OBJECTIVES: Studies have shown that preference-based measures of health-related quality oflife (utility measures) fail to provide interchangeable values, which raises concerns for the cross-study comparability of cost-effectiveness estimates. This study offers generalizable and condition-specific insight into why (rather than if) there are discrepancies between two widely used measures, the EQ-5D and SF-6D. METHODS: Comparisons focused on practical considerations and the respective descriptive and valuation components of the measures, addressing empirical and conceptual issues. More specifically, we addressed instrument-completion, item-completion, contextual framing of questions, dimension-to-dimension correlations, floor and ceiling effects, and construct validity. Data came from randomized controlled trial participants with nonspecific neck pain (n = 346). RESULTS: The descriptive classification systems do not permit respondents to describe their health state in the same manner, due, primarily, to contextual differences and the number of response options. Specific to neck pain populations, "vitality" was a unique contributor to the SF-6D, although both measures identified the same significant linear trends across theoretical constructs. Rates of instrument completion were significantly better for the EQ-5D. CONCLUSIONS: The EQ-5D and SF-6D do not provide interchangeable utility estimates for patients with nonspecific neck pain-a finding that is common to other clinical areas. However, this result, and the results from previous studies, should not be surprising given the extent of between-measure differences relating to the descriptive content of health dimensions across the two measures. Given the consistent messages emerging from method comparison studies for the EQ-5D and SF-6D, new and/or novel approaches are necessary to drive this research area forward.