Clinical Application and Psychometric Properties of a Norwegian Questionnaire for the Self-Assessment of Communication in Quiet and Adverse Conditions Using Two Revised APHAB Subscales

Abstract

Background:

Difficulty in following and understanding conversation in different daily life situations is a common complaint among persons with hearing loss. To the best of our knowledge, there is currently no published validated Norwegian questionnaire available that allows for a self-assessment of unaided communication ability in a population with hearing loss.

Purpose:

The aims of the present study were to investigate a questionnaire for the self-assessment of communication ability, examine the psychometric properties of this questionnaire, and explore how demographic variables such as degree of hearing loss, age, and sex influence response patterns.

Research Design:

A questionnaire based on the subscales of the Norwegian translation of the Abbreviated Profile of Hearing Aid Benefit was applied to a group of hearing aid users and normal-hearing controls.

Study Sample:

A total of 108 patients with bilateral hearing loss, and 101 controls with self-reported normal hearing.

Data Collection and Analysis:

The psychometric properties were evaluated. Associations and differences between outcome scores and descriptive variables were examined. A regression analysis was performed to investigate whether descriptive variables could predict outcome.

Results:

The measures of reliability suggest that the questionnaire has satisfactory psychometric properties, with the outcome of the questionnaire correlating to hearing loss severity, thus indicating that the concurrent validity of the questionnaire is good.

Conclusions:

The findings indicate that the proposed questionnaire is a valid measure of self-assessed communication ability in both quiet and adverse listening conditions in participants with and without hearing loss.

Key Words

communication - outcome - self-assessment

INTRODUCTION

The self-assessment of communication difficulties can provide valuable information to the clinician on how a hearing loss influences the client’s everyday life. This information is important be able to improve the rehabilitation quality and increase hearing aid uptake ([Beck, 2000]; [Knudsen et al, 2010]; [Saunders et al, 2012]). The identification and understanding of which variables are related to communication difficulties in persons with hearing loss help to provide knowledge on how to improve rehabilitation outcomes ([Hickson et al, 2010]). However, when using self-assessment questionnaires in clinical practice, it is important that the administration time is short, and that the analysis that the clinician has to do is both straightforward and easy to do ([Bagatto et al, 2011]). Patient-reported outcome measures are developed to directly capture the patient’s voice regarding his/her symptoms and functional status, but the ability to do so effectively relies on using an appropriate patient-reported outcome measure that addresses specific symptoms and functions ([Marshall et al, 2006]).

Difficulty with understanding conversation in different daily life situations is a common complaint among persons with hearing loss ([Gates and Mills, 2005]); such situations include those where individuals with hearing loss are known to have problems, such as during adverse listening conditions ([Humes et al, 2013]). These problems may lead to participation restrictions and activity limitations, which in turn may lead to withdrawal and a resulting poorer quality of life ([Stark and Hickson, 2004]; [Helvik et al, 2006]; [Öberg et al, 2007]) Hence, a main goal within audiological rehabilitation is to improve the hearing-impaired person’s perception of speech, for instance, through the use of amplification provided with hearing aids. To systematically measure the hearing-impaired person’s experience with communication before a hearing aid fitting should therefore be of primary interest ([Cox and Alexander, 1995]). This would be valuable insofar as involving the patient in the rehabilitative process and as a method for the verification of hearing aid fittings, not to mention the effect on self-reported difficulties with communication in both quiet and adverse listening situations. In addition, studies on the effect of hearing loss on general health and quality of life should include self-reported problems with communication and the effect of hearing aids on these problems. Such studies require a range of questionnaires to be answered. Hence, to ensure that the total burden of answering such surveys is tolerable, the size and complexity of each questionnaire should be critically evaluated ([Bowling, 2005]).

To the best of our knowledge, there is currently no published validated Norwegian questionnaire available that allows for specifically self-assessing communication ability in quiet and adverse listening situations in a hearing-impaired population. Consequently, there is need for a brief questionnaire for the assessment of self-perceived difficulties with communication. In the present study, we used two revised subscales of the Abbreviated Profile of Hearing Aid Benefit (APHAB; [Cox and Alexander, 1995]). Our aims were to examine the psychometric properties of this questionnaire for the self-assessment of communication, and explore how demographic variables such as degree of hearing loss, age, and sex all influence the questionnaire’s score.

The present article will focus on the self-assessment of unaided listening scenarios. The suggested questionnaire also has the potential to be the outcome measure for hearing aid fittings, which will be investigated in future studies.

METHOD

Participants and Procedures

Two groups of individuals participated in the current study: one group with patients diagnosed with hearing loss and one with control participants with self-reported normal hearing. Both groups completed a questionnaire for the self-assessment of communication in quiet and adverse listening situations.

