OBJECTIVES: Recently, a new index score to measure health status has been developed based on items from the SF36, the so-called SF6D. To assess its merits and validity we compared baseline and change statistics of the SF6D and the EQ5D in a group of patients undergoing coronary revascularisation. METHODS: A consecutive cohort of 535 patients enrolled in a randomised controlled trial comparing off-pump coronary artery bypass surgery (CABG) with on-pump CABG and percutaneous transluminalcoronary angioplasty (PTCA) was administered the SF-36 and the EQ5D questionnaires pre- and post-intervention. The SF6D "utility" was computed according to the scoring system that recently became available. RESULTS: We observed no differences in change score across the arms of trials. Therefore, data on all patients were analysed together. There were more missing in the SF6D than in the EQ5D: 12.7% versus 4.7%. All baseline and change variables had non-normal distributions. Although baseline mean values were comparable (SF6D 0.625; EQ5D 0.645, t-test: p=0.09) median values differed significantly (SF6D 0.603; EQ5D 0.691, Wilcoxon matched pairs test: p=0.0009). Median change from baseline as measured with the SF6D was significantly lower than change as measured with the EQ5D (0.024 versus 0.052, p=0.00002). CONCLUSIONS: The SF6D and the EQ5D appear to result in similar baseline scores using the recommended mean values. However, both the median baseline values and change scores differed significantly. This indicates that the SF6D and the EQ5D measure different concepts of health status and can not be used as equivalents.