International Journal of Audiology 2014; 53: 903–909
Original Article
Validation of a French (Québec) version of the Tinnitus Handicap Inventory Daniel Bolduc*, France Désilets*, Mireille Tardif* & Tony Leroux*,† *Institut Raymond-Dewar, Centre de recherche interdisciplinaire en réadaptation, Montréal, Québec, Canada, and †École d’orthophonie et d’audiologie, Faculté de médecine, Université de Montréal, Québec, Canada
Abstract Objective: To assess the psychometric properties of the Quebec French translation of the Tinnitus Handicap Inventory (THI-FQ). Design: The original version of the THI was translated into French by two different people, and then revised before being administered to 221 study participants. A subgroup of 75 participants also completed the French versions of the Beck Depression Inventory II (BDI-II) and the Tinnitus Reaction Questionnaire (TRQ). Study sample: The study participants used French as their home language, had completed an otolaryngology (ENT) assessment, and had sought help with disabling tinnitus. Results: No differences were found according to gender, age, or degree of hearing loss. The THI-FQ was very stable over a period of approximately 32 days. Internal consistency was 0.93 for the total score. Convergent validity with the TRQ was 0.81 (p ⬍ 0.001) while construct validity, as obtained with the BDI-II, was 0.74 (p ⬍ 0.001). As in other studies, the factor analysis did not confirm the structure proposed by Newman et al in the original study (1996). Conclusions: The THI-FQ presents very good reliability and validity, comparable to translations in other languages and the original version.
Key Words: Tinnitus; questionnaire; handicap; validity; reliability
Tinnitus is a frequently reported symptom in the general population. Various statistics have been published regarding its prevalence, with between 10 and 15% of people reporting having tinnitus (Davis & Rafaie, 2000). In Quebec (Canada), according to a large health survey (Paré & Levasseur, 2000), 13% of Quebecers aged 15 years and over answered ‘yes’ to the question: ‘Do you hear noises in your ears or your head that last for at least five minutes?’ Among those respondents who said they had tinnitus, 35% reported that they had it often or all the time. The impact of tinnitus on one’s daily life differs from one person to the next but sufferers generally report problems with concentration and sleep, anxious or depressive symptoms, irritability, and feeling isolated (Holmes & Padgham, 2009). Since tinnitus is invisible and usually inaudible to other people, this often increases the distress of sufferers who are afraid their problems will not be taken seriously. Clinically, it is a challenge for health professionals to assess tinnitus and its impact on peoples’ lives. At this time there are no proven objective ways to measure tinnitus. And even if there were, clinical tools are needed to measure
the difficulties that tinnitus can create because such difficulties are the main reason why sufferers seek help. Questionnaires are widely used to measure the impact of tinnitus in both research and clinical settings. They assess tinnitus-related annoyance via dimensions such as emotional distress, catastrophic thinking, concentration, and sleep problems. These instruments can also be used to measure progress during, at the end of treatment, and in follow-ups. Questionnaires are easy to use and, in principle, to understand, and do not take long to complete. Various instruments have been developed to meet these needs over the past twenty years. The best known and probably most widely used is the Tinnitus Handicap Inventory (THI) (Newman et al, 1996), a scale consisting of 25 questions to which the respondent answers ‘Yes’, ‘Sometimes’ or ‘No’, scored as 4, 2, or 0 points respectively. The total score gives a measure of the degree of tinnitus-related annoyance. The authors also grouped the questions into three subscales: ‘Functional’ (11 items), ‘Emotional’ (9 items) and ‘Catastrophic’ (5 items). The THI has been validated in many studies (Newman et al, 1996; 1998; Baguley et al, 2000; Baguley & Andersson, 2003) and it has been
Correspondence: Daniel Bolduc, Institut Raymond-Dewar, 3600, rue Berri, Montréal, Québec, H2L 4G9, Canada. E-mail: [email protected] (Received 9 May 2013; accepted 10 June 2014 ) ISSN 1499-2027 print/ISSN 1708-8186 online © 2014 British Society of Audiology, International Society of Audiology, and Nordic Audiological Society DOI: 10.3109/14992027.2014.935495
904
D. Bolduc et al.
Abbreviations Beck Depression Inventory, second version Catastrophic subscale of the Tinnitus Handicap Inventory CFI Comparative Fit Index dB HL Decibels referenced as hearing level E Emotional subscale of the Tinnitus Handicap Inventory ENT Ear, Nose, and Throat F Functional subscale of the Tinnitus Handicap Inventory HFPTA High Frequency Pure Tone Average Hz Hertz IRD Institut Raymond-Dewar KMO Kaiser-Meyer-Olkin LE Left ear NNFI Non-Normed Fit Index PTA Pure Tone average RE Right ear RMSEA Root Mean Square Error of Approximation THI Tinnitus Handicap Inventory THI-FQ Tinnitus Handicap Inventory - French Quebec version TRQ Tinnitus Reaction Questionnaire
2002. The French scale is called the Mesure du handicap relié aux acouphènes - THI-FQ and is shown in Table 1.
BDI-II C
translated into and validated in Spanish (Herràiz et al, 2001), Danish (Zachariae et al, 2000), Italian (Crocetti et al, 2009; Monzani et al, 2008; Passi et al, 2008), German (Kleinjung et al, 2007), Korean (Kim et al, 2002), Brazilian Portuguese (Alves Ferreira et al, 2005; Dias et al, 2006; Petersen Schmidt et al, 2006), Turkish (Aksoy et al, 2007), Cantonese (Kam et al, 2009), Hebrew (Oron et al, 2011), Mandarin (Meng et al, 2012), and Thai (Limviriyakul & Supavanich, 2012). Moreover, our clinical experience suggests that, unlike similar types of scales, the THI introduces very little bias related to hearing loss. The objective of this study was to assess the psychometric properties of the Quebec French translation (THI-FQ) of the THI (Newman et al, 1996). Another French translation and validation study was being done at the same time by a team in France (Ghulyan-Bédikian et al, 2010). Although the two translations share many similarities, the cultural differences in the use of terminology may justify two distinct versions that reflect particular cultural realities although they stem from the same language source. In addition, the present study examined more psychometric parameters, including test-retest reliability and convergent and construct validity.
Method Development of the Quebec French version of the Tinnitus Handicap Inventory The Tinnitus Handicap Inventory was first translated into Quebec French by an experienced Quebec psychologist working with tinnitus sufferers. A professional translator who had tinnitus independently translated the original version of the questionnaire into Quebec French. The two translated versions proved to be almost identical. The authors reconciled minor differences by retaining the wording that seemed closest to that used in the original version by Newman et al (1996). This version has been used in our clinic since
Participants The participants were recruited at the Institut Raymond-Dewar (IRD) from clients who had sought help with disabling tinnitus. The IRD is a rehabilitation centre in Montreal, Canada, that specializes in hearing and communication problems. To be eligible for the study, potential candidates had to be between 21 and 85 years of age, use French as their home language, and be able to complete the necessary questionnaires without help. All had previously had an ENT assessment and an audiogram plus additional medical assessments where appropriate. The original sample contained 221 participants, 55.2% of whom were men (mean age: 49.8 ⫾ 10.3 years) and 44.8% women (mean age: 49.3 ⫾ 9.1 years). Nearly two-thirds of the participants were employed while 12% were unemployed and another 12% were on sick leave (related or not to tinnitus). Figure 1 shows the mean audiogram of the participants. The large majority had constant tinnitus (92%) that had appeared on average nearly seven years earlier (6.9 ⫾ 9.0 years). The tinnitus was bilateral in 63.1% of the participants and unilateral in 27.2%. The mean total score obtained with the THI-FQ was 54.9 ⫾ 21.0, and 79.2% of the participants had a score suggesting at least a moderate handicap (ⱖ 38 points) according to the interpretation suggested by Newman et al (1996).