A total of 301 patients with hearing loss were initially invited to participate in the study, with 158 returning the completed questionnaire, thereby yielding a response rate of 52.5%. Patients with hearing loss were adults (aged ≥18 years) referred for a hearing aid fitting at Haukeland University Hospital in western Norway. Both first-time hearing aid users and experienced users referred for hearing aid renewal were included, whereas 23 patients with unilateral hearing loss were excluded. Unilateral hearing loss was defined by a better ear with pure-tone hearing thresholds equal to or better than 25 dB HL at octave frequencies of 0.125–8 kHz. Also, only patients with a hearing loss worse than 25 dB HL averaged over pure-tone frequencies of 0.5, 1, 2, and 4 kHz were included, excluding 27 patients referred for hearing aid fittings. Therefore, the results for patients with hearing loss presented in this article are based on the responses of 108 patients with bilateral hearing loss. Additionally, a group of adult participants (aged >18 years) with self-reported normal hearing was recruited to serve as controls (n = 101). These were recruited from among hospital staff, and from students and employees at a local university college. The hearing of the participants in the control group was not tested, with [Table 1] showing the demographic data for the normal-hearing controls. The Norwegian Regional Committees for Medical and Health Research Ethics provided advance approval for the project (project reference: 2013/1302).

Table 1 Descriptive Information for Patients with Hearing Loss and the Normal-Hearing Controls

	Hearing Impaired (n = 108)	Controls (n = 101)
Age (mean ± SD)	62 ± 11	29 ± 13
Males/females	62/46	42/59
Type hearing loss
Sensorineural	99
Mixed	9
Diagnosis
Presbyacusis	33
Noise induced	30
Hereditary	19
Congenital	12
Meningitis	1
Sudden hearing loss	2
Middle ear disease	7
Other trauma	1
Unknown	3
PTA
Best ear (mean ± SD)	42 ± 15
First-time users	45
Duration of hearing loss (mean ± SD) (n = 85)	26 ± 18

For patients with hearing loss, demographic data, as well as information regarding diagnosis and pure-tone audiograms, were collected from medical records. The authors inspected all records of the patients to control for the correct diagnosis and audiogram classifications, and any discrepancies between audiograms and a recorded diagnosis were revised. Based on their most recent audiogram, patients were grouped by better-ear pure-tone average (PTA; frequencies 0.5, 1, 2, and 4 kHz) as having slight/mild hearing loss (26–40 dB HL), moderate hearing loss (41–60 dB HL), or severe-to-profound hearing loss (61 dB HL or worse). Experienced hearing aid users (n = 45) were currently fitted with a variety of hearing aids: either in-the-ear (11.1%), behind-the-ear (80%), or BICROS systems (8.9%). A similar distribution of fittings was planned for first-time hearing aid users, with 19% in-the-ear, 76.2% behind-the-ear, and 4.8% BICROS systems. All hearing aids were digital, with descriptive information on the patients shown in [Table 1].

All participants were asked to answer the questionnaire describing unaided hearing scenarios. Patients completed the questionnaire by pen and paper at home, and the questionnaires were delivered and returned by regular mail. Questionnaires were typically mailed to patients three to four weeks before the hearing aid fitting, and they were encouraged to return the questionnaires within one week. Thus, data were collected within one month before the hearing aid fitting. The normal-hearing controls answered the questionnaire by pen and paper at their workstations, and the questionnaire was collected on the same day by a project representative carrying a sealed container.

Design of Questionnaire

This study used a questionnaire with 12 candidate items given in Norwegian on self-perceived difficulties with communication in different listening scenarios extracted from the Norwegian translation of the APHAB ([Cox and Alexander, 1995]), which served as a basis for the questionnaire. Because the aim of this study was to obtain a tool for the self-assessment of communication in quiet and adverse listening situations specifically, the 12 candidate items were extracted verbatim from the subscales “ease of communication” and “background noise” in the Norwegian version of the APHAB questionnaire ([Cox and Alexander, 1995]). The selection was made by a group of health-care professionals experienced in hearing aid fitting, following a discussion of previous experiences with existing questionnaires and the most common complaints among patients. Moreover, the use of APHAB subscales as a basis for the questionnaire was motivated by the potential of including additional questions from the APHAB if a more extensive questionnaire needed to be validated in later studies. Based on feedback from patients, and from their own experiences with questionnaire scoring, the group suggested that the seven-point Likert scale of the APHAB should be revised for ease of administration, completion, and scoring. This was done in order for respondents to perceive the questionnaire as quick and easy to use, especially when administered in combination with other questionnaires ([Preston and Colman, 2000]). Therefore, the response scale of each item in the questionnaire was revised using a four-point Likert format: “always/most of the time” (a score of 1), “half the time” (a score of 2), “sometimes” (a score of 3), and “never/very rarely” (a score of 4). A low score for each item translates to a frequent experience of difficulties with communication in such a situation. The questionnaire with the 12 candidate items used in this study is shown in [Table 2], together with the revised Likert scale.

