OBJECTIVES: To estimate and compare utility values obtained from the Spanish versions of SF-6D and EQ-5D in four different groups of patients, and to compare with those estimated using UK tariffs to test validity of the Spanish SF-6D tariffs. METHODS: Four groups of patients were recruited: 2.550 with Gastroesophageal Reflux Disease (GERD), 220 patients on Hemodialysis (HD) therapy, 214 HD-patient caregivers, and 501 suffering Generalized Anxiety Disorder (GAD). The Spanish versions of SF-6D and EQ-5D, and a quality of life visual analogue scale (VAS) were self-administered. Utility mean values were compared using a 2 way 4x3 ANOVA with repeated measures for the instrument factor. Sidak corrections for multiple comparisons were applied. RESULTS: Statistical interaction was found between the kind of measurement and the pathology group (F6,6880=12.7, p<0,001). Mean utility estimates using UK tariffs ( =0.690, SE=0.004) were significantly lower than those obtained with the SF6-D for Spanish population ( =0.727, SE=0.005) which in turn were lower from those obtained with the EQ-5D for Spanish population ( =0.744, SE=0.006). Spanish SF-6D did not detect significant differences between HD-caregivers and GERD patients (p=0.996) neither between GAD and HD patients (p=0.987). EQ-5D was not sensitive to differences between GAD and HD patients (p=0,899). UK SF-6D was not sensitive to differences between Caregivers and GERD patients (p=0.161). Correlations between Spanish SF-6D and EQ-5D scores ranged from 0.697 (Caregivers) to 0.808 (GERD). VAS measures were capable to detect the same differences as Spanish SF-6D although GAD patients showed better scores than HD patients. CONCLUSION: As it could be expected, groups of patients with varying degrees of severity present different subpopulation utility mean values. Nevertheless not all instruments used to estimate utilities give identical results. A special effort to gather culturally adjusted instrument specific norms for different subpopulations should be encouraged.