Data collection Starting in 2002 the Mesure du handicap relié aux acouphènes THI-FQ questionnaire (alpha version, with items in the usual order) has been mailed to people who come to the IRD for help with disabling tinnitus. They are asked to complete it before their first assessment. During this study, which lasted from 2006 to 2010, 75 individuals who made up the subsample completed, at the time of their first appointment, the beta version of the THI-FQ (order of the items changed) as well as the French versions of the Tinnitus Reaction Questionnaire (Wilson et al, 1991), validated by Meric et al (Meric et al, 1997, 2000), and the Beck Depression Inventory II (Beck et al, 1996). Clients who needed urgent help, such as those expressing suicidal ideations in their first request for services, had to be excluded from the subsample on ethical and humanitarian grounds. These urgent situations would not have allowed sufficient time for the testretest reliability study. Also, those few clients who exhibited a profile of major psychiatric problems, such as having incoherent discourse, were excluded from the subsample. The sociodemographic and clinical characteristics of the subsample were not statistically different from those of the participants in the original sample. The Tinnitus Reaction Questionnaire (TRQ) contains 26 questions designed to assess the degree of psychological distress related to the tinnitus. The total score ranges from 0 to 104 points, and a higher score indicates a greater degree of distress. The Beck Depression Inventory II (BDI-II) consists of 21 statements evaluating the presence and severity of depressive symptoms. The statements are scored on a scale from 0 to 3. The total score ranges from 0 to 63 points, and a higher score indicates more severe depressive symptoms. Theoretically, the depression score should be correlated with the level of tinnitus-related annoyance, at least that is what previous data indicated (Crocetti et al, 2009; Kim et al, 2002; Kleinjung et al, 2007; Newman et al, 1996; Petersen Schmidt et al, 2006; Zachariae et al, 2000).
Validation of French (Québec) version of the THI
905
Table 1. Mean, standard deviation (SD), and item-total correlation of each item in the THI-FQ.
1F 2F 3E 4F 5C 6E 7F 8C 9F 10E 11C 12F 13F 14E 15F 16E 17E 18F 19C 20F 21E 22E 23C 24F 25E
Items in the THI-FQ
Mean
SD
À cause de votre acouphène, avez-vous de la difficulté à vous concentrer? (Because of your tinnitus is it difficult for you to concentrate?) À cause de l’intensité de votre acouphène, avez-vous de la difficulté à entendre ce que dit une personne? (Does the loudness of your tinnitus make it difficult for you to hear people?) Est-ce que votre acouphène vous met en colère? (Does your tinnitus make you angry?) Est-ce que votre acouphène vous rend confus(e)? (Does your tinnitus make you feel confused?) Vous sentez-vous désespéré(e) à cause de votre acouphène? (Because of your tinnitus do you feel desperate?) Est-ce que vous vous plaignez beaucoup de votre acouphène? (Do you complain a great deal about your tinnitus?) Avez-vous de la difficulté à vous endormir le soir à cause de votre acouphène? (Because of your tinnitus do you have trouble falling to sleep at night?) Avez-vous l’impression que votre acouphène vous envahit, que vous ne pouvez pas y échapper? (Do you feel as though you cannot escape your tinnitus?) Est-ce que votre acouphène vous empêche de profiter d’activités sociales (par exemple, aller au restaurant, aller au cinéma)? (Does your tinnitus interfere with your ability to enjoy social activities (such as going out to dinner, to the movies)?) Vous sentez-vous frustré(e) à cause de votre acouphène? (Because of you tinnitus do you feel frustrated?) Avez-vous l’impression que vous avez une maladie grave à cause de votre acouphène? (Because of your tinnitus do you feel that you have a terrible disease?) Avez-vous de la difficulté à jouir de la vie à cause de votre acouphène? (Does your tinnitus make it difficult for you to enjoy life?) Est-ce que votre acouphène vous empêche de faire votre travail ou vos tâches domestiques? (Does your tinnitus interfere with your job or household responsibilities?) Êtes-vous souvent irritable à cause de votre acouphène? (Because of your tinnitus do you find that you are often irritable?) Avez-vous de la difficulté à lire à cause de votre acouphène? (Because of your tinnitus is it difficult for you to read?) Est-ce que votre acouphène vous agace? (Does your tinnitus make you upset?) Sentez-vous que votre acouphène a engendré du stress dans vos relations avec les membres de votre famille et avec vos amis? (Do you feel that your tinnitus problem has placed stress on your relationship with members of your family and friends?) Avez-vous de la difficulté à vous concentrer sur autre chose que votre acouphène? (Do you find it difficult to focus your attention away from your tinnitus and on other things?) Sentez-vous que vous n’avez aucun contrôle sur votre acouphène? (Do you feel that you have no control over your tinnitus?) Vous sentez-vous souvent fatigué(e) à cause de votre acouphène? (Because of your tinnitus do you often feel tired?) Vous sentez-vous déprimé(e) à cause de votre acouphène? (Because of your tinnitus do you feel depressed?) Vous sentez-vous anxieux(se) à cause de votre acouphène? (Does your tinnitus make you feel anxious?) Sentez-vous que vous n’arrivez plus du tout à vivre avec votre acouphène? (Do you feel that you can no longer cope with your tinnitus?) Est-ce que votre acouphène empire quand vous êtes stressé(e)? (Does your tinnitus get worse when you are under stress?) Est-ce que votre acouphène vous rend insécure? (Does your tinnitus make you feel insecure?)
2.61
1.26
1.99
1.54
2.01
1.54
1.17
1.47
2.12
1.43
1.95
1.50
2.40
1.58
3.10
1.28
1.72
1.68
2.84
1.43
1.42
1.62
2.32
1.50
1.24
1.53
2.16
1.38
1.65
1.58
3.56
0.90
1.87
1.64
2.00
1.24
3.35
1.50
2.58
1.54
2.18
1.45
2.38
1.46
1.23
1.30
3.21
1.37
2.16
1.70
Item-total correlation 0.59 (0.70) 0.21 (0.22) 0.62 (0.54) 0.56 (0.64) 0.75 (0.54) 0.51 (0.63) 0.56 (0.48) 0.60 (0.55) 0.51 (0.61) 0.72 (0.77) 0.55 (0.48) 0.73 (0.69) 0.66 (0.56) 0.61 (0.69) 0.52 (0.48) 0.54 (0.76) 0.60 (0.53) 0.64 (0.69) 0.33 (0.48) 0.68 (0.58) 0.73 (0.63) 0.66 (0.54) 0.69 (0.59) 0.33 (0.49) 0.64 (0.47)
All correlations are significant at the p ⬍.001 level. In column 1, F represents an item from the functional subscale, E an item from the emotional subscale and C an item from the catastrophic subscale. In column 2 the text in parentheses is the original THI item. In column 5 the values in parentheses are the corresponding original THI item-total correlation scores.
Data processing and analysis The data from the 75 participants in the subsample were used to assess reliability as well as construct and convergent validity. As for the reliability, the scores obtained on the alpha and beta versions of
the THI-FQ were compared using the Pearson intraclass correlation coefficient and student´s t-test. The average time between the administration of the alpha and beta versions was 32.1 ⫾ 13 days. This was considered long enough to minimize any memory effect and
D. Bolduc et al.
906 0
RE LE
10
Hearing threshold (dBHL)
20 30 40 50 60 70 80 90
500
1000
2000
4000
8000
Frequency (Hz)
Figure 1. Group mean hearing thresholds (n ⫽ 211 participants). short enough not to allow time for a possible spontaneous adaptation to the tinnitus. Construct validity was documented using a Pearson correlation coefficient by comparing the THI-FQ total score with the BDI-II score. The same analysis was done for convergent validity by comparing the THI-FQ total score with the TRQ score (Meric et al, 1997, 2000). Internal consistency was analysed with the Cronbach’s α coefficient using data from the 221 participants from the original sample. A confirmatory factor analysis testing the original threesubscale structure was also done using data from the same participants. The original model’s fit was estimated from three parameters: Comparative Fit Index (CFI) (Bentler, 1990), Non-Normed Fit Index (NNFI) (Tucker and Lewis, 1973), and Root Mean Square Error of Approximation (RMSEA) (Steiger, 2000). An exploratory principal components factor analysis was also done with a Varimax rotation and Kaiser normalization.
posed of participants with a mean hearing threshold less than or equal to 20 dB HL in each ear. The other participants made up the ‘hearing-impaired’ group. The statistical analysis was done using the pure tone average (PTA) at 500, 1000, and 2000 Hz, and then the high frequency pure tone average (HFPTA) at 1000, 2000, and 4000 Hz. The scores on the THI-FQ were not statistically different for the two groups (student´s t-test: p ⬎ 0.05), regardless of the frequencies used to calculate the mean hearing threshold. In addition, a correlation analysis comparing binaural mean hearing thresholds (PTA and HFPTA) and the THI-FQ total score did not indicate any significant relationship between these variables (PTA r ⫽ ⫺ 0.1, p ⫽ 0.11; HFPTA r ⫽ ⫺ 0.12, p ⫽ 0.09). However, when examining the correlation between specific items and hearing, a significant low to moderate positive relationship was observed between item 2 and the mean hearing thresholds (PTA r ⫽ 0.351, p ⬍ 0.001; HFPTA r ⫽ 0.423, p ⬍ 0.001). Other items possibly involving the sense of hearing were explored (6, 9, 10, 12, 13, 14, 17, 20, and 25) and very low negative correlations were found between items 6, 10, and 12 (from r ⫽ ⫺ 0.135 to ⫺ 0.164; p varies from 0.017 to 0.052) and mean hearing thresholds.
