT phone version was shown to have a sensitivity of 0.91 and a specificity of 0.93 at a cut-off of –4.1 dB SNR for distinguishing normal hearing from impaired hearing.10 Impaired hearing was further divided into insufficient and poor hearing ability, using a cut-off of −1.4 dB SNR. For a detailed explanation of these cutoffs see Smits and Houtgast.12 Because of the benefit of listening with two ears (binaural summation), the cut-off values for the internet version were adjusted with 1.4 dB to −5.5 dB and −2.8 dB SNR, respectively.5 In recent years the NHT has been adapted for worldwide use. It is used for screening purposes, for example, in the WHO hearing app hearWHO,13 as well as in research, for example, in the UK Biobank.14 It should be noted that monaural and diotic speech understanding in noise is less compromised in listeners with conductive hearing loss than in listeners with sensorineural hearing loss. Similarly, speech understanding in listeners with unilateral hearing loss is less affected by a diotic test. This means that participants with these types of hearing loss, even those normally using a hearing aid, may be classified as normal hearing. Participants using speakers may perform slightly worse in the test than if they had used headphones, resulting in underestimation of hearing ability and possible misclassification to the group with hearing loss for some of the participants with normal hearing.
Questionnaires
At each measurement round, participants filled out a web-based questionnaire with questions (table 1) about their hearing status, hearing aid use, hearing disability, communication strategies and personal adjustment, hearing-related health problems (tinnitus, hyperacusis and dizziness), employment status and work-related outcomes (work performance, need for recovery), psychosocial work characteristics, psychosocial health (depression, anxiety, distress, somatisation, loneliness), health-related quality of life, other chronic conditions, healthcare usage, lifestyle (smoking, alcohol) and social media use. Standardised validated Dutch-language questionnaires were chosen to measure these topics, supplemented by researcher-devised questions when standardised questionnaires were not available. A full overview of all topics on which data are collected, the questionnaires or researcher-devised questions used and the rounds in which they are measured can be found in table 1. Time required from the participants for performing the hearing test and filling out the questionnaires differs per measurement round (range 30–60 min).
Measurements at baseline (T0), 5-year follow-up (T1), 10-year follow-up (T2) and 15-year follow-up (T3)
Patient and public involvement
Involvement of patients and the public is important to the NL-SH. A patient advisory board was established at the start of the second measurement round in 2012. The board comprises 3–5 individuals with hearing loss, with no defined term of office. Two individuals have been engaged since the start, whereas three others have been replaced. The board members were involved in identifying research gaps, based on their own lived experience with hearing loss. They also comment on researcher-devised questions for follow-up measurement rounds, to help improve these.
Furthermore, the NL-SH website is publicly available and offers the broader public the option to comment on the study. Results of the study are communicated to the participants in the NL-SH newsletter. Participants are also encouraged to send in new research suggestions.
Findings to date
Participant characteristics at inclusion
Between 1 October 2006 and 1 February 2022, 3390 adults had enrolled in the NL-SH, of whom 590 gave no informed consent. Their personal data were removed, resulting in 2800 included participants. More than half of them, namely 1574 (56.2%) participants, were included in 2006 and 2007. Average enrolment per year since then was 87 participants. Figure 1 shows the number of participants included in each year.
Number of participants included per year from 1 October 2006 till 1 February 2022.
Baseline demographics and hearing-related characteristics of the 2800 participants, as well as the distribution of these characteristics over the hearing ability groups are presented in table 2. Mean age at baseline was 47.6 years (SD 13.2). The majority of participants were women, 62.0% (1762), and approximately half of the participants had a high level of education (bachelor’s degree or higher), 51.5% (1438). Based on the hearing test 45.5% (1274) had insufficient or poor hearing ability. A total of 646 (23.1%) participants reported using a hearing aid.
Baseline characteristics of participants for all participants, and split out to three categories of hearing ability
Response rate and study withdrawal
Considering inclusion is open-ended, meaning that follow-up measurement rounds are also open-ended, response can only be reported in relation to the year of inclusion. Up to 1 February 2022, 1379 participants took part in the 5-year follow-up, 833 participants in the 10-year follow-up and 422 participants in the 15-year follow-up. Of the 2099 participants who provided baseline data in 2006–2011, 1145 (54.5%) provided data at their 5-year follow-up in 2011–2016 and 825 (39.3%) at their 10-year follow-up in 2016–2021. It should be taken into account that participants may miss one measurement round.
