Hearing Aid Fitting in Tinnitus: A Scoping Review of Methodological Aspects and Effect on Tinnitus Distress and Perception

Current evidence on efficacy of hearing aids (HAs) on tinnitus perception and annoyance is considered insufficient due to the heterogeneity of tinnitus characteristics and of methods used in the relevant clinical studies. This is a scoping review focused on the methodological aspects of clinical studies evaluating the value of HA fitting as part of tinnitus management over the past 10 years. Thirty-four studies were included in the review, showing important heterogeneity in almost all aspects of inclusion criteria, comparators, outcome measures, follow-up time and HA fitting procedures. Although all studies show that HA fitting has a positive impact on tinnitus perception in patients with hearing loss, the methodological heterogeneity does not allow robust conclusions. Future studies taking into account the different nature and goals of each tinnitus therapeutic modality and adapting their methods, endpoints and timelines according to them could lay the groundwork for obtaining high-quality evidence on whether and how HA fitting shall be implemented in tinnitus management strategies.


Introduction
Tinnitus is traditionally defined as the perception of a sound in the absence of external stimuli; however, this definition has recently been updated in order to include patient's reaction and related annoyance as a determining factor [1]. Universally effective and accepted tinnitus treatment is currently pending, although a long list of substances and interventions, including but not limited to medicinal agents, sound treatment (ST), Transcranial Magnetic Stimulation (TMS), acupuncture and hearing amplification, has been tested for their efficacy in multiple studies [2].
Hearing loss and tinnitus are highly correlated, since it is estimated that up to 90% of patients experiencing tinnitus suffer from various degrees of hearing loss as well [3]. However, degree of hearing loss is not established as a prognostic factor for tinnitus existence and annoyance [4,5]. On top of this, approximately 10% of individuals with tinnitus have normal thresholds in Pure Tone Audiometry (PTA) [6]. This fact has triggered a wide interest in the literature with regards to cochlear synaptopathy which corresponds to a possible pathophysiological feature causing loss of the low spontaneous rate (low-SR) synapses without elevation of the PTA thresholds, as initially proposed by Schaette et al. (2011). Nevertheless, this concept has been recently questioned [7][8][9].
According to recent European guidelines, there is a weak recommendation for the hearing aids (HAs) in tinnitus treatment [1]. It is also stated that tinnitus presence should be taken into account during the hearing aid fitting procedure. However, supporting literature

Materials and Methods
This paper is a scoping review of the literature, aiming at pointing out the effect of hearing aid fitting on tinnitus perception and the methodological aspects of the relevant clinical studies. It is following the PRISMA Extension for Scoping Reviews (PRISMA-ScR) guidelines [12].
The main goals of this paper were to identify studies that describe hearing aid fitting in the case of people with tinnitus and evaluate their methodology as well as the effect of hearing aid fitting in tinnitus perception and related handicap, distress, annoyance and loudness.
Review questions were set as following: What are the methodological aspects of stu-dies evaluating the effect of HAs fitting on tinnitus perception? Is there an effect of the various HA fitting devices and methods on the perception of tinnitus characteristics in adults with hearing loss? More specifically, the question was formulated according to the PICO template as following: People: adults with tinnitus (bothersome or not) and hearing thresholds requiring amplification or not.
Intervention: hearing aid fitting with or without use of maskers and specific fitting techniques.
Comparator: not applicable. Outcomes: methodological aspects such as range of included hearing loss or outcome measures used and effect of HA fitting on tinnitus handicap, distress, annoyance and loudness as reported in questionnaires and scales used as outcome measures before and after hearing aid fitting.

Eligibility Criteria
Studies were selected according to the following criteria: Study Samples: Target population consists of adults with tinnitus, with hearing thresholds requiring amplification or not. Particulars: There was no restriction in tinnitus type. Studies that had a primary goal other than evaluating tinnitus were not excluded, as long as the results of hearing aid fitting on tinnitus perception were reported. There was no restriction in types of hearing loss. Sudden hearing loss, age-related hearing loss, noise trauma, hereditary hearing loss, otosclerosis etc, were all included to the review.
Intervention: hearing aid fitting, (no limitation on particular methodology, fitting technique, laterality, manufacturer or equipment).
Clinical experimental studies, case reports, case series, observational studies (longitudinal and cross-sectional), methodological papers, randomized clinical trials were included. Studies conducted during the past 10 years have been chosen for inclusion; stu-dies conducted with focus on hearing aid technology older than that has been consid-ered as out of the scope of this review. On-going studies, pediatric population studies, cochlear implantation and Tinnitus Retraining Therapy (TRT)-related studies, reviews and meta-analyses, experts' opinions and letters to the editor were all excluded. Articles in a language other than English were also considered non-eligible for this review.

