OBJECTIVES: Evidence on the validity of the generic, preference-based EQ-5D questionnaire in advanced melanoma is limited. Therefore, we evaluated the construct and content validity of both the EQ-5D-3L and EQ-5D-5L using data from the nationwide Dutch Melanoma Treatment Registry.

METHODS: Data on the EQ-5D and Functional Assessment of Cancer Therapy-Melanoma (FACT-M; a disease-specific questionnaire) were obtained for patients who completed the EQ-5D-3L (between 2012 and 2015) or the EQ-5D-5L (from 2016) at baseline (i.e., after their diagnosis). EQ-5D utility values were computed using Dutch value sets. Content validity was assessed comparing the proportions of ceiling effect at baseline of the EQ-5D-3L and EQ-5D-5L in both the overall scores and in each dimension as well as the FACT-M scores of the same patients. Spearman’s correlation was used to assess construct validity, comparing the strength of the relationship between the dimensions of the two version of the EQ-5D and the dimensions of the FACT-M for these patients at baseline.

RESULTS: At baseline, 416 patients filled in the EQ-5D-3L and 668 patients the EQ-5D-5L. The proportion of patient reporting perfect health was higher in the EQ-5D-3L (35.8% vs. 29.9%), while in both versions the ceiling effect was most obvious in the dimension of “Self-care”. Of these patients, approximately 100% that filled in the EQ-5D-3L and 60% that filled the 5L version, also filled in the FACT-M disease specific questionnaire for melanoma (score ranging 0-172). FACT-M baseline scores for these patients were higher (better health) and ranged less in patients that filled in the EQ-5D-5L (mean±SD: 153.92±9.57 versus 146.77±11.43). The Spearman’s coefficients, while following the same trends, showed more pronounced relationships between EQ-5D and FACT-M dimensions, in the 5L version.

CONCLUSIONS: The EQ-5D-5L showed better content and construct validity than the EQ-5D-3L in patients with advanced melanoma.