A preference-based index for the SF-12

Background: The SF-12 is a widely used generic measure of subjective health. As the scoring algorithms of the SF-12 do not include preference values, different approaches to assign a preference-based index are available that should be tested regarding their feasibility and validity. Objectives: To develop a concept for a preference-based index for the SF-12 on the basis of multi-attribute decision analysis and to perform initial tests of its feasibility and validity in an empirical study. Methods: A multi-attribute preference function for the SF-12 was developed, estimated and tested for validity. Two mail surveys (n = 100, 200) and an interview (n = 72) were conducted with women who had an operation for breast cancer. Visual analogue scale (VAS) and standard gamble (SG) measures elicited preference-based valuations. Results: Eight attributes were identified in the SF-12. Validity tests showed an average difference of 8 VAS score points between directly measured and predicted values for given health states. Conclusion: The initial results show that this approach might allow the direct assignment of a p reference-based valuation to the SF-12. The quality of the psychometric features of the multi-attribute value function is encouraging. Future studies should test this concept more extensively, especially by determining parameters for a representative sample of the general population and by comparing performance with other approaches to value the SF-12. Copyright (c) 2005 John Wiley & Sons, Ltd.