Information Sources
Four major databases (Medline, Central, Web of Science, ClinicalTrials.gov (accessed on 19 May 2021) and Scopus) have been searched for eligible studies by two reviewers independently. The results were then hand-searched [13].

Search
Typically, literature search includes three sets of terms: terms concerning the health condition of interest (in our case, tinnitus), terms describing the intervention/exposure (Hearing aid fitting) and terms for the type of eligible studies (not applicable in our case since we have no particular limitation in study type) [14]. In this context, the search syntax for this scoping review for Medline was: (amplification OR "hearing aid" OR ("hearing aids"[MeSH]) OR "hearing aid fitting") AND (tinnitus OR tinnitus[MeSH]) The rest of the databases have been searched in a similar manner, using the same keywords. Filters of "10 years" and "English" language have been applied in all databases. Filters excluding non-clinical studies and pediatric studies have been applied whenever available. The whole search procedure and results have been evaluated by means of PRESS Evidence-Based Checklist [15].

Selection of Sources of Evidence
Studies obtained from the aforementioned search were reviewed independently by two authors. In that stage of analysis, the authors identified duplicates or multiple reports of the same study. Then, they screened the relevance of yielded studies to the set research questions by first examining the titles and abstracts of the yielded studies and then their full text. No disagreements between the two authors occurred at this stage.

Data Charting Process
Two reviewers screened full-text articles and produced a matrix of relevant data independently [16]. Ambiguities on data charting have been discussed and resolved by the senior authors.

Data Items
Extracted data items concerning methodological aspects and results of the included tinnitus studies: Results of this scoping review are presented in the form of comprehensive tables. Detailed qualitative analysis and critical appraisal of included studies can be found in the Discussion section.

Selection of Sources of Evidence
Thirty-four studies were included in this scoping review. The process of their selection is provided in detail in Figure 1.

Synthesis of Results and Critical Appraisal of Individual Resources of Evidence
Results of this scoping review are presented in the form of comprehensive tables. Detailed qualitative analysis and critical appraisal of included studies can be found in the Discussion section.

Selection of Sources of Evidence
Thirty-four studies were included in this scoping review. The process of their selection is provided in detail in Figure 1.

Characteristics of Sources of Evidence and Synthesis of Results
Characteristics of each study, such as authors' names or year of publication, along with extracted data with regards to the methods used, are presented in Tables 1-3. Inclusion criteria (hearing loss range, tinnitus duration, tinnitus as primary complaint) and types of participants' groups are presented in Table 1. Fitting methods and whether tinnitus-specific counseling has been provided can be found in Table 2, while results of each study with regards to their effect on tinnitus perception in Table 3. A comprehensive list of the methodological limitations of the clinical studies included in this scoping review may be found in Table 4.

Characteristics of Sources of Evidence and Synthesis of Results
Characteristics of each study, such as authors' names or year of publication, along with extracted data with regards to the methods used, are presented in Tables 1-3. Inclusion criteria (hearing loss range, tinnitus duration, tinnitus as primary complaint) and types of participants' groups are presented in Table 1. Fitting methods and whether tinnitusspecific counseling has been provided can be found in Table 2, while results of each study with regards to their effect on tinnitus perception in Table 3. A comprehensive list of the methodological limitations of the clinical studies included in this scoping review may be found in Table 4.     Differences between initial and final measurements differ in a statistically significant level between groups, in favor of the group using spectral masking.
Henry, 2015 [25] Compare the use of combination instruments for tinnitus management with and without the use of broadband noise produced from the instruments. Relative efficacy of conventional receiver-in-the-canal hearing aids (HA), the same hearing aids with a sound generator (HA1SG), and extended-wear, deep fit hearing aids (EWHA) TFI 1-3 weeks after fitting, 2 months after fitting, 4-5 months after fitting All devices appear to offer clinically significant improvement in the functional effects of tinnitus but no statistical significance before-after, among devices or among groups was found No statistically significant differences between groups-however both showed significant improvements Parazzini, 2011 [33] TRT with HA vs. TRT with sound generator THI, VAS 3, 6, 12 months No significant differences between HA and sound masker There were no significant differences in primary or additional variables between hearing aid types at either 3 or 6 months.

