COMPARING THE PSYCHOMETRIC PERFORMANCE OF FOUR FATIGUE-RELATED EQ-5D-5L BOLT-ONS IN A MULTINATIONAL, LONGITUDINAL GENERAL POPULATION SETTING
Author(s)
Febe Brackx, MSc1, Fanni Rencz, PhD2, Sarah Dewilde, PhD1.
1SHE Bv, Brussels, Belgium, 2Department of Health Policy, Corvinus University of Budapest, Budapest, Hungary.
1SHE Bv, Brussels, Belgium, 2Department of Health Policy, Corvinus University of Budapest, Budapest, Hungary.
OBJECTIVES: EuroQol is developing the experimental EQ-5D Bolt-on Toolbox, including a fatigue bolt-on. We assessed the psychometric performance of four fatigue-related EQ-5D-5L bolt-ons: Tiredness, Energy, Fatigue and Exhaustion.
METHODS: We used data from GENESIS, a multinational survey of 10,500 adults representative of the general population. A subset was reassessed after two months (n=3,276 in total). Data collected included the EQ-5D-5L, four fatigue-related bolt-ons, FACIT-Fatigue, and the Rasch-Transformed Fatigue Severity Scale (RTF-SS). Participants also ranked the bolt-ons by perceived relevance to their vitality. Psychometric evaluation comprised distributional properties (ceiling), divergent and convergent validity, structural validity (CFA and IRT), test-retest reliability and responsiveness. Differential item functioning (DIF) was tested across languages.
RESULTS: In the relevance ranking exercise, Fatigue was perceived as most relevant (29.5%), whereas Tiredness was least relevant (20.9%). Overall, 70.8%, 58.6%, 51.4%, and 49.8% of respondents reported problems on the Tiredness, Fatigue, Exhaustion, and Energy bolt-ons, respectively. Each bolt-on reduced EQ-5D-5L ceiling (34.4%) to 18.6%-25.3%. Convergent validity was demonstrated through moderate correlations with RTF-SS (r=0.52-0.58) and strong correlations with FACIT-Fatigue (r=0.68-0.70). Divergent validity was supported by weak or moderate correlations with core EQ-5D-5L dimensions (r=0.26-0.52), with similar results across bolt-ons. CFA supported a unidimensional structure, suggesting the four bolt-ons measure the same underlying trait, with loadings from 0.87 (Energy) to 0.94 (Fatigue). IRT (graded response model) showed high discrimination for all bolt-ons, with parameters ranging from 3.33 (Energy) to 5.14 (Fatigue). Threshold parameters were ordered for all bolt-ons, indicating appropriate response category functioning. Test-retest reliability was adequate (level sum score ICC=0.80-0.81). Large Uniform DIF across languages was detected for Energy, Fatigue, and Exhaustion. All bolt-ons improved responsiveness.
CONCLUSIONS: All four fatigue-related bolt-ons demonstrated strong psychometric properties and improved the psychometric performance of the EQ-5D-5L. Their performance was broadly similar, although relative performance varied according to the measurement property assessed.
METHODS: We used data from GENESIS, a multinational survey of 10,500 adults representative of the general population. A subset was reassessed after two months (n=3,276 in total). Data collected included the EQ-5D-5L, four fatigue-related bolt-ons, FACIT-Fatigue, and the Rasch-Transformed Fatigue Severity Scale (RTF-SS). Participants also ranked the bolt-ons by perceived relevance to their vitality. Psychometric evaluation comprised distributional properties (ceiling), divergent and convergent validity, structural validity (CFA and IRT), test-retest reliability and responsiveness. Differential item functioning (DIF) was tested across languages.
RESULTS: In the relevance ranking exercise, Fatigue was perceived as most relevant (29.5%), whereas Tiredness was least relevant (20.9%). Overall, 70.8%, 58.6%, 51.4%, and 49.8% of respondents reported problems on the Tiredness, Fatigue, Exhaustion, and Energy bolt-ons, respectively. Each bolt-on reduced EQ-5D-5L ceiling (34.4%) to 18.6%-25.3%. Convergent validity was demonstrated through moderate correlations with RTF-SS (r=0.52-0.58) and strong correlations with FACIT-Fatigue (r=0.68-0.70). Divergent validity was supported by weak or moderate correlations with core EQ-5D-5L dimensions (r=0.26-0.52), with similar results across bolt-ons. CFA supported a unidimensional structure, suggesting the four bolt-ons measure the same underlying trait, with loadings from 0.87 (Energy) to 0.94 (Fatigue). IRT (graded response model) showed high discrimination for all bolt-ons, with parameters ranging from 3.33 (Energy) to 5.14 (Fatigue). Threshold parameters were ordered for all bolt-ons, indicating appropriate response category functioning. Test-retest reliability was adequate (level sum score ICC=0.80-0.81). Large Uniform DIF across languages was detected for Energy, Fatigue, and Exhaustion. All bolt-ons improved responsiveness.
CONCLUSIONS: All four fatigue-related bolt-ons demonstrated strong psychometric properties and improved the psychometric performance of the EQ-5D-5L. Their performance was broadly similar, although relative performance varied according to the measurement property assessed.
Conference/Value in Health Info
2026-11, ISPOR Europe 2026, Vienna, Austria
Value in Health, Volume 29, Issue 12S
Code
PCR264
Topic
Methodological & Statistical Research, Patient-Centered Research, Real World Data & Information Systems
Topic Subcategory
Health State Utilities, Instrument Development, Validation, & Translation, Patient-reported Outcomes & Quality of Life Outcomes
Disease
No Additional Disease & Conditions/Specialized Treatment Areas