OBJECTIVES The guideline on economic evaluations from the Swedish price and reimbursement authority (TLV) states that experience-based valuation of QALY weights are preferred before a hypothetical valuation. The UK hypothetical EQ-5D valuation set [Dolan 1997] has been used so far, but since 2013 an experience-based “tariff” exists for the general population in Sweden [Burström 2013]. This study explores the implications on ICERs by applying the two different value sets. METHODS FACT-P mapped EQ-5D responses from patients with mCRPC in the AFFIRM trial [Skaltsa 2013; Scher 2012] were converted into utility weights using both the hypothetical and experience-based value sets. A Markov cohort cost-utility model (with stable, progressed and dead health states) analyzing enzalutamide (Xtandi®) vs. best supportive care (BSC) was used for studying the implications of applying the different utility weights. RESULTS As the experience-based value set had a more compressed event space, and was therefore less sensitive to QoL changes, the “on treatment” utility for Xtandi was 0.03 vs. 0.06 (hypothetical). The stable and progressed health states attained utility weights of:  0.688 and 0.603, respectively (hypothetical); 0.826 and 0.784, respectively (experienced). The survival gain with Xtandi vs. BSC in combination with the lower hypothetical health state utility weights resulted in a QALY gain of 0.66 vs. 0.75 with the experienced weights. With an incremental cost of €51,100 (€1=9SEK) the resulting ICER was €77, 600 and €68,200 using the hypothetical and experience-based value sets, respectively. CONCLUSIONS The Swedish experience-based EQ-5D value set generates a greater absolute utility but with a more compressed event space compared to the UK hypothetical value set. In terms of ICERs, this tends to favour technologies that extend survival compared to QoL improving technologies, although Xtandi provides both. These implications exert great challenge on Swedish decision makers on how the Swedish value set should be implemented.