The discriminative power of the EuroQol visual analog scale is sensitive to survey language in Singapore

BACKGROUND Existing evidence for validity of the visual analog scale of the EQ-5D-3L questionnaire (EQ-VAS) is weak in Chinese-speaking respondents in Singapore. We therefore investigated the validity of the Chinese (Singapore) version of EQ-VAS in patients with diabetes. METHODS In a cross-sectional survey, patients with type 2 diabetes seen in a primary care facility completed an identical Chinese or English questionnaire containing the EQ-5D-3L and questions assessing other health and disease-related characteristics. Convergent and known-groups validity of the EQ-VAS was examined for Chinese- and English-speaking respondents separately. RESULTS The EQ-VAS was correlated with the EQ-5D-3L health index and a 5-point Likert-type scale for assessing global health in both Chinese-speaking (N = 335) and English-speaking respondents (N = 298), suggesting convergent validity. The mean EQ-VAS scores differed between English-speaking patients with differing duration of diabetes (< 10 years versus ≥ 10 years), comorbidity status (absence versus presence), and complications of diabetes (absence versus presence), providing evidence for known-groups validity. However, the EQ-VAS scores for Chinese-speaking respondents known to differ in these characteristics were similar, even among subgroups of relatively younger patients or those with formal school education. CONCLUSIONS Chinese- and English-speaking Singaporeans respond differently to the EQ-VAS. The Chinese version of EQ-VAS appears less sensitive than its English version for measuring global health in patient populations in Singapore.