Table 2 The Questionnaire with the 12 Candidate Items as Presented to Participants (English Translation)

1	Når jeg snakker med en ekspeditør i en dagligvareforretning med mange mennesker, er det vanskelig å følge med i samtalen. (When I am in a crowded grocery store, talking with the cashier, I have trouble following the conversation.)	1	2	3	4
2	Jeg har vanskeligheter med å høre, når jeg prater med en i familien min hjemme. (I have difficulty hearing a conversation when I'm with one of my family at home.)	1	2	3	4
3	Når jeg hører nyheter på bilradioen og andre prater samtidig, har jeg vanskeligheter med å høre nyhetene. (When I am listening to the news on the car radio, and family members are talking, I have trouble hearing the news.)	1	2	3	4
4	Når jeg sitter ved middagsbordet med flere mennesker og jeg forsøker å føre en samtale med en person, er det vanskelig å forstå hva som blir sagt. (When I am at the dinner table with several people, and am trying to have a conversation with one person, understanding speech is difficult.)	1	2	3	4
5	Når jeg er på et lite kontor og intervjuer noen, eller besvarer spørsmål, har jeg vanskeligheter med å følge samtalen. (When I am in a small office, interviewing or answering questions, I have difficulty following the conversation.)	1	2	3	4
6	Når jeg snakker med en venn, har jeg vanskeligheter med på forstå hva han/hun sier. (When I am having a quiet conversation with a friend, I have difficulty understanding.)	1	2	3	4
7	Når en person snakker til en liten gruppe og alle lytter, må jeg anstrenge meg for å forstå hva som blir sagt. (When a speaker is addressing a small group, and everyone is listening quietly, I have to strain to understand.)	1	2	3	4
8	Når jeg har en samtale med legen min i et undersøkelsesrom, er det vanskelig å følge med i samtalen. (When I'm in a quiet conversation with my doctor in an examination room, it is hard to follow the conversation.)	1	2	3	4
9	Det er vanskelig å følge med i samtalen når flere personer snakker samtidig. (It is difficult to understand conversations when several people are talking at the same time.)	1	2	3	4
10	Jeg har vansker med å kommunisere med andre når vi er i en gruppe. (I have trouble communicating when we are in a group.)	1	2	3	4
11	Jeg må be den som snakker om å gjenta det som ble sagt også i rolige Rom. (I have to ask people to repeat themselves in one-on-one conversations in a quiet room.)	1	2	3	4
12	Jeg har vanskeligheter med å forstå hva andre sier når et klimaanlegg eller vifte er på. (I have trouble understanding others when an air conditioner or fan is on.)	1	2	3	4

Notes: Vennligst velg det svar som kommer nærmest din daglige situasjon. Hvis du ikke har opplevd den aktuelle situasjon, tenk deg hvordan du ville oppleve en tilsvarende situasjon. (Please circle the answer that is closest to your everyday experience. If you have not experienced the situation that we describe, try to imagine how you would experience a similar situation.) 1 = Alltid/nesten alltid (Always/most of the time); 2 = Halvparten av tiden (Half of the time); 3 = Av og til (Sometimes); 4 = Sjelden/aldri (Never/very rarely).

Statistical Analysis

A statistical analysis was performed using the statistical program package IBM SPSS (Version 23.0 for Windows; IBM Corp. Armonk, NY). An exploratory factor analysis (principal component analysis) with Kaiser normalization was performed on results from the 12 candidate items to help explore the factor structure of the proposed questionnaire.

Internal consistency for the total questionnaire was tested using Cronbach’s α, which was also recalculated for the remaining scale following the removal of each individual item. Further testing of the psychometric properties of the scale included a calculation of the inter-item correlations, as well as the corrected item-total correlations. Pearson product-moment correlation coefficients (r) were used to investigate associations among hearing loss variables, age, single items, and total scale score.

Differences in single-item scores, total scale scores, subscale scores, age, and hearing loss variables for different demographic groups (male/female, first-time users/experienced users, controls/hearing aid users) were investigated using independent sample t tests with Bonferroni correction.

Stepwise linear regressions were calculated to help explore the associations among demographic variables and hearing loss characteristics (independent variables), in addition to the mean sum score of the entire questionnaire (dependent variable). Also, to test whether the outcome was different between groups of hearing impairment and the group with normal hearing, one-way analysis of variance and Bonferroni-corrected post hoc tests were carried out.

RESULTS

Mean scores with standard deviations (SDs) for single items, subscales, and total scale for groups of patients with hearing loss and normal-hearing controls are shown in [Table 3], whereas [Figure 1] shows the distribution of responders at different total scale scores for patients and normal-hearing controls.