Internal consistency The analysis of the internal consistency of the THI-FQ generated a Cronbach’s α coefficient of 0.92 compared to 0.93 for the original version. The coefficients for the functional, emotional and catastrophic subscales were 0.82, 0.85, and 0.74 respectively. These results are similar to those obtained for the original version, namely 0.86, 0.87, and 0.68.
Short-term test/retest reliability The total score was 53.6 ⫾ 18.6 at time 1 (alpha version) and 51.8 ⫾ 20.1 at time 2 (beta version). Student’s t-test [t(74) ⫽ 1.29; p ⫽ 0.20] did not show any statistically significant difference between the two administrations, 32 days apart on average. The analysis also indicates an intraclass correlation coefficient of 0.81 between the results at time 1 and time 2.
Ethical considerations This research project was approved by the Research Ethics Committee of the institutions that make up the Centre de recherche interdisciplinaire en réadaptation du Montréal métropolitain. The participants did not receive any compensation and they all completed a consent form.
Results Scores The mean and standard deviation of the scores obtained on each item of the THI-FQ and the item-total correlations are presented in Table 1. The item-total correlations obtained with the original version by Newman et al (1996) are also shown in parentheses. The item-total correlations are all statistically significant (p ⬍ 0.001) and vary between 0.21 (item 2) and 0.75 (item 5). As observed with the original version, the respondents’ age (r ⫽ ⫺ 0.078; p ⫽ 0.25) and gender (r ⫽ ⫺ 0.084; p ⫽ 0.22) were not correlated with the THI-FQ total score. The relationship between the THI-FQ total score and hearing loss was examined using the same approach as used by Newman et al (1996). The respondents were divided into two groups based on their mean hearing threshold. The ‘normal’ group was com-
Construct and convergent validity The mean total score on the BDI-II was 13.59 ⫾ 9.45 (range 0–54). This means that the participants presented depressive symptoms on the boundary of minimal and mild levels on the scale. A correlation of the total scores on the BDI-II and THI-FQ gave a construct validity index of 0.74. The TRQ was used to examine the THI-FQ’s convergent validity, and a very good correlation coefficient of 0.81 was obtained between the two measures. The mean total score on the TRQ was 38.12 ⫾ 22.01 (range 1–86). However, the authors of the TRQ (Wilson et al, 1991) and of the French version (Meric et al, 1997, 2000) did not provide any guidance as to the degree of severity.
Factor analyses A first principal components factor analysis was done to examine the original three subscales proposed by Newman et al (1996). It must be recalled that the original subscales were proposed solely after a content analysis of the items and were not based on a factor analysis. The confirmatory analysis of that model showed a χ2 of 598.7 (df: 272), giving a CFI of 0.84, NNFI of 0.81 and RMSEA of
Validation of French (Québec) version of the THI 10 9 8
Eigenvalue
7 6 5 4 3 2 1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Component
Figure 2. Scree plot of the THI-FQ. 0.073 (95%CI: 0.065–0.080). For a model’s fit to be considered acceptable, a CFI above 0.93 (Byrne, 1994) is suggested. To confirm the model’s fit, the NNFI must be above 0.90 (Byrne, 1994) or 0.95 (Schumacker & Lomax, 2004). Confirmation of a model must also be based on a RMSEA below 0.05 (Steiger, 1990) and a confidence interval upper limit not exceeding 0.08 (Hu & Bentler, 1995).
Therefore, as expected, our results did not confirm the original threesubscale structure proposed by Newman et al (1996). A second principal components factor analysis was done to explore the THI-FQ’s factor structure. A common variance of 0.91 was obtained, as measured by the Kaiser-Meyer-Olkin (KMO) ratio, confirming that factor analysis could be used. As shown by the scree plot in Figure 2, six factors were extracted with eigenvalues over 1.0 that explain 61.0% of the items’ total variance. The eigenvalue associated with each principal component indicates how much variation in the data set it explains. The first factor explains 35.7% of the total variance (eigenvalue ⫽ 8.92), the second factor explains 6.7% (eigenvalue ⫽ 1.67), the third 5.5% (eigenvalue ⫽ 1.38), the fourth 4.6% (eigenvalue ⫽ 1.15), the fifth 4.3% (eigenvalue ⫽ 1.08), and the sixth factor 4.2% (eigenvalue ⫽ 1.05). Table 2 presents the results of the six-factor explanatory solution with loadings over 0.35.
Discussion According to the results obtained on the THI-FQ, the participants in this study presented a higher level of handicap than the participants in the majority of other published studies. Apart from the study by Aksoy et al (2007) showing a mean score of 53.2 ⫾ 22.8 similar to our mean score of 54.9 ⫾ 21.0, all other studies referred to in this paper showed a mean score ranging from 25.4 ⫾ 20.5 (Newman et al, 1996) to 48.4 ⫾ 24.8 (Passi et al, 2008). Our
Table 2. Results of principal components factor analysis of the THI-FQ. Factors Items Emotional disturbance (35.7%) C 5 Despair E 10 Frustration C 11 Terrible disease E 21 Depression E 22 Anxiety C 23 Difficulty coping with tinnitus E 25 Insecurity Expression of suffering (6.7%) E 3 Anger E 6 Complaining E 14 Irritability E 17 Stress on relationships Functional disturbance (5.5%) F 1 Concentration F 7 Sleep F 15 Reading F 18 Ability to focus one’s attention C 19 No control over the tinnitus Feeling that tinnitus adversely affects one’s quality of life (4.6%) C 8 Inescapable F 9 Restriction in social activities F 12 Enjoyment of life Perceptual and cognitive disturbance (4.3%) F 2 Difficulty hearing F 4 Confusion F 13 Restriction in work/household responsabilities F 20 Fatigue Stress (4.2%) E 16 Being upset F 24 Tinnitus worse when under stress Only loadings above 0.35 are shown.
907
1 0.69 0.50 0.77 0.71 0.73 0.58 0.69
2
3
4
5
6
0.49
0.42 0.35
0.69 0.35 0.81 0.64
0.49
0.62 0.54 0.70 0.49 0.60
0.37
0.44
0.54 0.72 0.54
0.51
0.74 0.54 0.44 0.38
0.39 0.43 0.36 0.37
0.38 0.80
908
D. Bolduc et al.
participants generally reported a significant impact of the tinnitus on their lives. This may be because of the IRD’s function, namely as a second line institution providing rehabilitation services and not diagnostic consultations. People who come to the IRD may not have found relief or reassurance in their previous contacts within the medical field. A typical case history shows consultations with at least two ENT specialists and therapeutic efforts—usually medication or alternative medicine treatments—that were not satisfactory. As with the original version, item 2 presents the weakest itemtotal correlation. At first glance this item assesses a dimension related to hearing which is confirmed by our results showing a significant positive correlation between item 2 and mean hearing thresholds (r ⫽ 0.351 to 0.423). This finding confirms that scales designed to measure tinnitus-related annoyance should avoid questions about the respondents’ hearing ability. However, it is almost impossible to entirely remove the contribution of hearing loss to the assessment of tinnitus handicap as some items, not explicitly related to the hearing sense (items 6, 10, and 12) are weakly (r ⬍ 0.164) associated to hearing thresholds. Nevertheless, our data confirmed that the THI-FQ total score is not significantly affected by the respondents’ hearing thresholds. Like the original version and all subsequent validation studies, the THI-FQ presents excellent internal consistency. The test-retest reliability of the THI-FQ is also strong, even after 32 days on average. Previous studies had shown that the questionnaire was reliable over shorter periods (Aksoy et al, 2007; Kam et al, 2009; Newman et al, 1998; Alves Ferreira et al, 2005). It would be worthwhile repeating the reliability test over the same length of time as the intervention, which is 10 weeks at the IRD. A high correlation was obtained between the THI-FQ total score and the BDI-II and TRQ total scores. These results are very similar to those observed in other studies (Crocetti et al, 2009; Kleinjung et al, 2007; Oron et al, 2011; Petersen Schmidt et al, 2006; Zachariae et al, 2000). The correlation between the THI-FQ and the TRQ is stronger than that observed between the THI-FQ and the BDI-II, a finding that was also reported by others (Robinson et al, 2003). This also corresponds to what was expected theoretically. There should be a strong association between the THI and a different but related concept (construct validity), in this case depression measured with the BDI-II, and an even stronger association with a measure whose content is much closer to the THI, in this case the TRQ (convergent validity). As was the case with other researchers in the past (Baguley & Andersson, 2003; Kam et al, 2009; Monzani et al, 2008; Zachariae et al, 2000; Meng et al, 2012), the results obtained with the THI-FQ did not validate the three-factor structure proposed by Newman et al (1996). Our statistical analysis suggests a structure largely dominated by just one factor, coined emotional disturbance, which explains 36% of the variance, and five secondary factors, each of which explains less than 7% of the variance. The factor structure revealed by our data was not confirmed by the previously mentioned studies, where the secondary factors were grouped differently. However, the same profile as ours was observed, with one principal factor and several secondary factors. Although theoretically interesting, use of the three subscales proposed by Newman et al (1996) still does not have any empirical basis. Consequently, it seems wiser to consider the THI as presenting a one-factor structure.
