Relationship between self-reported listening and communication difficulties and executive function: a protocol for a systematic review and meta-analysis

Introduction Listening and communication difficulties can limit people’s participation in activity and adversely affect their quality of life. Hearing, as well as listening and communication difficulties, can be measured either by using behavioural tests or self-report measures, and the outcomes are not always closely linked. The association between behaviourally measured and self-reported hearing is strong, whereas the association between behavioural and self-reported measures of listening and communication difficulties is much weaker, suggesting they assess different aspects of listening. While behavioural measures of listening and communication difficulties have been associated with poorer cognitive performance including executive functions, the same association has not always been shown for self-report measures. The objective of this systematic review and meta-analysis is to understand the relationship between executive function and self-reported listening and communication difficulties in adults with hearing loss, and where possible, potential covariates of age and pure-tone audiometric thresholds. Methods and analysis Studies will be eligible for inclusion if they report data from both a self-report measure of listening difficulties and a behavioural measure of executive function. Eight databases are to be searched: MEDLINE (via Ovid SP), EMBASE (via Ovid SP), PsycINFO (via Ovid SP), ASSIA (via ProQuest), Cumulative Index to Nursing and Allied Health Literature or CINAHL (via EBSCO Host), Scopus, PubMed and Web of Science (Science and Social Science Citation Index). The JBI critical appraisal tool will be used to assess risk of bias for included studies. Results will be synthesised primarily using a meta-analysis, and where sufficient quantitative data are not available, a narrative synthesis will be carried out to describe key results. Ethics and dissemination No ethical issues are foreseen. Data will be disseminated via academic publication and conference presentations. Findings may also be published in scientific newsletters and magazines. PROSPERO registration number CRD42022293546.


INTRODUCTION
Listening and communication are crucial for a healthy life and difficulties in communication can limit people's participation and adversely affect their quality of life.Hearing loss plays a fundamental role in determining a person's ability to listen and communicate, although research over the years has shown that there are other factors, beyond hearing loss, that are also important.Both hearing and listening and communication can be measured using behavioural measures such as pure-tone audiograms and speech-in-noise tests, or via self-report questionnaires.Behavioural and self-report measures of hearing are generally well correlated, and behavioural measures of hearing are often well correlated with both behavioural and self-report measures of listening and communication, that is, questionnaires.When a listener has a behaviourally measured hearing loss, it is likely that they will also experience and report difficulties with listening and communicating.
Effective communication, which relies on good hearing, is instrumental for a high level of functioning and good quality of life. 1

STRENGTHS AND LIMITATIONS OF THIS STUDY
⇒ This systematic review is the first to investigate the relationship between self-reported listening and communication difficulties and executive function using meta-analysis to synthesise the available evidence.⇒ It uses an established framework (International Classification of Functioning and Disease) and taxonomy (Cattell-Horn-Carroll-Miyake) to define target domains and measures of listening and communication difficulties and executive functions, respectively.⇒ Grey literature (including unpublished study results) will be included.⇒ This protocol has been reported in accordance with the PRISMA-P statement.⇒ Only studies available in English are eligible for inclusion.

