OBJECTIVES: The EQ-5D has separate versions for children/adolescents and adults, covering the same health dimensions but with different wording. This study aims to compare the measurement performance of the EQ-5D-3L (adult version) and EQ-5D-Y-3L (youth version) in an adult general population sample.

METHODS: Two independent surveys, a computer-assisted personal interview (n=200) and an online questionnaire (n=996) were conducted in 2021, representative of the Hungarian adult general population by gender and age groups. We compared the following measurement properties: ceiling, floor, relative informativity (Shannon’s evenness index) and agreement (cross-tabulation and Kendall’s rank correlation). Known groups were defined based on general health status and chronic health conditions, with validity being tested for mean level sum scores (LSS) and utility values (using Hungarian value sets).

RESULTS: Almost twice as many different health state profiles occurred in the EQ-5D-Y-3L compared to the EQ-5D-3L (85 vs. 47 out of 243), with 59.0% of respondents reporting identical health profiles on both measures. Full health was less frequently reported on the EQ-5D-Y-3L (34.8%) compared to EQ-5D-3L (46.8%). Ceiling was significantly lower for all corresponding dimensions of the EQ-5D-Y-3L except for self-care, with the largest difference in EQ-5D-Y-3L worried/sad/unhappy (56.8%) vs. EQ-5D-3L anxiety/depression (71.6%). All EQ-5D-Y-3L dimensions (0.20-0.75) demonstrated better relative informativity than those of EQ-5D-3L (0.18-0.66). Strong agreement was found between corresponding dimensions with Kendall’s tau coefficients ranging from 0.636 (worried/sad/unhappy vs. anxiety/depression) to 0.841 (mobility). The two instruments were similar in terms of known-groups validity. Notably, EQ-5D-3L LSSs performed better in differentiating between respondents with and without physician-diagnosed anxiety.

CONCLUSIONS: This study highlights differences in measurement performance between EQ-5D-Y-3L and EQ-5D-3L in an adult general population sample, with the largest differences occurring between worried/sad/unhappy vs. anxiety/depression dimensions. Although EQ-5D-3L is still recommended for adult use, EQ-5D-Y-3L also seems to be a suitable alternative, offering valuable information for assessing mental health.