Searchfield, 2010 [38]
Effect of HAs combined with counseling compared to counseling only THQ 12 months post management THQ scores were reduced following intervention but only the HA group scores were found to differ significantly. The percentage improvement in total THQ score for the HA group (37%) was approximately twice that of counseling alone (13%).

Searchfield, 2016 [39]
Study 2: Determine masking preferences amongst participants. THI 2-2, 4 weeks There was a significant difference in THI change between the 3D and center masking Study 3: Provide preliminary evidence of the effectiveness of spatial masking with counseling relative to a therapy using TRT principles.

TFI, Tinnitus Severity Numeric Scale (TSNS) 2-4 weeks, 2 months
The 3D scores reduced slightly more (11.78) than TRT (6.89) but the treatment by time interaction was not statistically significant Shabana, 2018 [40] Effectiveness of counseling and using amplification and sound stimulation (Zen tones of fractal music) technology THI, TFI After counseling, 4 months after HAs fitting Statistically significant difference of THI scores between the post-counseling and following hearing aids fitting with or without Zen program, amount of improvement in the study group than in the control group except in THI emotional subscale score Table 3. Cont.

Shekhawat, 2013 [41]
To examine the effects of high frequency modification of the DSL(I/O) v5.0 prescriptive procedure on short-term tinnitus perception.

TFI Not determined
The higher the tinnitus pitch, the more the preferred real-ear output tended to match DSL(I/O) v5.0. For low-pitched tinnitus (<4 kHz) the preferred output tended to be lower than that of DSL(I/O) v5.0 across the entire frequency range.

Shekhawat, 2014 [42]
To assess if combination of tDCS and hearing aids may facilitate priming of the brain for sound therapy resulting in greater hearing aid benefit in a shorter period of time.
Tinnitus Case History Questionnaire (TCHQ), TFI, TSNS, THQ, VAS 3 and 6 months following hearing aid fitting The use of hearing aids led to a significant reduction in tinnitus handicap as measured with the TFI.
Shetty, 2019 [43] (1) to determine the effect of gain adjustment on tinnitus perception in low and high pitch tinnitus groups (2) to compare SNR 50 using NAL NL 2 and DSL (I/o) v 5.0 fitting formulae in high and low pitch tinnitus groups and (3) to compare tinnitus relief data and SNR-50 scores pre-and post-hearing aid use.

THI 30 days after fitting
In the low pitch tinnitus group, the significantly lesser gain adjustment was noted in DSL (I/o) v5 (0.5) than NAL-NL 2 (1.83). Similarly, for the high pitch tinnitus group, gain adjustment required was significantly less using DSL (I/o) v5 (1.16 dB) compared to NAL-NL 2 (5.6 dB). Additionally, speech perception in noise was unaffected by the adjusted gain at tinnitus pitch using either NAL NL 2 or DSL (I/o) v5 prescriptive formulae.

Strauss, 2017 [44]
Proof-of-concept that tinnitus distress can be reduced by the notch-induced lateral inhibition in NEST Tinnitus Questionnaire 12 (TQ12) 3 weeks post therapy Both TQ12 and τ factor improvement more prominent in notched environmental sound technology (NEST) group (both groups improved though) Sweetow, 2010 [45] To determine if the presence of various acoustic stimuli delivered through a hearing aid would reduce short-term tinnitus annoyance, and lower the subjective tinnitus handicap.

1, 3 and 6 months after fitting
The four fractal settings had similar median annoyance ratings, significantly better than the unaided (control) group TRQ: initial improvement, not consistent to 6 months. THI significant improvement at 6 months.

Discussion
This scoping review aimed to summarize current evidence on efficacy of HA fitting on tinnitus characteristics and patients' annoyance, along with the methodologies the relevant studies have used.