Table 3 Mean Scores (SD) in Groups of Participants, for Single Items, Total Scale, and Subscales “Quiet” and “Adverse”

					All Responders (n = 209)
Item	Slight Moderate (n = 63)	Moderate (n = 37)	Severe Profound (n = 8)	Controls (n = 101)	Mean (SD)	Range
1	2.44 (0.96)	2.27 (0.84)	1.87 (1.25)	3.60 (0.61)	2.94 (1.02)	1–4
2	2.87 (0.66)	2.41 (0.93)	2.25 (1.40)	3.84 (0.37)	3.22 (0.89)	1–4
3	2.11 (0.97)	1.89 (0.88)	1.50 (0.93)	3.08 (0.77)	2.51 (1.02)	1–4
4	2.30 (0.78)	1.95 (0.82)	1.75 (0.87)	3.51 (0.60)	2.79 (0.99)	1–4
5	3.11 (0.79)	2.70 (0.99)	2.25 (1.28)	3.96 (0.20)	3.41 (0.86)	1–4
6	3.16 (0.79)	2.84 (0.93)	2.63 (1.19)	3.92 (0.28)	3.44 (0.81)	1–4
7	2.52 (0.93)	2.27 (1.02)	1.75 (1.17)	3.86 (0.35)	3.08 (1.06)	1–4
8	3.25 (0.92)	2.76 (1.01)	2.13 (1.34)	3.99 (0.10)	3.47 (0.90)	1–4
9	1.98 (0.94)	1.76 (0.75)	1.38 (0.744)	2.92 (0.80)	2.36 (0.99)	1–4
10	2.44 (0.88)	2.03 (0.96)	1.75 (1.04)	3.63 (0.56)	2.91 (1.04)	1–4
11	2.71 (0.83)	2.41 (0.93)	2.38 (1.30)	3.76 (0.43)	3.14 (0.93)	1–4
12	2.52 (0.95)	2.41 (0.93)	1.75 (1.04)	3.62 (0.51)	3.00 (0.98)	1–4
Total scale	2.62 (0.62)	2.29 (0.72)	1.95 (0.92)	3.64 (0.26)	3.03 (0.80)	1–4
Quiet	2.88 (0.61)	2.54 (0.81)	2.16 (1.07)	3.84 (0.17)	3.26 (0.79)	1–4
Adverse	2.26 (0.74)	1.98 (0.69)	1.65 (0.84)	3.35 (0.49)	2.70 (0.89)	1–4

Figure 1 Number of responders at different mean sum scores for patients with hearing loss and normal-hearing controls.

Psychometric Properties

Several analyses were made to test the psychometric properties of the questionnaire. Inter-item correlations show that each item of the questionnaire correlated to all other items (r ≥ 0.563, p ≤ 0.01). As expected, all correlations were positive, since all items have the same direction in response values.

In patients with hearing loss, a Kaiser–Meyer–Olkin Test resulted in a value of 0.91, thereby confirming a sufficient sample size for factor analysis ([Kaiser, 1970]; [1974]). The exploratory principal component analysis identified two factors formed by responses to the 12 candidate items in patients with hearing loss ([Table 4]), with these factors accounting for 67.2% of the total variance. A Varimax rotation was applied, and coefficients with absolute values lower than 0.5 were suppressed. The two factors were formed similarly to the APHAB subscales “ease of communication” and “background noise” except for item 12, which was extracted from the APHAB subscale “background noise,” but which is part of the factor otherwise formed by items from the “ease of communication” subscale. Thus, the questionnaire seems to consist of two subscales “communication in quiet conditions” (items 2, 5, 6, 7, 8, 11, and 12) and “communication in adverse conditions” (items 1, 3, 4, 9, and 10), corresponding to the APHAB subscales “ease of communication” and “background noise,” with the exception of the placement of item 12.

Table 4 Distribution of Response to Items and Factor Analysis for Patients with Hearing Loss and for Control Participants