Conclusion The objective of this study was to validate the THI-FQ in order to have a valid and reliable tool to measure the degree of tinnitus-related
annoyance in a Quebec French population seeking help specifically for this symptom. Like the original version, the THI-FQ presents very good internal consistency and highly satisfactory convergent and construct validity. Nor is it affected by the degree of hearing impairment. The presence of three subscales was not confirmed, and only the total score should be considered. Finally, this study gives French-speaking clinicians and researchers a high quality measuring instrument that is used around the world. It is now possible to employ the THI-FQ with assurance of its validity to measure tinnitus-related annoyance.
Acknowledgments The authors thank Louise Desautels for her help with the data collection, Pascal Thibault for his help with the statistical analyses, and Martine Gendron for coordinating the research project. The authors also thank the Institut Raymond-Dewar and the Centre de recherche interdisciplinaire en réadaptation for their financial and logistical support.
Declaration of interest: The authors report no declarations of interest.
References Aksoy S., Firat Y. & Alpar R. 2007. The tinnitus handicap inventory: A study of validity and reliability. Int Tinnitus J, 13, 94–98. Alves Ferreira P.A., Cunha F., Tsuneo Onishi E., Alves Branco-Barreiro F.C. et al. 2005. Handicap inventory: Cross-cultural adaptation to Brazilian Portuguese. Pro Fono, 17, 303–310. Baguley D.M. & Anderson G. 2003. Factor analysis of the tinnitus handicap inventory. Am J Audiol, 12, 31–34. Baguley D.M., Humphriss R.L. & Hodgson C.A. 2000. Convergent validity of the tinnitus handicap inventory and the tinnitus questionnaire. J Laryngol Otol, 114, 840–843. Beck A.T., Steer R.A. & Brown G.K. 1996. Manual for Beck Depression Inventory-II (Manuel pour l’inventaire de dépression de Beck – II). San Antonio, USA: Psychological Corporation. Bentler P.M. 1990. Comparative fit indexes in structural models. Psychol Bull, 107, 238–246. Byrne B.M. 1994. Structural Equation Modeling with EQS and EQS/Windows. Thousand Oaks, USA: Sage Publications. Crocetti A., Forti S., Ambrosetti U. & Del Bo L. 2009. Questionnaires to evaluate anxiety and depressive levels in tinnitus patients. Otolaryngol Head Neck Surg, 140, 403–405. Davis A. & Rafaie E.A. 2000. Epidemiology of tinnitus. In: Richard T.S. (ed.) Tinnitus Handbook. San Diego: Singular. Dias A., Cordeiro R. & Correte J.E. 2006. Tinnitus annoyance assessed by the tinnitus handicap inventory. Rev Saude Publica, 40, 706–711. Ghulyan-Bédikian V., Paolino M., Giorgetti-D’Esclercs F. & Paolini F. 2010. Propriétés psychométriques d’une version française du Tinnitus Handicap Inventory. L’Encéphale, 36, 390–396. Herràiz, C., Hernandez Calvin, J., Plaza G., Tapia M.C. & De Los Santos G. 2001. Evaluacion de la incapacidad en pacientes con acufenos. Acta Otorrinolaringol Esp, 52, 534–538. Holmes S. & Padgham N.D. 2009. Review paper. More than ringing in the ears: A review of tinnitus and its psychosocial impact. J Clin Nurs, 18, 2927–2937. Hu L.T. & Bentler P.M. 1995. Evaluating model fit. In: R.H. Hoyle (ed.) Structural Equation Modeling: Concepts, Issues, and Applications. Thousand Oaks, USA: Sage Publications. Kam A.C.S., Cheung A.P.P., Chan P.Y.B., Leung E.K.S., Wong T.K.C. et al. 2009. Psychometric properties of the Chinese (Cantonese) tinnitus handicap inventory. Int J Audiol, 48, 568–575.
Validation of French (Québec) version of the THI Kim J.H., Lee S.Y., Kim C.H., Lim S.L., Shin J.N. et al. Reliability and validity of a Korean adaptation of the tinnitus handicap inventory. Korean J Otol Head Neck Surg, 45, 328–334. Kleinjung T., Fischer B., Langguth B., Sand P.G., Hajak G. et al. 2007. Validation of the German-version tinnitus handicap inventory (THI). Psychiatr Prax, 34, 140–142. Limviriyakul S. & Supavanich W. 2012. The validity and reliability of tinnitus handicap inventory Thai version. J Med Assoc Thai, 95, 1433–1440. Meng Z., Zheng Y., Shixi L., Wang K., Xiudan K. et al. 2012. Reliability and validity of the Chinese (Mandarin) tinnitus handicap inventory. Clin Exp Otorhinolaryngol, 5, 10–17. Meric C., Pham E. & Chéry-Croze S. 1997. Validation d’une traduction Française du questionnaire Mesure de la détresse liée à l’acouphène. L’Encéphale, 23, 442–446. Meric C., Pham E. & Chéry-Croze S. 2000. Validation Assessment of a French Version of the tinnitus reaction questionnaire: A comparison between data from English and French versions. J Speech Lang Hear Res, 43, 184–190. Monzani D., Genovese E., Marrara A., Gherpelli C., Pingani L. et al. 2008. Validity of the Italian adaptation of the tinnitus handicap inventory: Focus on quality of life and psychological distress in tinnitus-sufferers. Acta Otorhinolaryngol Ital, 28, 126–134. Newman C.W., Jacobson G.P. & Spitzer J.B. 1996. Development of the tinnitus handicap inventory. Arch Otolaryngol Head Neck Surg, 122, 143–148. Newman C.W., Sandridge S.A. & Jabson G.P. 1998. Psychometric adequacy of the tinnitus handicap inventory (THI) for evaluating treatment outcome. J Am Acad Audiol, 9, 153–160. Oron Y., Shushan S., Kreitler S. & Roth Y. 2011. A Hebrew adaptation of the tinnitus handicap inventory. Int J Audiol, 50, 426–430.
909
Paré L. & Levasseur M. 2000. Problèmes auditifs et problèmes visuels. In: QUÉBEC, I.D.L.S.D. (ed.) Enquête sociale et de santé 2000, 2e édition. Québec. Passi S., Ralli G., Capparelli E., Mammone A., Scacciatelli D. et al. 2008. The THI questionnaire: Psychometric data for reliability and validity of the Italian version. Int Tinnitus J, 14, 26–33. Petersen Schmidt L., Niemiec Teixeira V., Dall’Igna C., Dallagnol D. & Magnus Smith M. 2006. Brazilian Portuguese Language version of the tinnitus handicap inventory: Validity and reproducibility. Rev Bras Otorrinolaringol, 72, 808–810. Robinson S.K., McQuaid J.R., Viirre E.S., Betzig L.L., Miller D. et al. 2003. Relationship of tinnitus questionnaire to depressive symptoms, quality of well-being, and internal focus. Int Tinnitus J, 9, 97–103. Schumacker R.E. & Lomax R.G. 2004. A Beginner’s Guide to Structural Equation Modeling, Second edition. Mahwah, USA: Lawrence Erlbaum Associates. Steiger J.H. 1990. Structural model evaluation and modification: An interval estimation approach. Multivariate Behav Res, 25, 173–180. Steiger J.H. 2000. Point estimation, hypothesis testing and interval estimation using the RMSEA: Some comments and a reply to Hayduk and Glaser. Struct Equ Modeling, 7, 149–162. Tucker L.R. & Lewis C. 1973. The reliability coefficient for maximum likelihood factor analysis. Psychometrika, 38, 1–10. Wilson P.H., Henry J., Bowen M. & Haralambous G. 1991. Tinnitus reaction questionnaire: Psychometric properties of a measure of distress associated with tinnitus. J Speech Hear Res, 34, 197–201. Zachariae R., Mirs F., Johansen L.V., Anderson S.E., Bjerring P. et al. 2000. Reliability and validity of a Danish adaptation of the tinnitus handicap inventory. Scand Audiol, 29, 37–43.
Copyright of International Journal of Audiology is the property of Taylor & Francis Ltd and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.
