Comparison of the EQ-5D-5L and the EQ-5D-3L using individual patient data from the REFORM trial

Abstract

Background: This study compares the 5-level version of the EQ-5D (5L) with the 3-level version EQ-5D (3L) in older adults using individual patient data from the REFORM (REducing Falls with Orthoses and a Multifaceted podiatry intervention) trial. Methods: EQ-5D-5L and EQ-5D-3L were administered to men and women (n=151) over the age of 65 years alongside the REFORM trial. The two versions of the EQ-5D were assessed in terms of feasibility, level of consistency, ceiling effect and discriminatory power. We also undertook a comparison of the performance of different EQ-5D-3L and EQ-5D-5L value sets. Results: The proportion of participants that returned a complete questionnaire was higher for the 5L (96.7%) than for the 3L (92.7%). Missing values among dimensions were on average 1.59% (5L) and 1.45% (3L). The ceiling effect was reduced from 18.2% (3L) to 6% (5L). On average the proportion of inconsistent responses between both descriptive systems was 3.25%. Redistribution from 3L to 5L showed valid results for the majority of consistent level combinations, with slight inconsistency in the case of Anxiety/Depression. For the 5L, 67 unique health states were observed for the 5L compared to 27 for the 3L. The absolute informatively improved with the new classification system (5.48 for 5L versus 3.91 for 3L) and relative discriminatory power improved slightly on average (0.90 for 5L versus 0.84 for 3L). The mean difference between the EQ-5D-5L and EQ-5D-3L values was 0.091 (range -0.345 to 0.505); whilst the mean difference between the EQ-5D-5L and the crosswalk values was 0.082 (range -0.035 to 0.293). Conclusion: In the REFORM clinical trial involving an elderly population, our study supported the feasibility and convergent validity of both EQ-5D-3L and EQ-5D-5L. Results suggest that the 5L improves the ceiling effect and discriminatory power. The EQ-5D-5L scores were significantly higher than both EQ-5D-3L and crosswalk.