													Factor Analysis
Distribution over Response Alternatives for Groups of Hearing Loss Severity (Presented as Percentage of Responders)													Patients (n = 108)		Controls (n = 101)		All (n = 209)
													Rotated Component Matrix*		Rotated Component Matrix**		Component Matrix***
	Mild/Slight				Moderate				Severe/Profound				Component		Component		Component
	1	2	3	4	1	2	3	4	1	2	3	4	1	2	1	2	1
1	22.2	22.2	44.4	11.1	24.3	24.3	51.4	0	62.5	0	25	12.5	—	0.63	—	0.64	0.82
2	1.6	23.8	60.3	14.3	18.9	32.4	37.8	10.8	50	0	25	25	0.67		0.59	—	0.83
3	33.3	30.2	28.6	7.9	40.5	32.4	24.3	2.7	75	0	25	0	—	0.82	—	0.76	0.77
4	17.5	36.5	44.4	1.6	35.1	35.1	29.7	0	50	25	25	0	—	0.75	—	0.73	0.85
5	3.2	15.9	47.6	33.3	13.5	24.3	40.5	21.6	37.5	25.0	12.5	25.0	0.83	—	—	—	0.86
6	3.2	14.3	46.0	36.5	8.1	27	37.8	27	25	12.5	37.5	25	0.80	—	0.63	—	0.84
7	17.5	25.4	44.4	12.7	27	32.4	27	13.5	62.5	12.5	12.5	12.5	0.64	—	0.56	—	0.87
8	6.3	12.7	30.2	50.8	13.5	24.3	35.1	27	50	12.5	12.5	25	0.87	—	—	—	0.80
9	38.1	31.7	23.8	6.3	43.2	37.8	18.9	0	75	12.5	12.5	0	—	0.86	—	0.83	0.77
10	12.7	42.9	31.7	12.7	35.1	35.1	21.6	8.1	50	37.5	12.5	0	—	0.72	—	0.51	0.87
11	9.5	23.8	52.4	14.3	16.2	40.5	29.7	13.5	37.5	12.5	25	25	0.70	—	—	—	0.88
12	20.6	17.5	50.8	11.1	18.9	32.4	37.8	10.8	50	37.5	0	12.5	0.60	—	—	—	0.81

Notes: *Principal component factor analysis (Varimax rotation). Kaiser–Meyer–Olkin measure of sampling adequacy: 0.913. Bartlett's test of sphericity: χ² = 841.08; degrees of freedom: 66. p < 0.001. **Principal component factor analysis (Varimax rotation). Kaiser–Meyer–Olkin measure of sampling adequacy: 0.734. Bartlett's test of sphericity: χ² = 234.36; degrees of freedom: 66. p < 0.001. ***Principal component factor analysis (Varimax rotation). Kaiser-Meyer-Olkin measure of sampling adequacy: 0.953. Bartlett's test of sphericity: χ² = 2,271.24; degrees of freedom: 66. p < 0.001. 1 = Always/most of the time; 2 = Half of the time; 3 = Sometimes; 4 = Never/very rarely.

In the control group, the same distribution of items in the two factors was seen, though items 5, 8, 11, and 12 failed to load 0.50 or greater on a single factor. A factor analysis of responses from both patients and controls combined identified a single factor only.

The high Cronbach’s α (0.959) for all items suggests that the items measure the same basic construct. To further assess each item’s contribution to the internal consistency of the scale, each item was deleted, which in turn was followed by a recalculation of Cronbach’s α. The deletion of any item did not increase the Cronbach’s α of the total scale (0.953–0.957), therefore suggesting that each item is a valuable contribution to the questionnaire. Ranging from 0.735 to 0.851, the corrected item-total correlations also show that each item contributes to the total scale.

Items in the subscale “communication in quiet situations” yielded a Cronbach’s α of 0.943. If the item was deleted, the Cronbach’s α ranged from 0.931 to 0.940, with the corrected item-total correlations ranging from 0.753 to 0.855. For items in the subscale, “communication in adverse situations,” a Cronbach’s α of 0.922 was found, while the corrected item-total correlation and the Cronbach’s α if the item was deleted ranged from 0.765 to 0.825 and 0.903 to 0.911, respectively. As a result, both subscales show satisfactory psychometric properties when analyzed separately.

Based on these analyses, the averages of all candidate items were included in further analyses.

The 90th percentile of scores in normal-hearing controls was 3.24 points. This score was used to calculate the sensitivity and specificity of the questionnaire in regard to separating patients from controls. The sensitivity was found to be 81%, with a specificity of 91%.

Associations and Differences

[Table 3] shows mean and range scores for single items for all responders pooled into four groups: slight/mild hearing loss (26–40 dB HL), moderate hearing loss (41–60 dB HL), severe-to-profound hearing loss (61 dB HL or worse), and normal hearing. To avoid having groups with very low numbers of participants, those with a hearing loss worse than 61 dB formed a single group. It can be seen that the mean score of the total questionnaire, and for both subscales, decreases by hearing loss severity, indicating that those with worse hearing experience more frequent difficulties with communication, both in quiet and adverse listening situations.

To test if the responses were significantly different for these groups, a one-way analysis of variance was made, and it should be noted that the unequal size of groups could affect the power of this analysis. A significant difference in the mean total scale score between groups was found [F _(3.198) = 92.65, p = 0.000]. Bonferroni-corrected post hoc t tests revealed that the normal-hearing controls had significantly (p < 0.01) higher scores than all three groups of hearing-impaired participants, while participants with slight-to-mild hearing loss had significantly (p < 0.01) higher scores than the two groups with poorer hearing loss. No significant (p > 0.05) difference in scores was seen between the groups with moderate and severe-to-profound hearing loss. These findings also suggest that the concurrent validity of the questionnaire is good.

