OBJECTIVES: Different subgroups of normal populations may have different perceptions of health. In a valuation studies of the EQ-5D, these different preferences are averaged to produce the national value set. In this study we try to identify the determinants of the different health states preferences in the national EQ-5D-5L valuation in Indonesia. METHODS: Using composite time-trade off (C-TTO), each respondent provided values of 10 out of 86 hypothetical health states with computer-assisted face-to-face interviews. A multivariable random-effects linear regression model, adjusted for severity of health states, was used to determine the impact of the socio-demographic: i.e. gender, age, level of education, residence, data collection location, religion, ethnicity, monthly income, experience with illness and self-rated health on the C-TTO scores. RESULTS: There were 1054 respondents with demographic characteristics representative for the Indonesian population. In the models we found that being male [unstandardized regression coefficient (b) = 0.043], had middle and higher education (b = -0.063 &-0.091, respectively), being Sundanese/Sulawesi/others as ethnicity (b = 0.069, 0.078 & 0.035), and having 2500-5000 K IDR (± 180 - 365 USD) monthly income (b =0.087) were significantly associated with C-TTO scores. Concerning location of data collection, respondents interviewed in the Jogjakarta location (b = -0.087) and Medan (b = -0.192), had significantly lower C-TTO scores than those from Jakarta, while scores of the respondents from Surabaya (b =0.279) were the opposite direction. However, those interviewed in Makassar showed no significant differences with Jakarta respondents. CONCLUSIONS: Different socio-demographic characteristics groups had different preferences for EQ-5D-5L health states in this study.