This research project was approved by the Institutional Review Board of the University of Hong Kong/Hospital Authority Hong Kong West Cluster (IRB reference No., UW 06-089 T/1114 and trial registration No., HKCTR-151).

The Chinese (HK) translation of the CLDQ was developed by iterative translations, expert panel review and cognitive debriefing, as recommended guidelines by experts[1415]. The original CLDQ was translated into Chinese by two independent professional translators. Reconciliation of the forward translations into a single forward translation was carried out by a bilingual expert in HRQOL measures (Lam CLK) and the translators. The reconciled Chinese translation was back-translated into English by another professional translator. The back translation was reviewed by the original author and the bilingual expert to identify any non-equivalence in the Chinese translation, which was then revised. The first draft of the Chinese (HK) CLDQ was evaluated by cognitive debriefing interviews with six Southern Chinese patients with CHB infection and further revision was made to ensure item clarity and equivalence to become the final Chinese (HK) CLDQ (used in this study on psychometrics properties).

Patients with complicated CHB were recruited from outpatient hepatitis clinics of a regional hospital and patients with uncomplicated CHB were randomly selected from the computerized registers of three public primary care clinics serving over 100 000 people in one of five regions in Hong Kong. Patients aged 18 years or older who were hepatitis B surface antigen-positive for more than six months were included in the study. Patients were excluded if they could not communicate in Cantonese; had cognitive impairment shown by the patient’s inability to understand the study to give consent; were co-infected with HIV, HCV or hepatitis D virus; had undergone liver transplantation or had end-stage non-hepatitis B-related illnesses; were currently taking excessive alcohol (> 30 U/wk) or illegal drugs; or refused to give consent. Each patient completed the Chinese (HK) CLDQ, the Chinese (HK) SF-36v2 Health Survey and a structured questionnaire on morbidity and socio-demographics, administered by a trained interviewer. Each patient was asked if he/she had ever been diagnosed by a registered practitioner for more than four weeks to have hypertension, diabetes mellitus, heart disease, stroke, chronic lung disease, arthritis, psychological illness (i.e. depression, anxiety, neurasthenia or psychosis) or any other chronic diseases. Chronic co-morbidity was measured by the total number of diseases (summation of positive responses to the questions) and the presence of a specific diagnosis. Clinical data related to the CHB infection including Child’s staging for patients with cirrhosis and the biomarkers of liver disease (alanine aminotransferase, aspartate aminotransferase, α-fetoprotein and total bilirubin) in each patient were retrieved from medical records. Socio-demographic data including age, gender, education, marital status, occupation, household income and family history of liver disease were also collected.

The Chinese (HK) CLDQ was re-administered to the 46 subjects with uncomplicated CHB, whose condition was expected to be stable, by telephone two weeks from the first administration, in order to assess the test-retest reliability of the Chinese (HK) CLDQ. Sixty one percent of the repeat interviews were carried out by the same interviewer.

The Chinese (HK) CLDQ consists of 29 items measuring six subscales as described above. Each item is rated on a 7-point (1 = all of the time to 7 = none of the time) Likert scale. Scores for each of the six domains are calculated by the mean of the item scores within the subscale. A summary score is calculated by the mean of all subscale scores. The scores range from 1 to 7 with a higher score indicating better HRQOL.

The Chinese (HK) SF-36v2 Health Survey is a generic HRQOL measure that has been translated, validated and normed on the general Chinese population in Hong Kong[1617]. It measures eight domains of HRQOL on physical functioning (PF), role-physical (RP), bodily pain (BP), general health (GH), vitality (VT), social functioning (SF), role-motional (RE) and mental health (MH). Summations of item scores of the same domain give the domain scores, which are transformed into a range from 0 to 100. A higher score indicates better HRQOL. The eight domain scores are summarized to form the physical component (PCS) and mental component (MCS) summary scores.

All data analysis was carried out in SPSS for Windows 15.0. Statistical significant levels were set at P values less than 0.05.

The CLDQ item and subscale scores were calculated and tested against the following scaling assumptions: (1) Items should be substantially linearly correlated to the hypothesized subscale score with a coefficient of 0.4 or above by Spearman rank correlation test, to show the item is a significant indicator for the subscale concept. (2) An item should have a stronger correlation with its hypothesized subscale than other subscales indicating scaling success[18]. This is a test of item discriminant validity. The difference between correlations is statistically significant if it is greater than two standard errors (1 divided by the square root of sample size).