By 1 February 2022, 1068 participants were no longer participating. Out of this group, 398 were removed from the administrative database because they had not responded to both the 5-year and 10-year follow-up invitations, nor to a letter asking if they would like to continue their participation. Nine were removed because of defunct contact information, and 70 participants were reported as deceased by family members and consequently removed. Registration of death may be incomplete; there could be deceased participants among the participants with defunct contact details or among those missing two consecutive measurement rounds. Out of 591 participants who withdrew their consent, 320 participants declined to provide a reason for this. The most commonly reported reasons for withdrawal were loss of interest (n=87), believing they were no longer eligible for the study (n=54) or finding the questionnaire too time-consuming (n=40).
Baseline characteristics of participants who remained in the study and of those who dropped out are presented in table 3. Differences between these two groups were analysed using an independent samples t-test for normally distributed continuous data, and χ2 tests for categorical data. Results from these analyses are also shown in table 3. Dropouts were younger than remaining participants by on average 2.2 years (95% CI 1.2 to 3.2, p<0.001) and were less likely to have a high level of education (bachelor’s degree or higher), 42% versus 57% (p<0.001). Dropouts were also less likely to have parents with hearing loss (36% vs 42%, p=0.005) or siblings with hearing loss (12% vs 15%, p=0.033). Among those with self-reported hearing disability, participants with a gradual onset were more likely to be among the dropouts (75% vs 66%, p=0.001) and those with congenital hearing loss were less likely to quit the study (18% vs 12%, p=0.001). Among the same group with self-reported hearing disability, dropouts were less likely to be hearing aid users (17% vs 27%, p<0.001).
Baseline characteristics of remaining participants and dropouts, and results of the analyses of differences in characteristics of participants who were removed or withdrew consent as compared with those who remain in the study
Key findings
Results of the NL-SH have been published in peer-reviewed journals (27 papers3 4 6 7 15–37), trade journals (8 articles) and were reported by public media. Study findings have been used in the Dutch general practitioner and occupational medicine guidelines on hearing loss,38 39 for national policy documents concerning hearing loss and its treatment, and in statistics published by the Dutch Ministry for Public Health, Welfare and Sports.40 In this section, we provide a summary of the most notable published results so far, with references to the respective papers.
Risk factors for decline of hearing ability
We found that age-related decline in hearing ability over 5-year and 10-year intervals accelerated after the age of 50 years.20 27 Mean SRTn deterioration over 10 years was 0.89 dB SNR in all participants, but 1.37 dB SNR and 1.69 dB SNR in adults aged 51–60 and 61–70 years at baseline, respectively.27 Hearing ability also declined faster in adults with a history of smoking, but this relationship was not found for other risk factors of cardiovascular disease.27 29
Comorbidities and healthcare use
Co-occurrence of a number of chronic conditions was more common in adults with poor or insufficient hearing ability than in their normally hearing peers.19 Most notably, based on baseline data, those with poor hearing ability were more likely to have diabetes, falls caused by dizziness or an inflammatory arthritis other than rheumatoid arthritis. After the 5- year and 10-year measurement rounds were available, overall multilevel analyses of all rounds showed that participants with hypertension, rheumatoid arthritis and obesity had worse hearing ability.29 An increased use of non-hearing-related healthcare, including mental healthcare, was therefore suspected, but this association was not found in cross-sectional analyses.6
Psychosocial outcomes
