ASSESSING UTILITY VALUES IN RHEUMATOID ARTHRITIS PATIENTS: A COMPARISON BETWEEN THE TIME TRADE OFF AND THE EUROQOL

Background: Utility values can be assessed using specific instruments such as the time trade off (TTO) and the standard gamble, or indirectly by health profiles such as the EuroQoL-5D (EQ-5D). Quality-adjusted life years (QALYs) are calculated on utility scores and then, used in economic analysis. It remains uncertain if direct or indirect utility methods have the best performance.Objectives: To compare the utility values obtained by the TTO and the EQ-5D in a cohort of rheumatoid arthritis (RA) patients. To compare the gained QALYs based on these instruments. To analyze the association of the utility values with the Disease Activity Score (DAS28), and the Health Assessment Questionnaire (HAQ).Methods: A prospective, longitudinal, multi-centre, one-year, observational study. The setting: rheumatology clinics in tertiary care hospitals. Out-patients with RA (ACR criteria 1987) were randomly selected to participate in agreement with ethical committees. They were seen four times in a year. Two utility instruments were administered at baseline and each four months: a life expectancy-based TTO and the EQ-5D health profile. It gives 243 different health states, each of them with one utility value (social tariff). The social tariffs were assigned by the general population using TTO methods. QALYs were calculated by (utility value) x (time with the value). Utility scores were compared by Student''s t test. Correlation and regression analysis were used for assessing the associations between utility instruments and other variables.Results: 300 patients were included, 246 females (82%) with age (mean ±SD) 59.6 ± 13.34 years, DAS28 4.53 ±1.51, HAQ 1.18± 0.90, TTO 0.81 ± 0.22, and EQ-5D 0.5424 ± 0.3458, at baseline. TTO scores were significantly higher than those of EQ-5D, p<0.00001). A significant but poor association was found between TTO and EQ-5D (r=0.29, R2=0.08, p<0.00001). The mean changes ± SD along the year of the study were TTO 0.05 ±0.25, and EQ-5D 0.0120 ±0.3486, p=0.25. A significant but poor association was found between the mean changes of the TTO and the EQ-5D (r=0.34, R2=0.12, p<0.00001). Gained QALYs were 0.04 ±0.2 based on TTO, and 0.02 ±0.27 based on EQ-5D, p=0.48. The correlation between the gained QALYs measured by TTO and EQ-5D was significant but poor (r=0.38, R2=0.14, p<0.00001). At baseline, EQ-5D had good correlation with HAQ (r=-0.75, R2=0.56, p<0.00001) and moderate correlation with DAS28 (r=-0.48, R2=0.22, p<0.00001). The correlations between TTO and DAS28 (r=-0.27, R2=0.07, p<0.00001) and TTO and HAQ (r=-0.39, R2=0.15, p<0.00001) were significant but poor. The correlation between the change along one year in the EQ-5D and the change in the HAQ was moderate (r=-0.57, R2=0.32, p<0.00001). The correlations between the change in the TTO and the changes in DAS28 and HAQ were poor.Conclusion: Direct and indirect utility instruments did not yield the same results. The TTO scores were higher than those of the EQ-5D. RA patients gained similar QALYs with the two instruments but the correlation between them was low. EQ-5D had a higher association with the HAQ and DAS28 than the TTO. The data suggest that EQ-5D yield a more conservative estimation of the utility and it is more representative of the disease than the TTO. These findings should be considered in economic evaluations in rheumatoid arthritis.Citation: Ann Rheum Dis, volume 64, supplement III, year 2005, page 69Session: Advances in outcomes science and economics