Open access
Kiessling et al proposed a cascade linking hearing to effective communication, which in turn can be mapped to the International Classification of Functioning and Disease (ICF) framework (core set for hearing loss). 2 Both frameworks are displayed in table 1.
On the other hand, considering the activities and participation domain of listening and communication, behavioural and self-report measures are less robustly correlated.This is highlighted by the fact that two individuals who experience the same pure-tone average audiometric thresholds can experience and report substantially different degrees of listening and communication difficulties. 3One interpretation of this result could be that they assess slightly different concepts and/or highlight different contributing factors.One of those contributing factors whose role still remains to be fully understood is cognition. 4ognition has a complex relationship with auditory function depending on whether it is considered at the function (hearing) or activities/participation (listening/ communication) level.Specifically, hearing loss, both behaviourally measured and self-reported, has been shown to be associated with poorer cognitive performance across a range of cognitive domains including global cognition, episodic memory, processing speed, semantic memory, visuospatial ability, executive functions, and cognitive impairment and dementia. 5Indeed, Marrone and colleagues 6 reported adults reporting any trouble hearing were at nearly four times higher odds of reporting increased confusion and memory loss and half as likely to report good general health compared with adults reporting no hearing difficulty.These results are important to acknowledge because hearing loss has been identified as the leading potentially modifiable risk factor for dementia in midlife. 7or listening and communication, on the other hand, the type of assessment appears to play a role.For behavioural measures, the role of cognition for the ability to perceive speech (and in particular, speech in noise) has been reliably demonstrated for individuals with hearing loss, and this relationship is robust even when taking into consideration individuals' age and objective hearing levels (pure-tone average audiometric thresholds). 8Note that the cognitive ability most commonly assessed in studies of speech perception in noise is working memory.Other abilities such as attention and executive function are less regularly assessed and less robustly found to link to speech perception in noise.One reason for the less robust link might be that the speech in noise perception task needs to be a particular type or of more complexity in order to necessitate attentional and executive functions.
For self-report measures of listening or communication difficulties in quiet and in noise on the other hand, the role of cognition is much less clear and a clear link with cognition is not always shown. 8It is unclear why this link is so variable.Again, the cognitive ability most likely to be assessed is working memory.Maybe the listening situations most commonly assessed with self-report measures of communication are not of the type that require working memory or are more complex listening situations that would necessitate the involvement of executive functions.This idea would make sense given that listening and communicating in complex and noisy environments draw on the ability to shut out distractions and maintain focus.And thus it is conceivable that differences in executive functions may play a key role in the variation of individual experiences of listening and communication difficulties, regardless of absolute hearing levels.
Executive functions refer to 'higher order cognitive processes that control lower level cognitive processes in the service of goal-directed behaviour' (p.186). 9They enable the ability to think before acting, plan, meet novel and unanticipated challenges, resist temptations and maintain focus. 10ccording to Miyake and Friedman, 11 there are three core executive functions: mental-set shifting (shifting), information updating and monitoring (updating), and inhibition of prepotent responses (inhibition).Indeed, there is emerging evidence from large-scale cohort studies that individuals with self-reported hearing loss exhibit significantly poorer performance on tests of flexibility, psychomotor speed and executive function. 12Similarly, a systematic review of tinnitus research found individuals who reported tinnitus had poorer performance on measure of executive function compared with individuals who did not. 13Subsequent empirical research showed that for a population of adults with tinnitus, those reporting that their tinnitus was bothersome showed poorer performance on measures of executive function compared with those reporting non-bothersome tinnitus. 14n this review, we aim to synthesise the evidence assessing the relationship between self-reported listening and communication difficulties and objective measures of executive function, while controlling (where possible) for the potentially confounding factors of age and pure-tone audiometric hearing thresholds.To our knowledge, this independent relationship has yet to be extensively examined, despite data pertaining to both executive functions and self-reported listening and communication difficulties often being reported as part of wider research study methods.
When considering measures of self-reported listening and communication difficulties, it is important to clearly define what we mean, as there are well over a hundred self-report measures pertaining to listening, 15 and only a subset will be relevant to our current research question.For this reason, we adopt definitions of listening difficulties provided by the ICF as activity limitations and participation restrictions arising from hearing loss, and narrow our focus to self-report measures that align with ICF core set for hearing loss domains of listening (d115), communicating with -receiving -spoken messages (d310) and conversation (d350).Similarly, to definitively identify executive function domains and classify behavioural measures of executive function as either shifting, updating or inhibition, we will use the Cattell-Horn-Carroll-Miyake (CHC-M) taxonomy. 16view questions Primary research question Is there an association between self-reported listening and communication difficulties and performance on behavioural measures of executive function in adults with hearing loss?

Secondary research question
Is any association moderated by age and/or hearing loss (as measured using average pure-tone audiometric thresholds)?

Objectives
To review and synthesise evidence for the association between self-reported listening difficulties and performance on behavioural measures of executive function, in adults with hearing loss.

METHODS AND ANALYSIS Eligibility criteria Participants
Adults with hearing loss (with or without hearing devices), aged 18 years and over with no reported cognitive decline.We will accept a qualitative definition of hearing loss as 'mild', 'moderate', 'severe' or 'profound', or a quantitative definition where the group average pure-tone audiometric threshold is classed as mild hearing loss or greater using the WHO definition of mild (26-40 decibel [dB]  hearing level [HL] inclusive); moderate (41-60 dB HL inclusive); severe (61-80 dB HL inclusive) and profound (81+dB HL). 17 Studies that report on mixed populations (eg, children and adults or normal hearing participants and participants with hearing loss) will be included only if the data for the populations of interest are reported separately.

Intervention/interest
A correlation coefficient between self-reported listening or communication difficulties and executive function, either reported or calculated from other reported data.

Outcomes
Self-reported listening and communication difficulties can be measured by a single item or a questionnaire assessing the following ICF core set for hearing loss domains of listening (d115), communicating with-receiving-spoken messages (d310) and conversation (d350).
At least one behavioural measure of executive function must be included, defined according to the CHC-M taxonomy as tasks that measure updating (eg, verbal N-back), shifting (eg, trail making part B) and inhibition (eg, Stroop). 16here available, demographic information about the population (age, hearing device, group description) and objectively measured hearing loss (average puretone audiometric thresholds) will also be examined as subgroup descriptors and/or potential moderator(s).

Study design
Cross-sectional, longitudinal, experimental, quasiexperimental and observational studies will be included.