Primary Methodological Aspects of Included Studies
Half of the studies included in this review had less than 40 participants, only three had over 100, whereas there was only one large scale retrospective audit with 974 participants, comparing HAs and sound generators (Table 1). No sample size calculation is described in any of the studies. In addition, no power calculation was provided, neither ad hoc or post hoc.
A vast majority of studies (25 out of 34) did not set strict age criteria (Table 1). This fact might have an effect on the results since groups might not be adequately heterogenous. Although HL is far more common in older adults, three of the studies with certain age range as inclusion criterion set an upper limit (two of them 70 and one 80 years). Acceptance rates of HA are lower in younger adults and this could lead to lower representation of younger individuals compared to their actual proportion among tinnitus patients. On top of this, older adults might not be familiar with modern technologies recently implemented in HAs, like mobile applications and this could lead to selection bias and higher rates of drop outs, or sub-optimal use in terms of duration. None of the studies, even those who set a certain age range, took these potentially determining factors into consideration.
Only seven out of 35 studies (20.5%) were RCTs, a fact which highlights the relatively low level of evidence in the field. Eleven were case series with comparisons before and after treatment, whereas 13 were case control studies. Out of these, 5 compared different types of HA fitting, investigating the effect of additional or different features to tinnitus (open vs. classical fit, spectrally notched vs. unmodified, frequency transposition, addition of a sound generator or targeted counseling related to HA fitting), whereas 7 compared effect of HAs with various types of interventions (noise generators, maskers, TRT, notched environmental sound) ( Table 1). One of the studies examined the effect of HA aids in two different patient groups, with high and low tinnitus pitch [43]. Another study examined a "prototype" of 3-D masking [39].
Only eight out of 34 studies have clearly set tinnitus as a primary complaint as a certain inclusion criterion (Table 1). Hence identified as a drawback in tinnitus studies, a strict prerequisite of tinnitus as a primary complaint might not be absolutely relevant in studies targeting HA effectiveness in tinnitus. In patients with HL as a primary complaint, hearing aid fitting is indicated anyway. At the same time, a subgroup of patients with tinnitus as a primary complaint, also suffers from HL adequate to set a HA. In real conditions, there is a wide range of importance perception and level of annoyance correlated, between the two poles of HL and tinnitus. This means that patients belong to a wide spectrum between hearing loss and tinnitus as primary complaints-and all the shades in between.
This of course cannot eliminate the possibility of patients with HL as a primary complaint hence with a considerably bothersome tinnitus, even catastrophic, or vice versa: patients could mention tinnitus as a primary complaint and at the same time have important communication barriers due to severe hearing loss [51,52]. In conclusion, tinnitus as a primary complaint is of limited value in studies evaluating HA effectiveness in tinnitus, compared to studies targeting other interventions.
Related with this issue is the range of HL suitable for study inclusion. Although all but one studies reported that they targeted patients with HL requiring amplification, only six clearly determined specific thresholds per frequency as inclusion criteria ( Table 1). Eighteen of the studies provided some generic HL degrees (mild, moderate, severe), with a large heterogeneity, especially in regards to sever HL. Eleven studies did not have neither a broad determination of HL range, either with absence of any relevant information or with statements like "any type of HL" or "significant loss to warrant HA fitting". One study included only adults with hearing thresholds not requiring amplification [30]. Potential issues of this broad definitions are twofold. Different types of HAs are optimal for different types of HL. In most of the studies without specific HL inclusion criteria it can only be assumed that the relevant rules are applied, since a relevant statement is not made. Moreover, even if this was valid, it could lead to methodological discrepancies.
Less than half of the studies (14 out of 34) set clear inclusion criteria for tinnitus duration at the time of fitting (Table 1). This could potentially cause a difficulty to estimate effectiveness and clearly sets a potential selection bias. Absence of strict range of tinnitus onset is a draw of tinnitus literature in general, however in the evaluation of the HA effect, this might be even more influential, since there is a considerable proportion of patients who present total or intermittent elimination of tinnitus, which is not usually the case in other types of interventions. It is unknown though, whether a longer tinnitus duration might make more difficult tinnitus elimination or vice versa. In addition, none of the studies used tinnitus onset as a prognostic factor.

