Administration-order effects in head-to-head comparisons of the EQ-5D-3L and EQ-5D-5L descriptive systems: evidence from a randomised general-population panel in Great Britain.
Ling Jie Cheng, Xun Li
✦ AI-curated · Sources linked
In 30 seconds
This randomised study in a general-population panel in Great Britain examined whether administration order affects comparisons of the EQ-5D-3L and EQ-5D-5L. When each version was presented early, the full-health ceiling was 52.1% for the EQ-5D-3L and 35.4% for the EQ-5D-5L (difference 16.7% points, 95% CI 12.2 to 21.3); the 5L was more informative on every dimension. Prior exposure to the 3L, rather than later questionnaire position alone, substantially increased the 5L ceiling, consistent with response contamination.
Key findings
- In the early position, full-health ceilings were 52.1% for the EQ-5D-3L and 35.4% for the EQ-5D-5L (difference 16.7% points, 95% CI 12.2 to 21.3).
- The EQ-5D-5L was more informative than the EQ-5D-3L on every dimension when each was answered early.
- Moving the 5L to a late position without prior 3L exposure changed its ceiling by 2.5 points (95% CI -2.0 to 7.0); prior 3L exposure raised it by 12.8 points (8.2 to 17.4).
- Among interviewer-present respondents, the 5L ceiling was 34.5% without prior 3L, 43.0% after unnoticed repetition, and 53.2% when repetition was noticed.
Why it matters
Within-questionnaire comparisons may not reflect how each instrument performs independently: responses to the first version can influence answers to the second. Here, prior 3L exposure raised the 5L full-health ceiling, potentially narrowing the apparent difference between versions. Recognising this source of measurement bias is important when interpreting comparative instrument studies and their reported response distributions.
What to keep in mind
The abstract reports results from a general-population panel in Great Britain; it does not establish whether the order effects generalise to other populations or settings.