Original Article
Validation of a French (Québec) version of the Tinnitus Handicap Inventory Daniel Bolduc*, France Désilets*, Mireille Tardif* & Tony Leroux*,† *Institut Raymond-Dewar, Centre de recherche interdisciplinaire en réadaptation, Montréal, Québec, Canada, and †École d’orthophonie et d’audiologie, Faculté de médecine, Université de Montréal, Québec, Canada
Abstract Objective: To assess the psychometric properties of the Quebec French translation of the Tinnitus Handicap Inventory (THI-FQ). Design: The original version of the THI was translated into French by two different people, and then revised before being administered to 221 study participants. A subgroup of 75 participants also completed the French versions of the Beck Depression Inventory II (BDI-II) and the Tinnitus Reaction Questionnaire (TRQ). Study sample: The study participants used French as their home language, had completed an otolaryngology (ENT) assessment, and had sought help with disabling tinnitus. Results: No differences were found according to gender, age, or degree of hearing loss. The THI-FQ was very stable over a period of approximately 32 days. Internal consistency was 0.93 for the total score. Convergent validity with the TRQ was 0.81 (p ⬍ 0.001) while construct validity, as obtained with the BDI-II, was 0.74 (p ⬍ 0.001). As in other studies, the factor analysis did not confirm the structure proposed by Newman et al in the original study (1996). Conclusions: The THI-FQ presents very good reliability and validity, comparable to translations in other languages and the original version.
Key Words: Tinnitus; questionnaire; handicap; validity; reliability
Tinnitus is a frequently reported symptom in the general population. Various statistics have been published regarding its prevalence, with between 10 and 15% of people reporting having tinnitus (Davis & Rafaie, 2000). In Quebec (Canada), according to a large health survey (Paré & Levasseur, 2000), 13% of Quebecers aged 15 years and over answered ‘yes’ to the question: ‘Do you hear noises in your ears or your head that last for at least five minutes?’ Among those respondents who said they had tinnitus, 35% reported that they had it often or all the time. The impact of tinnitus on one’s daily life differs from one person to the next but sufferers generally report problems with concentration and sleep, anxious or depressive symptoms, irritability, and feeling isolated (Holmes & Padgham, 2009). Since tinnitus is invisible and usually inaudible to other people, this often increases the distress of sufferers who are afraid their problems will not be taken seriously. Clinically, it is a challenge for health professionals to assess tinnitus and its impact on peoples’ lives. At this time there are no proven objective ways to measure tinnitus. And even if there were, clinical tools are needed to measure
the difficulties that tinnitus can create because such difficulties are the main reason why sufferers seek help. Questionnaires are widely used to measure the impact of tinnitus in both research and clinical settings. They assess tinnitus-related annoyance via dimensions such as emotional distress, catastrophic thinking, concentration, and sleep problems. These instruments can also be used to measure progress during, at the end of treatment, and in follow-ups. Questionnaires are easy to use and, in principle, to understand, and do not take long to complete. Various instruments have been developed to meet these needs over the past twenty years. The best known and probably most widely used is the Tinnitus Handicap Inventory (THI) (Newman et al, 1996), a scale consisting of 25 questions to which the respondent answers ‘Yes’, ‘Sometimes’ or ‘No’, scored as 4, 2, or 0 points respectively. The total score gives a measure of the degree of tinnitus-related annoyance. The authors also grouped the questions into three subscales: ‘Functional’ (11 items), ‘Emotional’ (9 items) and ‘Catastrophic’ (5 items). The THI has been validated in many studies (Newman et al, 1996; 1998; Baguley et al, 2000; Baguley & Andersson, 2003) and it has been
Correspondence: Daniel Bolduc, Institut Raymond-Dewar, 3600, rue Berri, Montréal, Québec, H2L 4G9, Canada. E-mail: [email protected] (Received 9 May 2013; accepted 10 June 2014 ) ISSN 1499-2027 print/ISSN 1708-8186 online © 2014 British Society of Audiology, International Society of Audiology, and Nordic Audiological Society DOI: 10.3109/14992027.2014.935495
904
D. Bolduc et al.
Abbreviations Beck Depression Inventory, second version Catastrophic subscale of the Tinnitus Handicap Inventory CFI Comparative Fit Index dB HL Decibels referenced as hearing level E Emotional subscale of the Tinnitus Handicap Inventory ENT Ear, Nose, and Throat F Functional subscale of the Tinnitus Handicap Inventory HFPTA High Frequency Pure Tone Average Hz Hertz IRD Institut Raymond-Dewar KMO Kaiser-Meyer-Olkin LE Left ear NNFI Non-Normed Fit Index PTA Pure Tone average RE Right ear RMSEA Root Mean Square Error of Approximation THI Tinnitus Handicap Inventory THI-FQ Tinnitus Handicap Inventory - French Quebec version TRQ Tinnitus Reaction Questionnaire
2002. The French scale is called the Mesure du handicap relié aux acouphènes - THI-FQ and is shown in Table 1.
BDI-II C
translated into and validated in Spanish (Herràiz et al, 2001), Danish (Zachariae et al, 2000), Italian (Crocetti et al, 2009; Monzani et al, 2008; Passi et al, 2008), German (Kleinjung et al, 2007), Korean (Kim et al, 2002), Brazilian Portuguese (Alves Ferreira et al, 2005; Dias et al, 2006; Petersen Schmidt et al, 2006), Turkish (Aksoy et al, 2007), Cantonese (Kam et al, 2009), Hebrew (Oron et al, 2011), Mandarin (Meng et al, 2012), and Thai (Limviriyakul & Supavanich, 2012). Moreover, our clinical experience suggests that, unlike similar types of scales, the THI introduces very little bias related to hearing loss. The objective of this study was to assess the psychometric properties of the Quebec French translation (THI-FQ) of the THI (Newman et al, 1996). Another French translation and validation study was being done at the same time by a team in France (Ghulyan-Bédikian et al, 2010). Although the two translations share many similarities, the cultural differences in the use of terminology may justify two distinct versions that reflect particular cultural realities although they stem from the same language source. In addition, the present study examined more psychometric parameters, including test-retest reliability and convergent and construct validity.
Method Development of the Quebec French version of the Tinnitus Handicap Inventory The Tinnitus Handicap Inventory was first translated into Quebec French by an experienced Quebec psychologist working with tinnitus sufferers. A professional translator who had tinnitus independently translated the original version of the questionnaire into Quebec French. The two translated versions proved to be almost identical. The authors reconciled minor differences by retaining the wording that seemed closest to that used in the original version by Newman et al (1996). This version has been used in our clinic since
Participants The participants were recruited at the Institut Raymond-Dewar (IRD) from clients who had sought help with disabling tinnitus. The IRD is a rehabilitation centre in Montreal, Canada, that specializes in hearing and communication problems. To be eligible for the study, potential candidates had to be between 21 and 85 years of age, use French as their home language, and be able to complete the necessary questionnaires without help. All had previously had an ENT assessment and an audiogram plus additional medical assessments where appropriate. The original sample contained 221 participants, 55.2% of whom were men (mean age: 49.8 ⫾ 10.3 years) and 44.8% women (mean age: 49.3 ⫾ 9.1 years). Nearly two-thirds of the participants were employed while 12% were unemployed and another 12% were on sick leave (related or not to tinnitus). Figure 1 shows the mean audiogram of the participants. The large majority had constant tinnitus (92%) that had appeared on average nearly seven years earlier (6.9 ⫾ 9.0 years). The tinnitus was bilateral in 63.1% of the participants and unilateral in 27.2%. The mean total score obtained with the THI-FQ was 54.9 ⫾ 21.0, and 79.2% of the participants had a score suggesting at least a moderate handicap (ⱖ 38 points) according to the interpretation suggested by Newman et al (1996).