Hearing Aid Fitting Procedure
Only 12 out 34 studies gave a clear reference of the fitting formula used ( Table 2). The vast majority used NAL (5 used NAL-NL1 and 3 used NAL-NL2), whereas 3 used DSL. This parameter could be potentially important in regards to tinnitus suppression, on top hearing loss amplification, given that different formulas provide different gain patterns. Hence, it is interesting that no study presents a justification about the selected formula nor a predefined hypothesis that one might be more effective on tinnitus compared to another.
With regards to counseling, it is considered the cornerstone of most tinnitus treatments [53]. Majority of the studies included in this review do not describe what type of counseling was included in participants' workflow or whether they provided any counseling at all. Taking into account the confusion existing among counseling solely for HA fitting, counseling as part of TRT, long-term counseling through Cognitive Behavioral Therapy approaches and actual structured counseling targeting tinnitus, it is evident that the absence of this particular information in the included studies create a significant methodological limitation. Indeed, this is reflected in our results, as well. Only two studies present a clear description of the structured counseling they have conducted; Rocha et al. (2018) reported structured counseling using materials with videos and illustrations proposed by Siemens Audiology Solutions through counseling "Counseling Suite3.3". Newman and Sandridge (2012) also presented a detailed list of topics addressed during participants' education sessions. The rest of the studies reported some kind of counseling, mostly use components of TRT [23,26,33], while three of them do not provide any information [31,32,35]. Commercially available material which is integrated to specific hearing assistive devices (such as Widex Zen therapy) was also used in two studies, however it should be taken into account that this type of counseling deviates from the standard tinnitus counseling conducted by clinicians and its reproducibility is by default limited [40,46].

Tinnitus Assessment Methods
According to recent recommendations, outcome measures should be carefully chosen in tinnitus-related clinical studies, depending on the type of intervention. As for HAs, intrusiveness, sense of control, concentration and quality of sleep were among the dimensions that should be targeted by the outcome measures [52].
A sole outcome measure was used in 16 of the studies, whereas the rest used from two up to four outcome measures, with a moderate variance, since nine evaluation tools including validated questionnaires and scales were used in total. Most of the studies (19/34 or 55.9%) used Tinnitus Handicap Index (THI) as a primary outcome measure (Table 3), five Tinnitus Functional Index (TFI) [39][40][41][42]46], one Tinnitus Reaction Questionnaire (TRQ) [29] and two Tinnitus Handicap Questionnaire (THQ) [31,38]. Four studies used more than 2 scales [42,[45][46][47]. Scales were used as secondary outcome measures in nine studies and THQ, TQ and THI in two studies each (Table 3).
Although the selected outcome measures could be evaluated as satisfactory, given that they are both in line with the rest of the literature and with the recent recommendations, the main point in regards to evaluation is that these well-established tools are not designed for an intervention that has a far more binary nature compared to the rest. There is a considerable proportion of patients that experience total or close to total tinnitus elimination, at least during HA usage during the day. Questionnaires might globally reflect the change in quality of life, daily function or emotion due to these changes. The opposite could be valid for non-responders.
An additional limitation of currently used methods is that metrics targeting correlation between HA usage and tinnitus suppression in the time domain, usage duration and effect on tinnitus, comparison of HA usage individuals with and without tinnitus as well as some more trivial aspects like tinnitus relapse after HA removal and its effect are still missing. Ecological momentary assessment with use of mobile devices could be a very interesting research field towards this direction. Finally, the effect of total elimination is not well weighted though it might be the case that some patients respond very well and some not at all (binary response). Table 3 clearly shows that participants' follow up timeline ranged a lot across included studies. The majority of studies had a follow up time commonly used in the literature (13 of them had 3 months and 5 had 6 months). Only 1 study had a follow up period more than 12 months [28], whereas 3 had a follow up time less than 3 months (ranging from 3 weeks to 2 months) [39,43,44]. These latter are not considered adequate to draw conclusions in the context of tinnitus studies in general, however the effect of hearing aids is not as latent as in other interventions like sound therapy and CBT. On top of this, HA effect on tinnitus has two contradictory characteristics in the time domain: it is intermittent during the day, depending on whether the hearing aid is used or not and on the other hand, it is continuous in the largest time scale, since typically the effect, if present is not expected to substantially change during the period of HA usage. However, it is impressive that none of the studies report any evaluation of these parameters (effect during the day and stability of long-term effect). This latter could be an interesting research question for future studies.