Among patients with hearing loss, differences between males and females, and between first-time users and experienced users, were investigated. Total scale score, subscale scores, age, worse-ear PTA, better-ear PTA, and the duration of hearing loss were included in independent-sample t tests with a Bonferroni correction for multiple comparisons. No significant differences were found in any variable between first-time users and experienced hearing aid users, or between male and female patients with hearing loss.

For the patients with hearing loss, the variables age, better-ear PTA, worse-ear PTA, duration of hearing loss, mean score on the total scale, and subscales were all included in a correlation analysis using Pearson product-moment correlation coefficients, with the results presented in [Table 5]. The mean score of the questionnaire is shown to correlate significantly to better-ear PTA, worse-ear PTA, and duration of hearing loss. No significant correlation is seen between age and mean score, while hearing loss duration correlates to both PTA values, hence indicating that an increase in hearing loss duration is associated with worse hearing.

Table 5 Pearson’s Correlation Coefficients between Examined Variables in Patients with Hearing Loss (n = 108)

	Age	Duration	Best PTA	Worst PTA	Quiet	Adverse
Duration	−0.17
Best PTA	−0.01	0.35*
Worst PTA	−0.13	0.33*	0.76*
Quiet	0.02	−0.19	−0.23**	−0.21**
Adverse	0.13	−0.22**	−0.20**	−0.23**	0.72**
Total scale	0.07	−0.23	−0.23**	−0.24**	0.95*	0.91*

Notes: *Correlation is significant at the 0.01 level (two-tailed); **correlation is significant at the 0.05 level (two-tailed).

In the control group, no significant (p > 0.05) difference was seen in the mean score of the total scale between male (3.67) and female (3.66) responders, and there was no significant correlation between mean score and age. Thus, an increasing age does not seem to cause more frequent reports of difficulties with communication in quiet and adverse listening conditions in persons who self-report normal hearing.

Prediction of Outcome

For the patients with hearing loss, hearing loss variables that showed significant correlations in the correlation analysis were selected, and a stepwise linear regression was performed using these variables (better-ear PTA, worse-ear PTA, and duration of hearing loss) as independent values, and the mean score of the total scale as the dependent value. PTA in the worse ear was identified as a single predictor of the mean score of the questionnaire (R ² = 0.072, p = 0.01, β = −0.290). The β values for the variables better-ear PTA and duration of hearing loss were −0.159 and −0.173, respectively.

DISCUSSION

The proposed questionnaire is a validated Norwegian translation of revised questions from the “ease of communication” and “background noise” subscales of the APHAB, with a modified Likert format. Testing in a group of patients with hearing loss and controls with self-reported normal hearing, followed by measures of reliability, suggest that the questionnaire has satisfactory properties.

Questionnaire Design, Psychometric Properties, and Limitations

This study describes a questionnaire with 12 candidate items given in Norwegian on self-perceived difficulties with communication in quiet and adverse listening conditions. The concurrent validity of the questionnaire is good; analyses of the psychometric properties of the questionnaire show that all items measure the same basic construct, and that all single items contribute to the total scale. Therefore, we suggest that all 12 candidate questions are used when administering this questionnaire, as they seem to independently contribute important information regarding self-reported difficulties with communication. The findings indicate that the proposed questionnaire is a valid measure of self-assessed communication ability in quiet and adverse listening conditions in participants both with and without hearing loss.

In clinical practice, it is important to reduce administration time and facilitate analysis when using self-assessment questionnaires ([Bagatto et al, 2011]); thus, the initial number of candidate questions was low. Another reason was that the proposed questionnaire consists of revised subscales from APHAB ([Cox and Alexander, 1995]). Including a number of novel candidate questions to be part of the factor analysis could have possibly provided a more thorough tool. On the other hand, a main aim of the study was to ensure that the final proposed questionnaire was brief and handy for day-to-day clinical uses, as well as a potential part of larger surveys.

A larger number of patients with hearing loss could have been included, which would have provided valuable information on the distribution of scores over different grades of hearing loss severity. This is especially true for our group of participants with severe-to-profound hearing loss, in which only eight participants were included. It may also be discussed how accurately experienced hearing aid users are able to evaluate their hearing in unaided scenarios. Nevertheless, the lack of significant differences in the scores of first-time users and the experienced hearing aid users suggests that the experienced hearing aid users are able to remember or imagine unaided performance in the various scenarios. This could also be influenced by the daily use of hearing aids in the experienced users, but this information was not collected in the present study. Future studies should include this variable.

We did not test the hearing of our control group, but instead relied on self-reported normal hearing. This is a possible weakness in the data collected from this group. By contrast, it allows us to collect self-assessed difficulties with communication from a group of participants who self-assess their hearing to be normal. Consequently, we can compare these data to those from a group of participants we know to be hearing impaired. In the process of developing a tool for self-assessment in patients who think they need a hearing aid, this seems relevant. Still, we acknowledge the value of incorporating both pure-tone audiometry and speech-in-noise tests in future studies of this kind, even though previous studies have found only weak correlations between speech-in-noise tests and the APHAB subscales ([Cox et al, 2003]). In addition, there was a significant difference in age between the hearing aid users and the control group. Despite the lack of correlation between age and outcome in our data, future studies should attempt to match patients and controls regarding age.