Factor analysis: Exploratory factor analysis using principal components with varimax rotation was performed to evaluate the factor structure of the Chinese (HK) CLDQ. The criterion for factor extraction was an eigenvalue greater than one. The highest factor loading was identified for each item. The scree plot was also used to determine the number of factors.

Convergent validity: Construct validity was also tested by convergent validity determined by Spearman correlations between corresponding CLDQ and SF-36v2 Health Survey domain scores. It was hypothesized that moderate (r = 0.4 to 0.7) to strong (r > 0.7) correlations should exist between CLDQ FA and SF-36v2 VT; between CLDQ SS and SF-36v2 BP; between CLDQ AC and SF-36v2 PF, RP and RE; and between CLDQ EF and SF-36v2 MH scores.

The mean CLDQ scores were compared between two CHB patient groups, and the difference was tested by independent t to evaluate its sensitivity in detecting a difference between patients with complicated and uncomplicated infections. The sensitivity of the CLDQ was also assessed by the effect size (difference between group mean scores/overall standard deviation). According to Cohen[19], effect sizes of 0.3, 0.5 and 0.8 were considered small, medium and large differences, respectively. An effect size of less than 0.3 was considered not significant.

Different methods were used to assess reliability, including internal consistency and test-retest reliability. Internal consistency was measured by Cronbach’s α, which is a measure of the extent to which items in a questionnaire are homogeneous (correlated) in supporting the same concept[20]. Test-retest reliability refers to the stability of an instrument over time[21], which was measured by the intra-class correlation (ICC) between the two-week test-retest results. Reliability coefficients ≥ 0.7 and 0.9 are usually expected for group comparisons and individual comparisons, respectively[21].

All items except item 11 (level of energy) were found to be understood by 6 patients. Three out of six patients did not understand item 11. Five patients (83%) interpreted the meaning of all except four items (11, 13, 19 and 28) correctly. Four out of six patients misinterpreted the meaning of item 11 with three interpreting it as decreased physical strength. Two patients (33.3%) had difficulty in differentiating the meaning of “sleepy” and “drowsy”; and did not seem to have interpreted the words “mood swings” (item 19). Two out of six patients did not include the meaning of “worried about never feeling better” (item 28) in their interpretation. The Chinese (HK) translation was revised based on the results of cognitive debriefing and the revised questionnaire was then field tested on 23 CHB patients before this study. The final Chinese (HK) CLDQ was formed and its back-translation is shown in the appendix.

One hundred and eighty four CHB patients were identified; 6 patients were excluded (3 had hepatitis B infection less than 6 mo, 2 had communication problems and 1 had co-infection with HCV) and 28 patients refused to participate in this study. One hundred and fifty Chinese adults consisting of 72 uncomplicated (normal liver function defined as liver enzymes persistently within the normal range and without any history of cirrhosis or HCC) and 78 complicated (cirrhosis or HCC) completed the study, giving a response rate of 84.3% (150/178). Table 1 shows their characteristics, overall and by disease severity groups. There were 8 patients in the complicated CHB group who had HCC without any cirrhosis and had normal liver function. There were no statistical differences in demographics between the uncomplicated and complicated CHB groups, except age and sex (P < 0.001). Complicated CHB patients were older and there were more men than those in the uncomplicated group which was expected because CHB complications were more common in men than in women and the median age for the development of complications was 57.2 years[2223].

Table 2 shows the distribution of the Chinese (HK) CLDQ and SF-36v2 scores. There was practically no floor effect but there were significant ceiling effects in the Chinese (HK) CLDQ AS, AC and WO subscales, more so in the uncomplicated than the complicated group. Significant ceiling effects were also found in most SF-36v2 Health Survey scales. Sub-group analysis showed that the mean Chinese (HK) CLDQ scores were significantly lower in the complicated group than the uncomplicated group in all subscale and overall scores.

Figure 1 compares the distribution of the Chinese (HK) CLDQ scores with those from other countries. The distribution pattern of the Chinese (HK) CLDQ subscale scores was very similar to those of other countries[391113], except Italy, supporting cross-cultural conceptual equivalence.

Open in New Tab   Full Size Figure   Download Figure

Figure 1 The Chinese (HK) CLDQ scores by countries. AS: Abdominal Symptoms; FA: Fatigue; SS: Systemic Symptoms; AC: Activity; EF: Emotional Function; WO: Worry.

Table 3 shows the mean item scores and standard deviation of the 29 CLDQ items grouped under their hypothesized subscales. All correlations between items and their hypothesized subscales score exceeded the standard of 0.4.

