A Comparison of the SF-6Dv2 and SF-6D UK Utility Values in a Mixed Patient and Healthy Population

Background This paper describes the first evaluation of the construct validity and performance of the newly developed preference-based measure of health, the SF-6D version 2 (SF-6Dv2). Method Utilising data from the Multi-Instrument Comparison (MIC) project (n = 7932), we explored the descriptive differences in utility values between the SF-6Dv2 and the SF-6D and evaluated the known group validity of both measures by testing the statistical significance of differences in utility values and calculating the effect sizes across known groups. The convergent validity of the SF-6Dv2 was explored by examining whether the SF-6Dv2 is related to alternative validated measures, including the EQ-5D-5L and AQoL-8D. Results Differences between the utility values of the SF-6Dv2 and SF-6D were evident; utilities were generally lower for the SF-6Dv2, with larger standard deviations resulting in larger absolute differences between groups. The SF-6Dv2 performed well in terms of known-group validity and successfully distinguished disease severity and between the disease and healthy groups, outperforming the SF-6D in some but not all groups. Convergent validity analyses indicated strong associations between the SF-6Dv2 and the SF-6D, EQ-5D-5L and AQoL-8D utilities. Conclusions The psychometric performance of the SF-6Dv2 is favourable with respect to known group validity and convergent validity, but does not seem to have improved, compared with the SF-6D. However, the new method of valuation has had a substantial impact on the size of absolute differences in utility values, which could impact quality-adjusted life-year results. The economic evaluation of health interventions may therefore be influenced by the choice of the SF-6Dv2 over the SF-6D.