Propriétés psychométriques d’une version française du Tinnitus Handicap Inventory

Abstract

Tinnitus is a subjective phenomenon and in the majority of cases it is impossible to objectify it. The patient's self-evaluation using questionnaires and rating scales is the only way to quantify this symptom. Among the most widely used tools is the Tinnitus Handicap Inventory (THI). This questionnaire permits the practitioner to quantify the severity of tinnitus and measure the psycho-emotional and functional components of the tinnitus related handicap. It has been translated and validated in many languages. However, there is no French version of the THI. The reliability of the THI as a powerful clinical tool for tinnitus evaluation, and the lack of a French version of this questionnaire led us to translate it into French. The goal of this study was to validate our French adaptation of the THI (fTHI) for the management of French-speaking tinnitus sufferers. This is also essential for an effective international cooperation. The fTHI was presented to a sample of 174 patients (63% men and 34% women) who had consulted our IMERTA Tinnitus Centre from June 2007 to July 2008. All patients reported tinnitus that had persisted for at least six months (average duration: 5.9 ± 4.3 years) as their primary complaint. The mean age of the sample was 54.2 ± 16.4 years, ranging from 22 to 84. The patients did not present any linguistic or mental problems likely to compromise their capacities to answer French questionnaires. During the first visit, before the consultation with the doctor, the patients were asked to complete the fTHI, and to evaluate the clearness of items and the ease of answering the questionnaire. Using rating scales, the patients were also asked to evaluate the severity of tinnitus and its impact on their quality of life. Internal consistency reliability of the fTHI and each sub-scale was evaluated using Cronbach's alpha coefficient (α). The average values, standard errors and correlations between the items, the scores of fTHI and those of the rating scales were calculated using "Statistica 6" software. The level of significance was set at 0.05. The homogeneity was considered to be satisfactory for α greater than 0.8. Ninety-seven percent of the patients answered the totality of the fTHI. Ninety-six percent of them evaluated it as simple, comprehensible and easy to answer. Cronbach's α for the fTHI was 0.90 with item/total correlations (p<0.001) ranging from r=0.21 (item n(o) 19) to r=0.71 (items n(o) 17 and 21). Cronbach's α for the functional sub-scale was 0.80 with item/total correlations (p<0.001) ranging from r=0.38 (item n(o) 24) and 0.69 (item n(o) 4). Cronbach's α for the emotional sub-scale was 0.85 with item/total correlations (p<0.001) ranging from r=0.52 (item n(o) 6) and r=0.78 (item n(o) 21). Cronbach's α for the catastrophic sub-scale was 0.49 with item/total correlations (p<0.001) ranging from r=0.51 (item n(o) 19) and 0.63 (item n(o) 23). The total score of the fTHI is correlated significantly (p<0.001) with the functional (r=0.91), emotional (r=0.90) and catastrophic (r=0.69) sub-scales. Correlations between the fTHI and the rating scales vary between 0.26 (p<0.001) and 0.61 (p<0.001). The highest coefficient of correlation is obtained for the "Total score of fTHI"/"Tinnitus impact on quality of life". Our results show that the French adaptation of the THI and its sub-scales has good internal consistency reliabilities, comparable to those of the original version. Significant inter-correlations are found between the sub-scales items, even if they represent three distinct factors. In addition, the fTHI is correlated with the self-report rating scales of the tinnitus's impact on the patient's quality of life and the severity of tinnitus. Thus, the fTHI is a psychometrically robust, reliable and valid measure of tinnitus related handicaps and can be used in a clinical setting to quantify the impact of tinnitus on French-speaking patients' quality of life, and to monitor their progress with specific therapies.