Poorer hearing ability was associated with higher distress, somatisation, depression and loneliness in cross-sectional analyses.3 Different age groups exhibited different associations: poorer speech-in-noise recognition in adults aged 18–30 years was associated with higher loneliness whereas in adults aged 40–49 years it was associated with distress, depression and anxiety. Over a 5-year period, a reduction in hearing ability was longitudinally associated with increased levels of loneliness in adults with a high educational level, those who entered into matrimony and those with a stable pattern of hearing aid non-use.22
Socioeconomic and work-related outcomes
In cross-sectional analyses we found that hearing ability was associated with lower levels of education, lower income and less participation in work.18 Cross-sectionally, working adults with worse hearing ability had a higher need for recovery from work, meaning that it takes them more time to recuperate from work-induced effort.16 Comparable results were found in longitudinal analyses.33
Outcomes in users of hearing solutions
In cross-sectional analyses, no differences in psychosocial health were found for those with insufficient or poor hearing status compared with good hearing status. This was also the case for those with hearing aids compared with those without hearing aids.3 However, a 5-year deterioration in hearing ability was only associated with increased loneliness among participants who did not use hearing aids during that period, in contrast with those with a stable pattern of hearing aid use and those who took up a hearing aid.22 Further study should clarify if hearing aids prevent an increase in loneliness among individuals with deteriorating hearing ability and if characteristics associated with uptake and continued use of hearing aids play a role in this relationship. The uptake of hearing aids and one or more alternative hearing assistive technologies during a 5-year period was 21.6% among working participants with insufficient and poor hearing ability.28 This uptake did not have a significant effect on their 5-year change in need for recovery from work.33
We also studied psychosocial health of participants who use a cochlear implant.23 Among adults with self-reported hearing loss, cochlear implant users had better scores on emotional loneliness than both individuals who used hearing aid users and those who did not. Cochlear implant users also had lower anxiety scores than hearing aid users. No differences in psychosocial health were found between cochlear implant users and adults who reported hearing normally.
Determinants and outcomes of tinnitus
Development of tinnitus was longitudinally associated with higher levels of somatisation and a history of smoking reported 5 years earlier.34 Anxiety and poor speech recognition in noise were associated with higher tinnitus annoyance in participants with newly developed tinnitus.34 Having tinnitus and higher tinnitus annoyance were both not associated with a higher need for recovery after work.36
Future plans
The NL-SH dataset comprises up to 15 years of follow-up on a broad set of outcomes relevant to adults with hearing loss. This abundance of information provides the opportunity to answer a wealth of research questions for years to come. Several publications are currently submitted or in preparation, among which studies on the impact of the COVID-19 measures on adults with hearing loss compared with normal hearing adults, and the longitudinal association of SRTn with risk of falling.
Additionally, the NL-SH participates in the Geoscience and Health Cohort Consortium (GECCO)24 32 as well as in an interdepartmental collaboration of several cohort studies within the Amsterdam Public Health Institute. The GECCO collaboration enables a linkage with a broad range of environmental characteristics. These data linkages will enrich the NL-SH data to allow research questions on relations between environmental factors, hearing ability and psychosocial outcomes. For example, whether hearing ability is associated with noise pollution or the association between hearing aid ownership and distance living from the nearest hearing aid dispenser. A one-time linkage to the national database on mortality and cause of death from Statistics Netherlands is in preparation. These data will be used to study differences in mortality rates between normal hearing adults and those with hearing loss.
