A QALY IS NOT A QALY... WITH EQ-5D, YOU HAVE 70% MORE QALYS GAINED THAN WITH SF-6D FOR THE SAME TREATMENT IN EARLY ARTHRITIS: RESULTS OF THE ESPOIR COHORT

Background: The explosion of drug development for rheumatoid arthritis and the revolution of early aggressive therapy for the disease have fuelled the search for better approaches to establish cost-effectiveness in early arthritis (EA). Unfortunately consensus on the choice of utility instrument is still lacking. Objectives: We aimed to compare the EQ-5D and SF-6D, 2 indirect utility measures widely used to calculate quality-adjusted life-years (QALYs), in terms of their utility score changes, and QALYS gain obtained, in a large prospective cohort of patients with EA, according to different therapeutic strategies. Methods: – Patients: included in the French nationwide ESPOIR cohort of EA (at least 2 swollen joints for less than 6 months and suspicion of RA). – Data available: SF-6D and EQ-5D utility measures were longitudinally assessed in 813 patients with EA (at baseline, 6 months, 1year). Bio-clinical variables and X-rays were also recorded. – Analysis: The change in SF-6D and EQ-5D utility scores and the QALYS gain obtained according to each utility measure change at 1 year (area under the curve (AUC) of the change in utility score between baseline and 6 months + AUC of the change in utility score between 6 and 12 months) were calculated and compared using paired t-test for the entire sample. This analysis was also conducted in patients treated by methotrexate (MTX) within the first 3 months without biological treatment. Results: 813 patients were included: mean age=48.1±12.6 years, 76.7% were female, mean DAS28=5.11±1.31; 372 patients (45.8%) were RF-positive and 315 (38.8%) were ACPA-positive. At 12 months, the majority of patients had improved (91.4%): mean DAS change= -1.95 (SD=1.5). Whereas the distribution of utility scores was bimodal for the EQ-5D and near normal for the SF-6D, the distribution of utility change was almost normal for both. The EQ-5D provided larger absolute mean change estimates with greater change variance than the SF-6D: 0.154±0.315 for EQ-5D vs 0.095±0.129 for SF-6D at 1 year. Regarding the 313 patients treated by MTX within the first 3 months without biological treatment, the mean QALYs gain for the EQ-5D was 0,118±0,213 and for the SF-6D=0,071±0,082 with a significant difference of 0,050±0,190 QALYs in favour of EQ-5D (70% more QALYs gained than with SF-6D) (p<0.0001). Conclusions: The high mean change of the EQ-5D compared to the SF-6D has consequences for cost-effectiveness analyses. In this example, the change estimated with the EQ-5D would result in a cost per QALY gained 40% lower than the cost per QALY gained calculated with the SF-6D. Furthermore, the smaller variance of the SF-6D would result in less uncertainty in estimating the relative cost-effectiveness of 2 treatments. The SF-6D may be more appropriate for use in RCTs of treatments for EA patients. Disclosure of Interest: None declared DOI: 10.1136/annrheumdis-2014-eular.5708Citation: Annals of the Rheumatic Diseases, volume 73, supplement 2, year 2014, page 332Session: Epidemiology, health services and outcome research (Poster Presentations )