Comparing the properties of the EQ-5D-5L and EQ-5D-3L in general population in Iran

Little information exists on properties of EQ-5D-5L (the 5L hereafter) compared to the EQ-5D-3L (the 3L hereafter) measures in the context of Iran. This study aims to compare the measurement properties of the two versions of the EQ-5D (the 3L vs. the 5L) using data obtained from general population in Iran.

A total of 886 adults aged 18 years and above from September to November 2020 participated in this cross-sectional analysis. The required data collected using self-administered and-standard questionnaire and multistage sampling method was used to select the samples. The 3L and 5L measures compared in terms of celling effect, distribution and redistribution pattern, feasibility, convergent validity, know-groups validity and informativity.

From September to November 2020, 886 adults (mean aged = 44.6 years; 55% male and 87.1% married) included in the study. The study indicated that the 5L had lower celling effects compared to the 3L (45% vs. 46%). A better convergent validity and known-groups validity was found for the 5L version compared to the 3L and significantly stronger association found between the 5L measure with both the Visual Analogue Scale and the 5-point health status scale. The 5L index score showed higher relative efficiency (RE) in 9 of 11 condition (mean RE = 1.36). Compared to the 3L, the 5L classification system had higher Shannon index (H') in all dimensions: mobility (0.52 vs. 0.40), self-care (0.23 vs. 0.20), usual activities (0.61 vs. 0.47), pain/discomfort (1.19 vs. 0.89) and anxiety/depression (1.22 vs. 0.47).

The study demonstrated that the measurement properties of 5L version in terms of celling effects, convergent validity, known-groups validity, RE and informativity similar or better than the 3L among general population; suggesting the use of 5L in the context of Iran. Hence, we suggested the use of the 5L in economic evaluation, clinical and public health studies in Iran.