Background The EQ-5D, a 5-dimensional multi-attribute questionnaire, and the SF-6D, derived from the SF36, are 2 indirect utility measures widely used to calculate Quality-adjusted life-years (QALYs) in order to assess health benefits.
Objectives 1) To assess which variables at baseline are associated with EQ-5D and SF-6D in early arthritis 2) To assess which variables at baseline are associated with substantial utility difference between EQ-5D and SF-6D
Methods - Patients: included in the French nationwide cohort of early arthritis ESPOIR (at least 2 swollen joints for less than 6 months and suspicion of RA). - Data available: EQ-5D and SF-6D utility measures were assessed in 813 patients with EA. Bio-clinical variables and X-rays were also recorded. - Analysis: The determinants of EQ-5D and SF-6D at baseline were assessed by multivariate linear regression. Multivariate linear and logistic regressions were used to determine which specific aspects of early arthritis were independently associated with the difference between the 2 utility measures (continuous and dichotomous: |SF-6D – EQ5D|>0.03, minimal important difference, MID) [ref].
Results At baseline, mean values were 0.50±0.32 (range -0.53 to 1) for EQ-5D and 0.58±0.11 (range 0.30 to 0.92) for SF-6D. In the multivariate linear regression model, higher HAQ, lower mental component of the SF-36 (MCS), higher CRP, higher patient global assessment and be employed were significant determinants of a lower EQ-5D. HAQ and MCS scores explained 56 and 4% of the variance respectively. In the multivariate linear regression model, higher HAQ, lower MCS, higher DAS28, higher pain at rest, higher patient global evaluation, and be employed were significantly associated with lower SF-6D. HAQ and MCS scores explained 49 and 22% of the variance respectively. Higher difference between the 2 utility values (mean 0.083 [range -0.469 to 0.891]) was explained by higher HAQ, lower number of swollen joint count, and higher CRP. The majority of patients presented a difference between the 2 utility values>0.03=MID (88%). In the logistic regression model, only higher disability was associated with |SF-6D –EQ5D|>0.03: OR=1.68 [1.13;2.50].
Conclusions there is systematic disagreement between EQ-5D and SF-6D in EA, especially in patients with worse clinical outcomes. Higher functional disability was the key element leading to substantial difference between the 2 utility measures in EA patients.
Marra CA, et al. Soc.Sci.Med. 2005;60:1571-1582
Disclosure of Interest None Declared