OBJECTIVES To evaluate the relationship of self reported general health [GH] on the EQ-5D preference scores (time trade-off [TTO]) as well as its independent influence when demographic, socioeconomic, and risk factors are included. METHODS 41,392 adults from the first general population survey from a nationally representative sample were included: National Risk Factor [RF] Survey (2005). Variables included GH, sociodemographic parameters (SD) and RF (diabetes [DBT], high blood pressure [HBP], dyslipemia [DLP], obesity [OB], smoking [SMK], healthy diet [HD], physical activity [PA]), and the EQ-5D instrument. EQ-5D profiles were mapped to TTO preference values using a local EQ-5D Value study. We assessed the independent relationships between EQ-5D scores with GH, SD, and RF. We used multivariable linear regression modeling. RESULTS Final models included 33,964 subjects (representing 17,586,759 subjects). Before including GH the following variables were statistically, consistently, and clinically associated with EQ-5D scores: age, gender, income, selected provinces, health coverage, DBT, HBP, DLP, PA,OB (R2 0.16). Nevertheless, when GH was incorporated, it showed to be the most important explanatory variable (R2 0.35) and most health variables lost their association (DBT, HBP, DLP, OB). Moreover, a significant interaction was found in relation with age (>50), gender and income. Although GH was still significant in all the eight subgroups defined by the three variables with positive interactions, the magnitude of the coefficients corresponding with each category of GH significantly varied (subgroup numbers ranged between 1501 and 8637, and R2from 0.16 to 0.46). CONCLUSIONS In Argentina, GH is the strongest parameter associated with scores on the EQ-5D. The magnitude of this association varied according to age, gender an income level. Based on these findings, we suggest that the influence of GH on individuals' preferences should be evaluated by subgroup since its relevance seems to vary according to selected individual's attributes.