Strengths and limitations
The main strength of the NL-SH is that up to 15 years of follow-up data on adults aged 18–70 years at inclusion are available. It is the only cohort with this duration of follow-up on recognition of speech-in-noise, self-reported hearing disability and coping, psychosocial health and work-related outcomes. The repeated measurements have so far allowed for research on associations between decline in recognition of speech-in-noise with psychosocial outcomes, which had not been done previously. The longitudinal design also made it possible to study risk factors for a decline in recognition of speech-in-noise and for tinnitus, thus helping in building evidence for (lack of) causality. Another strength of the NL-SH is the use of a validated speech-in-noise test that does not require in-person testing, improving ease of participation and decreasing the burden of participation. While the majority of cohort studies that have followed hearing ability and health trajectories of working-age adults for at least 5 years requested their participants to travel a study site to undergo pure-tone audiometry, this is labour intensive, costly, and requires time and effort from participants. It can also be argued that the ability to recognise speech-in-noise is more relevant for measuring the disabling effects of hearing loss than pure-tone measurements. The first and major complaint of adults with hearing loss is difficulty in understanding speech in background noise, which is not fully captured by pure-tone audiometry.41 42 Finally, the web-based design allows participants to perform measurements at a time and place that is convenient to them. Internet sampling is more effective for enrolling groups that are hard to reach or relatively healthy at baseline compared with traditional methods.43 The Netherlands has always been a forerunner in internet use. At the start of recruitment to the study in 2005, 78% of households had access to the internet at home,44 with a nearly similar proportion of adults aged 50–65 years having access to it as in the general population.3 This reach was deemed sufficient to ensure comprehensive inclusion. Internet access at home has since increased to 98% in 2021, ensuring an even broader reach of the Dutch population.45
As the limitations of the NL-SH are also described in our published papers, we focus on points for improvement not previously mentioned. First, the cohort concerns adults who actively sought to test their hearing ability and volunteered to participate in the study after hearing their results. Results may therefore not be completely generalisable to adults who have no interest in their own hearing ability and/or taking part in hearing-related research. This particularly affects analyses on (factors related to) likeliness of seeking hearing healthcare. NL-SH participants with hearing impairment may also have more favourable outcomes than non-participants with hearing impairment, due to their more active approach to understanding their hearing problem. Selection bias is further exacerbated by differential dropout rates, with participants with a gradual decline in hearing ability (in majority these are adults with age-related hearing loss) being more likely to dropout, and those with congenital hearing loss as well as those using a hearing aid being more likely to continue. A limitation of the design with measurements 5 years apart is that this is not suitable for studying outcomes which are realised in a short time after exposure, or to identify specific trajectories of decline. Finally, response rate at 10-year follow-up is acceptable, considering the NL-SH is a fully web-based cohort. Still, analyses may have insufficient power when dichotomous outcomes are uncommon and effect sizes for continuous outcomes are modest.
Collaborations
NL-SH data are available for collaborative research with other non-profit institutes. Interested investigators should contact the principal investigator, Professor Sophia E Kramer (se.kramer@amsterdamumc.nl), to obtain additional information, discuss collaborative opportunities and request a project proposal form. Further information can be found online (www.hooronderzoek.nl).
Data availability statement
Data are available upon reasonable request. Access to NL-SH data is available through collaborative agreements. Please contact Professor Sophia E Kramer (se.kramer@amsterdamumc.nl) for further information.
Ethics statements
Patient consent for publication
Ethics approval
This study involves human participants but at the start of the NL-SH in 2006, the study protocol has been approved by the Medical Ethics Committee of the Amsterdam Medical Center, location VUmc in Amsterdam, The Netherlands (2006/83; NL12015.029.06). As of 27 July 2016 this committee decided that the Medical Research Involving Human Subjects Act (WMO) does not apply to the NL-SH and that an official approval of this study by this committee is not required anymore. Participants gave informed consent to participate in the study before taking part.
Acknowledgments
We are grateful to the participating individuals for their commitment to this ongoing study.
References
Footnotes
Twitter @Mareeky
Contributors SEK, CS, UL, JRA, JN, MS,TG, LAJ and MFvW were involved in the conception and design of the study. JN, MS, MFvW, LAJ and JHMvB were responsible for data collection, data management and continued management of the cohort. MFvW performed the analyses presented in this paper, with a contribution from BIL-W. MFvW, LAJ, TG and SEK were responsible for interpretation of the data, and drafted and edited the manuscript. All authors have critically revised the manuscript, have approved the final version and agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy of integrity of any part of the work are appropriately investigated and resolved. MFvW is the guarantor.
Funding This cohort has been supported by the Heinsius-Houbolt Foundation, the Netherlands, grant number N/A; Sonova AG, Switzerland, grant number N/A; the EMGO+ Institute for Health and Care Research, The Netherlands, grant number N/A; and the Amsterdam Public Health research institute, Quality of Care, Amsterdam, The Netherlands, grant number N/A.
Competing interests MFvW, LAJ, TG, MS, JN, JHMvB, JRA, BIL-W, CS and SEK none. UL is affiliated with Sonova AG, one of the funding entities.
Patient and public involvement Patients and/or the public were involved in the design, or conduct, or reporting, or dissemination plans of this research. Refer to the Cohort description section for further details.
Provenance and peer review Not commissioned; externally peer reviewed.