Chinese version of TPFQ
A Chinese translation of the scale is provided on the website of University of Iowa, but the website does not provide details of the translation. The original TPFQ was translated by two independent otologists, who were Chinese native-speakers with an advanced level of competence in English and had not read the Chinese translation of TPFQ on the website. Three senior otologists who had extensive experience in the management of tinnitus compared and discussed the three translated editions. Each expert received an evaluation questionnaire. Each item of the questionnaire was evaluated separately, which ranged from 0 (inaccurate translation) to 10 (excellent translation). If all three raters assigned a score less than 8, the translations was rejected and a new translation was required. The highest score of each item was greater than 8. Then final translation of each item was agreed upon. This version was then back translated by two Chinese native-speaker otologists who was unaware of the original questionnaire. The authors compared them with the original version of TPFQ and concluded that there had no differ significantly in any aspect.
Ten adult tinnitus outpatients were randomly selected to fill in the Chinese version of TPFQ. The average time to fill in the questionnaire was 5 minutes. The patients reported no difficulty in understanding the items of the questionnaire and they stated that they believed that the questionnaire contained important issues related to tinnitus. This resulted in the final Chinese version of TPFQ (Appendix 2 of Supplementary material) to be evaluated.
Subjects
The study was conducted on 116 patients, and consisted of 54 (46.6%) males and 62 (53.4%) females, with mean age of 48.3±15.7 years. The mean duration of tinnitus was 18 (range of 3-360) months. The other subject characteristics are presented in Table 1.
Reliability
Internal consistency of the Chinese TPFQ was examined using Cronbach coefficient alpha. The total TPFQ Cronbach’s alpha was 0.94. The Cronbach’s alpha of factor Concentration, factor Emotion, factor Hearing, factor Sleep were respectively 0.88, 0.80, 0.87, 0.91. Table 2 shows the item score and item-total correlations for the 20-item version. The item-total correlations ranged from 0.48 to 0.74, and the median value was 0.66.
Factor Structure
One step was to examine the factor structure of the questionnaire using factor analysis techniques (Table 3). The Kaiser-Meyer-Olkin (KMO) measure was performed and a KMO ratio of 0.97 was obtained for the set of 20 items. The first factor included item 2, 3, 9, 14, 17 and explained 20.65% of the total variance (eigenvalue=4.13); the second factor included item 5, 13, 16, 18, 20 and explained 19.14% of the total variance (eigenvalue=3.83); the third factor included item 7, 10, 11, 12, 15, 19 and explained 19.05% of the total variance (eigenvalue=3.81); the fourth factor included item1, 4, 6, 8 and explained 11.79% of the total variance (eigenvalue=2.36). The communality of each variable explained by these four factors ranged from 0.37 to 0.83.
Construct Validity
A spearman correlation coefficient was calculated by comparing the total score of TPFQ with tinnitus loudness magnitude estimation and THI. As shown in Table 4, the correlation between TPFQ and loudness magnitude estimation was 0.42, and a high correlation of 0.73 was found between TPFQ and THI. We also compared the scores of PSQI, BAI, BDI and other characters with the total TPFQ and subscale scores. The total TPFQ, Concentration, Emotion and Sleep subscale correlated significantly with the PSQI (r =0.43, 0.32, 0.35, 0.56, p < .01). The total TPFQ, Concentration, Emotions, Hearing and Sleep subscales correlated with the BAI and the BDI (p < .01)). The Hearing subscale was correlated with the bilateral average PTA (r =0 .21, p < .05). There was no significant correlation between tinnitus pitch and loudness matching with the TPFQ scores.
The 12-Item Version of the Chinese TPFQ
We also evaluated the 12-Item Version TPFQ. In the 12-item version, questions 7, 11, and 15 were chose for Concentration; questions 4, 8, and 10 were chose for Emotion; questions 2, 14, and 17 were chose for Hearing; questions 16, 18, and 20 were chose for Sleep.
Table 5 displays the mean total and subscale scores of the 12-item version. The mean total score of 26.58% was found for the 12 items, which is slightly below the mean score for all 20 items (31.87%). Mean subscale scores ranged from 23.13% to 32.07%.
Reliability
Internal consistency of the Chinese 12-item version TPFQ was also examined using Cronbach coefficient alpha. The total 12-item version TPFQ Cronbach’s alpha was 0.92. The Cronbach’s alpha of factor concentration, factor emotional, factor hearing, factor sleep were respectively 0.90, 0.79, 0.88, and 0.84. The 12-item version of the questionnaire also demonstrates good reliability.
Construct Validity
A spearman correlation coefficient was calculated to compare results for tinnitus loudness magnitude estimation, THI, PSQI, BDI, BAI, and the bilateral average PTA with the short version total score and subscale scores (Table 6). The total score and subscale scores from the 12-item version were significantly correlated with tinnitus loudness magnitude estimation, THI, PSQI, BDI and BAI. The bilateral average PTA was significantly correlated with the Hearing subscale (r =0 .24, p < .05).