Presentation (CTI)

OBJECTIVES: EQ-5D is currently the main tool used for measuring health-related quality of life in health technology assessments in Europe and beyond. A systematic literature review from 2015 found that the EQ-5D-3L was responsive in fewer than half of the conditions studied (45%), while evidence was inconclusive or absent in the rest (Payakachat, et al., 2015). Given the recent introduction of the EQ-5D-5L version into NICE UK guidelines, and the time elapsed since the previous study, we conducted a new search during the spring of 2026 to provide a more updated picture of EQ-5D's reliability for decision making across disease areas.
METHODS: We searched the Medline and Embase databases to identify studies published in English from January, 2014 to May 2026 using keywords that were related to EQ-5D-5L responsiveness. Abstracts and full texts were screened according to inclusion criteria. We summarized responsiveness evidence across nine disease areas using reported standardized response mean (SRM), effect size (ES), area under the receiver-operating characteristic curve (AUROC). Cohen’s categories were used to identify small (<0.5), moderate (0.5 - 0.79), and large (≥0.8) responsiveness for SRM and ES. AUROC values were interpreted as small (< 0.59), moderate (0.6 - 0.69) and large (0.7 - 1.0) responsiveness.
RESULTS: 1 384 records were retrieved, of which 398 were duplicates, resulting in 986 screened at the abstract review. 70 publications were screened at full-text review, with 51 publications meeting the inclusion criteria. Evidence for EQ-5D-5L responsiveness was strongest after clear clinical change, such as surgery or acute recovery, whereas weaker responsiveness was documented in mental health, neurology, oncology, multimorbidity and for carergivers.
CONCLUSIONS: EQ-5D-5L appears sufficiently responsive in most disease areas but should be complemented with other measures in disease areas with documented low responsiveness to better inform medical decision making.