OBJECTIVES: This study aims to examine whether the test-retest reliability of SF-36, WHOQOL-BREF, and EQ-5D questionnaires will be changed by different administration approaches by which patients with chronic liver disease self-report their quality of life. METHODS: Patients with chronic liver disease were recruited from the outpatient department of a medical centre in Taiwan. Their self-reported questionnaires were collected by two approaches. The first approach patients received was an interview and questionnaire in hospital. They returned the retest questionnaire by mail two weeks later. In the second approach, patients were instructed to fill out both test and retest questionnaires at home and send back by two separate mails. The time gap was also two weeks. After scoring questionnaires, a paired-t test was conducted to compare test-retest reliability for three questionnaires. The mean score difference between two approaches was examined by independent t test. An analyses of mean score differences of different domains were performed by multiple linear regressions. RESULTS: Of 69 patients recruited for the first approach; 52 persons completed both questionnaires (75%), while the response rate of the second approach was 84% (127 of 151). The results indicate that there is no statistically significant difference in the test-retest reliability of SF-36, WHOQOL-BREF, and EQ-5D questionnaires. There was also no significant difference in the test-retest results between two approaches, except in the dimension ‘pain/discomfort' mean difference (0.3 ± 1.2 and -0.1 ± 1.1, p=0.03) by EQ-5D. Similar results (p=0.04) were also found by multiple linear regression, after controlling age, sex, and education. This reflects that greater pain/discomfort was more likely to present in the first approach as compared to that in the second one. CONCLUSIONS: Alternative administration approaches did affect the results of test-retest questionnaires, which idicated that the Hawthorne Effect occurred in the interview in hospital.