BACKGROUND/PURPOSE: We know of no validated Taiwanese-language instrument to measure a utility of the patient's health. Our aim was to evaluate the reliability and validity of a Taiwanese version of the EuroQol instrument (EQ-5D) in a Taiwanese population. METHODS: Questionnaires containing the Taiwanese versions of the EQ-5D and the Short-Form 12 Health Survey (SF-12) were sent to 12,923 people in Taiwan in December 2002. Concurrent validity of the EQ-5D was analyzed by assuming that subjects with problems in any EQ-5D dimensions had decreased SF-12 scores. Discriminant validity of the EQ-5D was analyzed by assuming that subjects with the following characteristics had lowered EQ-5D indexes and scores on the EQ-5D visual analog scale (VAS): more chronic diseases than others, serious illness, more hospitalizations in the past year than others, poor general health, and more outpatient visits than others. Test-retest reliability was analyzed in a subgroup of respondents who were evaluated twice within a month by using the intraclass correlation coefficient and the kappa method. RESULTS: The general survey response rate was 12.7% (1644 of 12,923). SF-12 scores were lower in subjects reporting problems on EQ-5D dimensions than in others without such problems (p < 0.01). Subjects with more health problems than others had lower EQ-5D indexes and VAS scores (p < 0.01). The physical dimension of the EQ-5D was more strongly correlated with the SF-12 Physical Component Summary than with the Mental Component Summary; this finding satisfied the a priori hypothesis. For test-retest reliability of items on the EQ-5D, kappa values ranged from 0.49 to 1 (p < 0.001). CONCLUSION: The Taiwanese EQ-5D instrument appears to be a moderately valid and reliable tool for measuring the health status of the general population in Taiwan.