OBJECTIVES: Cost-effectiveness analysis has been recommended by national agencies in the UK and USA. The National Institute of Health and Clinical Excellence (NICE) supports the use of generic health-related quality of life instruments such as EuroQol EQ-5D when quality-adjusted life years (QALYs) are used to measure health benefits. Despite the urgent need for appropriate methodologies to support decision-making processes and to improve the use of scare resources in the healthcare sector in Latin American countries, little is know about how health is valued. This study reports on the construction of a set of social preference weights for use in Latin America. METHODS: A national population survey was conducted in the United States in 2002, based on a sample of 1603 Whites and 1115 Hispanics. Participants provided TTO utilities for a subset of 43 EQ-5D health states. Hispanic respondents were grouped according to their language preferences (Spanish or English speaking). Mean utilities were compared for each health state. A random effects model was used to determine whether real population differences exist after adjusting for socio-demographic characteristics. A population value set for all 243 EQ-5D health states was developed using only the data from Spanish-speaking Hispanics. RESULTS: Mean valuations differed slightly between Whites and English-speaking Hispanics. However, Spanish-speaking Hispanics tended to give higher valuations when compared to Whites (P<0.05) corresponding to an average 0.034 point difference on a 0-1 scale. The regression model was developed for Spanish-speaking Hispanics with a mean absolute error of 0.031. CONCLUSION: The availability of a Hispanic model for health state valuations for EQ-5D is a breakthrough in providing decision-makers with a set of social preference weights that may be applicable in Latin American countries that presently lack their own domestic value set.