Results
Before being able to interpret the results of clinical studies comparing two different tinnitus therapeutic modalities, one should take into account that different outcome measures may be more suitable for specific types of tinnitus treatment than others. In the context of COMiT'ID study, Hall et al. (2018) identified and reported the widest approved outcome measures for clinical trials of Sound-, Psychology-, and Pharmacology Based interventions for chronic subjective tinnitus. In the case of HA related clinical studies, COMiT'ID suggests that the minimum set of outcome measures should contain the following: "ability to ignore", "concentration", "quality of sleep" and "sense of control", while psychology-based ones should include endpoints such as "mood". Although, "what" each clinical study should include as endpoint is clearly stated, to date, no consensus on "how" this endpoint should be obtained and compared between treatment groups exists [11]. In clinical studies comparing a HA fitting with CBT-based therapy, a valid approach would be the inclusion of endpoints relevant to both treatments and the estimation of the effect of each treatment separately. This approach may enable comparisons between treatments in a more binary way, where each treatment has failed or succeeded at creating a significant tinnitus benefit with regards to its corresponding outcomes.
Studies included in this review seem to have adequately covered all the aforementioned core outcome domains through validated tools such as THI or through VAS. However, the same tools are used for all different types of treatment under evaluation, so the interpretation and generalization of their results should be done with caution.
The first important question that has direct research and clinical implications is whether there is evidence that HA fitting may have positive impact in tinnitus perception, distress or annoyance and if yes, what the possibility of improvement that should be expected after HA fitting is. Previous systematic review on HA fitting effect in terms of tinnitus benefit in adults with hearing loss and tinnitus concluded that only one clinical study was of adequate quality and thus no safe conclusion could be reached [54].
According to the outcome measures used, HA fitting-related improvement was up to 50%, while the proportion of participants showing at least some improvement ranged from 40% to 85% (Table 3).
As for the effectiveness, a clear trend can be identified among the case series included. All studies claimed a persistent significant improvement, regardless of the type of HA provided and the outcome measures used. Consistency of this finding among eleven studies is important, however it should be evaluated taking into account, on top of low methodological level and heterogeneity, that two of them did not conduct statistical analysis [21,31]. Moreover, one study shows that this improvement may elapse if patients abandon their HA.
In regards to superiority, only two case control studies compared hearing aid fitting with other types of interventions., namely TRT [33] and counseling [38]. In the latter study, HA fitting was claimed to be superior, however this obviously cannot be considered adequate evidence to draw conclusions. The rest of the case control studies compared either different types of devices and fitting techniques or HA fitting alone versus some kind of maskers or sound generators.
Whether HA fitting alone is inferior to HA fitting and sound therapy combined still remains unclear. Some studies have shown no statistically or clinically significant difference between HA and combined therapy [22,26], whereas others concluded that the overall benefit was significantly higher in those patients having undergone HA fitting and sound masking [24,25,40,44]. Finally, Jalilvand et al. (2015) showed that amplification alone was superior to sound therapy by means of a noise generator, while Newman and Sandridge (2012) compared sound therapy with broad band noise (BBN) and with Neuromonics Tinnitus Treatment (NTT) observing no statistically significant difference. This review has identified seven RCTs, out of which three evaluated the additional effect of sound generator, maskers and fitting techniques (frequency transposition) [22,26]. In none of these trials an additional effect of these interventions was concluded. Other RCTs compared HA fitting against gingko biloba and motivational interviewing [35,49], hence their findings were contradictory and the superiority could not be established.
To conclude, hearing aid fitting, itself, should be considered a valid tinnitus management approach for patients with HL. Current evidence implies that the size of its effect is clinically non-negligible. However, whether it should be combined with sound therapy or not needs further investigation through large scale longitudinal controlled studies. Future studies should overcome the heterogeneity of tinnitus assessment tools and outcome measures used, the different follow-up timelines and the conflicts of interest in those studies using commercially available tools so that they can lead safely to robust conclusions. Future studies with adequate study design and sample sizes, clearly set demographic inclusion criteria, clear ranges of hearing loss and tinnitus characteristics of the included subjects are warranted.

Conflicts of Interest:
The authors declare no conflict of interest.