LONGITUDINAL CHANGES IN QUALITY OF LIFE IN RHEUMATOID ARTHRITIS COMPARED WITH A HEALTHY US NATIONAL SAMPLE
Kaleb Michaud, MS, Stanford University, Stanford, CA
Purpose. Quality of life (QOL) is impaired in rheumatoid arthritis (RA), but there have been no studies comparing RA QOL with US national data.
Methods. I evaluated 12,353 patients with RA in a long-term observational study of arthritis outcomes. Between July 2002 and January 2006, patients contributed semiannual measurements of indirect QOL utilities, including the EQ5D (US & UK scoring), EQ-VAS and SF-6D; a 1,355 patient-subset completed the HUI3 in January 2003. Coons et al (in preparation) provided national EQ5D-US values for comparison. QOL measures were also adjusted to US-norms in ethnicity, income, marital status and education. Fixed effects (FE) models with overlapping age splines measured within-patient change over time and accounted for changes in QOL that controlled for participation bias and survivor effect.
Results. Mean (SD) characteristics of RA subjects were: age 62.2 (13.1) years, male 21.4%, non-Hispanic white 92.8%, education 13.7 (2.3) years and HAQ 1.03 (0.71). Unadjusted QOL utilities were EQ5D-US 0.743 (0.173), EQ5D-UK 0.653 (0.256), SF-6D 0.626 (0.087), EQ-VAS 0.678 (0.177) and HUI3 0.657 (0.246). The mean EQ5D-US QOL in RA increased slightly over age in contrast to the US mean, but the FE showed a similar to US data QOL by age relationship until post-retirement with increased worsening in RA. While the average male QOL with RA was greater than the female (0.723 vs 0.707), the FE curve showed there is a much greater QOL decline in males over time. Adjusted EQ5D-US for RA was 0.709, 0.163 lower than US average of 0.872. The EQ5D-US correlated with the HAQ (r=-0.66), increased on average (SD) by 0.030 (0.133) for a 6-month 0.25-unit improvement in HAQ and ranged from 0.894 (0.100) at 0-0.125 HAQ to 0.332 (0.221) at 2.75-3.0 HAQ.
Conclusions. Lifetime QOL in RA parallels US-norms but is consistently and significantly reduced. The HAQ can be an approximate utility measure in effectiveness analyses.