Information sources
Articles must be available in English.No restrictions on publication dates will be applied.
Databases to be searched (see box Open access literature including PhD theses, unpublished datasets and conference proceedings are eligible for inclusion.Unpublished data will be accessed by contacting the corresponding authors of identified records.Literature searches were carried out on 11 May 2022.

Article selection process
Two reviewers will independently screen titles and abstracts, and full texts of retrieved records, against the inclusion and exclusion criteria.If insufficient information is provided in the titles and abstracts to know if it should be included or if there is disagreement between the two reviewers, the article will be included in the fulltext screening.Disagreement at the full-text screening will be resolved by a third reviewer.

Data extraction process
A data extraction form will be created and improved by pilot testing before data extraction starts.The data from each study will be extracted separately by two reviewers and then compared.A third reviewer will be involved if there is any disagreement.Article selection and data extraction will be carried out using Covidence review management software (https://www.covidence.org/).

DATA ITEMS
The data to be extracted are the aim, study design, setting, conflicts of interest, demographic information about the population (age, hearing device, group description), sample size, bibliographic information (publication year, authors, journal), correlation coefficients between selfreported listening difficulties and executive function (and (if reported) between pure-tone audiometric thresholds and self-reported listening difficulties/executive function), type of executive function measure, type of self-reported listening difficulty measure and (where relevant) procedure of puretone audiometric assessment, as well as documenting any missing outcome data.We will note if both self-report and behavioural measures have been completed while wearing a hearing device.The authors will be contacted via email if sufficient detail is not reported.If data are only reported via figures, then WebPlotDigitizer (http://arohatgi.info/WebPlotDigitizer/app/) will be used to extract data.A third reviewer will be involved if there is any disagreement between data extracted.

Study risk of bias assessment
Two reviewers will assess risk of bias for each study identified by for each study using the appropriate JBI critical appraisal tool.If disagreements arise, a third reviewer will be involved.The JBI critical appraisal tools include assessment of methodological quality, and different checklists are used depending on the design of the study (eg, crosssectional, longitudinal, randomised controlled trial).For each criteria, studies will be assessed for fulfilment (yes, no, unclear or not applicable).

Data synthesis
Key study characteristics will be described, including study design, sample size and type of executive function measures used.The effect to be synthesised is the relationship between self-reported listening difficulties and behavioural measures of executive function defined by correlation coefficients.If correlation coefficients cannot be calculated or extracted for meta-analysis, study authors will be contacted to request the required information.Subgroup analyses will examine (where reported) key factors of age, category/measure of self-reported listening difficulty, type of executive function measure, and type of hearing device, and pure-tone audiometric thresholds.Meta-analyses will be conducted for subgroups where data for a minimum of n=5 effects are available.
The meta-analysis will be carried out using the correlation coefficient as the outcome measure.A random-effects model will be fitted to the data.We will calculate at maximum one correlation coefficient per type of executive function by type of listening difficulty.If a study reports multiple different correlations, there is likely to be some level of dependency in the data that needs to be dealt with.To handle any dependency between effect sizes in the analyses, a multilevel random-effects meta-analysis approach, as recommended by Assink and Wibbelink, will be applied. 18This approach includes one random effect for each study as an addition to the random effect for each effect size.Likelihood ratio tests will compare the fit of the multilevel model to the fit of the reduced models.If the multilevel random-effects analysis has better fit, it will be used in all analyses, otherwise the randomeffects model will be used.
The amount of heterogeneity (ie, τ 2) will be estimated using the restricted maximum-likelihood estimator. 19In addition to the estimate of τ 2, the Q-test for heterogeneity and the I 2 statistic will be reported. 20 21In case any amount of heterogeneity is detected (ie, τˆ2>0, regardless of the results of the Q-test), a prediction interval for the true outcomes will also be provided. 22Studentized residuals and Cook's distances will be used to examine whether studies can be defined as outliers and potentially influential in the context of the model. 23Studies with a studentized residual larger than the 100 × (1-0.05/(2×k))th percentile of a standard normal distribution will be considered potential outliers (ie, using a Bonferroni correction with two-sided α=0.05 for k studies included in the meta-analysis).Studies with a Cook's distance larger than the median plus six times the IQR of the Cook's distances will be considered influential.The analysis will be carried out using R and the metafor package. 24 25he analysis for the secondary research question will be carried out in the same way as above, with the difference that meta-regressions will be used to investigate potential moderator effects.The moderators will be evaluated one by one, and if both get significant effects, they will be evaluated together.The meta-regression procedure will follow the tutorial for meta-regression on the metafor package home page. 26

Table 1
Frameworks describing hearing and effective communication ICF, International Classification of Functioning and Disease.
1 for search terms): MEDLINE (via Ovid SP), EMBASE (via Ovid SP), PsycINFO (via Ovid SP), ASSIA (via ProQuest), Cumulative Index to Nursing and Allied Health Literature or CINAHL (via EBSCO Host), Scopus, PubMed and Web of Science (Science and Social Science Citation Index).