Data collection Starting in 2002 the Mesure du handicap relié aux acouphènes THI-FQ questionnaire (alpha version, with items in the usual order) has been mailed to people who come to the IRD for help with disabling tinnitus. They are asked to complete it before their first assessment. During this study, which lasted from 2006 to 2010, 75 individuals who made up the subsample completed, at the time of their first appointment, the beta version of the THI-FQ (order of the items changed) as well as the French versions of the Tinnitus Reaction Questionnaire (Wilson et al, 1991), validated by Meric et al (Meric et al, 1997, 2000), and the Beck Depression Inventory II (Beck et al, 1996). Clients who needed urgent help, such as those expressing suicidal ideations in their first request for services, had to be excluded from the subsample on ethical and humanitarian grounds. These urgent situations would not have allowed sufficient time for the testretest reliability study. Also, those few clients who exhibited a profile of major psychiatric problems, such as having incoherent discourse, were excluded from the subsample. The sociodemographic and clinical characteristics of the subsample were not statistically different from those of the participants in the original sample. The Tinnitus Reaction Questionnaire (TRQ) contains 26 questions designed to assess the degree of psychological distress related to the tinnitus. The total score ranges from 0 to 104 points, and a higher score indicates a greater degree of distress. The Beck Depression Inventory II (BDI-II) consists of 21 statements evaluating the presence and severity of depressive symptoms. The statements are scored on a scale from 0 to 3. The total score ranges from 0 to 63 points, and a higher score indicates more severe depressive symptoms. Theoretically, the depression score should be correlated with the level of tinnitus-related annoyance, at least that is what previous data indicated (Crocetti et al, 2009; Kim et al, 2002; Kleinjung et al, 2007; Newman et al, 1996; Petersen Schmidt et al, 2006; Zachariae et al, 2000).
Validation of French (Québec) version of the THI
905
Table 1. Mean, standard deviation (SD), and item-total correlation of each item in the THI-FQ.
1F 2F 3E 4F 5C 6E 7F 8C 9F 10E 11C 12F 13F 14E 15F 16E 17E 18F 19C 20F 21E 22E 23C 24F 25E
Items in the THI-FQ
Mean
SD
À cause de votre acouphène, avez-vous de la difficulté à vous concentrer? (Because of your tinnitus is it difficult for you to concentrate?) À cause de l’intensité de votre acouphène, avez-vous de la difficulté à entendre ce que dit une personne? (Does the loudness of your tinnitus make it difficult for you to hear people?) Est-ce que votre acouphène vous met en colère? (Does your tinnitus make you angry?) Est-ce que votre acouphène vous rend confus(e)? (Does your tinnitus make you feel confused?) Vous sentez-vous désespéré(e) à cause de votre acouphène? (Because of your tinnitus do you feel desperate?) Est-ce que vous vous plaignez beaucoup de votre acouphène? (Do you complain a great deal about your tinnitus?) Avez-vous de la difficulté à vous endormir le soir à cause de votre acouphène? (Because of your tinnitus do you have trouble falling to sleep at night?) Avez-vous l’impression que votre acouphène vous envahit, que vous ne pouvez pas y échapper? (Do you feel as though you cannot escape your tinnitus?) Est-ce que votre acouphène vous empêche de profiter d’activités sociales (par exemple, aller au restaurant, aller au cinéma)? (Does your tinnitus interfere with your ability to enjoy social activities (such as going out to dinner, to the movies)?) Vous sentez-vous frustré(e) à cause de votre acouphène? (Because of you tinnitus do you feel frustrated?) Avez-vous l’impression que vous avez une maladie grave à cause de votre acouphène? (Because of your tinnitus do you feel that you have a terrible disease?) Avez-vous de la difficulté à jouir de la vie à cause de votre acouphène? (Does your tinnitus make it difficult for you to enjoy life?) Est-ce que votre acouphène vous empêche de faire votre travail ou vos tâches domestiques? (Does your tinnitus interfere with your job or household responsibilities?) Êtes-vous souvent irritable à cause de votre acouphène? (Because of your tinnitus do you find that you are often irritable?) Avez-vous de la difficulté à lire à cause de votre acouphène? (Because of your tinnitus is it difficult for you to read?) Est-ce que votre acouphène vous agace? (Does your tinnitus make you upset?) Sentez-vous que votre acouphène a engendré du stress dans vos relations avec les membres de votre famille et avec vos amis? (Do you feel that your tinnitus problem has placed stress on your relationship with members of your family and friends?) Avez-vous de la difficulté à vous concentrer sur autre chose que votre acouphène? (Do you find it difficult to focus your attention away from your tinnitus and on other things?) Sentez-vous que vous n’avez aucun contrôle sur votre acouphène? (Do you feel that you have no control over your tinnitus?) Vous sentez-vous souvent fatigué(e) à cause de votre acouphène? (Because of your tinnitus do you often feel tired?) Vous sentez-vous déprimé(e) à cause de votre acouphène? (Because of your tinnitus do you feel depressed?) Vous sentez-vous anxieux(se) à cause de votre acouphène? (Does your tinnitus make you feel anxious?) Sentez-vous que vous n’arrivez plus du tout à vivre avec votre acouphène? (Do you feel that you can no longer cope with your tinnitus?) Est-ce que votre acouphène empire quand vous êtes stressé(e)? (Does your tinnitus get worse when you are under stress?) Est-ce que votre acouphène vous rend insécure? (Does your tinnitus make you feel insecure?)
2.61
1.26
1.99
1.54
2.01
1.54
1.17
1.47
2.12
1.43
1.95
1.50
2.40
1.58
3.10
1.28
1.72
1.68
2.84
1.43
1.42
1.62
2.32
1.50
1.24
1.53
2.16
1.38
1.65
1.58
3.56
0.90
1.87
1.64
2.00
1.24
3.35
1.50
2.58
1.54
2.18
1.45
2.38
1.46
1.23
1.30
3.21
1.37
2.16
1.70
Item-total correlation 0.59 (0.70) 0.21 (0.22) 0.62 (0.54) 0.56 (0.64) 0.75 (0.54) 0.51 (0.63) 0.56 (0.48) 0.60 (0.55) 0.51 (0.61) 0.72 (0.77) 0.55 (0.48) 0.73 (0.69) 0.66 (0.56) 0.61 (0.69) 0.52 (0.48) 0.54 (0.76) 0.60 (0.53) 0.64 (0.69) 0.33 (0.48) 0.68 (0.58) 0.73 (0.63) 0.66 (0.54) 0.69 (0.59) 0.33 (0.49) 0.64 (0.47)
All correlations are significant at the p ⬍.001 level. In column 1, F represents an item from the functional subscale, E an item from the emotional subscale and C an item from the catastrophic subscale. In column 2 the text in parentheses is the original THI item. In column 5 the values in parentheses are the corresponding original THI item-total correlation scores.
Data processing and analysis The data from the 75 participants in the subsample were used to assess reliability as well as construct and convergent validity. As for the reliability, the scores obtained on the alpha and beta versions of
the THI-FQ were compared using the Pearson intraclass correlation coefficient and student´s t-test. The average time between the administration of the alpha and beta versions was 32.1 ⫾ 13 days. This was considered long enough to minimize any memory effect and
D. Bolduc et al.
906 0
RE LE
10
Hearing threshold (dBHL)
20 30 40 50 60 70 80 90
500
1000
2000
4000
8000
Frequency (Hz)
Figure 1. Group mean hearing thresholds (n ⫽ 211 participants). short enough not to allow time for a possible spontaneous adaptation to the tinnitus. Construct validity was documented using a Pearson correlation coefficient by comparing the THI-FQ total score with the BDI-II score. The same analysis was done for convergent validity by comparing the THI-FQ total score with the TRQ score (Meric et al, 1997, 2000). Internal consistency was analysed with the Cronbach’s α coefficient using data from the 221 participants from the original sample. A confirmatory factor analysis testing the original threesubscale structure was also done using data from the same participants. The original model’s fit was estimated from three parameters: Comparative Fit Index (CFI) (Bentler, 1990), Non-Normed Fit Index (NNFI) (Tucker and Lewis, 1973), and Root Mean Square Error of Approximation (RMSEA) (Steiger, 2000). An exploratory principal components factor analysis was also done with a Varimax rotation and Kaiser normalization.
posed of participants with a mean hearing threshold less than or equal to 20 dB HL in each ear. The other participants made up the ‘hearing-impaired’ group. The statistical analysis was done using the pure tone average (PTA) at 500, 1000, and 2000 Hz, and then the high frequency pure tone average (HFPTA) at 1000, 2000, and 4000 Hz. The scores on the THI-FQ were not statistically different for the two groups (student´s t-test: p ⬎ 0.05), regardless of the frequencies used to calculate the mean hearing threshold. In addition, a correlation analysis comparing binaural mean hearing thresholds (PTA and HFPTA) and the THI-FQ total score did not indicate any significant relationship between these variables (PTA r ⫽ ⫺ 0.1, p ⫽ 0.11; HFPTA r ⫽ ⫺ 0.12, p ⫽ 0.09). However, when examining the correlation between specific items and hearing, a significant low to moderate positive relationship was observed between item 2 and the mean hearing thresholds (PTA r ⫽ 0.351, p ⬍ 0.001; HFPTA r ⫽ 0.423, p ⬍ 0.001). Other items possibly involving the sense of hearing were explored (6, 9, 10, 12, 13, 14, 17, 20, and 25) and very low negative correlations were found between items 6, 10, and 12 (from r ⫽ ⫺ 0.135 to ⫺ 0.164; p varies from 0.017 to 0.052) and mean hearing thresholds.