Our decision to modify the Likert-scale format of the original APHAB affects our ability to compare scores to other studies that have used the APHAB or its subscales ([Cox and Alexander, 1995]; [Cox et al, 2003]; [Johnson et al, 2010]; [Kam et al, 2011]; [Löhler et al, 2016]). The revision of the Likert-scale format was motivated by the desire to reduce response burden and administration time. Also, previous studies on the number of points in Likert-type scales show little or no effect on psychometric properties from increasing the number of points on the scale beyond four points ([Lozano et al, 2008]; [Lee and Paek, 2014]). Additionally, we use the mean sum score of the questionnaire as an outcome measure, instead of the percentage values extracted from the original APHAB. To test the influence of the number of possible sum scores on our results, we have reduced the mean sum score to 4 and 8 Ntiles, and these show correlations of 0.95 and 0.98 to the mean sum score of the scale. For this reason, splitting the scale into additional response alternatives does not seem to provide any additional information to the mean sum score. Thus, as additional response alternatives do not seem to affect the mean sum score, we suggest using a four-point Likert scale. In line with previous studies ([Lozano et al, 2008]; [Lee and Paek, 2014]), and as shown by analysis of responses to the proposed questionnaire, this provides satisfactory psychometric properties while reducing response burden. Even so, some similarities and differences with previous studies seem relevant to discuss. The APHAB originally consists of four subscales with a total of 24 questions ([Cox and Alexander, 1995]). We have used questions from the two subscales “ease of communication” and “background noise.” Some previous studies have failed to identify the proposed APHAB subscales after a factor analysis, with inconsistent findings between studies ([Kam et al, 2011]). In patients with hearing loss, we seem to reproduce the subscale structure from the APHAB, except for the placement of item 12: “I have trouble understanding others when an air conditioner or fan is on.” We suggest that this may be due to cultural and/or geographical differences, given that the use of air conditioners and fans is not as common in Norway. Patients may therefore have trouble imagining how they would perceive such a listening scenario. Or, they may not consider this type of background noise as troublesome in the same way as those described in the questions forming the “adverse conditions” subscale.

Differences and Associations

The proposed questionnaire aims to assess the difficulty with understanding conversation in different daily life situations, which is a common complaint among persons with hearing loss ([Gates and Mills, 2005]). To help improve rehabilitation outcomes, it is important to identify and understand which variables are related to communication difficulties in persons with hearing loss ([Hickson et al, 2010]). This information provides knowledge on how to improve the rehabilitation quality and increase hearing aid uptake ([Beck, 2000]; [Knudsen et al, 2010]; [Saunders et al, 2012]). In the current study, a lower score (more frequent difficulties with communication) is associated with worse hearing. This is shown both by the correlation between the total scale score and PTA, as well as in the differences between groups of hearing loss severity and those with normal hearing. Hence, it seems that worse hearing is associated with more frequent problems with verbal communication, and that the proposed questionnaire is able to measure this relationship in a systematic manner. Although there is a significant correlation between PTA in the worse ear and total scale score, the correlation is weak. This is in line with previous studies using other questionnaires to assess similar constructs ([Cox et al, 2003]; [Kam et al, 2011]). Thus, it may be that the difference between groups of hearing loss severity, and between the patients with hearing loss and normal-hearing controls, are primarily explained by variables not included in this study, and that self-reported difficulties with communication could offer information on the effect of hearing loss on communication in quiet and adverse conditions that cannot be measured with pure-tone audiometry. These findings resemble those of previous applications of the APHAB questionnaire ([Cox et al, 2003]; [Kam et al, 2011]), who found that self-reported difficulties with daily life hearing situations is only partially explained by the most frequently used audiometric and psychoacoustic measurements. In a recent study of unaided APHAB scores and the influence of hearing-loss characteristics, [Löhler et al (2016)] also demonstrated weak correlations to hearing loss severity. They suggest that the lack of strong correlations between subjective hearing difficulties and the measured hearing loss may be due to a large variability of how well individuals are able to compensate for their hearing loss in different listening scenarios. Our future studies will also investigate the effect of hearing aids on the outcome of this questionnaire, and how it performs as a measure of hearing aid satisfaction.

Possible Applications of Data from the Normal-Hearing Controls

Data from the normal-hearing control group have many potential applications. It could be used as a measure of how many problems with communication a particular patient with hearing loss experiences compared to group of normal-hearing participants when unaided to determine severity. We also believe it could be of value when evaluating the outcome of hearing aid fittings. Although providing the possibility for measuring aided/unaided scores, the questionnaire could also be useful for persons with hearing loss to compare his/her aided problems with communication to that of those with normal hearing. As seen in [Table 3] and [Figure 1], most normal-hearing controls have scores higher than three points, thereby suggesting a potential cutoff point for how many problems a normal-hearing group experiences in various listening situations. The proposed questionnaire will be applied in a group of hearing aid users as an outcome measure after hearing aid fittings in our future studies. This might offer valuable information to the process of ensuring that a hearing aid fitting is optimal.

