Comparing the psychometric properties of the EQ-5D-3L and EQ-5D-5L in cancer patients in Korea

This study was conducted to assess the redistribution properties of the EQ-5D-3L when using the EQ-5D-5L and to compare the
validity, informativity, and reliability of both EQ-5D-3L and EQ-5D-5L in Korean cancer patients.

Patients visiting one ambulatory cancer center self-administered the two versions of the EQ-5D and the EORTC QLQ-C30 questionnaire.
Redistribution properties in each dimension of EQ-5D were analyzed between EQ-5D-3L and EQ-5D-5L. Informativity was evaluated
using the Shannon entropy and ceiling effect. Convergent validity was evaluated by comparing the EQ-VAS, ECOG performance
status, and EORTC QLQ-C30 subscales. Reliability was also evaluated in terms of test–retest reliability.

All levels of the EQ-5D-3L substantially partitioned into associated levels of the EQ-5D-5L. The average inconsistency rate
of the two versions was 3.5%. Absolute informativity was higher for the EQ-5D-5L than for the EQ-5D-3L, but their informative
efficiency tended to be similar. The proportion of ‘perfect health’ (11111) decreased from 16.8% in the EQ-5D-3L to 9.7% in
the EQ-5D-5L. EQ-5D-5L demonstrated similar or higher correlations with the EQ-VAS, ECOG performance status, and EORTC QLQ-C30,
than the EQ-5D-3L. The intraclass correlation coefficient of the EQ-5D-5L index was 0.77.

The EQ-5D-5L had greater informativity and lower rate in the ceiling effect than those values of the EQ-5D-3L. The EQ-5D-5L
showed good construct validity and reasonable reliability. Therefore, considering these findings, the EQ-5D-5L may be preferable
to the EQ-5D-3L.