All but six items had a higher correlation with its hypothesized subscale than other subscales, i.e. 100% scaling success. Four items of the SS subscale and two items of the AC subscale correlated more highly with some other subscales than their own. Scaling success was the lowest in item 3 “bodily pain”, which correlated more highly with four other subscales than with the SS subscale, with the highest found for EF, but the differences were not statistically significant. Items 6 “shortness of breath”, 23 “dry mouth”, 27 “itching”, 7 “not able to eat as much as you would like” and 9 “trouble in lifting or carrying heavy objects” correlated higher with one to three other subscales than its own, but the differences in the correlations were not statistically significant.

The overall scaling success rate on discriminant validity was 100% for four scales (AS, FA, EF and WO), but it was 73% for the AC subscale and 64% for the SS subscale.

Table 4 illustrates the rotated factor loadings between the 29 items and 6 factors with eigenvalue > 1. The six factors explained 70.1% of total variance. The factor loadings of the items were not entirely consistent with the scaling hypothesis. Items of the FA and SS subscales, except bodily pain (item 3), decreased strength (item 8) and decreased energy (item 11), seemed to load on the same factor (factor 3). Two FA subscale items (8 “decreased energy” and 11 “decreased strength”) loaded more strongly on AC than its hypothesized factor. A new factor (factor 6) was found with the highest loading from two items assessing sleep (items 16 and 20). The items of EF, WO, AS and AC subscales loaded nicely on their hypothesized factors.

Table 5 shows the correlations between the scores of the CLDQ and SF-36v2 Health Survey. As hypothesized, moderate to strong correlations were found between CLDQ FA and SF-36v2 VT scores; and between CLDQ SS and the SF-36v2 BP scores. The CLDQ AC score correlated significantly with all SF-36v2 Health Survey domain scores and the strongest was found with the SF-36v2 RP and SF scores. The CLDQ EF score correlated strongly not only with the SF-36v2 MH score but moderately with the SF-36v2 VT, RE, RP and GH scores.

As shown in Table 2, the CLDQ overall and subscale mean scores were all significantly higher in the uncomplicated than the complicated CHB group. The effect sizes of the group differences in the CLDQ scores all exceeded 0.4 (range 0.4-0.6). Only three of the eight SF-36v2 domain scores (PF, RP and SF) and the PCS score detected a significant difference between the uncomplicated and the complicated groups. However, the greatest effect size difference between the two groups was found in the SF-36v2 RP score.

Across all subscales, the Cronbach’s α coefficients of the internal consistency reliability were higher than the recommended value of 0.7 (Table 5). ICC coefficients measuring the two-week test-retest reliability exceeded 0.7 in all but the AS (0.58) and AC (0.66) subscales. The reliability coefficients were comparable to those of the SF-36v2 Health Survey.

Table 5 also shows that the correlations (range 0.50-0.87) between the CLDQ subscales were smaller than the subscale internal reliability coefficients for all subscales, showing that each subscale measures a distinct concept. The overall CLDQ scores correlated strongly with all CLDQ subscale scores.

Chronic hepatitis B (CHB) virus infection remains a global problem and a public health threat. CHB patients may suffer or die from liver-related complications, posing a threat to both mental and physical health, leading to impairment of quality of life.

Health-related quality of life (HRQOL) outcomes should supplement traditional clinical outcomes in the evaluation of the impact and the effectiveness of treatment for patients with CHB infection.

The Chronic Liver Disease Questionnaire (CLDQ) had been applied mainly to patients with hepatitis C virus infection in Western countries. This study was the first to show that a Chinese (Hong Kong) translation of the CLDQ was valid, reliable and sensitive for Southern Chinese patients with CHB infection. The CLDQ can be applied to millions of Southern Chinese CHB patients to evaluate their HRQOL. Some modifications might further improve its validity, reliability and sensitivity.

The Chinese (Hong Kong) CLDQ can be used to evaluate the impact of CHB virus infection and assess the effectiveness of anti-viral drug treatments in Cantonese-speaking Southern Chinese. The CLDQ can be used as a cross-cultural HRQOL measure in international studies that include Southern Chinese.

CHB virus infection refers to those who are hepatitis B surface antigen-positive for more than six months. Validity is defined as the extent to which a test measures what it is intended to measure. Reliability refers to the consistency or stability of the measurement process across time, patients or observers.

The authors validated and tested the psychometric properties of a Southern Chinese translation of the CLDQ and determined that their questionnaire was valid, reliable, and sensitive for southern Chinese patients with hepatitis B virus infection. The study was well done and used appropriate methodology to validate and test the questionnaire.