CONCLUSION

A revised and shortened version of the Norwegian translation of APHAB for the self-assessment of difficulties with communication in quiet and adverse listening conditions is proposed. The findings indicate that the proposed questionnaire is a valid measure of self-assessed communication ability in quiet and adverse listening conditions in participants both with and without hearing loss. The questionnaire has potential as a stand-alone questionnaire, with subscales separately describing problems with communication in quiet and adverse situations. Yet, further testing of the questionnaire is required to investigate its potential as an outcome measure for hearing aid fittings.

Abbreviations

APHAB: Abbreviated Profile of Hearing Aid Benefit

PTA: pure-tone average

Die Autoren geben an, dass kein Interessenkonflikt besteht.

• REFERENCES

• Bagatto MP, Moodie ST, Seewald RC, Bartlett DJ, Scollie SD. 2011; A critical review of audiological outcome measures for infants and children. Trends Amplif 15 (01) 23-33
• Beck LB. 2000; The role of outcomes data in health-care resource allocation. Ear Hear 21 (04) (Suppl) 89S-96S
• Bowling A. 2005; Just one question: if one question works, why ask several?. J Epidemiol Community Health 59 (05) 342-345
• Cox RM, Alexander GC. 1995; The abbreviated profile of hearing aid benefit. Ear Hear 16 (02) 176-186
• Cox RM, Alexander GC, Gray GA. 2003; Audiometric correlates of the unaided APHAB. J Am Acad Audiol 14 (07) 361-371
• Gates GA, Mills JH. 2005; Presbycusis. Lancet 366 (9491) 1111-1120
• Helvik AS, Jacobsen G, Hallberg LR. 2006; Psychological well-being of adults with acquired hearing impairment. Disabil Rehabil 28 (09) 535-545
• Hickson L, Clutterbuck S, Khan A. 2010; Factors associated with hearing aid fitting outcomes on the IOI-HA. Int J Audiol 49 (08) 586-595
• Humes LE, Kidd GR, Lentz JJ. 2013; Auditory and cognitive factors underlying individual differences in aided speech-understanding among older adults. Front Syst Neurosci 7: 55
• Johnson JA, Cox RM, Alexander GC. 2010; Development of APHAB norms for WDRC hearing aids and comparisons with original norms. Ear Hear 31 (01) 47-55
• Kaiser HF. 1970; A second-generation little jiffy. Psychometrika 35 (04) 401-415
• Kaiser HF. 1974; Little Jiffy, Mark IV. Educ Psychol Meas 34: 111-117
• Kam ACS, Tong MCF, van Hasselt A. 2011; Cross-cultural adaptation and validation of the Chinese abbreviated profile of hearing aid benefit. Int J Audiol 50 (05) 334-339
• Knudsen LV, Öberg M, Nielsen C, Naylor G, Kramer SE. 2010; Factors influencing help seeking, hearing aid uptake, hearing aid use and satisfaction with hearing aids: a review of the literature. Trends Amplif 14 (03) 127-154
• Lee J, Paek I. 2014; In search of the optimal number of response categories in a rating scale. J Psychoed Assess 32 (07) 663-673
• Löhler J, Akcicek B, Wollenberg B, Kappe T, Schlattmann P, Schönweiler R. 2016; The influence of frequency-dependent hearing loss to unaided APHAB scores. Eur Arch Otorhinolaryngol 273 (11) 3587-3593
• Lozano LM, Garcia-Cueto E, Muñiz J. 2008; Effect of the number of response categories on the reliability and validity of rating scales. Methodology 4: 73-79
• Marshall S, Haywood K, Fitzpatrick R. 2006; Impact of patient-reported outcome measures on routine practice: a structured review. J Eval Clin Pract 12 (05) 559-568
• Öberg M, Lunner T, Andersson G. 2007; Psychometric evaluation of hearing specific self-report measures and their associations with psychosocial and demographic variables. Audiol Med 5 (03) 188-199
• Preston CC, Colman AM. 2000; Optimal number of response categories in rating scales: reliability, validity, discriminating power, and respondent preferences. Acta Psychol (Amst) 104 (01) 1-15
• Saunders GH, Chisolm TH, Wallhagen MI. 2012; Older adults and hearing help-seeking behaviors. Am J Audiol 21 (02) 331-337
• Stark P, Hickson L. 2004; Outcomes of hearing aid fitting for older people with hearing impairment and their significant others. Int J Audiol 43 (07) 390-398