Internal consistency The analysis of the internal consistency of the THI-FQ generated a Cronbach’s α coefficient of 0.92 compared to 0.93 for the original version. The coefficients for the functional, emotional and catastrophic subscales were 0.82, 0.85, and 0.74 respectively. These results are similar to those obtained for the original version, namely 0.86, 0.87, and 0.68.
Short-term test/retest reliability The total score was 53.6 ⫾ 18.6 at time 1 (alpha version) and 51.8 ⫾ 20.1 at time 2 (beta version). Student’s t-test [t(74) ⫽ 1.29; p ⫽ 0.20] did not show any statistically significant difference between the two administrations, 32 days apart on average. The analysis also indicates an intraclass correlation coefficient of 0.81 between the results at time 1 and time 2.
Ethical considerations This research project was approved by the Research Ethics Committee of the institutions that make up the Centre de recherche interdisciplinaire en réadaptation du Montréal métropolitain. The participants did not receive any compensation and they all completed a consent form.
Results Scores The mean and standard deviation of the scores obtained on each item of the THI-FQ and the item-total correlations are presented in Table 1. The item-total correlations obtained with the original version by Newman et al (1996) are also shown in parentheses. The item-total correlations are all statistically significant (p ⬍ 0.001) and vary between 0.21 (item 2) and 0.75 (item 5). As observed with the original version, the respondents’ age (r ⫽ ⫺ 0.078; p ⫽ 0.25) and gender (r ⫽ ⫺ 0.084; p ⫽ 0.22) were not correlated with the THI-FQ total score. The relationship between the THI-FQ total score and hearing loss was examined using the same approach as used by Newman et al (1996). The respondents were divided into two groups based on their mean hearing threshold. The ‘normal’ group was com-
Construct and convergent validity The mean total score on the BDI-II was 13.59 ⫾ 9.45 (range 0–54). This means that the participants presented depressive symptoms on the boundary of minimal and mild levels on the scale. A correlation of the total scores on the BDI-II and THI-FQ gave a construct validity index of 0.74. The TRQ was used to examine the THI-FQ’s convergent validity, and a very good correlation coefficient of 0.81 was obtained between the two measures. The mean total score on the TRQ was 38.12 ⫾ 22.01 (range 1–86). However, the authors of the TRQ (Wilson et al, 1991) and of the French version (Meric et al, 1997, 2000) did not provide any guidance as to the degree of severity.
Factor analyses A first principal components factor analysis was done to examine the original three subscales proposed by Newman et al (1996). It must be recalled that the original subscales were proposed solely after a content analysis of the items and were not based on a factor analysis. The confirmatory analysis of that model showed a χ2 of 598.7 (df: 272), giving a CFI of 0.84, NNFI of 0.81 and RMSEA of
Validation of French (Québec) version of the THI 10 9 8
Eigenvalue
7 6 5 4 3 2 1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Component
Figure 2. Scree plot of the THI-FQ. 0.073 (95%CI: 0.065–0.080). For a model’s fit to be considered acceptable, a CFI above 0.93 (Byrne, 1994) is suggested. To confirm the model’s fit, the NNFI must be above 0.90 (Byrne, 1994) or 0.95 (Schumacker & Lomax, 2004). Confirmation of a model must also be based on a RMSEA below 0.05 (Steiger, 1990) and a confidence interval upper limit not exceeding 0.08 (Hu & Bentler, 1995).
Therefore, as expected, our results did not confirm the original threesubscale structure proposed by Newman et al (1996). A second principal components factor analysis was done to explore the THI-FQ’s factor structure. A common variance of 0.91 was obtained, as measured by the Kaiser-Meyer-Olkin (KMO) ratio, confirming that factor analysis could be used. As shown by the scree plot in Figure 2, six factors were extracted with eigenvalues over 1.0 that explain 61.0% of the items’ total variance. The eigenvalue associated with each principal component indicates how much variation in the data set it explains. The first factor explains 35.7% of the total variance (eigenvalue ⫽ 8.92), the second factor explains 6.7% (eigenvalue ⫽ 1.67), the third 5.5% (eigenvalue ⫽ 1.38), the fourth 4.6% (eigenvalue ⫽ 1.15), the fifth 4.3% (eigenvalue ⫽ 1.08), and the sixth factor 4.2% (eigenvalue ⫽ 1.05). Table 2 presents the results of the six-factor explanatory solution with loadings over 0.35.
Discussion According to the results obtained on the THI-FQ, the participants in this study presented a higher level of handicap than the participants in the majority of other published studies. Apart from the study by Aksoy et al (2007) showing a mean score of 53.2 ⫾ 22.8 similar to our mean score of 54.9 ⫾ 21.0, all other studies referred to in this paper showed a mean score ranging from 25.4 ⫾ 20.5 (Newman et al, 1996) to 48.4 ⫾ 24.8 (Passi et al, 2008). Our
Table 2. Results of principal components factor analysis of the THI-FQ. Factors Items Emotional disturbance (35.7%) C 5 Despair E 10 Frustration C 11 Terrible disease E 21 Depression E 22 Anxiety C 23 Difficulty coping with tinnitus E 25 Insecurity Expression of suffering (6.7%) E 3 Anger E 6 Complaining E 14 Irritability E 17 Stress on relationships Functional disturbance (5.5%) F 1 Concentration F 7 Sleep F 15 Reading F 18 Ability to focus one’s attention C 19 No control over the tinnitus Feeling that tinnitus adversely affects one’s quality of life (4.6%) C 8 Inescapable F 9 Restriction in social activities F 12 Enjoyment of life Perceptual and cognitive disturbance (4.3%) F 2 Difficulty hearing F 4 Confusion F 13 Restriction in work/household responsabilities F 20 Fatigue Stress (4.2%) E 16 Being upset F 24 Tinnitus worse when under stress Only loadings above 0.35 are shown.
907
1 0.69 0.50 0.77 0.71 0.73 0.58 0.69
2
3
4
5
6
0.49
0.42 0.35
0.69 0.35 0.81 0.64
0.49
0.62 0.54 0.70 0.49 0.60
0.37
0.44
0.54 0.72 0.54
0.51
0.74 0.54 0.44 0.38
0.39 0.43 0.36 0.37
0.38 0.80
908
D. Bolduc et al.
participants generally reported a significant impact of the tinnitus on their lives. This may be because of the IRD’s function, namely as a second line institution providing rehabilitation services and not diagnostic consultations. People who come to the IRD may not have found relief or reassurance in their previous contacts within the medical field. A typical case history shows consultations with at least two ENT specialists and therapeutic efforts—usually medication or alternative medicine treatments—that were not satisfactory. As with the original version, item 2 presents the weakest itemtotal correlation. At first glance this item assesses a dimension related to hearing which is confirmed by our results showing a significant positive correlation between item 2 and mean hearing thresholds (r ⫽ 0.351 to 0.423). This finding confirms that scales designed to measure tinnitus-related annoyance should avoid questions about the respondents’ hearing ability. However, it is almost impossible to entirely remove the contribution of hearing loss to the assessment of tinnitus handicap as some items, not explicitly related to the hearing sense (items 6, 10, and 12) are weakly (r ⬍ 0.164) associated to hearing thresholds. Nevertheless, our data confirmed that the THI-FQ total score is not significantly affected by the respondents’ hearing thresholds. Like the original version and all subsequent validation studies, the THI-FQ presents excellent internal consistency. The test-retest reliability of the THI-FQ is also strong, even after 32 days on average. Previous studies had shown that the questionnaire was reliable over shorter periods (Aksoy et al, 2007; Kam et al, 2009; Newman et al, 1998; Alves Ferreira et al, 2005). It would be worthwhile repeating the reliability test over the same length of time as the intervention, which is 10 weeks at the IRD. A high correlation was obtained between the THI-FQ total score and the BDI-II and TRQ total scores. These results are very similar to those observed in other studies (Crocetti et al, 2009; Kleinjung et al, 2007; Oron et al, 2011; Petersen Schmidt et al, 2006; Zachariae et al, 2000). The correlation between the THI-FQ and the TRQ is stronger than that observed between the THI-FQ and the BDI-II, a finding that was also reported by others (Robinson et al, 2003). This also corresponds to what was expected theoretically. There should be a strong association between the THI and a different but related concept (construct validity), in this case depression measured with the BDI-II, and an even stronger association with a measure whose content is much closer to the THI, in this case the TRQ (convergent validity). As was the case with other researchers in the past (Baguley & Andersson, 2003; Kam et al, 2009; Monzani et al, 2008; Zachariae et al, 2000; Meng et al, 2012), the results obtained with the THI-FQ did not validate the three-factor structure proposed by Newman et al (1996). Our statistical analysis suggests a structure largely dominated by just one factor, coined emotional disturbance, which explains 36% of the variance, and five secondary factors, each of which explains less than 7% of the variance. The factor structure revealed by our data was not confirmed by the previously mentioned studies, where the secondary factors were grouped differently. However, the same profile as ours was observed, with one principal factor and several secondary factors. Although theoretically interesting, use of the three subscales proposed by Newman et al (1996) still does not have any empirical basis. Consequently, it seems wiser to consider the THI as presenting a one-factor structure.
Conclusion The objective of this study was to validate the THI-FQ in order to have a valid and reliable tool to measure the degree of tinnitus-related
annoyance in a Quebec French population seeking help specifically for this symptom. Like the original version, the THI-FQ presents very good internal consistency and highly satisfactory convergent and construct validity. Nor is it affected by the degree of hearing impairment. The presence of three subscales was not confirmed, and only the total score should be considered. Finally, this study gives French-speaking clinicians and researchers a high quality measuring instrument that is used around the world. It is now possible to employ the THI-FQ with assurance of its validity to measure tinnitus-related annoyance.
Acknowledgments The authors thank Louise Desautels for her help with the data collection, Pascal Thibault for his help with the statistical analyses, and Martine Gendron for coordinating the research project. The authors also thank the Institut Raymond-Dewar and the Centre de recherche interdisciplinaire en réadaptation for their financial and logistical support.
Declaration of interest: The authors report no declarations of interest.
References Aksoy S., Firat Y. & Alpar R. 2007. The tinnitus handicap inventory: A study of validity and reliability. Int Tinnitus J, 13, 94–98. Alves Ferreira P.A., Cunha F., Tsuneo Onishi E., Alves Branco-Barreiro F.C. et al. 2005. Handicap inventory: Cross-cultural adaptation to Brazilian Portuguese. Pro Fono, 17, 303–310. Baguley D.M. & Anderson G. 2003. Factor analysis of the tinnitus handicap inventory. Am J Audiol, 12, 31–34. Baguley D.M., Humphriss R.L. & Hodgson C.A. 2000. Convergent validity of the tinnitus handicap inventory and the tinnitus questionnaire. J Laryngol Otol, 114, 840–843. Beck A.T., Steer R.A. & Brown G.K. 1996. Manual for Beck Depression Inventory-II (Manuel pour l’inventaire de dépression de Beck – II). San Antonio, USA: Psychological Corporation. Bentler P.M. 1990. Comparative fit indexes in structural models. Psychol Bull, 107, 238–246. Byrne B.M. 1994. Structural Equation Modeling with EQS and EQS/Windows. Thousand Oaks, USA: Sage Publications. Crocetti A., Forti S., Ambrosetti U. & Del Bo L. 2009. Questionnaires to evaluate anxiety and depressive levels in tinnitus patients. Otolaryngol Head Neck Surg, 140, 403–405. Davis A. & Rafaie E.A. 2000. Epidemiology of tinnitus. In: Richard T.S. (ed.) Tinnitus Handbook. San Diego: Singular. Dias A., Cordeiro R. & Correte J.E. 2006. Tinnitus annoyance assessed by the tinnitus handicap inventory. Rev Saude Publica, 40, 706–711. Ghulyan-Bédikian V., Paolino M., Giorgetti-D’Esclercs F. & Paolini F. 2010. Propriétés psychométriques d’une version française du Tinnitus Handicap Inventory. L’Encéphale, 36, 390–396. Herràiz, C., Hernandez Calvin, J., Plaza G., Tapia M.C. & De Los Santos G. 2001. Evaluacion de la incapacidad en pacientes con acufenos. Acta Otorrinolaringol Esp, 52, 534–538. Holmes S. & Padgham N.D. 2009. Review paper. More than ringing in the ears: A review of tinnitus and its psychosocial impact. J Clin Nurs, 18, 2927–2937. Hu L.T. & Bentler P.M. 1995. Evaluating model fit. In: R.H. Hoyle (ed.) Structural Equation Modeling: Concepts, Issues, and Applications. Thousand Oaks, USA: Sage Publications. Kam A.C.S., Cheung A.P.P., Chan P.Y.B., Leung E.K.S., Wong T.K.C. et al. 2009. Psychometric properties of the Chinese (Cantonese) tinnitus handicap inventory. Int J Audiol, 48, 568–575.
Validation of French (Québec) version of the THI Kim J.H., Lee S.Y., Kim C.H., Lim S.L., Shin J.N. et al. Reliability and validity of a Korean adaptation of the tinnitus handicap inventory. Korean J Otol Head Neck Surg, 45, 328–334. Kleinjung T., Fischer B., Langguth B., Sand P.G., Hajak G. et al. 2007. Validation of the German-version tinnitus handicap inventory (THI). Psychiatr Prax, 34, 140–142. Limviriyakul S. & Supavanich W. 2012. The validity and reliability of tinnitus handicap inventory Thai version. J Med Assoc Thai, 95, 1433–1440. Meng Z., Zheng Y., Shixi L., Wang K., Xiudan K. et al. 2012. Reliability and validity of the Chinese (Mandarin) tinnitus handicap inventory. Clin Exp Otorhinolaryngol, 5, 10–17. Meric C., Pham E. & Chéry-Croze S. 1997. Validation d’une traduction Française du questionnaire Mesure de la détresse liée à l’acouphène. L’Encéphale, 23, 442–446. Meric C., Pham E. & Chéry-Croze S. 2000. Validation Assessment of a French Version of the tinnitus reaction questionnaire: A comparison between data from English and French versions. J Speech Lang Hear Res, 43, 184–190. Monzani D., Genovese E., Marrara A., Gherpelli C., Pingani L. et al. 2008. Validity of the Italian adaptation of the tinnitus handicap inventory: Focus on quality of life and psychological distress in tinnitus-sufferers. Acta Otorhinolaryngol Ital, 28, 126–134. Newman C.W., Jacobson G.P. & Spitzer J.B. 1996. Development of the tinnitus handicap inventory. Arch Otolaryngol Head Neck Surg, 122, 143–148. Newman C.W., Sandridge S.A. & Jabson G.P. 1998. Psychometric adequacy of the tinnitus handicap inventory (THI) for evaluating treatment outcome. J Am Acad Audiol, 9, 153–160. Oron Y., Shushan S., Kreitler S. & Roth Y. 2011. A Hebrew adaptation of the tinnitus handicap inventory. Int J Audiol, 50, 426–430.
909
Paré L. & Levasseur M. 2000. Problèmes auditifs et problèmes visuels. In: QUÉBEC, I.D.L.S.D. (ed.) Enquête sociale et de santé 2000, 2e édition. Québec. Passi S., Ralli G., Capparelli E., Mammone A., Scacciatelli D. et al. 2008. The THI questionnaire: Psychometric data for reliability and validity of the Italian version. Int Tinnitus J, 14, 26–33. Petersen Schmidt L., Niemiec Teixeira V., Dall’Igna C., Dallagnol D. & Magnus Smith M. 2006. Brazilian Portuguese Language version of the tinnitus handicap inventory: Validity and reproducibility. Rev Bras Otorrinolaringol, 72, 808–810. Robinson S.K., McQuaid J.R., Viirre E.S., Betzig L.L., Miller D. et al. 2003. Relationship of tinnitus questionnaire to depressive symptoms, quality of well-being, and internal focus. Int Tinnitus J, 9, 97–103. Schumacker R.E. & Lomax R.G. 2004. A Beginner’s Guide to Structural Equation Modeling, Second edition. Mahwah, USA: Lawrence Erlbaum Associates. Steiger J.H. 1990. Structural model evaluation and modification: An interval estimation approach. Multivariate Behav Res, 25, 173–180. Steiger J.H. 2000. Point estimation, hypothesis testing and interval estimation using the RMSEA: Some comments and a reply to Hayduk and Glaser. Struct Equ Modeling, 7, 149–162. Tucker L.R. & Lewis C. 1973. The reliability coefficient for maximum likelihood factor analysis. Psychometrika, 38, 1–10. Wilson P.H., Henry J., Bowen M. & Haralambous G. 1991. Tinnitus reaction questionnaire: Psychometric properties of a measure of distress associated with tinnitus. J Speech Hear Res, 34, 197–201. Zachariae R., Mirs F., Johansen L.V., Anderson S.E., Bjerring P. et al. 2000. Reliability and validity of a Danish adaptation of the tinnitus handicap inventory. Scand Audiol, 29, 37–43.
Copyright of International Journal of Audiology is the property of Taylor & Francis Ltd and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.
